A Multiplex Serological Assay for the Detection of Antibody Responses to Arboviruses

Arboviruses collectively account for thousands of deaths every year. These viruses are transmitted to humans via infected arthropods, mostly mosquitoes and ticks. The severity of arboviral infections is diverse and ranges from asymptomatic infection to severe disease that can lead to death. In 2024, the World Health Organization (WHO) recorded over 10,000 deaths caused by the four serotypes of the dengue virus (DENV) alone¹. The same year, the virus infected over 14 million people, including 50,000 severe cases¹. While DENV is the most prominent arbovirus, the group includes other life-threatening pathogens that have been responsible for multiple outbreaks worldwide. A recent summary review of arboviral infections in Africa found that these viruses, including DENV, yellow fever virus (YFV), chikungunya virus (CHIKV), Crimean-Congo hemorrhagic fever virus (CCHFV), Rift Valley fever virus (RVFV), West Nile virus (WNV), and Zika virus (ZIKV), caused at least 29 outbreaks across the continent in 2023². In 2022, a report published by the WHO Regional Office for Africa highlighted the current limitations in arbovirus surveillance across the region, emphasizing the urgent need to develop methods for arbovirus surveillance and control³.

Clinical diagnosis of arboviral infections is challenging due to the similarities in symptoms, such as fever, arthralgia, and general malaise, that are common to other pathogens present in endemic regions⁴. In addition, due to the co-circulation and potentially co-infection of the viruses, distinction of the etiological pathogen is difficult⁵. Laboratory confirmation is crucial for differential diagnosis. Several methods exist for the detection of arborviruses in human samples. Direct methods, including molecular diagnostic tests, viral antigen detection, and viral isolation, detect the presence of the pathogen, while indirect methods, such as serological diagnosis, measure the immune response to the viral infection⁶.

A plethora of methods for the detection of antibody responses to arboviruses has been developed. These include enzyme-linked immunosorbent assays (ELISAs), complement fixation, hemagglutination inhibition, and microsphere immunoassays^6,7. This report presents a multiplex fluorescence microsphere-based immunoassay (FMIA) for the detection of antibody responses to arbovirus antigens. Microsphere-based assays⁸ are extensively used in serology and offer several advantages, most importantly, the ability for the detection of multiple analytes in a single reaction with minimal sample volume. Optically coded microspheres (magnetic polystyrene beads) coated with recombinant antigens of interest are used in an indirect serological immunoassay. The coupled beads are then incubated with samples, and antigen-specific antibodies are detected using an anti-human IgG secondary antibody conjugated to a Phycoerythrin (PE) fluorophore reporter. Several analyzers are available for the detection of emitted fluorescence, which is reported as MFI.

The principle of the assay is detailed in Angeloni et al.⁸. Briefly, the antigens of interest are chemically bound to color-coded microspheres. For this, the beads are initially treated with 1-ethyl-3-[3-dimethylaminopropyl] Carbodiimide (EDC) and N-hydroxysulfosuccinimide (sulfo-NHS), which activate the carboxyl group on the surface of the beads, creating an amine-reactive compound. This intermediate semi-stable compound is then mixed with the protein of interest, whereby amine moieties of the protein are covalently bound to the beads, generating stable microbeads coated with the protein of interest at the surface. The antigen-bound beads are incubated with samples, allowing antibodies to bind to specific antigens. Antigen-bound antibodies are detected using an anti-immunoglobulin secondary antibody conjugated to a fluorescent dye. We use PE-labelled secondary antibody. MFIs are measured using a light imager (details of the instrument are provided in the Table of Materials). The instrument uses two LED lights, a red LED and a green LED, which separately excite the beads' internal dyes and a reporter fluorescence on the surface of the beads⁹. Filters are then used to classify beads based on the color and to quantify fluorescence emitted by the analytes.

We have previously applied the technology for the detection of antibody responses to SARS-CoV-2¹⁰ and vaccine-preventable diseases¹¹ and showed that multiplex assays constitute an invaluable tool for seroprevalence and surveillance studies. The goal of this project is to extend the use of the FMIA to the detection of antibody responses to multiple antigens belonging to arboviruses with an aim to apply the assay to screen samples from the African region.

Arbovirus multiplex assays using multiplex technology have been described elsewhere using different antigen targets such as non-structural protein 1 (NS1)^12,13 or envelope domain III (EDIII)¹⁴. A 50-plex assay, which includes 26 arbovirus antigens (including NS1, EDIII, viral-like particles (VLPs), and multiple other antigens to test for pathogens that are co-circulating in the region, is presented. The assay will allow assessment of transmission intensity for these pathogens. A summary of recombinant antigens used for the development of the assay is provided in Table 1. The protocol presented in this report was developed in complement to an extensive method development study focused on optimizing the detection of arboviruses.

Ordinarily, the coupling (or fixation) of antigens to the microsphere beads is carried out in single tubes⁸ (manual coupling). This process is time-consuming, cumbersome, and prone to reproducibility errors. This article reports a protocol using an automated instrument to perform the successive steps involved in the chemical binding of antigens to the beads. In this manuscript, we report steps undertaken for the validation of the new protocol. The validation of the new protocol was performed by comparing IgG responses measured with a 50-plex FMIA using antigen-coupled beads prepared with either the manual method or the automated protocol. Further validation of the method was performed by analyzing human serum samples from African and French cohorts.

The automated beads coupling protocol is fast, allowing simultaneous coupling of up to 96 analytes in a single run. More importantly, results are comparable to those obtained using the established manual protocol, with reduced batch-to-batch variation.

Appropriate ethics approval was obtained for the use of human samples. Samples were anonymized residuals from previous studies. The protocol for the use of "standard pool" ORPAL samples was approved by the National Ethics Board of Suriname (CMWO (Commissie voor Mensgebonden Wetenschappelijk Onderzoek), Opinion Number VG 25-17). The use of samples from Madagascar was approved as part of the PvSTATEM project by The London School of Hygiene & Tropical Medicine's Ethics Committee (28541) and the Madagascar Comité d'Éthique de la Recherche Biomédicale (135MSANP/SG/AMM/CERBM), while the use of Senegalese samples was approved as part of the ongoing Dielmo and Ndiop project by the Senegalese National Health Research Ethics Committee (CNERS Sénégal). Approval to measure antibodies to NTDs in these samples was granted by CNERS Sénégal (N 00000007 MSAS/CNERS/Sec). Non-endemic French samples were derived from routine medical testing within French hospital laboratories and processed in accordance with existing regulations and guidelines of the French Commission for Data Protection (Commission Nationale de l'Informatique et des Libertes). Human serum samples were obtained from subjects from regions with high prevalence of arboviruses and from a naïve French population. Serum samples were heat-inactivated prior to use by incubation at 56 °C for 30 min. If samples are not inactivated, manipulate them under a safety cabinet. Most recombinant antigens were commercially obtained except for EDIII recombinant proteins, which were synthesized in-house. An imaging-based analyzer instrument⁹ was used for data acquisition. The reagents and equipment used are provided in the Table of Materials.

1. Determine optimal antigen concentrations for coupling

NOTE: To prevent photobleaching, minimize exposure of microsphere beads to light throughout the procedure. Prior to bulk coupling, optimize coupling concentrations for each antigen using a manual coupling protocol. The protocol was adapted from Angeloni et al.⁸. Buffers and solutions are detailed in Supplementary Table 1.

1. Select bead regions to use for coupling. Thoroughly vortex for 20 s. Sonicate the tube for a minimum of 60 s.
2. Transfer 200 µL (2.5 M beads) of beads to a 1.5 mL tube.
3. Wash the beads with 200 µL 100 mM Monobasic sodium phosphate/0.125x Triton.
   1. Place the tube on a magnetic tube separator for 60 s. Carefully remove the supernatant without touching the beads.
   2. Add Monobasic sodium phosphate buffer. Leave on the magnetic rack for 60 s. Remove the supernatant.
4. Activate the beads with a buffer containing 80 µL of 100 mM monobasic sodium phosphate, 0.125x Triton, 10 µL sulfo-NHS, and 10 µL EDC.
   1. Add monobasic sodium phosphate -sulfo-NHS -EDC to tubes. Incubate in the dark for 20 min at room temperature on a roller.
   2. Wash the beads in 250 µL PBS (1x) and 0.125x Triton. Repeat washing three times.
5. Bind the antigens.
   1. Prepare the different concentrations (serial dilutions) of antigens in a total volume of 1 mL 1x PBS and 0.125x Triton. Test at least two concentrations of antigens.
   2. Add antigen solutions to activated beads. Incubate in the dark 2 h at room temperature or overnight at 4 °C on a roller.
6. Wash coupled beads in 500 µL PBS-TBN and 0.125x Triton. Repeat the wash three times.
7. Resuspend the coupled beads in 1 mL of PBS-TBN and 0.125x Triton.
   NOTE: Store the coupled beads at 4 °C for at least 24 h before use. Coupled beads showed variability in MFIs when tested immediately after coupling in comparison to testing after 24 h.
8. Select an optimal concentration to use in subsequent large-scale coupling. Use different coupled beads for the immunoassay on a standard pool. Select the lowest concentration at which a linear standard curve was obtained.
   NOTE: An example of linearity of antibody response to a standard pool is provided in the results section, Figure 1.

2. Large-scale automated bead coupling

1. Prepare Pro Magnetic COOH beads.
   1. Calculate required beads: Using 500 beads/sample, 2.5 million beads are sufficient for 5000 samples.
   2. Prepare bead regions to use for coupling as previously described.
   3. Transfer the desired volume of beads to the allocated wells. For 2.5 million, 200 µL of beads (standard Luminex concentration is 1.25e7 beads/mL) were used.
2. Prepare the antigens. This is a critical step to ensure accurate quantities of antigens.
   1. Calculate quantities of required antigen, following optimization. The concentrations used for each antigen are detailed in Table 1.
   2. In a 1.5 mL microcentrifuge tube, thoroughly mix an adequate volume of antigens of interest with 1x PBS /0.125x Triton to a total of 1 mL.
3. Prepare plates with reagents. Prepare 96-deep well plates as follows. Volumes are per well and calculated for 2.5 million beads coupling.
   1. Plate 1 (Tip Combs): load tip combs.
   2. Plate 2 (Beads): Add 8 µL of 0.125x Triton and 200 µL of the bead regions of interest.
   3. Plate 3 (Wash1): Add 200 µL of 100 mM Monobasic sodium phosphate / 0.125x Triton.
   4. Plate 4 (Activation): Add 80 µL of 100 mM Monobasic sodium phosphate / 0.125x Triton + 10 µL sulfo-NHS + 10 µL EDC.
   5. Plate 5 (Wash2): Add 250 µL of 1x PBS /0.125x Triton.
   6. Plate 6 (Coupling): Transfer 1 mL of antigen solution to the allocated well.
   7. Plate 7 (Wash3): Add 500 µL of PBS-TBN/0.125x Triton.
   8. Plate 8 (Final plate, contains coupled beads): Add 1 mL of PBS-TBN + 0.125x Triton.
4. Set up coupling cycle run in the automated processor.
   1. Follow the manufacturer's instructions to switch on the machine and set up the protocol. The necessary software is downloadable on the PC¹⁵. Run-specific parameters are detailed in Supplementary Table 2.
   2. Load plates into the allocated slot on the machine. Select the protocol. Start the cycle. The estimated time to complete the programme is 2:50 h.
5. End of cycle
   1. Follow the manufacturer's instructions to switch off the machine. Unload plates as instructed. Visually check that plate 8 contains the beads (brown pellet). Additionally, a run report is exported from the instruments.
   2. Transfer coupled beads from plate 8 to individual 1.5 mL tubes for storage. Store coupled beads at 4 °C until use.
6. Count the beads.
   NOTE: A cell counter was used. For manual counting using a hemacytometer, refer to the Luminex cookbook manual⁸.
   1. Thoroughly vortex coupled beads. Transfer 10 µL to a cell counting chamber. Follow the manufacturer's instructions to read the bead count per mL.
   2. Check aggregation by visualizing the image on the cell counter. See Supplementary Figure 1 for a representative image of non-aggregated beads.
   3. If aggregates are present, thoroughly vortex and sonicate for at least 5 min. Repeat counting
      ​NOTE: After bead count, coupling efficiency was evaluated by performing an immunoassay of coupled beads on a standard pool to assess the linearity of the antibody response.

Table 1: Panel of antigens included in the 50-plex. The assay includes multiple antigens for the detection of arboviral infections and co-circulating pathogens. Quantities of antigens used for the coupling of 1.25 million beads are provided. Please click here to download this Table.

3. Multiplex immunoassay

NOTE: Detailed steps of the protocol for measurement of MFI are provided in the respective manual⁹. To minimize exposure of the beads to light and risk of photobleaching, all solutions and samples are prepared before taking the beads out of the fridge.

1. Set up the plate layout. 96-well plates were used and included serial dilutions of a standard pool as well as a blank (buffer only) in the first row.
2. Dilute samples and standard in PBT buffer in a separate non-binding plate. A 1:200 dilution (final concentration of 1:400) was used for the samples. For the standard curve, a 1:50 dilution of the standard serum pool (resulting in a final concentration of 1:100) and ten 2-fold serial dilutions were used.
3. Determine the quantity of coupled beads to use in a total volume of 50 µL/well. 500 beads of each bead per well were used and adjusted to volume with PBT buffer.
4. Thoroughly vortex the coupled beads mix for 30 s. Sonicate for at least 60 s.
5. Calculate the volume of each coupled bead to add to the multiplex premix to ensure an equal beads for each analyte. Volumes of beads were calculated as follows: Volume of beads per well = number of beads per well/bead count.
   NOTE: An example for the calculation of beads and buffer volumes necessary for the premix is illustrated in Supplementary Table 3.
6. Transfer adequate bead-antigen volumes to a centrifuge tube.
7. Adjust the volume with PBT.
8. Thoroughly mix the premix.
9. Distribute 50 µL of the premix to the imaging microplate (e.g, ps, f-bottom plates).
10. Add 50 µL of the diluted samples to the beads.
11. Mix on a plate shaker for 30 min at room temperature, 700 rpm.
12. Wash the plates three times as follows (manual wash).
    1. Place the plate on a magnetic rack for 60 s at room temperature. Hold the plate tightly on the rack and discard the supernatant.
    2. Add 100 µL PBT buffer. Leave on the magnetic rack for 60 s. Repeat the washes
       NOTE: Alternatively, a plate washer¹⁶ can be used. The washer was set as follows: 1 cycle soak, followed by 3 cycles (aspiration, dispense, and soak) and a final cycle aspirate.
13. Discard the supernatant and add 50 µL of diluted secondary antibody.
    1. Follow supplier's instructions to prepare a stock solution of the secondary antibody. A 1/120 dilution solution of secondary antibody was used. Volumes of secondary antibody were calculated as follows: volume secondary antibody = total volume/dilution factor.
    2. Mix the secondary antibody with an adequate volume of PBT buffer. Thoroughly mix. Add 50 µL of diluted secondary antibody to the washed beads.
14. Incubate on a shaker for 15 min at room temperature, 700 rpm.
15. Perform three washes as previously described.
16. Resuspend beads in 150 µL PBT buffer.
17. Incubate on a shaker for 5 min at room temperature, 700 rpm. Follow the manufacturer's instructions to set up the instruments⁹.
    NOTE: The xPONENT software¹⁷ provided with the imager instrument was used. The settings for the fluorescence reading were set as follows: collection volume was set to 50 µL, with a minimum measure of 50 beads, without XY heater, and analysis type as none.

4. Data analysis

1. Run validation
   1. Check that the minimum bead count was reached. At least 35 beads per well are recommended to obtain a statistically valid MFI⁸.
   2. Check blank and negative control sample MFI values are minimal.
   3. Check MFI values for standards by plotting a standard curve. These are higher than the blank and/or negative control sample.
2. Sample data analysis
   1. Export the data as CSV files and externally analyze using R software.

To illustrate the process followed for validation of the protocol, we have selected three antigens that represent three families of arboviruses: Flaviviruses (DENV1 NS1 protein), Alphaviruses (CHIKV VLP protein), and Bunyaviruses (RVFV NP). The antigens were tested on a standard pool of a pre-exposed population (ORPAL), French (non-exposed) population sera, and two cohorts from Africa (Senegal and Madagascar).

Dilution linearity of antibody response
The concentration of antigens to use for bead coupling was determined using serial dilutions of the ORPAL pool. The optimal concentration was selected as the lowest concentration at which a linear standard curve was obtained (data not shown). Figure 1 shows the linearity of selected arbovirus antigens at optimal concentration for each antigen. For quality control of coupling, each batch of coupled beads was tested for dilution linearity on the ORPAL standard pool using the pre-selected concentration.

Comparability to the standard manual coupling assay
To assess the comparability of data obtained using the new automated protocol with the well-established manual coupling, we compared standard curves generated from results obtained with manual versus automated coupling. Figure 2 shows parallel linearity for both assays. MFI values vary by antigen.

Bead coupling batch variation
The coupling batch variation, as the reproducibility of tests from the same coupling batch was assessed, was assessed using the Levey-Jennings chart18. The representative antigens and control BSA-coupled beads were used and tested on the ORPAL standard pool. Figure 3 shows MFIs from 19 different plates, tested on different days for each antigen. The resulting MFIs fall within the range of the mean ± 3 SD for the three antigens and the control.

Effect of multiplexing on MFI
To further assess the performance of the assay, we tested whether multiplexing of several antigens affected the measured antibody response. The response was compared to a single antigen (or single-plex bead) using the ORPAL standard pool. Figure 4 shows a strong correlation (R = 0.99, p < 2.2e-16) between results obtained when MFIs were measured on single antigens or from a premix containing 50 antigens.

Antibody responses from human serum samples in African populations
To evaluate the applicability of the assay to samples from diverse cohorts, we compared antibody responses in human serum samples from African populations, which have a high likelihood of exposure to arboviruses, with those from a French cohort that presumably has lower exposure to these viruses. In total, 1,596 samples were tested for Madagascar, 1,428 for Senegal, and 84 from the French cohort. Figure 5 presents density plots for the different cohorts. For the Senegal cohort, a clear bimodal distribution of data is observed for the responses to CHIKV VLP, suggesting two distinct responses to CHIKV in the population. A subtle bimodal tail is also observed for DENV1 NS1 and RVFV NP in the Senegal cohort, while there is no difference between the French and Madagascar cohorts.

Figure 1: Linearity of antibody response tested on ORPAL standard pool. Selected arbovirus antigens are shown as an example. Median fluorescence intensities (MFI) are shown on the y-axis (log scale) and serial dilution of the standard pool on the x-axis. All antigens in the panel were pre-tested on the standard pool prior to bulk testing on samples. Please click here to view a larger version of this figure.

Figure 2: Parallel linearity of antibody response by coupling methods. Log of MFI (y-axis) of representative antigens (facet) tested on serial dilutions of the ORPAL standard pool (x-axis). The colors represent different coupling methods: automation (red, green, blue) and manual (purple). Please click here to view a larger version of this figure.

Figure 3: Plate-to-plate variation. Levey_Jennings plots for BSA coupled bead (control), DENV1 NS1, CHIKV VLP, and RVFV are shown for ORPAL standard pool at a concentration of 1:400. Results represent 19 plates run on different days. MFIs are presented on y-axis, the number of tests is on the x-axis. Please click here to view a larger version of this figure.

Figure 4: Correlation of antibody response to mono and multiplex assay. The standard pool was diluted at a final concentration of 1:400 and tested on a premix containing the 50-plex antigens in parallel with single antigen-coated beads. Pearson correlation was used for analysis. Three representative antigens are shown as an example. Please click here to view a larger version of this figure.

Figure 5: Application of the assay on the endemic and non-endemic cohort of samples. The assay was tested on samples from Madagascar (green), Senegal (blue), and France(red). Samples were diluted at a final concentration of 1:400. The X-axis represents normalized MFIs (nMFI), calculated as a ratio of sample MFI to standard pool MFI at the same concentration (1:400). Please click here to view a larger version of this figure.

Supplementary Figure 1: Example representative image of bead aggregation. Panel (A) shows non-aggregated beads, and panel (B) shows bead aggregates. Please click here to download this File.

Supplementary Table 1: Table of buffers. Please click here to download this File.

Supplementary Table 2: KingFisher Apex beads coupling protocol settings. Please click here to download this File.

Supplementary Table 3: Example for calculation of bead volumes for premix. Please click here to download this File.

The use of FMIA has changed the landscape of serological testing. The strength of the assays lies in the ability to evaluate antibody responses to multiple antigens in one reaction, reducing time to results and decreasing the cost of testing. This is particularly important in serological surveillance. Routine use of FMIA assays is, however, hindered by the labour-intensive, time-consuming manual bead coupling step, which requires handling single antigens per reaction. This report presents an automated method for chemical coupling of antigens to beads. The automation allows simultaneous coupling of up to 96 antigens in a single run, considerably reducing the time required for high-throughput antigen coupling.

Figure 1 shows that the antigens were successfully bound to the beads. Results also show a strong parallelism between the new automated protocol and the well-established manual protocol (Figure 2), indicatingnon-inferior accuracy in detecting IgG responses to the evaluated antigens. Results of plate-to-plate (Figure 3) and bead coupling batch variation (Figure 2) show comparability of data when using the automated method for coupling. High correlation between multiplex and singleplex MFIs (Figure 4) shows that multiplexing a high number of antigens does not affect antibody response to the set of individual antigens used in this assay.

For validation of the assay, samples from both endemic and non-endemic cohorts were tested (Figure 5). While seroprevalence studies of arboviruses are still scarce in Africa, available data^19,20,21,22 suggest a nuanced picture with variable IgG seroprevalences between and within individual countries. Notably, the existing data suggest higher prevalences of the three viruses (chikungunya, dengue, and Rift Valley) in Senegal compared to Madagascar, for example¹⁹. These prevalences are consistent with the observation in this report, which showed a clear, distinct response to the different antigens in Senegal when compared to the other cohorts. This supports the assay's ability to distinguish antibody profiles in the context of relatively high seroprevalence. The observed data for the Madagascar cohort may be explained by the reported low prevalence of the viruses in the sample collection site²², while the comparability of data to the non-endemic cohort may partially be explained by the small sample size of the French cohort. In addition, reported variability between regions might explain the differences observed in the datasets. An extensive study to evaluate IgG seroprevalence is necessary; however, the data presented in this report show that the developed assay is specific for detecting IgG responses to the evaluated arbovirus antigens, distinguishing data from endemic regions as well as nuanced, distinct responses within the same population.

Key technical challenges included achieving homogeneous antigen solutions despite their use in very small quantities, ensuring equimolar representation of bead regions in the premix, and minimizing bead aggregation. To overcome the difficulty of achieving a homogeneous solution in the 96-well plate, a step was implemented where all antigenic solutions were thoroughly premixed in a 1.5 mL tube before being distributed into the plate. Additionally, a step was added to count the coupled beads and adjust the volumes to ensure that an equal number of bead regions are added to the premix. A sonication step was also included to facilitate bead mixing and minimize aggregation in the premix.

In this limited study, the stability of the coupled beads and their impact on assay performance were not evaluated, as all samples were tested within one month of bead coupling. This evaluation was beyond the scope of this report and will be assessed in future studies. Despite these limitations, this report presents a rapid and reliable method for detecting antibody responses to a variety of pathogens, including arboviruses. Serological surveillance of arboviral infections is crucial for controlling the diseases, and the availability of a reproducible, rapid method for detecting antibody responses to these viruses represents a significant advancement in global effortsto assess heterogeneities in transmission and target interventions to mitigate the burden of these infections. In addition, although the focus of this report is on arboviruses, this protocol is broadly applicable to a wide range of antigens and represents a significant advancement in the field.