key: cord-0826379-azc95rbs
authors: Berg, Dr. Michael G; Zhen, Wei; Lucic, Danijela; Degli-Angeli, Emily J; Anderson, Mark; Forberg, Kenneth; Olivo, Ana; Sheikh, Farah; Toolsie, Dan; Greninger, Alexander L; Cloherty, Gavin A; Coombs, Robert W; Berry, Gregory J
title: Development of the RealTime SARS-CoV-2 quantitative Laboratory Developed Test and correlation with viral culture as a measure of infectivity
date: 2021-08-17
journal: J Clin Virol
DOI: 10.1016/j.jcv.2021.104945
sha: 629eb94f0b4e6c649fb3881c466d07cbecb49a94
doc_id: 826379
cord_uid: azc95rbs

While diagnosis of COVID-19 relies on qualitative molecular testing for the absence or presence of SARS-CoV-2 RNA, quantitative viral load determination for SARS-CoV-2 has many potential applications in antiviral therapy and vaccine trials as well as implications for public health and quarantine guidance. To date, no quantitative SARS-CoV-2 viral load tests have been authorized for clinical use by the FDA. In this study, we modified the FDA emergency use authorized qualitative RealTime SARS-CoV-2 assay into a quantitative SARS-CoV-2 Laboratory Developed Test (LDT) using newly developed Abbott SARS-CoV-2 calibration standards. Both analytical and clinical performance of this SARS-CoV-2 quantitative LDT was evaluated using nasopharyngeal swabs (NPS). We further assessed the correlation between Ct and the ability to culture virus on Vero CCL81 cells. The SARS-CoV-2 quantitative LDT demonstrated high linearity with R(2) value of 0.992, high inter- and intra-assay reproducibility across the dynamic range (SDs ±0.08-0.14 log(10) copies/mL for inter-assay reproducibility and ±0.09 to 0.19 log(10) copies/mL for intra-assay reproducibility). Lower limit of detection was determined as 1.90 log(10) copies/mL. The highest Ct at which CPE was detected ranged between 28.21-28.49, corresponding to approximately 4.2 log(10) copies/mL. Quantitative tests, validated against viral culture capacity, may allow more accurate identification of individuals with and without infectious viral shedding from the respiratory tract.

The diagnosis of COVID-19 relies primarily on nucleic acid amplification tests for SARS-CoV-2 RNA targets. To date, these tests have been authorized by the FDA as qualitative tests, reporting the presence or absence of viral RNA in a clinical specimen. Quantitative assays of SARS-CoV-2 viral load (RNA copies/mL) have clear applications in preclinical and clinical trials of antiviral therapies and may have public health implications for contact tracing and quarantine guidance for individual patients. However, while numerous studies have reported the dynamics of SARS-CoV-2 Ct following infection using PCR-based assays, the relationship between viral load and infectivity, symptom severity and mortality remains unclear (1, 2) .

For PCR-based tests, the cycle threshold (Ct), defined as the number of PCR cycles needed to amplify the target viral RNA so that it can be detected over background, is inversely correlated with viral load. Due to different assay design strategies used by manufacturers, there are concerns about accurately comparing Ct values across qualitative assays (3) (4) (5) . Once there is a universal standard for SARS-CoV-2 such as a World Health Organization International Standard, viral load could be estimated based on the Ct value from the FDA Emergency Use Authorization (EUA) SARS-CoV-2 assays. Recently, several studies have shown an inverse correlation between the Ct and the ability to culture SARS-CoV-2 in vitro as a measure of infectivity (6) (7) (8) .

Here, we describe the modification of the EUA-approved qualitative RealTime SARS-CoV-2 assay (Abbott Molecular, Des Plaines, IL) (9, 10) into a SARS-CoV-2 quantitative LDT using Abbott SARS-CoV-2 calibration standards that correlate Ct and viral load. The analytical performance of the SARS-CoV-2 quantitative LDT was evaluated using commercially available 5 SARS-CoV-2 material or using the Abbott SARS-CoV-2 material. Analytical analysis consisted of linearity, limit of detection, the inter-run and intra-run reproducibility. Using nasopharyngeal swab samples, the clinical performance was compared to three EUA SARS-CoV-2 qualitative assays which also included inter-laboratory agreement of the SARS-CoV-2 quantitative LDT.

Using measurements from the SARS-CoV-2 quantitative LDT, we also assessed the correlation between Ct (viral load) and infectivity in clinical samples based on cytopathic effects (CPE) observed in Vero cells grown in culture.

Open mode functionality on m2000sp/rt system was utilized to develop this SARS-CoV-2 quantitative LDT by using EUA Abbott SARS-CoV-2 qualitative reagents. Identical extraction and amplification/detection protocols developed for the RealTime SARS-CoV-2 qualitative EUA assay were also used for the development of the RealTime SARS-CoV-2 quantitative LDT.

Specifically, 500 µl of sample was used for sample extraction, viral nucleic acid bound to the microparticles was eluted with 90 µl of elution buffer and 40 µl of the eluate was used for the amplification and detection reaction. The RealTime SARS-CoV-2 qualitative EUA assay and the RealTime SARS-CoV-2 quantitative LDT both utilize 10 unread cycles as part of their amplification and detection. In this assay, two calibrator levels (3 log 10 RNA copies/mL and 6 log 10 RNA copies/mL) tested in triplicate were used to generate a calibration curve and three control levels (negative, low positive at 3 log 10 RNA copies/mL and high positive at 5 log 10 RNA copies/mL) were included in each run for quality management. 6 

A calibrator used for SARS-CoV-2 quantitative LDT was derived from virus propagated in culture on Vero cells (ATCC CCL-81). Briefly, nasopharyngeal (NP) swabs (New York Biologics, NY) containing viable SARS-CoV-2 were eluted into viral transport media (VTM, Becton Dickenson) and three specimens with the lowest Ct values were selected for lot production. Initially, 100 µL of VTM diluted in 1.9 mL of fresh MEM media (10-009-CV) without FBS was overlaid on a 10-cm plate with 2 x10 6 cells and incubated for 2 hours with rocking every 30 min to keep the monolayer from drying. The inoculum was removed by aspiration, monolayers were washed with 5 mL of 1X PBS, and 10 mL of fresh complete media containing 10% FBS was added. Cytopathic effects (CPE) developed within 4-5 days and both cells and supernatant were harvested. Primary lysates were freeze/thawed once and used to repeat infections on ten 10-cm plates, expanding each sample to 100 mL of virus lysate.

In order to assess viral stock concentration, heat-inactivated viral stock was tested in triplicate with the RealTime SARS CoV-2 qualitative assay. Quantitation was determined by plotting the Verification Panel. The above dilution process was also followed for the preparation of SARS-CoV-2 high positive control.

Linearity of the RealTime SARS-CoV-2 quantitative LDT was evaluated with the BEI SARS-CoV-2 viral stock (NR-52281; USA-WA1/2020), SeraCare AccuPlex SARS-CoV-2 Verification panel, and the Abbott SARS-CoV-2 material. The copy number of the BEI stock (lot #70033175) was 2.07x10 9 genome equivalents/mL determined by the BioRad QX200 Droplet Digital PCR (ddPCR™) System. For the BEI and the Abbott SARS-CoV-2 material, an intermediate concentration targeting 8 log 10 copies/mL was diluted using RNA Storage solution. This material was then serially diluted to 1.7 log 10 copies/mL and tested in 2 to 6 replicates at each dilution on different days. SeraCare verification panel was tested neat along with one additional dilution level at 2 log 10 RNA copies/mL over three days.

The limit of detection (LOD) was determined by testing serial dilutions of the Abbott SARS-CoV-2 material on different days. Inactivated SARS-CoV-2 whole genome virus was diluted in the RNA Storage solution and used to prepare serial dilutions in Log 10 RNA copies/mL: 5.00, 4.00, 3.00, 2.70, 2.40, 2.00, and 1.70 and replicates ranging from 6 to 22 at each dilution. LOD was defined as the concentration of the lowest dilution that could be detected with >95% probability. 8 Inter-and intra-run reproducibility was assessed with the Abbott SARS-CoV-2 material diluted to two different viral load concentrations near the LOD (2.40 and 2.70 log 10 copies/mL) and another at a higher viral load (5.0 Log 10 copies/mL). The prepared dilutions were tested with a minimum of three replicates across three different days. quantitative LDT was evaluated using 89 remnant clinical specimens (26 quantifiable specimens, 3 specimens <2.0 log 10 copies/mL, 35 negative specimens and 25 specimens which were greater than the upper limit of quantitation (7.0 log 10 copies/mL)) collected after standard of care SARS-CoV-2 testing on the above platforms. All samples from the University of Washington were deidentified, excess clinical material and were deemed to be non-human subjects by the respective Institutional Review Board. Northwell Health Laboratories study was approved by the respective Institutional Review Board (IRB number 21-0284).

De-identified frozen NP swabs (n=459) in VTM were sourced from multiple hospital systems across the United States (Supplementary Table S1 ). A 1 mL aliquot of VTM was heat-9 inactivated at 65°C for 30 min and then processed with the RealTime SARS-CoV-2 quantitative LDT. The first of 5 consecutive runs included Abbott SARS-CoV-2 standards in triplicate. Noninactivated specimens were kept at 4 o C to avoid a freeze/thaw cycle until culturing of selected positive samples was performed later the same day. Duplicate wells of Vero cells (3x10 5 cells/100 µl) were co-plated with 100 µl of patient VTM overnight and cultured for ≥96 hours.

CPE was determined by microscopy and a corresponding decrease in Viral Tox Glo luminescence (Cat# G8941, Promega, Madison WI).

Linear regression was used to evaluate assay's linearity, Ct method and inter-lab assay comparison. LOD and reproducibility was assessed by evaluating mean and standard deviation.

(Microsoft Office Excel 365 software, Microsoft, Redmond, WA).

LDT functionality (open mode on m2000sp/rt) was utilized to develop the modified RealTime SARS-CoV-2 quantitative LDT. This modified LDT is an automated and quantitative real-time PCR assay that uses extraction and amplification/detection protocols developed for the RealTime SARS-CoV-2 qualitative EUA assay (9, 10). Data reduction parameters were updated to include three control levels (one high positive SARS-CoV-2 control, one low positive SARS-CoV-2 control and one negative control) that were used for quality management of the assay, and two calibration levels (3.00 log 10 copies/ml and 6.00 log 10 copies/ml) that were used to generate calibration curves for the quantitative assay.

Excellent assay linearity was demonstrated across BEI SARS-CoV-2 viral stock (R 2 = 0.9985; Figure 1A) , Abbott SARS-CoV-2 material (R 2 = 0.9992; Figure 1B) and SeraCare Verification Panel (R 2 > 0.99; Figure 1C) . These results demonstrated that the RealTime SARS-CoV-2 quantitative LDT can produce an accurate quantification value across the range of 1.7 to 8.0 log 10 copies/ml. Using the dilution series of the Abbott SARS-CoV-2 material, the LOD was determined to be 1.90 log 10 copies/mL (79 copies/mL in NPS) having a 100% detection rate with a SD±0.23 ( Table 1) . High inter-and intra-assay reproducibility was seen across the dynamic range ( Table 2 ) with the SD of 0.14 at the lowest concentration tested (2.4 log 10 copies/mL) and 0.08 at the highest concentration tested (5.0 log 10 copies/mL) in the inter-run analysis. We observed an SD of 0.19 and 0.09 at the same concentrations in the intra-run analysis.

Ct's obtained with the RealTime SARS-CoV-2 quantitative LDT correlated well with those obtained by the EUA CDC 2019-nCoV assay (R 2 = 0.9444; Figure 2a) , cobas SARS-CoV-2 assay (R 2 = 0.9495; Figure 2b ) and Alinity m SARS-CoV-2 assay (R 2 = 0.9584; Figure 2c ).

Inter-lab comparison of the RealTime SARS-CoV-2 quantitative LDT demonstrated excellent correlation (R 2 =0.955; Figure 2d ) with a mean bias of 0.29 log 10 RNA copies/mL (data not shown). 11 Numerous groups have sought to determine the Ct or viral load at which SARS-CoV-2 is transmissible (6-8, 12-17). One approximation is the ability to culture virus in vitro. A pilot experiment was performed with Vero CCL81 cells in 12 well plates using 36 nasopharyngeal swabs in VTM spanning a range of Cts (7.5-32) previously determined by the m2000 RealTime SARS-CoV-2 Qualitative assay (9, 10). CPE readily developed within 2-4 days as assessed by microscopy, however, only in those specimens with the lowest Cts (Figure 3a) . Non-quantitative results indicated a cutoff of 18.27, or 28.27 when adjusting for the 10 unread cycles and qPCR/culture input volumes. SARS-CoV-2 culture was then adapted to a 96-well format to enable a quantitative measure of CPE using the Viral Tox Glo system (Figure 3b) . We validated the assay by plating serial dilutions of the high titer calibrators in quadruplicate. After 96 hours in culture, CPE was evaluated by both microscopy and luminescence, with a decrease in RLUs corresponding to the cell death observed visually. Co-plating of trypsinized cells with VTM at 1:1 v/v (Method 1) resulted in greater reproducibility and sensitivity compared to 2-hr infections (Method 2) with cells seeded the night before. For Method 1, an apparent cut-off of 62,500 copies/mL (4.8 log 10 copies/mL) was determined, wherein 2/4 replicates were infected for both calibrators, although CPE could still be detected in 1 of 4 replicates at 15,625 copies/mL (4.2 log 10 copies/mL), which equates to a Ct of 18.5 (28.5 without unread cycles; Figure 3b , left panel). By contrast, a cut-off of 125,000 copies/ml (5.1 log 10 copies/ml) at Ct = 15.65 (25.65) was observed for Method 2, with no dilutions lower than this showing evidence of successful infection (Figure 3b, right panel) . Note that large error bars near these cutoffs indicate variability in replicates, wherein high RLU values (uninfected) are averaged with low values (infected).

Using the in vitro culture method and NP clinical specimens (n=459) sourced from multiple hospital systems across the United States (Supplementary Table S1) , we explored the relationship between Ct, viral copy number/mL, and infectivity. An aliquot of each VTM was heat-inactivated, extracted, and measured using the newly developed SARS-CoV-2 RealTime quantitative LDT. Five successive experiments of 93 samples with calibrators and controls were run per day and positive specimens were infected in duplicate on Vero cells that evening. A total of 51 positive specimens were identified (11.1%) by qPCR, the majority (>75%) having a Ct > Table S1 ). Only 9 specimens (17.6%) induced CPE: 8/9 had Cts ≤ 12.4 (22.4) , corresponding to ≥ 6.2 log 10 copies/mL or >1,000,000 copies/mL, and 1/9 had a Ct = 18.21 (28.21) , corresponding to 4.31 log 10 copies/mL, or 20,417 copies/mL (Figure 3c) . These results mirrored our prior experimental data and confirmed that a Ct ≤ 28.5 with a titer ≥ 4.2 log 10 RNA copies/mL (>16,000 cp/ml) is required for successful Vero cell culture of SARS-CoV-2.

Using Abbott SARS-CoV-2 calibrators to correlate viral load with Ct, we calibrated and modified the existing RealTime SARS-CoV-2 EUA assay into a quantitative measure of viral load. The SARS-CoV-2 quantitative LDT demonstrated broad assay linearity, reproducibility across the dynamic range, and an LOD of approximately 1.9 log 10 copies/mL (79 copies/mL) at a Ct of 27 (37 without unread cycles). This study demonstrated that Ct values had a high degree of correlation (R 2 >0.94) between different SARS-CoV-2 EUA assays. Inter-lab comparison of the RealTime SARS-CoV-2 quantitative LDT demonstrated excellent correlation (R 2 =0.955) with a mean bias of 0.29 log 10 RNA copies/mL. Additionally, quantitation of samples between 7-8 13 log 10 RNA copies/mL (n=13) demonstrated good correlation with a mean bias of 0.07 log 10 RNA copies/mL, thus suggesting that the dynamic range could be expanded to 8 log 10 RNA copies/mL. This change may have significant workflow and turnaround time impact for the laboratory as fewer dilutions and repeat testing would be required prior to result reporting. Viral culture was performed as a final step to correlate the viral load measured with the quantitative assay in clinical samples and potential infectivity, measured as cytopathic effect in culture. We conducted three separate culture studies that achieved consistent results.

Accounting for unread cycles (referred to as dark cycles), the range of Ct at which CPE was detected was 28.21-28.49 corresponding to approximately 16,000 RNA copies/mL (4.2 log 10 copies/mL). While a culture cutoff of Ct = 33 (500 copies/mL, 2.7 log 10 copies/mL) has been reported by others, our data consistently shows that titers more than 30 times greater (4.5 Cts) than this are required for culturing SARS-CoV-2 (14) . Numerous reports in the literature corroborate our findings that viral replication is seen from samples with Ct ≤28 (18, 19) . In most reports, no replication was observed from samples above Ct > 30 (range of Ct 24 -34) (6, 7, 14, 20, 21) . Of note, Ct values can be highly platform-and assay-specific, which may account for the large ranges of values seen across studies. The number of freeze-thaw cycles, how samples were sourced, and differences in cell lines used can also play a role in recovery. Here, we infected Vero CCL81 cells, whereas Vero E6 or Vero E6 lines over-expressing TMPRSS2 or ACE2 are considered more sensitive (22, 23) . Some studies have reported a small percentage of samples with Ct ≥ 30 are able to be cultured, though the probability drops with each additional Ct (13) (14) (15) 17) . Other studies show a complete inability to culture virus (12, 24, 25) . Some of the variability may be due to the ratio of genomic to sub-genomic RNA being measured (26) . 14 Our finding that a large viral load is needed to culture SARS-CoV-2 suggests that the ability to culture virus may be predictive of transmissibility. RT-PCR is extremely sensitive and can detect low levels of RNA shed persistently during a period when individuals may no longer be infectious (27) . Studies have reported no symptomatic COVID-19 cases with viral loads below 4 log 10 copies/mL (8, (28) (29) (30) . Indeed, the presence of viral RNA by itself is not proof of infectivity, which can only be determined by clinical investigation outside the lab, but a Ct > 30 found in air and surface samples has been associated with non-infectious samples (25) . Thus, a qualitative approach to interpreting RT-PCR assays will likely report as positive individuals with low levels of virus who may no longer be infectious (31) . Quantitative assays, validated against viral culture capacity, may allow more accurate identification of individuals with and without infectious viremia. It is conceivable that mutations that increase transmissibility will influence this value (32) (33) (34) . Nevertheless, given the close correspondence of rapid antigen test limits of detection (~40,000 copies/swab; Ct ~30) with what can be successfully cultured, this would argue that a cheaper, less sensitive test may be acceptable from a public health perspective to increase testing and reduce transmission (35, 36) .

As SARS-CoV-2 infection typically resolves in weeks, the application of a quantitative assay is limited in terms of treatment decision-making, though it may have major public health implications for decisions about patient quarantine/isolation time. Quantification of SARS-CoV-2 may also be useful for therapeutic clinical trials and vaccine development, and in screening efforts to direct contact tracing. Future research is needed to determine whether viral load is a clinical measure of disease severity. A total of 9/51 (11.1%) demonstrated CPE, the majority having a Ct<12 (>1,000,000 copies/mL). The lowest titer demonstrating CPE in culture had a Ct=18.21 (28.21 w/o dark cycles), corresponding to 4.31 log 10 copies/mL. 

SARS-CoV-2 detection, viral load and infectivity over the course of an infection

SARS-CoV-2 viral load on admission is associated with 30-day mortality

College of American Pathologists (CAP) Microbiology Committee Perspective: Caution must be used in interpreting the Cycle Threshold (Ct) value

Comparison of commercially available and laboratory-developed assays for in vitro detection of SARS-CoV-2 in clinical laboratories

Comparative performance of SARS-CoV-2 detection assays using seven different primer-probe sets and one assay kit

Predicting infectious SARS-CoV-2 from diagnostic samples

Repeat COVID-19 molecular testing: correlation of SARS-CoV-2 culture with molecular assays and cycle thresholds

Culture-based virus isolation to evaluate potential infectivity of clinical specimens tested for COVID-19

Abbott RealTime SARS-CoV-2 Assay

Validation and verification of the Abbott RealTime SARS-CoV-2 assay analytical and clinical performance

CDC 2019-Novel Coronavirus (2019-nCoV) Real-Time RT-PCR Diagnostic Panel (EUA) Instructions for Use

SARS-CoV2 RNA detected in blood samples from patients with COVID-19 is not associated with infectious virus

Cell-based culture of SARS-CoV-2 informs infectivity and safe de-isolation assessments during COVID-19

Viral RNA load as determined by cell culture as a management tool for discharge of SARS-CoV-2 patients from infectious disease wards

Regular mass screening for SARS-CoV-2 infection in care homes already affected by COVID-19 outbreaks: Implications of false positive test results

Investigation of SARS-CoV-2 outbreaks in six care homes in London

Duration of infectiousness and correlation with RT-PCR cycle threshold values in cases of COVID-19

Culture-competent SARS-CoV-2 in nasopharynx of symptomatic neonates, children, and adolescents

SARS-CoV-2 virus culture and subgenomic RNA for respiratory specimens from patients with mild coronavirus disease

Virological assessment of hospitalized patients with COVID-2019

Singapore Novel Coronavirus Outbreak Research t. 2020. Viral dynamics and immune correlates of COVID-19 disease severity

Severe Acute Respiratory Syndrome Coronavirus 2 from patient with coronavirus disease

Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells

2020. Clinical, immunological and virological characterization of COVID-19 patients that test repositive for SARS-CoV-2 by RT-PCR

Investigating SARS-CoV-2 surface and air contamination in an acute healthcare setting during the peak of the COVID-19 pandemic in London

SARS-CoV-2 genomic and subgenomic RNAs in diagnostic samples are not an indicator of active replication

virus during an outbreak of SARS-CoV-2 in an inpatient geriatric psychiatry unit in King County, Washington

Presymptomatic SARS-CoV-2 infections and transmission in a skilled nursing facility

Asymptomatic and pre-symptomatic transmission of SARS-CoV-2: a systematic review

Temporal dynamics in viral shedding and transmissibility of COVID-19

Viral cultures for COVID-19 infectivity assessment. Systematic review

Resurgence of COVID-19 in Manaus, Brazil, despite high seroprevalence

Emergence and rapid spread of a new severe acute respiratory syndrome-related coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in South Africa

Transmission of SARS-CoV-2 Lineage B.1.1.7 in England: Insights from linking epidemiological and genetic data

Antigen-based testing but not real-time PCR correlates with SARS-CoV-2 virus culture

Analytical sensitivity of the Abbott BinaxNOW COVID-19 Ag Card

We thank Erin Quaco and Christopher Lark of Abbott Laboratories for technical assistance and Meei-Li Huang PhD for providing de-identified clinical specimens for the University of Washington studies. RWC is financially supported by the following grants: UM1-AI-106701; UM1-AI-068618; UW CFAR P30-AI-027757. ALG reports contract funding from Abbott and Gilead for testing, research funding from Merck, all outside of the submitted work.