In the present study, we established a MERS-CoV-specific MNA using staining of virus foci to demonstrate neutralization. We compared this newly developed quantitative assay with our standard virus neutralization test (VNT100), in which we evaluate neutralization by an analysis of CPE. Both neutralization tests use authentic virus for multiple replication cycles and are therefore considered the gold standard as they best mimic the natural situation [6]. Neutralization assays based on pseudotypes, such as vesicular stomatitis virus or lentiviruses that contain only the foreign viral surface glycoprotein, have major limitations. They are often restricted to one viral replication cycle, and since only the major surface protein of the virus is incorporated, the expression, the density, and geometry of the protein may differ from those of authentic virus. This can affect the ability of antibodies to bind and neutralize the virus [6].

In addition to the advantages, our VNT100 has limitations: Due to the microscopic evaluation, it is subject to operator-dependent variability and the results cannot be easily documented. The VNT100 is less sensitive as the test material must contain sufficient antibodies to neutralize almost 100% of the virus used [6]. Furthermore, several replicates and thus relatively large amounts of serum are required for the analysis. By developing our new assay, we have overcome these limitations. Now less serum is required, the sensitivity has been improved, and we automatically document the results with a spot analyzer. However, this assay also has a limitation. For historical reasons, we work with MERS-CoV under BSL4 conditions, even though it is a BSL3 pathogen. Due to the time constraints of BSL4 work (4 h per entry), the plates must be inactivated and removed from the BSL4 for staining and analysis as these works would take longer than 4 h or tie up more staff. The practical procedure of the assay limits the number of samples that can be analyzed as well because in the same working time only 24 samples can be analyzed in the MNA, while 80 samples can be analyzed in the VNT100. In this regard, the VNT100 has an advantage. Since the assay is evaluated microscopically in the BSL4 and practical procedures are more rapid, a larger number of samples can be analyzed.

We have shown that the MNA is more sensitive than the VNT100 most likely because the IC50 is determined instead of the IC100. And this is the case even though about 2000 PFUs are used in the MNA compared to 100 PFUs in the VNT100 to infect the cells of a well. However, as shown by others, such a high amount of virus is required in the MNA [8] to achieve the desired amount of 100–150 virus foci. Amanat et al. have shown that approximately 10,000 TCID50 need be used to generate 100 foci per well in their MNA for SARS-CoV-2 [8]. Since other differences between the two assays can also contribute to the MNA being more sensitive than the VNT, such as the cell lines or the incubation time used, we have provided a detailed comparison of the main characteristics of the two tests in the supplement (Online resource, Table 1).

To compare our assays, we used a receptor-binding domain-specific monoclonal antibody, m336. The IC50 of m336 to neutralize authentic MERS-CoV was determined to be 70 ng/ml [10]. In our assay, the IC50 of m336 was determined to be 3.1 ng/ml. We cannot definitively explain this discrepancy. However, it could be due to differences in the viruses used, the experimental procedures, as well as differences in the storage and solvent of the antibody. Ying et al. used 200 TCID50 of a clinical isolate of MERS-CoV (not specified in more detail) and incubated it with serial dilutions of m336 for 2 h before transfer to confluent Vero cells. Evaluation was performed at 3 days post infection by microscopy of CPE. In terms of antibody stability, we have found that when m336 is stored in phosphate-buffered saline, a single freeze–thaw cycle severely impairs the function of the antibody (personal observation, VK, unpublished).

Neutralization tests are often not comparable. Therefore, it is of particular importance to use standardized materials such as the WHO IS, to be able to compare the results. In 2019, a collaborative study to investigate the comparability of serologic assays for MERS-CoV demonstrated that the use of a reference reagent, such as WHO IS, greatly improves the agreement between assays, enabling more consistent and therefore more meaningful comparisons between results [13]. Therefore, in addition to the neutralizing titers, we determined the IC50 and IC80 of WHO IS in our MNA with 0.67 IU/ml and 2.6 IU/ml, respectively. The WHO IS has been available since 2020 [14]. Unfortunately, we have not found any published results on MERS-CoV neutralization tests to compare the results of our test with other tests. Probably because tests were published before the WHO IS was available. Furthermore, we analyzed the inhibitory effect of the pan-coronavirus inhibitor EK1C4 in our MNA. Xia et al., 2020, showed that EK1C4 can inhibit multiple CoVs, including MERS-CoV. They have shown that MERS-CoV infection was inhibited with an IC50 of 4.2 nM [12]. In our assay, the IC50 was determined to be 50 nM. Xia et al. used a plaque reduction neutralization test with 100 TCID50 MERS-CoV per well for 72 h of infection of VeroE6 cells. In comparison, we use 2000 PFU for 24 h. This leads us to assume that the results were so different because Xia et al. used fewer virus.

We demonstrated that the correlation between our assays is good and that both are suitable to determine the neutralizing capacity of serum samples as well as antibodies. As both assays have limitations, a combination of both assays is optimal for the reliable detection of vaccine-induced neutralizing antibodies in clinical trials.