Systematic evaluating and modeling of SARS-CoV-2 UVC disinfection

Abstract

The ongoing COVID-19 global pandemic has necessitated evaluating various disinfection technologies for reducing viral transmission in public settings. Ultraviolet (UV) radiation can inactivate pathogens and viruses but more insight is needed into the performance of different UV wavelengths and their applications. We observed greater than a 3-log reduction of SARS-CoV-2 infectivity with a dose of 12.5 mJ/cm² of 254 nm UV light when the viruses were suspended in PBS, while a dose of 25 mJ/cm² was necessary to achieve a similar reduction when they were in an EMEM culture medium containing 2%(v/v) FBS, highlighting the critical effect of media in which the virus is suspended, given that SARS-CoV-2 is always aerosolized when airborne or deposited on a surface. It was found that SARS-CoV-2 susceptibility (a measure of the effectiveness of the UV light) in a buffer such as PBS was 4.4-fold greater than that in a cell culture medium. Furthermore, we discovered the attenuation of UVC disinfection by amino acids, vitamins, and niacinamide, highlighting the importance of determining UVC dosages under a condition close to aerosols that wrap the viruses. We developed a disinfection model to determine the effect of the environment on UVC effectiveness with three different wavelengths, 222 nm, 254 nm, and 265 nm. An inverse correlation between the liquid absorbance and the viral susceptibility was observed. We found that 222 nm light was most effective at reducing viral infectivity in low absorbing liquids such as PBS, whereas 265 nm light was most effective in high absorbing liquids such as cell culture medium. Viral susceptibility was further decreased in N95 masks with 222 nm light being the most effective. The safety of 222 nm was also studied. We detected changes to the mechanical properties of the stratum corneum of human skins when the 222 nm accumulative exposure exceeded 50 J/cm².The findings highlight the need to evaluate each UV for a given application, as well as limiting the dose to the lowest dose necessary to avoid unnecessary exposure to the public.

Figure 1

Inactivation of SARS-CoV-2 by UVC irradiation. (A) A schematic diagram of the UVC dose testing device. (B) A plaque-forming assay of SARS-CoV-2 infectivity after a high dose (12.5–200 mJ/cm²) UVC treatment. (C) A plaque-forming assay of SARS-CoV-2 infectivity after a low dose (2.5–10 mJ/cm²) UVC treatment performed in 1X PBS. Controls were treated exactly the same, but did not receive any UV light exposure. (D) UVC 254 nm inactivation curves from the plaque-forming assay for SARS-CoV-2 in 1 × PBS or MEM + 2%(v/v) FBS. (E) Susceptibility factors calculated from the UVC inactivation (*p value < 0.05).

Figure 2

GFP-lentivirus experimental workflow. (A) Lentiviral samples are prepared either as a liquid medium or fomite. (B) A small representative portion of the sample is placed in a polystyrene 96-well plate and exposed to UVC light. (C) Treated samples are used to infect HEK293-T cells and incubated for two days. (D) A nuclear stain is added to facilitate cell counting using a high-content imaging device to determine a log reduction in virus infectivity after UVC irradiation.

Figure 3

Effect of a medium on UVC virus disinfection. (A) Comparison of virus inactivation under different wavelengths (222 nm, 254 nm, and 265 nm) of UVC irradiation. A portion of 300 µL of lentivirus suspension in either DMEM + 10%(v/v) FBS (top panel) or DPBS without calcium or magnesium (bottom panel) was exposed to 4 J/cm² of either 222, 254 or 265 nm UV light. Scale bar: 1000 µm. (B) Attenuation coefficient for suspension liquids used in UVC experiments. Attenuation coefficients for (C) amino acids and (D) vitamins in DMEM formulation.

Figure 4

Lentiviral UVC susceptibility when being suspended in three different media. A portion of 300 µL of a lentivirus suspended in DPBS, an artificial saliva, or complete DMEM medium supplemented with 10%(v/v) FBS were exposed to various doses of 222, 254, or 265 nm UVC light. Lentivirus was exposed either to (A) 0.15–2 J/cm² in DMEM + 10%(v/v) FBS or a (B) lower dose range of 0.05–0.2 J/cm² before being used to infect HEK293-T cells. Cells were stained with a 1:2000 diluted Hoechst 33,342 at two days post infection. Scale bar: 1000 µm. Image analysis quantified the transduction efficiency of HEK293-T cultures for a (C) lower dose range and (D) the higher low dose range, and from these data (E) the susceptibility factors for lentivirus exposed to 222 nm, 254 nm, and 265 nm was calculated and compared. p value < 0.05. (F) Absorbance spectra of the three liquids showed DPBS absorbed both light sources negligibly across the 230–300 nm range. DMEM + 10%(v/v) FBS contains a secondary peak around 277 nm created by its components as well as the addition of the FBS itself. Error bars denote standard deviation. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Figure 5

Environmental effect on virus UVC susceptibility. (A) Viruses suspended in DMEM + 10%(v/v) FBS with different liquid depths created by adding different volumes (0.1, 0.2, and 0.3 mL) of virus suspension to a 96-well plate. Viruses were exposed to various does of 222 nm, 254 nm, or 265 nm UVC light from which susceptibility factors were calculated. p values < 0.05. (B) The correlation of virus susceptibility to liquid absorbance of UVC light at 254 nm. The 254 nm viral susceptibilities were negatively correlated to the measured liquid absorbance at the three liquid heights. (C) Mixtures containing different percentages of DMEM + 10%(v/v) FBS (25, 50, 75, and 100%) to DPBS were exposed to 222 nm, 254 nm, and 265 nm UVC light. (D) The 254 nm viral susceptibility factors were negatively correlated to the measured absorbances of the different mixtures. (E) N95 mask cut-out samples were contaminated with 10 µL of a concentrated viral suspension and then irradiated to 254 nm and 222 nm at various doses. Virus was significantly more susceptible to 222 nm in DPBS. (F) N95 masks decreased the susceptibility of virus tenfold when compared to a pure liquid environment. Error bars denote standard deviation. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Figure 6

UVC light induces changes in the mechanical properties of SC equilibrated to 25% RH. Average (A) elastic modulus, $E$, (B) fracture stress, ${\sigma }_f$, (C) fracture strain, ${\gamma }_f$, and (D) work of fracture, $W_f$ for unirradiated controls (Control; light grey), 254 nm irradiated samples (dosage range: 50–400 J/cm²; dark grey), and 222 nm irradiated samples (dosage range: 50–400 J/cm²; white). Bars denote average values of $4\le n\le 5$ individual sample measurements for each range and dosage condition. Error bars denote standard deviations. Confidence intervals are indicated by *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Figure 7

UVC light induces changes in the mechanical properties of SC equilibrated to 100% RH. Average (A) elastic modulus, $E$, (B) fracture stress, ${\sigma }_f$, (C) fracture strain, ${\gamma }_f$, and (D) work of fracture, $W_f$ for unirradiated controls (Control; light grey), 254 nm irradiated samples (dosage range: 50–200 J/cm²; dark grey), and 222 nm irradiated samples (dosage range: 50–200 J/cm²; white). Bars denote average values of $4\le n\le 5$ individual sample measurements for each range and dosage condition. Error bars denote standard deviations. Confidence intervals are indicated by *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Figure 8

Loss of SC plastic deformability with UVC irradiation. Representative stress–strain plots of an unirradiated SC sample (dark grey circle), a sample irradiated with 200 J/cm² of incident 254 nm light (light grey triangle) and a sample irradiated with 200 J/cm² of incident 222 nm light (white square). All samples are equilibrated to 100% RH.