Introduction

Emergence of SARS-CoV-2 in Wuhan, China, in December 2019 led to the SARS-CoV-2-associated respiratory disease (COVID-19) pandemic, which was a major global health concern1,2. The omicron variant B.1.1.529, first reported in November 20213, and its descendent lineages BA.1, BA.2, BA.4/BA.5, showed increased transmissibility and escape from neutralizing antibodies, which complicated efforts to control the spread of SARS-CoV-2 infections4,5. Fortunately, rapid development of vaccines, as well as other drugs, to control COVID-19, coupled with general hygiene and preventive measures, have brought the virus under control. On May 2023, the World Health Organization (WHO) declared that COVID-19 no longer qualifies as a public health emergency of international concern6,7.

Although the COVID-19 pandemic has officially ended, SARS-CoV-2 continues to circulate among human populations, and its genome continues to acquire mutations. In fact, by the end of 2023, several omicron sub-lineages derived from BA.2 (BA.2.86 and JN.1), which carry mutations at key antigenic sites in the spike protein, were reported as having potential immune-evasion characteristics8. Therefore, continued surveillance and rapid phenotypic characterization of newly emerging variants/sub-lineages is critical if we are to better understand the mechanisms underlying viral transmission, pathogenicity, and immune evasion, and to assess potential threats to public health.

Usually, COVID-19 is diagnosed by detecting viral nucleic acids or antigens in specimens taken from the upper respiratory tract (i.e., nasopharyngeal/oropharyngeal swabs, nasal washes, or saliva)9. The high sensitivity and specificity of reverse transcription polymerase chain reaction (RT-PCR) makes it the most reliable test for this virus10. However, virus isolation from cell culture of clinical specimens provides surrogate information about viral transmission, and is necessary for characterizing the virological properties of live SARS-CoV-2 isolates. Vero E6 cells showing constitutive expression of transmembrane protease serine 2 (TMPRSS2; required for viral entry into cells), called VeroE6/TMPRSS2 cells, are highly susceptible and permissive to infection by SARS-CoV-2 because they also express the angiotensin converting enzyme 2 (ACE2) receptor11,12. Appearance of cytopathic effects (CPEs), which include rounding and detachment of VeroE6/TMPRSS2 cells inoculated with specimens, are a reliable indicator of the presence of infectious SARS-CoV-212. The VeroE6/TMPRSS2 cell line has been useful for isolating SARS-CoV-2 from clinical respiratory specimens13. However, the omicron variant replicates more slowly than the Delta variant (B.1.617.2 lineage first identified in October 2020) in these cells14, suggesting that this line is less effective as a means of isolating omicron variants. Furthermore, SARS-CoV-2 variants exhibit more profound CPEs and higher TCID50 values when infecting Vero E6 cells overexpressing human ACE2 and TMPRSS2 (i.e., Vero E6-TMPRSS2-T2A-ACE2 cells) than when infecting Vero E6/TMPRSS2 cells15. Thus, it appears that Vero E6-TMPRSS2-T2A-ACE2 cells are superior to VeroE6/TMPRSS2 cells for isolating of SARS-CoV-2, especially omicron variants, from clinical respiratory specimens. The aim of this study was to examine whether this is actually the case. We compared the efficiency of Vero E6-TMPRSS2-T2A-ACE2 cells with that of VeroE6/TMPRSS2 cells by using these lines to isolate SARS-CoV-2 from clinical respiratory specimens from COVID-19 patients.

Materials and methods

Cells

Monkey kidney-derived Vero E6 cells (ATCC CRL-1586) and human cervical adenocarcinoma-derived HeLa229 cells (ATCC CCL-2.1) were maintained at 37℃/5% CO2 in Dulbecco’s-modified Eagle’s medium (DMEM, Fujifilm Wako Chemicals) containing 10% heat-inactivated fetal bovine serum (FBS, Cytiva) and 100 U/mL penicillin/streptomycin (Merck). VeroE6/TMPRSS2 cells (JCRB1819, Japanese Collection of Research Bioresources Cell Bank) were maintained at 37℃/5% CO2 in low-glucose DMEM containing 10% heat-inactivated FBS, 1.0 mg/mL Geneticin (Thermo Fisher Scientific), and 100 U/mL penicillin/streptomycin. Vero E6-TMPRSS2-T2A-ACE2 cells (NR-54970, BEI Resources) were maintained at 37℃/5% CO2 in DMEM containing 10% heat-inactivated FBS, 100 U/mL penicillin/streptomycin, 1% MEM non-essential amino acids (Thermo Fisher Scientific), 1.0 mM sodium pyruvate (Thermo Fisher Scientific), and 1.0 µg/mL of puromycin dihydrochloride (Thermo Fisher Scientific).

Western blot analysis

Cell lysates obtained from HeLa229, Vero E6, VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells were resolved in sodium dodecyl sulfate–polyacrylamide gels electrophoresis and the proteins transferred to polyvinylidene difluoride membranes by electroblotting. ACE2 protein was detected by a goat anti-human ACE2 ectodomain antibody (AF933, R&D systems). β actin (loading control) was detected by a mouse anti-β actin antibody (C4, Santa Cruz Biotechnology). The protein bands on a western blot were quantified by using Image J software (https://imagej.net/software/imagej/).

Clinical specimens

A total of 288 upper respiratory samples (nasopharyngeal/nasal swabs or saliva) suspended in phosphate-buffered saline or viral transport medium were collected from individuals positive for viral RNA or antigen, and from individuals suspected as having SARS-CoV-2 infection (detected as part of a COVID-19 surveillance at international airport quarantine stations, under the quarantine act in Japan, in January 2022). Another set of 192 upper respiratory samples collected from COVID-19-suspected patients from October to November 2022 were obtained from the Repository of Data and Biospecimen of Infectious Disease (REBIND) project, which was launched jointly by the National Center for Global Health and Medicine (NCGM) and the National Institute of Infectious Diseases (NIID) commissioned from the Ministry of Health, Labor and Welfare (MHLW), Japan. Specimens were obtained with permission from the REBIND project (No. EX20004). Written informed consent was obtained from all participants before experimental procedures were performed. All samples were anonymized, examined by quantitative RT-PCR (qRT-PCR), and subjected to virus isolation. All experimental procedures were performed in accordance with Ethical Guidelines for Medical and Biological Research Involving Human Subjects in Japan.

Next-generation sequencing of the SARS-CoV-2 genome

Whole genome sequences of SARS-CoV-2 isolates were obtained using a modified protocol based on The Wellcome Trust ARTIC Network protocol16; however, several primers replaced or others added, as described in previous studies17,18. Genomic sequencing was performed using the iSeq 100 or MiSeq (Illumina) platforms. The trimmed reads were mapped to the SARS-CoV-2 reference genome MN908947.3, and consensus sequences were obtained using CLC Genomics Workbench (version 22, Qiagen). The obtained viral genome consensus sequences were subjected to lineage determination using Nextclade v.2.9.1 (https://clades.nextstrain.org) and Pangolin COVID-19 Lineage Assigner v.4.1.3 (https://pangolin.cog-uk.io/). The sequences were also registered with the Global Initiative on Sharing All Influenza Data (GISAID; https://www.gisaid.org/). The original viral genome sequences obtained from clinical specimens were obtained from the REBIND project, with a permission from the REBIND project (No. EX20004).

Viruses

The ancestral SARS-CoV-2 virus (hCoV-19/Japan/TY-WK-521/2020, Wuhan strain), the BA.5 variant (hCoV-19/Japan/TY41-702-P1/2022), the BF.7.4.1 variant (hCoV-19/Japan/TY41-828-P1/2022), the BF.7 variant (hCoV-19/Japan/TY41-820-P1/2022), the XBB.1.5 variant (hCoV-19/Japan/23–018-P1/2022), and the CH.1.1.3 (hCoV-19/Japan/TY41-832-P1/2022) were prepared using VeroE6/TMPRSS2 cells, as described previously19. The viral titer was determined in a 50% tissue culture infectious dose (TCID50) assay using VeroE6/TMPRSS2 cells. All procedures using infectious SARS-CoV-2 was performed in a Biosafety-level (BSL)-3 laboratory at NIID, Japan.

Neutralization assay

Neutralization of SARS-CoV-2 by rabbit serum raised against the anti-receptor binding ___domain (RBD) of SARS-CoV-2 ancestral virus13 was performed as previously described19. Briefly, serially diluted anti-RBD (twofold, starting from 1:5) in DMEM containing 2% heat-inactivated FBS and 100 U/mL penicillin/streptomycin, mixed with 100 TCID50 of SARS-CoV-2 ancestral or BA.5, and incubated at 37℃ for 1 h. Then, the mixture was inoculated onto Vero E6, VeroE6/TMPRSS2, or Vero E6-TMPRSS2-T2A-ACE2 cells seeded in 96 well culture plates. Cells were then cultured at 37℃ for 5 days, fixed with 10% formalin, and stained with crystal violet solution. The neutralization (NT) titer was defined as the geometric mean of the reciprocal of the highest sample dilution that protected at least 50% of the cells from a CPE (based on duplicate wells).

qRT-PCR

The N2 set qRT-PCR for SARS-CoV-2 was performed as described previously10. Briefly, the 20 μL reaction mixture comprised 2 × Master mix, the RT mix from the QuantiTect RT-PCR kit (Qiagen), the N2 primer/probe set (NIID_2019-nCoV_N_F2 primer, NIID_2019-nCoV_N_R2 primer and NIID_2019-nCoV_N_P2 probe), and 5 μL of RNA extracted from upper respiratory specimens. Reactions were performed on a LightCycler 480 (Roche) under the following conditions: 50℃ for 30 min; 95℃ for 15 min; and 45 cycles for 95℃ for 15 s and 60℃ for 1 min. Ct values were obtained for all qRT-PCR positive samples. The RNA copy number in each reaction was also determined using a standard curve based on measurements obtained using standard RNA10.

The qRT-PCR to detect TMPRSS2 messenger RNA (mRNA) was performed as essentially described12. Briefly, 200 ng of total RNA extracted from HeLa229, Vero E6, VeroE6/TMPRSS2, or Vero E6-TMPRSS2-T2A-ACE2 cells were subjected to qRT-PCR assay using a primer/probe set specific for TMPRSS2 (Hs00237175_m1, Thermo Fisher Scientific); the reaction conditions were as recommended by the manufacturer.

Virus isolation

Upper respiratory samples positive for SARS-CoV-2 qRT-PCR were stored at -80℃ before processing for cell culture. VeroE6/TMPRSS2 or Vero E6-TMPRSS2-T2A-ACE2 cells were inoculated with upper respiratory samples, as described previously13. Briefly, 50 μL of respiratory sample was mixed with isolation medium [DMEM supplemented with 2% FBS, 1% Antibiotic–Antimycotic (Thermo Fisher Scientific)] and inoculated onto VeroE6/TMPRRSS2 or Vero E6-TMPRSS2-T2A-ACE2 cells seeded in 96-well culture plates. Cells were observed daily under a microscope. The virus isolation assay was considered positive when CPEs were observed. To confirm the presence of SARS-CoV-2 virus in CPE-positive culture wells, the N2 set qRT-PCR for SARS-CoV-2 was performed using RNA extracted from culture supernatants. Virus isolation was considered negative when no CPEs were observed by 6 days post-inoculation (dpi). All virus isolation procedures were performed in a BSL-3 laboratory at NIID, Japan.

VSV pseudoviruses

VSV pseudoviruses bearing SARS-CoV-2 spike proteins were generated as previously described20. Briefly, 293 T cells were grown on collagen-I-coated tissue culture plates (Corning) and co-transfected with an expression plasmid (1.0 μg/well in a 12-well plate) encoding the C-terminal 19 amino acid-truncated spike protein of SARS-CoV-2 strains WK-521 (ancestral), TY4-920 (D614G), QHN002 (Alpha), TY8-612 (Beta), TY7-501 (Gamma), TY11-927 (Delta), TY26-717 (Mu), TY38-873 (BA.1), TY38-871 (BA.1.1), TY40-158 (BA.2), TY41-702 (BA.5), TY41-795 (XBB), or TY41-796 (BQ.1.1) using the TransIT-LT1 transfection reagent (Mirus Bio LLC). At 48 h after transfection, the cells were washed once with PBS and inoculated with G-complemented (*G) VSV∆G/Luc (VSV∆G/Luc-*G). After 1.5 h at 37 °C, the cells were washed three times with PBS and resuspended in 5% FBS-DMEM. After 24 h, the supernatants (containing VSV pseudoviruses) were centrifuged to remove cell debris and stored at − 80 °C until required. The infectivity of the pseudoviruses was assessed by measuring luciferase activity using the Bright-Glo luciferase assay system (Promega) and a GloMax Discover microplate reader (Promega); infectivity was expressed as relative light units (RLUs). An equivalent infectious dose of each pseudoviruses (~ 1 × 106 RLU), determined using VeroE6/TMPRSS2 cells, was inoculated onto Vero E6, VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells to compare susceptibility of these cells to infection. A negligible amount of seed virus VSV∆G/Luc-*G contamination in the SARS-CoV-2 pseudovirus preparation was preliminary confirmed by using anti-VSV-G antibody prior to inoculation of SARS-CoV-2 pseudovirus (Suppl. Fig. S1).

Statistical analysis

Data analysis and visualization were performed using GraphPad Prism 8 (GraphPad software). An unpaired t-test was used to determine statistically significant differences between groups. The threshold for statistical significance was defined as a P-value < 0.005.

Results

Expression of TMPRSS2 and ACE2

Previous studies show that VeroE6/TMPRSS2 cells are superior with respect to isolation of SARS-CoV-2 from clinical specimens due to constitutive expression of the serine-protease TMPRSS2, in addition to ACE212. First, we examined expression of TMPRSS2 mRNA in Vero E6-TMPRSS2-T2A-ACE2 cells relative to that in VeroE6 and VeroE6/TMPRSS2 cells. Expression of TMPRSS2 mRNA in Vero E6-TMPRSS2-T2A-ACE2 cells was approximately 100-fold higher than that in VeroE6/TMPRSS2 cells (Fig. 1A). By contrast, a very low, nearly undetectable, level of TMPRSS2 mRNA was observed in Vero E6 cells, consisted with previous findings12,21. Next, we examined expression of ACE2. As shown in Fig. 1B, a dense protein band corresponding to ACE2 was detected in Vero E6-TMPRSS2-T2A-ACE2 cells by western blotting with an anti-ACE2 antibody. The expression level of ACE2 in VeroE6/TMPRSS2 cells was comparable with that in Vero E6 cells. These results indicate that expression of both TMPRSS2 and ACE2 in Vero E6-TMPRSS2-T2A-ACE2 cells is higher than that in VeroE6/TMPRSS2 cells.

Fig. 1
figure 1

Increased expression of TMPRSS2 and ACE2 by Vero E6-TMPRSS2-T2A-ACE2 cells. (A) Analysis of TMPRSS2 mRNA levels by qRT-PCR. mRNA levels (-fold) in Vero E6, VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells were compared with those in HeLa229 cells (set to 1.0). *: p < 0.005, **: p < 0.0001. (B) Western blot analysis with an anti-human ACE2 antibody was performed to detect ACE2 protein in HeLa229, Vero E6, VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells. β-actin was used as a loading control. The amount of ACE2 protein detected in each cells was normalized with the that of β-actin and expressed as a fold change (the amount in HeLa cells was set to 1.0). The transferred membrane was cut prior to hybridization with anti-ACE2 or anti-β-actin. The original of blots with membrane edges visible was shown in the Supplementary figure S4.

Full size image

Isolation of SARS-CoV-2 from clinical specimens

In total, SARS-CoV-2 was isolated from 288 clinical specimens (collected at a quarantine station in Japan in January 2022) using Vero E6-TMPRSS2-T2A-ACE2 and VeroE6/TMPRSS2 cells in parallel. Cell morphology was monitored daily under a microscope. Obvious CPEs were observed within 5 dpi, with no additional CPEs at 6. dpi. The number of specimens showing CPEs in inoculated cells is shown in Fig. 2. An apparent CPEs were observed in eight and 57 samples at 1 and 2 dpi, respectively, in Vero E6-TMPRSS2-T2A-ACE2 cells, whereas there were no visible CPEs at 1 dpi, and only 10 samples were CPE-positive at 2 dpi, in VeroE6/TMPRSS2 cells. Thus, up until 5 dpi, 88 (30.6%) and 52 specimens (18.1%) were virus isolation-positive using Vero E6-TMPRSS2-T2A-ACE2 and VeroE6/TMPRSS2 cells, respectively. Among the 88 specimens positive in Vero E6-TMPRSS2-T2A-ACE2 cells, 47 (53.4%) were also isolation-positive, and 41 (46.6%) were negative, using VeroE6/TMPRSS2 (Table 1). These results indicate that Vero E6-TMPRSS2-T2A-ACE2 cells are more suitable for isolation of SARS-CoV-2 than VeroE6/TMPRSS2 cells.

Fig. 2
figure 2

Time course showing the number of inoculated wells with visible CPEs after virus inoculation. A total of 288 clinical specimens were inoculated onto VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells. Appearance of CPEs was monitored daily under a microscope. The number of inoculated wells showing visible CPEs at 1, 2, 3, 4, 5 dpi is shown.

Full size image
Table 1 Comparison of SARS-CoV-2 isolation using VeroE6-TMPRSS2-T2A-ACE2 cells with that using VeroE6/TMPRSS2.
Full size table

To examine whether virus isolation is dependent on the viral load in the clinical specimens, we measured the viral RNA copy number. The viral RNA copy number for 210 specimens was assessed by qRT-PCR. When using VeroE6/TMPRSS2 cells for virus isolation, the viral load among isolation-positive (n = 52) samples varied significantly (p < 0.0001), ranging from 5.0 to 8.2 (median 6.4) log10 copies/mL. The copy number among isolation-negative (n = 158) samples ranged from 4.1 to 8.1 (median 5.5) log10 copies/mL (Suppl Fig. S2). When using Vero E6-TMPRSS2-T2A-ACE2 cells, there is also a significant difference (p < 0.0001) in the viral load among isolation-positive samples (n = 85), ranging from 4.2 to 8.2 (median 6.2) log10 copies/mL, and among isolation-negative samples (n = 125), ranging from 4.1 to 7.5 (median 5.5) log10 copies/mL (Suppl Fig. S2). Among specimens that were isolation-positive in Vero E6-TMPRSS2-T2A-ACE2 cells (n = 88), three were negative by qRT-PCR, probably due to a very low infectious dose in these specimens. There was no significant difference in the viral load between the specimens positive for virus isolation in VeroE6/TMPRSS2 cells and those that were positive in Vero E6-TMPRSS2-T2A-ACE2 cells (Suppl Fig. S2). Furthermore, there was no significant difference in the viral load between the specimens positive for isolation only in VeroE6/TMPRSS2 cells (n = 5) and those positive for isolation only in Vero E6-TMPRSS2-T2A-ACE2 cells (n = 41) (Suppl Fig. S3).

Among specimens that contained a viral load of ≧106 copies/mL, 52.1% (38/73) and 67.1% (49/73) were positive in VeroE6/TMPRSS2 and in Vero E6-TMPRSS2-T2A-ACE2 cells, respectively, whereas among specimens that contained < 106 copies/mL, 10.2% (14/137) and 26.3% (36/137) were positive in VeroE6/TMPRSS2 and in Vero E6-TMPRSS2-T2A-ACE2 cells, respectively. There was a 2.6-fold increase in the percentage of virus isolation-positive specimens containing < 106 copies/mL in Vero E6-TMPRSS2-T2A-ACE2 cells (Fig. 3). These results indicate that Vero E6-TMPRSS2-T2A-ACE2 cells are superior to VeroE6/TMPRSS2 with respect to SARS-CoV-2 isolation, even though some specimens had a low viral load.

Fig. 3
figure 3

Virus isolation is superior when using Vero E6-TMPRSS2-T2A-ACE2 cells. Isolation-positive rate (%) according to viral genome copy number in the specimens (< 106 and ≧ 106 copies/mL), or in the total specimens (total).

Full size image

Genome sequences of isolated SARS-CoV-2 strains

To investigate whether SARS-CoV-2 isolated from clinical specimens acquires specific mutation(s), we used another set of clinical specimens obtained from the REBIND project and isolated SARS-CoV-2 using both VeroE6/TMPRSS2 and Vero E6-TMPRSS2-T2A-ACE2 cells, followed by whole genome sequencing. Among 192 clinical specimens, 40 (20.8%) and 57 (29.7%) were virus isolation-positives in VeroE6/TMPRSS2 and in Vero E6-TMPRSS2-T2A-ACE2 cells, respectively (Table 2). The isolated sequences were compared with those obtained from original clinical specimens. Among 17 viruses isolated from VeroE6/TMPRSS2 cells, two (11.8%) harbored mutations when compared with virus obtained directly from our clinical specimens; the remaining 15 isolates (88.2%) did not exhibit mutations (Table 3). By contrast, among 38 viruses isolated from Vero E6-TMPRSS2-T2A-ACE2 cells, eight (21.1%) harbored mutations compared with viruses obtained directly from original clinical specimens; the remaining 30 isolates (78.9%) did not exhibit any mutations (Table 3). The rate mutation in isolates derived from Vero E6-TMPRSS2-T2A-ACE2 cells seemed to be greater than that in isolates obtained from VeroE6/TMPRSS2 cells. Next, we examined whether isolating the same SARS-CoV-2 positive clinical sample independently from VeroE6/TMPRSS2 and Vero E6-TMPRSS2-T2A-ACE2 resulted in diverse mutations across the genome. By comparing a set of SARS-CoV-2 whole genome sequences from clinical specimens, VeroE6/TMPRSS2 cells, and Vero E6-TMPRSS2-T2A-ACE2 cells, that was available for 17 samples, we found that Vero E6-TMPRSS2-T2A-ACE2 cell-derived isolates did not acquire any specific mutation (Table 4). Mutations at nucleotide positions 6459/8898 or 11,750/27,513 were seen only in isolates derived from VeroE6/TMPRSS2 or Vero E6-TMPRSS2-T2A-ACE2 cells, respectively. An insertion between nucleotide 27,691 and 27,692 was observed in isolates from both cell lines (Table 4). These mutations caused amino acid substitutions in ORF1ab or ORF7 protein (Table 4), although undetermined whether each amino acid substitution affected the protein function.

Table 2 SARS-CoV-2 isolation from 192 samples obtained in REBIND project in 2022 using Vero E6-TMPRSS2-T2A-ACE2 cells and VeroE6/TMPRSS2.
Full size table
Table 3 Number of SARS-CoV-2 isolates having mutations/or no mutation compared to viral genome sequences directly from clinical specimens.
Full size table
Table 4 Comparison of whole genome sequence of SARS-CoV-2 obtained by virus isolation with those obtained directly from clinical specimensa.
Full size table

Infectivity of VSV pseudoviruses

Next, we investigated whether SARS-CoV-2 showed enhanced entry into Vero E6-TMPRSS2-T2A-ACE2 cells. To do this, we constructed VSV pseudoviruses bearing SARS-CoV-2 spike protein (VSV/SARS-2pv). Variations in the spike protein affect the ability of the virus to bind ACE222,23, and alters dependency on TMPRSS2 during cell entry24. Therefore, VSV/SARS-2pv constructed using various SARS-CoV-2 strains (i.e., D614G, Alpha, Beta, Gamma, Delta, Mu, BA.1, BA.1.1, BA.2, BA.5, XBB, and BQ.1.1 variants in addition to the ancestral strain) were inoculated to Vero E6, VeroE6/TMPRSS2, or Vero E6-TMPRSS2-T2A-ACE2 cells. The ability of VSV/SARS-2pv to infect VeroE6/TMPRSS2 and Vero E6-TMPRSS2-T2A-ACE2 cells was then compared with the ability to infect Vero E6 cells (set as 1.0) (Fig. 4A). We observed a 5- to tenfold increase in the infectivity of the ancestral, D614G, Alpha, Beta, Delta, and Mu variants in VeroE6/TMPRSS2 cells; however, infectivity of these variants increased even further (more than 20-fold) in Vero E6-TMPRSS2-T2A-ACE2 cells, along with several-fold increases in the infectivity of the remaining variants. The relative infectivity in Vero E6-TMPRSS2-T2A-ACE2 cells (VeroE6/TMPRSS2 set as 1.0) revealed a 2- to fourfold increase in the infectivity of VSV pseudoviruses bearing the SARS-CoV-2 spike protein (Fig. 4B). The omicron variants did not show any increase in infectivity in VeroE6/TMPRSS2 when compared with Vero E6 cells (Fig. 4A,B), which is consistent with findings that omicron variant uses TMPRSS2 less efficiently than the ancestral virus and Delta variant24.

Fig. 4
figure 4

Infectivity of VSV/SARS-2pv in Vero E6, VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells. VSV/SARS-2pv bearing spike proteins from the indicated SARS-CoV-2 variants were inoculated simultaneously onto Vero E6, VeroE6/TMPRSS2, or Vero E6-TMPRSS2-T2A-ACE2 cells. (A) Relative infectivity was calculated by dividing the RLU for VeroE6/TMPRSS2 (left) or Vero E6-TMPRSS2-T2A-ACE2 (right) by that for VeroE6. (B) Relative infectivity was calculated by dividing the RLU for Vero E6 (left) Vero E6-TMPRSS2-T2A-ACE2 (right) by that for VeroE6/TMPRSS2. Infection was performed in triplicate wells, using VSV∆G/Luc-*G (VSVpv) as a control pseudovirus.

Full size image

Virus titration and serum neutralization assays in Vero E6-TMPRSS2-T2A-ACE2 cells

The data described above indicate early appearance of CPEs and a significant increase in infectivity in Vero E6-TMPRSS2-T2A-ACE2 cells. Consistent with these observations, we found higher viral TCID50 values for all of the SARS-CoV-2 strains tested in Vero E6-TMPRSS2-T2A-ACE2 cells than in VeroE6/TMPRSS2 cells (Fig. 5A). It is speculated that the serum neutralization titer (which depends on the appearance of CPEs in inoculated cells) determined using Vero E6-TMPRSS2-T2A-ACE2 cells might be different from that determined using VeroE6/TMPRSS2 cells. To investigate this possibility, we performed a SARS-CoV-2 neutralization assay using Vero E6, VeroE6/TMPRSS2, or Vero E6-TMPRSS2-T2A-ACE2 cells. Rabbit anti-serum raised against the RBD protein of ancestral SARS-CoV-2 was mixed with SARS-CoV-2 ancestral or BA.5 viruses, and then used as the inoculum. The serum neutralization titer in Vero E6-TMPRSS2-T2A-ACE2 cells (Fig. 5B) was lowest among the cell lines in experiments using both ancestral and BA.5 virus. As in the virus isolation experiments (Fig. 2), clear CPEs were observed in the virus titration and neutralization assays within 5 dpi, with no additional CPE at 6 dpi (data not shown). These results indicate that the results of virus titration and neutralization assays might depend on the cell type used, and that a relatively high virus titer and a low serum neutralization titer might be obtained when using Vero E6-TMPRSS2-T2A-ACE2 cells.

Fig. 5
figure 5

Virus titration and serum neutralization assays using Vero E6-TMPRSS2-T2A-ACE2 cells. (A) Titration of the indicated SARS-CoV-2 strains on VeroE6/TMPRSS2 or Vero E6-TMPRSS2-T2A-ACE2 cells. The viral titer was determined in a TCID50 assay. (B) Neutralization assay using anti-RBD serum specific for the SARS-CoV-2 ancestral and BA.5 was performed on VeroE6, VeroE6/TMPRSS2, and Vero E6-TMPRSS2-T2A-ACE2 cells. The neutralization (NT) titers, defined as the geometric mean of the reciprocal of the highest sample dilution that protected from CPE (using duplicate wells for each measurement) are shown.

Full size image

Discussion

Isolating SARS-CoV-2 from clinical specimens is an essential for obtaining live virus to characterize new variants, and for virological diagnosis of COVID-19. The Vero cell line and its derivatives (e.g., Vero E6, VeroE6/TMPRSS2 cells) have been used to isolate SARS-CoV-212,25,26,27. Previous studies report efficient isolation of SARS-CoV-2 from clinical specimens using VeroE6/TMPRSS2; this is because these cells stably express both the receptor molecule ACE2 and the serine-protease12,13. This cell line has also been used to prepare virus stocks, for virus titration, and/or for examination of serum neutralization antibodies in neutralization assays19,28. Successful virus isolation is dependent on the viral load within a clinical sample13. Furthermore, it is very difficult to isolate SARS-CoV-2 from samples that contain < 106 viral genome copies/mL29.

Our novel finding is that Vero E6-TMPRSS2-T2A-ACE2 cells are superior to VeroE6/TMPRSS2 cells with respect to isolation of SARS-CoV-2 from clinical specimens. The improvement (a 2.6-fold increase in efficacy) was particularly noticeable when using clinical specimens with a relatively low viral load (Fig. 3). In addition, we found that SARS-CoV-2 isolated using VeroE6-TMPRSS2-T2A-ACE2 cells was genetically stable. When we compared the whole genome sequence of SARS-CoV-2 obtained by virus isolation with sequences obtained directly from clinical specimens, we found no specific or hot spot mutations in genomes isolated fromVeroE6/TMPRSS2 or Vero E6-TMPRSS2-T2A-ACE2 cells (Table 4). Therefore, we concluded that Vero E6-TMPRSS2-T2A-ACE2 cells might be useful for isolating SARS-CoV-2 from clinical specimens with high sensitivity. During the preparation of this manuscript, we became aware of a study by Yoshida et al.30, who showed that the isolation rate of SARS-CoV-2 omicron variants from Vero E6-TMPRSS2-T2A-ACE2 cells was higher than that from VeroE6/TMPRSS2; however, in our study, we measured the viral load in the clinical specimens, and found that Vero E6-TMPRSS2-T2A-ACE2 cells were better for isolating viruses from clinical samples with a low viral load.

We assumed that these results are due to the fact that viral entry into Vero E6-TMPRSS2-T2A-ACE2 cells is more efficient. Therefore, we examined viral entry using a VSV pseudovirus system (VSV/SARS-2pv), and obtained evidence that supports our assumption. Omicron variants were inefficient at using TMPRSS2 for cell entry (Fig. 4A), which is consistent with previous findings24; however, we observed 2– to fourfold increases in the infectivity of VSV/SARS-2pv bearing the spike proteins of all SARS-CoV-2 strains tested in Vero E6-TMPRSS2-T2A-ACE2 cells (Fig. 4B) when compared with VeroE6/TMPRSS2 cells. This might be due to higher expression of TMPRSS2 and ACE2 by Vero E6-TMPRSS2-T2A-ACE2 cells (Fig. 1). We could not differentiate the contributions of abundant expression of ACE2 and that of TMPRSS2 with respect to the efficient SARS-CoV2 isolation by using Vero E6-TMPRSS2-T2A-ACE2 cells. However, as for the omicron variants, ACE2 rather than TMPRSS2 seemed to have had a remarked strong effect on isolation since the TMPRSS2 was inefficiently used for the cell entry of omicron variants (Fig. 4).

One limitation of this study is that we performed single experiments (inoculation of clinical specimens into a single well containing VeroE6/TMPRSS2 and Vero E6-TMPRSS2-T2A-ACE2 cells) for the virus isolation experiments; this was due to the limited volume of the clinical specimens. We noted that among the 52 specimens that were virus isolation-positive in VeroE6/TMPRSS2 cells, five were negative in Vero E6-TMPRSS2-T2A-ACE2 cells (Table 1). Although the exact reason for this unexpected observation is unclear, it may be that these specimens had a very low infectious virus load. Another limitation is that we did not use human-derived cell lines. A recent study showed that human-derived Caco-2/AT and Huh-6/AT cells overexpressing ACE2 and TMPRSS2 are highly permissive for SARS-CoV-2 infection, and allow efficient isolation of SARS-CoV-2 from clinical specimens31. Chen et al.31, used 20 clinical specimens, whereas we tested a much larger number, and confirmed that the genomes of the isolated viruses were stable. Future studies with human-derived cells that overexpress ACE2 and TMPRSS2 might have the best prospects for isolating SARS-CoV-2 from clinical specimens.

Vero E6-TMPRSS2-T2A-ACE2 cells have been used for various SARS-CoV-2 studies (e.g., virus isolation, titration, stock preparation, inactivation validation, and neutralization assays). Kumar et al.15 showed that CPEs are more pronounced in Vero E6-TMPRSS2-T2A-ACE2 cells than in VeroE6/TMPRSS2 cells, making it possible to use a highly stringent anti-virus drug screening approach in cultured Vero E6-TMPRSS2-T2A-ACE2 cells. Our study shows that Vero E6-TMPRSS2-T2A-ACE2 cells have advantages over VeroE6/TMPRSS2 not only in terms of isolating SARS-CoV-2 from clinical specimens, but also in terms of measuring viral titers and detecting neutralizing antibodies with a high sensitivity. However, care should be taken when using cell culture systems to determine SARS-CoV-2 titers because the infectivity of SARS-CoV-2 may differ among cell culture systems. It is possible that the ratio of particles per plaque-forming unit, often used to define specific infectivity32,33,34, of SARS-CoV-2 could differ depending on the cell culture system. Although we did not measure the number of viral particles in each SARS-CoV-2 preparation, the viral titration data that we obtained from the two different cell lines could have a significant impact on how future studies of SARS-CoV-2 are designed with respect to measuring viral infectivity in vitro.