key: cord-1028221-xwno2vt6
authors: Welch, Stephen R.; Davies, Katherine A.; Buczkowski, Hubert; Hettiarachchi, Nipunadi; Green, Nicole; Arnold, Ulrike; Jones, Matthew; Hannah, Matthew J.; Evans, Reah; Burton, Christopher; Burton, Jane E.; Guiver, Malcolm; Cane, Patricia A.; Woodford, Neil; Bruce, Christine B.; Roberts, Allen D. G.; Killip, Marian J.
title: Analysis of Inactivation of SARS-CoV-2 by Specimen Transport Media, Nucleic Acid Extraction Reagents, Detergents, and Fixatives
date: 2020-10-21
journal: J Clin Microbiol
DOI: 10.1128/jcm.01713-20
sha: a2f6a6ea9800da87e1eafb3e598ad917e96f3efb
doc_id: 1028221
cord_uid: xwno2vt6

The COVID-19 pandemic has necessitated a multifaceted rapid response by the scientific community, bringing researchers, health officials, and industry together to address the ongoing public health emergency. To meet this challenge, participants need an informed approach for working safely with the etiological agent, the novel human coronavirus SARS-CoV-2. Work with infectious SARS-CoV-2 is currently restricted to high-containment laboratories, but material can be handled at a lower containment level after inactivation. Given the wide array of inactivation reagents that are being used in laboratories during this pandemic, it is vital that their effectiveness is thoroughly investigated. Here, we evaluated a total of 23 commercial reagents designed for clinical sample transportation, nucleic acid extraction, and virus inactivation for their ability to inactivate SARS-CoV-2, as well as seven other common chemicals, including detergents and fixatives. As part of this study, we have also tested five filtration matrices for their effectiveness at removing the cytotoxic elements of each reagent, permitting accurate determination of levels of infectious virus remaining following treatment. In addition to providing critical data informing inactivation methods and risk assessments for diagnostic and research laboratories working with SARS-CoV-2, these data provide a framework for other laboratories to validate their inactivation processes and to guide similar studies for other pathogens.

(MSCIII) in a CL3 laboratory. Working virus stocks were generated by infecting Vero E6 cells at a multiplicity of infection (MOI) of 0.001 in the presence of 5% FCS. Cell culture supernatants were collected at 72 h postinfection, clarified for 10 min at 3,000 ϫ g, aliquoted, and stored at Ϫ80°C until required. Viral titers were calculated by either plaque assays or 50% tissue culture infectious doses (TCID 50 ). For plaque assays, 24-well plates were seeded the day before the assay (1.5 ϫ 10 5 cells/well in MEM-10% FCS). Tenfold dilutions of virus stock were inoculated onto plates (100 l per well) at room temperature for 1 h, overlaid with 1.5% medium-viscosity carboxymethylcellulose (Sigma-Aldrich), and incubated at 37°C in 5% CO 2 for 3 days. For TCID 50 , 10-fold dilutions of virus stock (25 l) were plated onto 96-well plates containing Vero E6 cell suspensions (2.5 ϫ 10 4 cells/well in 100 l of MEM-5% FCS) and incubated at 37°C in 5% CO 2 for 5 to 7 days. Plates were fixed with 4% (vol/vol) formaldehydephosphate-buffered saline (PBS) and stained with 0.2% (vol/vol) crystal violet-water. TCID 50 titers were determined by the Spearman-Kärber method (22, 23) .

Reagents and chemicals used for SARS-CoV-2 inactivation. The commercial reagents evaluated in this study, along with their compositions (if known) and manufacturers' instructions for use (if provided), are given in Table S1 in the supplemental material. Specimen transport reagents tested were the following: Sigma molecular transport medium (MM; Medical Wire), eNAT (Copan), Primestore molecular transport medium (MTM; Longhorn Vaccines and Diagnostics), Cobas PCR medium (Roche), Aptima specimen transport medium (Hologic), DNA/RNA Shield (Zymo Research), guanidine hydrochloride (GCHl) and guanidine thiocyanate (GITC) buffers containing Triton X-100 (both, Oxoid/Thermo Fisher), and virus transport and preservation medium (inactivated) (BioComma). Molecular extraction reagents tested were the following: AVL, RLT, and AL (all Qiagen); MagNA Pure external lysis buffer and Cobas Omni LYS used for on-board lysis by Cobas extraction platforms (Roche); viral PCR sample solution (VPSS) and lysis buffer (both, E&O Laboratories); NeuMoDx lysis buffer (NeuMoDx Molecular); Samba II SCoV lysis buffer (Diagnostics for the Real World); NucliSENS lysis buffer (bioMérieux); Panther Fusion specimen lysis tubes (Hologic); and an in-house extraction buffer containing guanidine thiocyanate and Triton X-100 (PHE Media Services). Detergents tested were the following: Tween 20, Triton X-100, and NP-40 Surfact-Amps detergent solution (all, Thermo Scientific) and UltraPure SDS 10% solution (Invitrogen). Other reagents assessed included polyhexamethylene biguanide (PHMB; Blueberry Therapeutics), formaldehyde and glutaraldehyde (both, TAAB), and ethanol and methanol (both, Fisher Scientific).

Removal of reagent cytotoxicity. Specimen transport tube reagents were assessed undiluted unless otherwise indicated. For testing of molecular extraction reagents, mock samples were generated by diluting reagent in PBS at ratios given in the manufacturers' instructions. Detergents, fixatives, and solvents were all assessed at the concentrations indicated in Table 1 . All methods were evaluated in a spin column format for ease of sample processing within the high-containment laboratory. Pierce detergent removal spin columns (0.5 ml; Thermo Scientific), Microspin Sephacryl S400HR (GE Healthcare), and Amicon Ultra 0.5-ml 50-kDa centrifugal filters (Merck Millipore) were prepared according to the manufacturers' instructions. Sephadex LH-20 (GE Healthcare) and Bio-Beads SM2 resin (Bio-Rad) were suspended in PBS and poured into empty 0.8-ml Pierce centrifuge columns (Thermo Scientific) and centrifuged for 1 min at 1,000 ϫ g to remove PBS immediately before use. For all matrices aside from the Amicon Ultra columns, 100 l of treated sample was added to each spin column, incubated for 2 min at room temperature, and then eluted by centrifugation at 1,000 ϫ g for 2 min. For Amicon Ultra filters, 500 l of sample was added and centrifuged at 14,000 ϫ g for 10 min, which was followed by three washes with 500 l of PBS. Sample was then collected by resuspending contents of the filtration device with 500 l of PBS. To assess remaining cytotoxicity, a 2-fold dilution series of treated filtered sample was prepared in PBS, and 6.5 l of each dilution was transferred in triplicate to 384-well plates containing Vero E6 cells (6.25 ϫ 10 3 cells/well in 25 l of MEM-5% FCS) and incubated overnight. Cell viability was determined by a CellTiter Aqueous One Solution cell proliferation assay (Promega) according to the manufacturer's instructions. Normalized values of absorbance (relative to untreated cells) were used to fit a four-parameter equation to semilog plots of the concentration-response data and to interpolate the concentration that resulted in 80% cell viability (20% cytotoxic concentration [CC 20 ]) in reagent-treated cells. All analyses were performed using GraphPad Prism (version 8.4.1; GraphPad Software).

SARS-CoV-2 inactivation. For commercial products, virus preparations (tissue culture fluid, titers ranging from 1 ϫ 10 6 to 1 ϫ 10 8 PFU/ml) were treated in triplicate with reagents at concentrations and for contact times recommended in the manufacturers' instructions for use, where available, or for concentrations and times specifically requested by testing laboratories. Where a range of concentrations was given by the manufacturer, the lowest ratio of product to sample was tested (i.e., lowest recommended concentration of test product). Specimen transport tube reagents were tested using a ratio of 1 volume of tissue culture fluid to 10 volumes of reagent, unless a volume ratio of sample fluid to reagent was specified by the manufacturer. Detergents, fixatives, and solvents were tested at the concentrations and times indicated in the tables. All inactivation steps were performed at ambient room temperature (18 to 25°C). For testing of alternative sample types, virus was spiked into the indicated sample matrix at a ratio of 1:9 and then treated with test reagents as described above. All experiments included triplicate control mock-treated samples with equivalent volumes of PBS in place of test reagent. Immediately following the required contact time, 1 ml of treated sample was processed using the appropriately selected filtration matrix. Reagent removal for inactivation testing was carried out in a larger spin column format using Pierce 4-ml detergent removal spin columns (Thermo Fisher) or by filling empty Pierce 10-ml-capacity centrifuge columns (Thermo Fisher) with SM2 Bio-Beads, Sephacryl S400HR, or Sephadex LH-20 to give 4 ml of packed beads/resin. For purification using Amicon filters, two 500-l samples were purified using two centrifugal filters by the method previously described and then pooled. For formaldehyde and formaldehyde with glutaraldehyde removal, one filter was used with one 500-l Values represent the dilution factor required after one purification process to achieve the CC 20 . CI, confidence interval; NA, not available; NC, not able to be calculated.

sample volume, which was resuspended after processing in 500 l of PBS and added to 400 l of MEM-5% FBS. For inactivation of infected monolayers, 12.5-cm 2 flasks of Vero E6 cells (2.5 ϫ 10 6 cells/flask in 2.5 ml of MEM-5% FBS) were infected at an MOI of 0.001 and incubated at 37°C in 5% CO 2 for 24 h. Supernatant was removed, and cells were fixed using 5 ml of formaldehyde or formaldehyde and glutaraldehyde at room temperature for 15 or 60 min. The fixative was removed, and monolayers were washed three times with PBS before cells were scraped into 1 ml of MEM-5% FBS and sonicated (three cycles of 10 s on and 10 s off at 100% power and amplitude) using a UP200St with VialTweeter attachment (Hielscher Ultrasound Technology). Supernatants were clarified by centrifugation at 3,000 ϫ g for 10 min. SARS-CoV-2 quantification and titer reduction evaluation. Virus present in treated and purified samples or in mock-treated and purified samples was quantified by either TCID 50 or plaque assay. As additional assay controls, unfiltered mock-treated sample was titrated to determine virus loss during filtration, and filtered test reagent-only (no virus) sample was titrated to determine residual test buffer cytotoxicity. For TCID 50 assays, neat to 10 Ϫ7 10-fold dilutions were prepared, and for plaque assays, neat to 10 Ϫ5 10-fold dilutions were prepared, both in MEM-5% FCS. TCID 50 titers were determined by the Spearman-Kärber method (22, 23) . When the levels of detected virus were so low that the TCID 50 could not be calculated by the Spearman-Kärber method, the TCID 50 was calculated by the Taylor method (24) . When no virus was detectable, values are given as less than or equal to the Taylor-derived TCID 50 titer given by a single virus-positive well at the lowest dilution at which no cytotoxicity was observed. Titer reduction was calculated by subtracting the mean logarithmic virus titer for test buffer-treated, purified conditions from the mean logarithmic virus titer for the PBS-treated, purified condition, with standard errors calculated according to Spearman (22) .

Serial passages of treated samples. In parallel to virus quantification, 12.5-cm 2 flasks of Vero E6 cells (6.25 ϫ 10 4 cells/flask in 2.5 ml of MEM-5% FBS) were inoculated with either 500 l or 50 l of treated filtered sample. Flasks were examined for cytopathic effect (CPE), and 500 l of culture medium from each flask was used to inoculate new 12.5-cm 2 flasks of Vero E6 cells after 7 days. If no CPE was observed, this process was continued for up to four serial passages. For the duration of the passage series, a flask of untreated cells was included as a control for cross-contamination between flasks, and a SARS-CoV-2-infected control was included to ensure suitable conditions for virus propagation. To distinguish CPE from any residual cytotoxicity associated with test reagents, samples of cell culture medium were taken from each flask at the beginning and end of each passage. Nucleic acid was extracted from cell culture medium manually using a QIAamp Viral RNA Minikit (Qiagen) or by using NucliSENS easyMAG or EMAG platforms (both, bioMérieux). Viral RNA levels were quantified by quantitative reverse transcriptase PCR (qRT-PCR) specific for the SARS-CoV-2 E gene (25) using TaqMan Fast 1-Step Master Mix (Applied Biosystems) on a 7500 Fast real-time PCR system (Applied Biosystems). A positive result for virus amplification was recorded if effects on the monolayer consistent with CPE and a decrease in the threshold cycle (C T ) value across the course of a passage were observed.

Reagent filtration optimization to minimize cytotoxicity and maximum virus recovery. Prior to evaluating the effectiveness of the reagents at inactivating SARS-CoV-2, we investigated the cytotoxicity of each reagent before and after filtration though one of five matrices: Sephadex LH-20, Sephacryl S400HR, Amicon Ultra 50-kDamolecular-weight-cutoff centrifugal filters, Pierce detergent removal spin columns (DRSC), and Bio-Beads SM2 nonpolar polystyrene adsorbents. Reagents were diluted with PBS to the working concentrations recommended by the manufacturer (for commercial sample transport and molecular extraction reagents) or to the concentrations indicated in the tables (for all other chemicals), followed by a single reagent removal step with each filtration matrix. Dilution series of filtered and unfiltered samples were generated to determine concentration-dependent cytotoxicity, from which the CC 20 value for each combination of reagent and filtration method was interpolated (see Fig. S1 in the supplemental material). The CC 20 was chosen as cells retain 80% viability at this concentration and enable distinction of active SARS-CoV-2 replication by visualization of CPE in the monolayer. Table 1 shows the dilution factor of reagent-treated sample required to achieve the CC 20 after filtration, with a value of Ͻ1 indicating complete removal of cytotoxicity. These data were used to determine the relative cytotoxicity removed by one filtration step for each combination of reagent and matrix (Fig. 1A ).

All unfiltered reagents tested here were cytotoxic, but the degree of cytotoxicity varied considerably as did the optimal filtration matrix for each reagent. The detergent Tween 20 used at a 1% concentration was the least cytotoxic unfiltered, requiring a dilution factor of only 7.7 to reach the CC 20 although only the Bio-Bead SM2 filters were effective at removing all cytotoxicity. The chemical fixative combination of 2% formaldehyde and 1.5% glutaraldehyde was the most cytotoxic unfiltered, requiring a dilution of over 4,000 to reach the CC 20 , with only the Amicon Ultra columns able to remove 100% of the cytotoxicity. However, for the majority reagents (27/34) tested, filtration through at least one matrix type removed 100% of cytotoxicity, allowing neat eluate to be used directly in cell culture without further dilution. There were several exceptions to this: DNA/RNA Shield (maximum 99.4% cytotoxicity removal using SM2), 40% GHCl (99.1% using Pierce DRSC), 4 M GITC (99.7% using Pierce DRSC), MagNA Pure (99.7% using SM2), AL buffer (87.4% using S400HR), Cobas Omni LYS (97.0% using SM2), and NeuMoDx (93.4% using S400HR). For these reagents, filtered eluate was still cytotoxic when used undiluted in cell culture. However, CC 20 values indicated that this remaining cytotoxicity would be removed by first or second (10 Ϫ1 to 10 Ϫ2 ) dilutions in the TCID 50 assay, allowing evaluation of titer reduction using these reagents, with the caveat that the effective assay limit of detection (LOD) would be higher. Passing treated samples through more than one column or increasing the depth of the resin/bead bed within the spin column can also improve cytotoxicity removal for some reagents (unpublished data).

In addition to cytotoxicity removal, a successful filtration method must also purify virus without adversely affecting titer or integrity. We therefore assessed SARS-CoV-2 recovery after each filtration method. Using an input titer of 1.35 ϫ 10 6 TCID 50 /ml, triplicate purifications of virus through Sephadex LH-20 or Pierce detergent removal spin columns resulted in recovery of 100% of input virus (Fig. 1B) . In contrast, the recoverable titer after one filtration through Amicon Ultra filters was 2.13 ϫ 10 5 TCID 50 / ml, an 84.5% reduction from input. Purification with S400HR and Bio-Beads SM2 matrices resulted in recoverable titers of 1.08 ϫ 10 6 TCID 50 /ml and 8.99 ϫ 10 5 TCID 50 / ml, losses of 20.1% and 33.6% of input virus, respectively.

SARS-CoV-2 inactivation by specimen transport and molecular extraction reagents. Specimen transport tubes are designed to inactivate microorganisms present in clinical specimens prior to sample transport while preserving the integrity of nucleic acids for molecular testing. If effective, these products have the potential to streamline SARS-CoV-2 diagnostic processing in testing laboratories by eliminating the requirement for CL3 processing or, for activities carried out at CL2, permitting processing outside an MSC. The BS EN 14476 standard requires demonstration of a Ͼ4-log 10 titer reduction for virucidal suspension tests (24) , and we were able to demonstrate a Ն4-log 10 TCID 50 titer reduction for all specimen transport media evaluated in a tissue culture fluid matrix (Table 2) . However, infectious virus remained recoverable in treated samples after inactivation with most reagents tested (by either TCID 50 assay or blind passage). The exceptions to this were PrimeStore MTM and 4 M GITC, from which no residual virus was detectable by either TCID 50 assay or by the passaging of treated purified sample. While several contact times were evaluated for all of these reagents, length of contact time had no effect on either the level of virus titer reduction or whether virus remained detectable upon passage.

We also sought to inform sample processing by examining inactivation by molecular extraction lysis buffers used in several manual and automated extraction protocols within SARS-CoV-2 diagnostic and research laboratories. We could demonstrate a Ն4-log 10 reduction in TCID 50 titer for all but two molecular extraction reagents when they were evaluated using tissue culture fluid ( Table 3) . The exceptions to this were AL and Cobas Omni LYS, where remaining cytotoxicity in the filtered eluate increased the TCID 50 LOD to a level such that the maximum calculable titer reductions were Ն3.5 and Ն3.9 log 10 TCID 50 , respectively. However, given that no virus was detected at any passage, it is likely that infectious virus was effectively inactivated by these two reagents. For reagents tested with multiple contact times (NucliSENS and Panther Fusion), shorter times (10 min) were as effective at reducing virus titers as longer contact times. Most reagents reduced viral titers to around the TCID 50 assay LOD, indicating that any remaining virus posttreatment was present only at very low titers (Ͻ10 TCID 50 /ml), but higher levels of virus were recoverable from samples treated with some extraction buffers. For NeuMoDx lysis buffer, although titers were reduced by Ն4 log 10 TCID 50 , an average of 91 (Ϯ38) TCID 50 /ml remained detectable. Similarly, buffer AVL reduced virus titers by 5.1 log 10 TCID 50 , but after treatment virus was detectable in all treated sample replicates (average, 54 Ϯ 18 TCID 50 /ml). However, after addition of four sample volumes of absolute ethanol following a 10-min contact time with AVL (the next step in the Qiagen Viral RNA Minikit manual), a Ն5.9-log 10 titer reduction was recorded, with no virus recoverable following passages in cell culture.

Panther Fusion lysis buffer was further tested against a relevant clinical sample matrix, pooled fluid from oropharyngeal (OP) and nasopharyngeal (NP) swab specimens, resulting in a Ն5.1-log 10 titer reduction with no remaining infectious virus detectable. We additionally evaluated the tissue lysis buffer RLT using homogenized ferret lung as sample material, with treatment resulting in a Ն4.8-log 10 titer reduction with no residual infectious virus detectable.

Detergents can be used to inactivate lipid enveloped viruses such as coronaviruses by disrupting the viral envelope, thereby rendering them unable to attach or enter cells (26) (27) (28) (29) . Here, we evaluated Triton X-100, SDS, NP-40, and Tween 20 for their ability to inactivate SARS-CoV-2. SDS treatment at 0.1% or 0.5% reduced titers by Ն5.7 and Ն6.5 log 10 TCID 50 , respectively, while both concentrations of NP-40 reduced titers by Ն6.5 log 10 TCID 50 with no residual virus detectable following NP-40 treatment. In contrast, up to 0.5% Tween 20 had no effect on viral titers. Triton X-100 is commonly used in viral inactivation reagents, and here we show that at both 0.1% and 0.5% (vol/vol) concentrations, virus titers in tissue culture fluid were reduced by Ն4.9 log 10 TCID 50 , even with less than 2 min of contact time (Table 4 ). Furthermore, we were unable to recover infectious virus from samples treated with 0.5% Triton X-100 for 10 min or longer. We also saw effective inactivation of SARS-CoV-2 by SDS, NP-40, and Triton X-100 in spiked NP and OP swab specimen fluid, but, importantly, we were not able to replicate this in spiked serum; 1% Triton X-100 reduced titers in human serum by only a maximum of 2 log 10 TCID 50 with contact times of up to 2 h.

In addition to evaluating inactivation efficacy by detergents, we assessed the effects of treatment on RNA integrity to determine their suitability for inactivation prior to nucleic acid testing. Extracted RNA from treated samples was tested using a SARS-CoV-2-specific qRT-PCR, and the C T difference between detergent-treated samples and a TM, transport medium; MM, molecular transport medium; GHCl, guanidine hydrochloride; GITC, guanidinium isothiocyanate; Tx, Triton X-100. b All specimen transport media were evaluated in a tissue culture fluid matrix. c Samples were titrated by TCID 50 , with a limit of detection of 5 TCID 50 /ml (0.7 log 10 TCID 50 /ml) unless otherwise indicated as follows: *, limit of detection of 50 TCID 50 /ml (1.7 log 10 TCID 50 /ml) due to cytotoxicity in neat wells of TCID 50 assay; **, limit of detection of 504 TCID 50 /ml (2.7 log 10 TCID 50 /ml) due to cytotoxicity in neat and Ϫ1 wells of TCID 50 assay; †, titration by plaque assay with a limit of detection of 3.3 PFU/ml (0.5 log 10 PFU/ml). d Values in parentheses represent the number of positive samples/number of replicates.

mock-treated controls was determined (Table 4) . A time-dependent increase in C T value following treatment with 0.5% Triton X-100 was observed, indicating a detrimental effect on RNA stability with increasing treatment times. Treatment with NP-40 had a marked effect, with a 30-min treatment leading to an increase in 9 to 10 C T s. While we saw no increase in C T in tissue culture fluid samples treated with 0.5% SDS, we observed an increase in the C T value for SDS-treated swab fluid samples, likely due to an increased concentration of RNases in clinical samples.

Fixation and inactivation of viruses by the addition of formaldehyde or of a combination of formaldehyde and glutaraldehyde is a well-established protocol, particularly for diagnostic electron mi- Values represent AVL/virus/ethanol. c Samples were titrated by TCID 50 , with a limit of detection of 5 TCID 50 /ml (0.7 log 10 TCID 50 /ml) unless otherwise indicated as follows: *, limit of detection of 50 TCID 50 /ml (1.7 log 10 TCID 50 /ml) due to cytotoxicity in neat wells of TCID 50 assay; **, limit of detection of 504 TCID 50 /ml (2.7 log 10 TCID 50 /ml) due to cytotoxicity in neat and Ϫ1 wells of TCID 50 assay; †, titration by plaque assay with a limit of detection of 3.3 PFU/ml (0.5 log 10 PFU/ml). croscopy (30, 31) . A treatment with 4% or 2% formaldehyde for 15 or 60 min reduced virus titers by Ն4.8 log 10 TCID 50 when evaluated against a tissue culture fluid matrix, with no remaining infectious virus detectable (Table 5 ). When infected monolayers were subjected to the same treatment protocol, all titer reductions were Ն6.8 log 10 TCID 50 , with 60 min of contact time moderately more effective than 15 min. However, in this format, a 60-min 4% formaldehyde treatment was the only one from which no infectious virus was detectable. A mixture of 2% formaldehyde and 1.5% glutaraldehyde tested on infected monolayers reduced virus titers by Ն6.7 log 10 TCID 50 , with no remaining infectious virus detectable for both 15-and 60-min contact times. Polyhexanide biguanide (PHMB) is a polymer used as a disinfectant and antiseptic, evaluated here as a potential lysis buffer, but it was able to reduce viral titers by only 1.6 log 10 TCID 50 at the highest concentration tested (2%).

Samples containing infectious SARS-CoV-2 require an initial inactivation step in primary containment (e.g., in an MSC) before further processing; given the rapid emergence of SARS-CoV-2, these inactivation protocols have been guided by existing data for other coronaviruses, and there is an urgent need to both confirm these historical data using the new virus and to validate new approaches for inactivating SARS-CoV-2. We therefore analyzed numerous commercially and commonly available reagents used by public health agencies and research laboratories around the world in their response to the pandemic. In addition, to address challenges of reagent cytotoxicity in inactivation evaluation, we provide data on the effectiveness of filtration methods for removing cytotoxicity from chemically treated samples.

Knowledge of the expected amount of infectious virus in clinical samples obtained from COVID-19 patients is important when viral inactivation study data are applied to diagnostic sample processing, allowing end users to interpret whether the material they are handling is likely to represent an infectious risk to themselves and others. These values are dependent on several factors, including time post-symptom onset, duration of symptoms, time elapsed between sampling and testing, the presence of neutralizing antibody responses, and immunocompetency of the individual (32) . Data regarding quantitative infectious viral levels in typical clinical specimens are minimal, with most studies reporting viral loads as determined by qRT-PCR only (33) (34) (35) . However, one study investigating infectious titers in 90 qRT-PCR-positive NP or endo- tracheal samples from COVID-19 patients estimated a median titer of 3.3 log 10 TCID 50 /ml (32) . Although here we were able to demonstrate a Ͼ4-log 10 reduction in titer for all specimen transport reagents, the observation that virus could be recovered from most treated samples indicates that while these reagents can effectively reduce viral titers, they cannot be assumed to completely inactivate SARS-CoV-2 in clinical specimens.

Limited SARS-CoV-2 inactivation data on molecular extraction reagents used in nucleic acid detection assays are currently available. We demonstrate here that the majority of commonly used reagents evaluated were effective at reducing viral titers by more than 4 logs, with several treatments completely removing all infectivity. For two reagents, buffer AL and Cobas Omni LYS buffer, we were not able to show a Ͼ4-log reduction. However, this was due to an increase in the effective limit of detection in the TCID 50 assay as no purification system was able to remove all of the cytotoxicity. Given that no virus was detected in serial passage of the treated samples, it is probable that treatment with either of these buffers is effective at inactivating SARS-CoV-2. A previous study reported that buffer AVL either alone or in combination with ethanol was not effective at completely inactivating SARS-CoV-2 (17) . In contrast, we could not recover any infectious virus from samples treated with AVL plus ethanol, consistent with previous studies indicating that AVL and ethanol in combination is effective at inactivating MERS and other enveloped viruses (10, 36) and indicating that both AVL and ethanol steps of manual extraction procedures should be performed before removal of samples from primary containment for additional assurance. Our detergent inactivation data indicating that SDS, Triton X-100, and NP-40, but not Tween 20, can effectively inactivate SARS-CoV-2 both in tissue culture fluid and also in pooled NP and OP swab fluid, which more accurately mimic authentic clinical specimen types, corroborate findings of a recent study (19) . However, as has been demonstrated for other viruses (33), we observed an inhibitory effect of serum on virus inactivation by detergent, highlighting the importance of validating inactivation methods with different sample types.

Based on our findings comparing filtration matrices, we found that the optimum method for reagent removal for inactivation studies is determined by evaluating three factors: (i) effectiveness of cytotoxicity removal, (ii) efficiency of virus recovery, and (iii) the ease of performing these methods within a containment space. Methods permitting complete removal of cytotoxic reagent components with no or little effect on virus recovery give assurance that low levels of residual virus, if present, could be detected in virus inactivation studies. During reagent testing, there were several instances where we noted residual cytotoxicity in the neat eluate, which is contrary to what was expected based on the initial reagent removal data and is likely due to the extended incubation period required for inactivation testing (up to 7 days, compared with overnight for cytotoxicity evaluation). In all cases, however, we were still able to enhance the levels of titer reduction detectable compared with what would have been achieved by sample dilution alone.

In conclusion, we have evaluated methods for straightforward, rapid determination of purification options for reagents prior to inactivation testing, enabling establishment of effective methods for sample purification while minimizing virus loss. This is applicable to inactivation studies for all viruses (known and novel), not only SARS-CoV-2. We have applied these methods to obtain SARS-CoV-2 inactivation data for a wide range of reagents in use (or proposed for use) in SARS-CoV-2 diagnostic and research laboratories. In addition to guiding laboratory risk assessments, this information enables laboratories to assess alternative reagents that may be used for virus inactivation and nucleic acid extraction, particularly considering concerns about extraction reagent availability due to increased global demand caused by the COVID-19 pandemic. Furthermore, chemical treatments evaluated here are commonly used for inactivation of a wide range of different viruses and other pathogens, and the results presented may be used to directly inform and improve the design of future inactivation studies.

Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 1.6 MB.

The species Severe acute respiratory syndromerelated coronavirus: classifying 2019-nCoV and naming it SARS-CoV-2

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

A new coronavirus associated with human respiratory disease in China

COVID-19: guidance for sampling and for diagnostic laboratories

World Health Organization. 2020. Laboratory biosafety guidance related to the novel coronavirus

Interim laboratory biosafety guidelines for handling and processing specimens associated with coronavirus disease 2019 (COVID-19

SARS-coronavirus (SARS-CoV) and the safety of a solvent/detergent (S/D) treated immunoglobulin preparation

Efficacy of various disinfectants against SARS coronavirus

Heat inactivation of the Middle East respiratory syndrome coronavirus

Inactivation and safety testing of Middle East Respiratory Syndrome Coronavirus

Inactivation of the coronavirus that induces severe acute respiratory syndrome, SARS-CoV

Evaluation of inactivation methods for severe acute respiratory syndrome coronavirus in noncellular blood products

Inactivation of SARS coronavirus by means of povidone-iodine, physical conditions, and chemical reagents

Heat inactivation of different types of SARS-CoV-2 samples: what protocols for biosafety, molecular detection and serological diagnostics?

Effective heat inactivation of SARS-CoV-2

Efficacy of a novel iodine complex solution, CupriDyne, in inactivating SARS-CoV-2

Evaluation of chemical protocols for inactivating SARS-CoV-2 infectious samples

Rapid in-vitro inactivation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) using povidone-iodine oral antiseptic rinse

Methods of inactivation of SARS-CoV-2 for downstream biological assays

Practical guidelines for collection, manipulation and inactivation of SARS-CoV-2 and COVID-19 clinical specimens

Propagation, inactivation, and safety testing of SARS-CoV-2

The method of right and wrong cases (constant stimuli) without Gauss's formulae

Beitrag zur kollektiven Behandlung pharmakologischer Reihenversuche

Chemical disinfectants and antiseptics. Quantitative suspension test for the evaluation of virucidal activity in the medical area

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR

Diagnostic performances of clinical laboratory tests using Triton X-100 to reduce the biohazard associated with routine testing of Ebola virus-infected patients

Inactivation of human immunodeficiency virus type 1 by nonoxynol-9, C31G, or an alkyl sulfate, sodium dodecyl sulfate

Effective chemical virus inactivation of patient serum compatible with accurate serodiagnosis of infections

Inactivation of human and avian influenza viruses by potassium oleate of natural soap component through exothermic interaction

Electron microscopy for rapid diagnosis of infectious agents in emergent situations

Evaluation of virus inactivation by formaldehyde to enhance biosafety of diagnostic electron microscopy

Predicting infectious SARS-CoV-2 from diagnostic samples

Ebola virus inactivation by detergents is annulled in serum

Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study

Virological assessment of hospitalized patients with COVID-2019

Buffer AVL alone does not inactivate Ebola virus in a representative clinical sample type

We thank the Respiratory Virus Unit at PHE Colindale and the Virology Laboratories at PHE Cambridge and PHE Bristol for donation of pooled respiratory samples, Julia Tree at PHE Porton for donation of lung tissue, and Ayoub Saei at the Statistics Unit, PHE Colindale, for statistical advice.