Estimating clinical SARS-CoV-2 infectiousness in Vero E6 and primary airway epithelial cells

Understanding the window of infectiousness for SARS-CoV-2 is essential for infection control measures. RT-PCR remains the gold standard for diagnosis but cannot inform on the presence of infectious virus, which can only be determined via inoculating cultured cells. Such findings are crucial for estimating infectiousness.¹ Although the process of virus transmission is multifactorial, viral load and successful isolation of the virus are most closely associated with the likelihood of transmission. However, studies estimating the probability of virus isolation were primarily done in Vero E6 cells, and found a strongly reduced success of isolation when the viral load was below 5–7 log10 RNA copies per mL or after more than 1 week of symptoms.2, 3, 4 Although Vero E6 cells are highly susceptible to SARS-CoV-2 and widely used for isolation, they do not mimic the primary site of entry in the human respiratory tract.

To assess presence of infectious SARS-CoV-2 in a more relevant model, we investigated virus isolation on Vero E6 and human primary airway epithelial cells in parallel, with viral load quantified with the WHO International Standard for SARS-CoV-2 RNA (National Institute for Biological Standards and Control code: 20/146).⁵ In 39 clinical samples (nasopharyngeal swabs of adults positive for SARS-CoV-2 within 5 days of symptom onset) with viral loads of 4·5–8·8 log10 SARS-CoV-2 international units (IUs) per mL, virus isolation was successful for 27 (69%) samples in Vero E6 cells and 12 (31%) samples in airway epithelial cells (appendix).

Using Probit analysis, the probability of virus isolation was below 5% when viral load was lower than 4·8 log10 IU/mL (95% CI 4·6–5·3) in Vero E6 cells, and 5·5 log10 IU/mL (4·9–6·1) in airway epithelial cells (p<0·05). Differences in the probability of virus isolation were highest between 5·5 and 7·5 IU/mL in Vero E6 cells versus airway epithelial cells (appendix). Overall Vero E6 cells were more permissive for SARS-CoV-2 infection than airway epithelial cells, allowing virus isolation in samples with lower viral load (appendix). This finding could indicate that actual infectiousness, transmissibility, and virus shedding in human cells in vivo are slightly overestimated when the presence of infectious virus is determined using Vero E6 cells. This conclusion is limited by the fact that viral loads and virus isolation do not fully equate to infectiousness in vivo. Because the assessment of successful virus isolation and viral load quantification can vary between laboratories, depending on protocols, samples, and materials used, the strength of our study is the use of viral load standardised by IU and parallel isolation of the same clinical sample in two cell culture systems. Our study emphasises the importance of the cell lines used for SARS-CoV-2 culture and supports the use of models closely mimicking the in-vivo situation for better understanding of SARS-CoV-2 transmission risks.

We declare no competing interests. We thank Catia Alvarez and Pascale Sattonnet-Roche for excellent technical support and Erik Boehm for language editing. This work was supported by the Private HUG Foundation, by the Pictet Charitable Foundation and by the Swiss National Science Foundation (196644, 196383).