key: cord-0747020-lrvwkosk
authors: Humbert, Maria Victoria; Opurum, Precious Chinonyerem; Brendish, Nathan J; Poole, Stephen; He, Peijun; Katis, Ioannis; Quaye, Jerry; Bediako, Yaw; Duriez, Patrick Jacques; Eason, Robert W; Sones, Collin; Quaye, Osbourne; Awandare, Gordon A; Christodoulides, Myron; Clark, Tristan W; Quashie, Peter K; McCormick, Christopher J
title: A SARS-CoV-2 nucleocapsid ELISA represents a low-cost alternative to lateral flow testing for community screening in LMI countries
date: 2021-10-01
journal: J Infect
DOI: 10.1016/j.jinf.2021.08.049
sha: 51f35385bd35e03139ab4911640a1b671fc2e8cb
doc_id: 747020
cord_uid: lrvwkosk

Background Controlling the spread of SARS-CoV-2 is problematic because of transmission driven by asymptomatic and pre-symptomatic individuals. Community screening can help identify these individuals but is often too expensive for countries with limited health care resources. Low-cost ELISA assays may address this problem, but their use has not yet been widely reported. Methods We developed a SARS-CoV-2 nucleocapsid ELISA and assessed its diagnostic performance on nose and throat swab samples from UK hospitalised patients and sputum samples from patients in Ghana. Results The ELISA had a limit of detection of 8.4 pg/ml antigen and 16 pfu/ml virus. When tested on UK samples (128 positive and 10 negative patients), sensitivity was 58.6% (49.6–67.2) rising to 78.3% (66.7–87.3) if real-time PCR Ct values > 30 were excluded, while specificity was 100% (69.2–100). In a second trial using the Ghanaian samples (121 positive, 96 negative), sensitivity was 52% (42.8–61.2) rising to 72.6% (61.8–81.2) when a > 30 Ct cut-off was applied, while specificity was 100% (96.2–100). Conclusions: Our data show that nucleocapsid ELISAs can test a variety of patient sample types while achieving levels of sensitivity and specificity required for effective community screening. Further investigations into the opportunities that this provides are warranted.

Coronavirus disease 2019 (COVID-19) is an acute life-threatening disease that can leave survivors with long-term sequelae [1] . COVID-19 deaths stand at over 3.7 million [2] , and although there now are effective vaccines to prevent disease [3] [4] [5] , it will likely take several years for them to be administered on a large scale in economically disadvantaged parts of the world. In the interim other measures will be required to control transmission, particularly in vulnerable settings such as hospitals.

Testing plays an important role in managing disease outbreaks, particularly in the case of SARS-CoV-2, where both pre-symptomatic and asymptomatic individuals are able to transmit virus to others [6] . Deciding which test to use in any particular setting depends on a trade-off between test sensitivity, turn-around times, logistic requirements and costs. The gold standard assay for detecting SARS-CoV-2 is the polymerase chain reaction (PCR); a nucleic acid-based test requiring expensive infrastructure and reagents. Loop-mediated isothermal amplification (LAMP) also detects nucleic acid and benefits from reduced infrastructure and reagent costs [7, 8] . However, reagent and infrastructure costs associated with LAMP are still significant, in particular those associated with RNA purification -a requirement for maximum sensitivity to be achieved [7] .

In comparison to nucleic acid test modalities, tests detecting abundant viral protein antigens such as nucleocapsid (Nc) typically have lower infrastructure and reagent costs. Several commercial lateral flow device (LFD) tests detecting SARS-CoV-2 Nc already exist [9] . With short turnaround times, LFDs offer the advantage of point-of-care testing. They are also relatively cheap to run as there is no need for lab and/or additional equipment to process tests and the more basic ones do not require a reader. However, LFDs suffer from moderate to low sensitivity compared to PCR [10] [11] [12] [13] [14] . Also, while those LFD tests that do not require additional instrumentation are cheaper to perform than nucleic acid tests, there remains uncertainty as to their exact cost for low and middle-income countries (LMICs). Gates Foundation support has allowed commercial LFDs to be provided at no more than $5/test [15] . However, if this upper limit cost is close to the actual cost the limited health care budgets in LMICs will still be stretched, particularly as pharma companies prefer to use thirdparty agents for distribution in LMICs, driving up costs as much as 5-10 fold.

Enzyme linked immunoassays (ELISAs) are another means by which Nc antigen can be detected. While ELISAs are slower to generate test results, they are typically able to detect lower levels of antigen compared to LFDs. Also while ELISAs do require lab processing, the laboratory infrastructure needed for this is minimal, with reading of the plates even being possible using mobile phone apps [16] . Importantly, ELISAs can be manufactured at a much lower unit cost/test than LFDs, making their use in certain LMIC settings attractive. In this study, we have developed an in-house ELISA and tested it using a variety of upper and lower respiratory tract samples from different population settings in both the UK and Ghana.

The gene sequences for SARS-CoV-2 nucleocapsid phosphoprotein (Nc, Gene ID: 43740575), HKU1 Nc (Gene ID: 3200423), OC43 Nc (Gene ID: 39105221), 229E Nc (Gene ID: 918763) and NL63 Nc (Gene ID: 2943504) were purchased as codon-optimized DNAs for expression in E. coli (GeneArt -ThermoFisher Scientific). For details of their cloning and expression see the methods section in Supplementary materials.

Recombinant SARS-CoV-2 Nc protein was used to immunize 2 rabbits and the polyclonal antibody (pAb) from each resultant sera was affinity-purified against the immunizing antigen (Eurogentec; Speedy Rabbit immunization programme). Pooled pAb was biotinylated with a 25 molar excess of N-hydroxysuccinimide (NHS)-4 x polyethylene glycol (PEG4)-Biotin (Sigma) for 2 hours at 4°C, dialysed overnight at 4°C against PBS and stored as aliquots at -20°C until used.

MAXISORP NUNC-IMMUNO plates were coated with 50 μl/well anti-SARS-CoV-2 pAb at 1.5 µg/ml overnight at 4°C. Plates were washed 5 x 5 min with Tris-buffered saline supplemented with 0.05% Tween-20 (TBS-T, pH 7.2), before blocking with 300 μl/well of 0.1 M NaH 2 PO 4 /Na 2 HPO 4 (pH 7.2), 0.15 M NaCl, 0.05% (v/v) Tween-20 and 1% (w/v) bovine serum albumin (BSA) for 1 hr at room temp. The blocking buffer was removed and 50 μl/well sample added. Each plate also included a series of dilutions of recombinant SARS-CoV-2

Nc as an internal control. Antigen incubation was performed at RT for 1 hr. The plates were then washed 5 x 5 min with TBS-T before addition of 1.25 μg/ml (50 μl/well) of biotinylated, affinity-purified rabbit α-SARS-CoV-2 IgG, diluted in 0.1 M NaH 2 PO 4 /Na 2 HPO 4 buffer (pH 7.2), 0.15 M NaCl, 0.05% (v/v) Tween-20 and 10% (v/v) rabbit serum (Sigma). Plates were incubated for 1 hr at room temp and washed again 5 x 5 min with TBS-T. Horseradish peroxidase (HRP) -Avidin conjugate (BioLegend) was diluted 1:2,500 in blocking buffer and added at 50 μl/well. After 1 hr incubation at RT, plates were washed 6 x 5 min with TBS-T before addition of 50 μl/well of 3,3',5,5'-tetramethylbenzidine substrate solution (SeraCare), which was left to develop for 1 hr at room temperature. Reactions were stopped with 50 μl/well of 1 M H 2 SO 4 solution and a plate reader used to measure absorbance at 450 nm. Samples obtained through genomic surveillance were stored upon collection (and after a positive-PCR result) at -80ºC or -20ºC depending on the facility. They were transported to WACCBIP in dry ice and either transferred to a -20ºC freezer or processed immediately for the ELISA. Samples collected by WACCBIP were either processed immediately or stored at

Respiratory samples were processed into a viral lysis buffer (VLB) comprising 1% (v/v) Triton X-100 / 0.3% (v/v) tri butyl phosphate, 1X Complete EDTA-free protease inhibitor and 5 mM EDTA in Dulbecco Phosphate Buffered Saline (PBS) before use. Briefly, dry nasal swabs were immersed in VLB (0.4ml and 1.0ml for UK and Ghana samples respectively) for 15 minutes. Wet oropharyngeal swabs stored in viral transport medium (VTM) were supplemented with 0.1 volumes of 11X VLB. Sputum samples were initially mixed with 1mL of PBS to reduce the viscosity and 300 µL of this solution supplemented with 30µl volumes of 11X VLB. Sample processing of nasal lavage was similar to VTM from wet swabs. All samples processed for ELISA were either used immediately or stored frozen until needed.

The Southampton clinical arm of the study was approved by the South Central -Hampshire 

Recombinant SARS-CoV-2 Nc expressed in E.coli was used to raise antisera in New Zealand rabbits. The resultant polyclonal antibody (pAb) displayed strong reactivity to SARS-CoV-2 Nc by indirect ELISA, producing a robust signal that reached saturation at concentrations ~1 µg/ml (Fig 1a) . Western blot showed pAb recognised SARS-CoV-2 Nc but did not cross-react with recombinant Nc from the other 4 coronaviruses endemic within the human population (Fig 1b) .

Using the pAb, a capture ELISA was developed that relied on a non-biotinylated form of the antibody to capture Nc from solution and a biotinylated form to detect captured antigen.

After optimization the ELISA demonstrated a dynamic response signal up to concentrations of 1600 pg/ml using recombinant Nc and a linear dose-response below 400 pg/ml (R 2 = 0.999, SE = 0.005) (Fig 2a, b) . Based on the regression analysis at this lower dose range the limit of detection (LoD) was 8.4 pg/ml (LoD = 3.3 x σ/S) and the limit of quantification was 25.5 pg/ml (LoQ = 10 x σ/S). Consistent with our previous Western blot data, the presence of recombinant Nc from other coronavirus species did not generate a signal in the ELISA except at very high concentrations ≥ 4 g/ml (Fig 2c) .

When expressed in an infected cell, Nc is phosphorylated. To ensure the capture ELISA recognised native viral Nc, we tested its ability to detect virus from uv-inactivated SARS-CoV-2 viral supernatants in the presence of detergent. Using a series of two-fold dilutions of three independently generated cell culture supernatants, the capture ELISA was found to have a LoD between 8-16 pfu/ml, depending on the viral supernatant used (Fig 3a) . Further analysis established that the ELISA recognised Nc from both membrane-protected and unprotected compartments within the viral supernatants, consistent with it recognising both virion-associated and non-virion-associated antigen (Fig 3b) .

To assess the capacity of the ELISA to diagnose SARS-CoV-2 infection, nose and throat swab samples were collected from patients admitted to the Southampton University Trust Hospital within 16 hours of testing positive for COVID-19 by quantitative syndromic PCR [17] . In total 138 samples were tested, 128 of which were from confirmed PCR positive cases (median Ct value = 29.8, interquartile range (IQR) = 23.0 to 32.2) and 10 of which came from PCR negative controls. The median (IQR) age of the infected patients was 64 years (60-74). The median (IQR) time from symptom onset to recruitment was 7 days (3-10).

From the ELISA read values it was clear that the signal from the negative controls was universally low but the signal from the positive sample varied considerably (Fig 4a) , with there being both negative signals as well as many strong signals exceeding the linear range of the Nc dose-response curve, present as an internal control for all assays. Specificity of the assay was excellent but given the low numbers of negative controls in the trial, confidence levels were wide (100% (95% CI, 69.2-100)). The sensitivity of the assay was lower but with tighter confidence intervals (58.6% (95% CI, 49.6-67.2). Stratifying the data against Ct values obtained from the earlier syndromic PCR values showed that the sensitivity of the assay increased if the analysis was restricted to patients with low Ct values (i.e. high viral loads) (Fig 4b) suggested that there was a relationship between RNA concentration and Nc levels (Fig 4c) .

To analyse this relationship in more detail, concentrations of Nc were determined using the internal standard curves run within the assay. For those samples where the signal exceeded the linear range of the standard curve additional ELISA analysis was done using serial dilutions of the samples. This revealed a significant inverse linear relationship between Ct values and log(10) Nc concentration (p<0.0001), albeit with a modest R 2 value of 0.54 (Fig   4d) .

To examine the performance of the ELISA in an LMIC setting, the assay was evaluated in (Fig 5a) . Again specificity was seen to be high, with much tighter confidence intervals because of the higher numbers in the negative control group (100% (95% CI, 96.2-100)). Sensitivity was broadly similar to that seen for the Southampton samples, either when all positive samples were included in the analysis (52% (95% CI, 42.8-61.2)) or when the Ct cut off threshold was reduced to 30 and below (72.6% (95% CI, 61.8-81.2; 84 positive samples)) (Fig 5b) .

Restricting the analysis to samples where there was information on clinical presentation (n=48), no obvious difference in sensitivity could be observed between the 21 symptomatic (81.0% (95% CI, 58.1-94.6)) versus 27 asymptomatic (74.1% (95% CI 53.7-88.9)) cases.

These relatively high sensitivity values compared to the cohort overall likely reflect the high viral loads in this subsection of the cohort (mean Ct values of 20.9 and 22.7 respectively).

Although no assessment of Nc concentration was undertaken, it was clear that ELISA signal strength increased with decreasing Ct values of the samples (Fig 5c) . However, it was also apparent that there were small numbers of samples with very low Ct values that that failed to generate a signal in the ELISA assay (i.e. 4 samples with a Ct ≤ 20 proved test negative).

There is currently an unmet need for low-cost SARS-CoV-2 diagnostic tests in LMICs. In this study, we investigated whether a standard ELISA-based platform offered a potential solution to this problem. Initial development work produced an Nc ELISA specific for SARS-CoV-2 that detected levels of recombinant Nc protein as low as 8.4 pg/ml and infectious virus from cell culture supernatants at levels between 8-16 pfu/ml. Assessment of ELISA performance using nose and throat swab samples from a UK cohort with severe COVID found the test to have a sensitivity of 58.6% (95% CI, 49.6-67.2) and specificity of 100% (95% CI, 69.2-100).

A repeat assessment of performance using sputum samples from Ghana produced similar results, with a sensitivity of 52.0% (95% CI, 42.8-61.2) and specificity of 100% (95% CI,

Accessibility, frequency and the time taken from sampling to obtaining a result are all key factors in ensuring effective population screening for SARS-CoV-2, with analytical limits of detection being of secondary importance [18] . The low cost and minimal lab equipment requirements needed to run ELISAs make them both accessible and cost-effective when performing frequent repeated sampling. Analysis of ELISA performance in large-scale SARS-CoV-2 community screening programmes has yet to be undertaken. However, there is reason to think that the findings reported here are of relevance and support the use of ELISAs in such settings. Firstly, ELISA performance was robust given it produced broadly similar results using samples from Ghana and the UK even though sample types were different and different lab personnel were running the two arms of the trial. Secondly, although the positive samples were predominately from symptomatic patients, looking at Ct values as an inverse proxy for the viral load it appears that these were either higher (UK . Ct values vary between studies due to variables such as the efficiency of the PCR reaction, target sequence, sample volume and sample handling so using them to make comparisons has to be approached with caution. However, at least for the UK cohort, the high Ct values are also likely due to the time of sampling. Viral titres are considered to peak around the time individuals first become symptomatic [20] . Increasing the time between symptom onset and testing for the presence of Nc in patients still positive by PCR also reduces Nc-based test sensitivity [21] . As the UK cohort samples came from patients admitted to hospital with severe COVID-19, many had been symptomatic for a week or more (median time 7 days, IQR range of 3-10 days) and thus would be expected to have relatively low titres. Consequently, the overall sensitivity of the ELISA calculated from using our UK hospital cohort may be a slight underestimate of what might be seen if the ELISA was used in community screening.

A key aim of community screening is to identify individuals with high viral loads, as these individuals will be the most infectious [22] . While there is a correlation between Ct values and recovery of infectious virus, there remains some ambiguity as to what the Ct threshold might be above which individuals can be considered non-infectious, with different studies reporting widely varying results [23] [24] [25] [26] [27] . However, based on current WHO guidance those individuals most infectious have test Ct values lower than 25-30 [28] . When analysis of our ELISA test results was limited to Ct values of 30 or below, the sensitivity of the ELISA was 78.3% (95% CI, 66.7-87.3) and 72.6% (95% CI, 61.8-81.2) respectively for the UK and Ghana arms of the study. When the analysis was restricted to Ct values of 26 or below sensitivity increased to 87.0% (95% CI, 73.7-95.1) and 74.6% (95% CI, 62.5-84.5). These aforementioned sensitivities are broadly similar to that of a commercial direct LAMP saliva assay (83% and 94% with Ct cut-offs of <33 and <25) currently employed as a screening tool in the UK [29] . They also compare favourably to the sensitivities reported in a head-tohead comparison of 5 commercial LFDs looking at asymptomatic individuals with Ct values < 30 [30] . Nc ELISA performance also appears similar to that seen for the Innova LFD [13] , a test system widely employed by the UK government for community surveillance; see Table   S1 for comparison. An added advantage to the ELISA format is that it is compatible with sputum, the preferred sample type in LMICs due to difficulties and costs associated with swab collection. Community screening requires robust test performance for both symptomatic and asymptomatic individuals with high viral loads. While we found no obvious difference in test sensitivity between symptomatic and asymptomatic cases in the Ghanaian arm of our study, a potential limitation of this study was the relatively few individuals for which clinical presentation data was available.

Ct values and Nc concentration have previously been reported to be tightly correlated [31, 32] . Therefore, one of the more puzzling observations from this study was the more limited correlation that we observed. Part of the explanation will be due to separate swabs being taken for PCR analysis and then ELISA analysis, with there being up to 24 hours between the two sampling events. Even if the two samples had been taken simultaneously, evidence suggests as much as a 100-fold variation can be observed in sampling when taking consecutive nasal swabs [33] . Nonetheless, it is harder to explain a similar phenomenon seen with the sputum samples, where occasional very low Ct value samples appeared to contain no detectable Nc. One explanation could be that protease inhibitors in sample processing buffer failed to completely suppress protease activity in some samples and Nc was subsequently destroyed. Antigen masking resulting from the presence of Nc antibodies in the sample is another possible explanation. However, antigen masking has been discounted as a significant problem by others [34] and use of pAb in our capture ELISA reduces the likelihood of it further. Failure to detect Nc in low Ct samples may of course reflect genuine absence of the antigen in the sample. However, given that virions contain a very high molecular ratio of Nc to viral RNA [35] , the RNA for the PCR signal would have to then come from elsewhere. Replication associated vesicles released from infected cells could be this source [36] , as although not established, it seems likely that the ratio of Nc to viral RNA in them would be lower that that of virions. However there would still need to be conditions in the lung, such as the presence of opsonizing antibodies directed against viral envelope proteins [37] , for Nc levels to drop more precipitously than viral RNA levels. Given this later possible explanation and the implications it would have for antigen-based testing, it would be interesting to see whether there is a downward shift in the ratio of Nc to viral RNA in both infected individuals that had been previously vaccinated as well as in re-infected individuals.

When obtaining 2 reads per sample and including Nc controls in each plate, the experimental cost of reagents and consumables for running a Nc sputum-based ELISA was 0.31 GBP/sample in the UK and 0.46 GBP/sample in Ghana (Table S2 and 

TWC has received discounted equipment and consumables from QIAGEN to support the independent research study that collected the patient samples used in this study. He has received speaker fees, honoraria, travel reimbursement, equipment and consumables free of charge for research outside of this submitted study from BioFire Diagnostics and BioMerieux; consultancy fees from Synairgen Research, Randox Laboratories, and Cidara Therapeutics; is a member of an advisory board for Roche and a member of two independent data monitoring committees for trials sponsored by Roche; and has acted as the UK chief investigator for an investigational medicinal product study sponsored by Janssen. RWE, CS, IK and PH are co-founders of Highfield Diagnostics Ltd. All other authors declare no competing interests.

The data analysed and presented in this study are available from the corresponding author on reasonable request, providing the request meets local ethical and research governance criteria. 

Pathophysiology, Transmission, Diagnosis, and Treatment of Coronavirus Disease 2019 (COVID-19): A Review

Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine

Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK

Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: A living systematic review and meta-analysis

A colorimetric RT-LAMP assay and LAMP-sequencing for detecting SARS-CoV-2 RNA in clinical samples

Development and evaluation of a rapid and simple diagnostic assay for COVID-19 based on loop-mediated isothermal amplification

A critical review of point-of-care diagnostic technologies to combat viral pandemics

Comparison of seven commercial SARS-CoV-2 rapid point-of-care antigen tests: a single-centre laboratory evaluation study. The Lancet Microbe

Diagnostic performance of a SARS-CoV-2 rapid antigen test in a large

Diagnostic performance of CerTest and Panbio antigen rapid diagnostic tests to diagnose SARS-CoV-2 infection

COVID-19: Rapid antigen detection for SARS-CoV-2 by lateral flow assay: A national systematic evaluation of sensitivity and specificity for mass-testing

Comparison of Rapid Antigen Tests for COVID-19

Novel versatile smart phone based Microplate readers for on-site diagnoses

Clinical impact of molecular point-of-care testing for suspected COVID-19 in hospital (COV-19POC): a prospective, interventional, non-randomised, controlled study. The Lancet Respiratory medicine 2020

Ct threshold values, a proxy for viral load in community SARS-CoV-2 cases, demonstrate wide variation across populations and over time

load and infectivity over the course of an infection

Multi-site assessment of rapid, point-of-care antigen testing for the diagnosis of SARS-CoV-2 infection in a lowprevalence setting: A validation and implementation study

S gene variants and demographic factors and the utility of lateral flow devices to prevent transmission

Predicting Infectious Severe Acute Respiratory Syndrome Coronavirus 2 From Diagnostic Samples

Viral RNA load as determined by cell culture as a management tool for discharge of SARS-CoV-2 patients from infectious disease wards. European journal of clinical microbiology & infectious diseases : official publication of the European Society of

Ct value is not enough to discriminate patients harbouring infective virus

Euro surveillance : bulletin Europeen sur les maladies transmissibles

Assessment of SARS-CoV-2 infectivity of upper respiratory specimens from COVID-19 patients by virus isolation using VeroE6/TMPRSS2 cells

SARS-CoV-2 Antigen detecting rapid diagnostic test implementation projects

Rapid evaluation of OptiGene RT-LAMP assay (direct and RNA formats

Performance characteristics of five antigen-detecting rapid diagnostic test (Ag-RDT) for SARS-CoV-2 asymptomatic infection: a head-to-head benchmark comparison

Quantitative DiaSorin Liaison Antigen Test to Reverse Transcription-PCR for the Diagnosis of COVID-19 in Symptomatic and Asymptomatic Outpatients

CoV-2 Nucleocapsid Antigen and RNA Concentrations in Nasopharyngeal Samples from Children and Adults Using an Ultrasensitive and Quantitative Antigen Assay

Sampling variability between two mid-turbinate swabs of the same patient has implications for influenza viral load monitoring

N-protein presents early in blood, dried blood and saliva during asymptomatic and symptomatic SARS-CoV-2 infection

A structural analysis of M protein in coronavirus assembly and morphology

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

Fc receptor-mediated phagocytosis makes a significant contribution to clearance of influenza virus infections

We thank patients, clinical teams and testing personnel at the various hospital sites, all study volunteers and various team members whose contributions, aided either sample collection or processing, especially the WACCBIP Next Generation Sequencing team, Dr. Gloria