Background
Epstein-Barr virus (EBV) is the leading aetiological agent of infectious mononucleosis and several malignancies including nasopharyngeal carcinoma (NPC), gastric carcinoma (GC), Burkitt lymphoma (BL), Hodgkin lymphoma (HL) and natural killer (NK) /T cell lymphoma [
1]. EBV is also associated with autoimmune diseases such as systemic lupus erythematosus, Sjögren’s syndrome and rheumatoid arthritis [
2]. Recently, a longitudinal study reported that the risk of multiple sclerosis (MS) increased 32-fold after EBV infection [
3]. EBV causes heavy global public health burdens with 113,205, 105,554, 40,109 and 6,318 new cases per year of GC, NPC, HL and BL worldwide, respectively [
4]. However, no vaccines against EBV infection or therapeutic agents for EBV-linked diseases are available.
During its infectious cycle, EBV exhibits two distinct tropisms toward epithelial cells and B cells. Virions derived from epithelial cells tend to infect B cells while virions produced by B cells more effectively infect epithelial cells [
5]. These tropisms depend on the viral surface density of different glycoprotein complexes, gHgLgp42 and gHgL. During virus entry, additional glycoproteins, are involved in attachment to target cells (gp350 and BMRF2) and execution of membrane fusion (gB). The most abundant glycoprotein on the viral surface is gp350, which interacts with complement receptor 2 (CR2) [
6] or CR1 [
7] to initiate the infection of B cells. Following attachment, gHgLgp42 binds to human leukocyte antigens class II (HLA-II) and further triggers gB conformational changes to finish fusion [
8]. Epithelial cell infection is initiated by BMRF2 binding to cellular integrins [
9]. This is followed by gHgL binding to ephrin receptor A2 and activation of gB to execute membrane fusion [
8,
10,
11].
Many serological studies have attempted to correlate antibodies elicited by EBV with infection status or disease outcomes [
12‐
16]. Three parameters are commonly used to distinguish acute infection from past infection: viral capsid antigen (VCA)-IgG, VCA-IgM and EBV nuclear antigen 1 (EBNA1) -IgG [
12]. The serological characteristics of healthy EBV carriers are VCA-IgM (−), VCA-IgG ( +), EBNA1-IgG ( +) and EBNA2-IgG (weak or -) [
13]. Furthermore, levels of antibodies against various EBV proteins are predictive markers for the risk of developing NPC, GC, HL, BL and NK/T lymphoma [
14,
15]. Another study revealed that high titers of antibodies targeting glycoproteins were detected in both NPC patients and healthy carriers, and sera from each group have similar neutralizing abilities [
16]. Anti-gp350 antibodies are the major contributors to B cell neutralization, while anti-gHgL antibodies play an important role in epithelial cell neutralization [
17].
Many prophylactic vaccine formulations against EBV infection have been studied since the 1980s. gp350 has been considered an ideal candidate for the development of prophylactic vaccines to prevent the initial EBV infection. Various gp350-based vaccine modalities including soluble recombinant proteins (multimeric and monomeric), viral vectors, nucleic acids, virus-like particles and nanoparticles were developed and evaluated in animal models [
18]. Besides, clinical trials have been launched to evaluate gp350-based vaccines, including a recombinant vaccinia virus (Tien Tan strain) expressing gp350 [
19], gp350 adjuvanted with alum or AS04 [
20‐
22] and ferritin nanoparticles displaying gp350 (NCT0464514). Recently, a phase I clinical trial for an mRNA-based vaccine consisting of four mRNAs encoding gH, gL, gp42 and gp220 has also been initiated (mRNA-1189; NCT05164094).
EBV infection is a complicated process and humoral immune responses are important for EBV primary infection control. Neutralizing monoclonal antibodies are potential therapeutic agents and useful guides to improve vaccine design. To date, various neutralizing monoclonal antibodies targeting EBV envelope glycoproteins have been reported, including 72A1 (gp350) [
23], AMMO1 (gHgL) [
24], 6H2 (gHgL) [
25], 1D8 (gHgL) [
26], CL40 (gHgL) [
27], CL59 (gHgL) [
27], E1D1 (gL) [
28], F-2–1 (gp42) [
29], AMMO5 (gB) [
24],3A3 (gB) [
30], 3A5 (gB) [
30], 8A9 (gB) [
31] and 8C12 (gB) [
31].
Determination of neutralizing titers is considered the critical index for serological studies and vaccine-induced humoral responses and is essential for monoclonal antibody screening. However, available approaches including inhibition of human B cell transformation, immunofluorescence-based assay, competition enzyme-linked immunosorbent assay (ELISA) and flow cytometry-based (FCM) based assay to determine neutralizing titers are time-consuming and unsuitable for testing large-scale clinical samples in high-throughput settings as discussed below [
32‐
35]. A classical method to measure neutralizing titers is based on inhibition of human B cell transformation, which requires a 6–8 week detection period [
32]. An immunofluorescence-based assay to detect EBV-positive stained Raji cells was developed but this early approach was limited by manual counting [
33]. Alternatively, a competition ELISA using the neutralizing monoclonal antibody 72A1 provides a surrogate approach to detect the presence of neutralizing antibodies, but this assay does not determine actual titers [
34]. Furthermore, an FCM neutralization assay utilizing B cells infected by EBV-GFP (green fluorescence protein) was developed, which is limited by the relatively low throughput at data collection and analysis [
35]. Recently, a higher-throughput fluorescent imaging assay (FIA) using Akata-EBV-GFP to infect SVK-CR2 cells (an epithelial cell line overexpressing CR2) was reported, but it may not truly reflect the natural infection process [
36,
37].
High content imaging system (HCIS) uses a high-throughput live cell imaging format and applies automated microscopy, fluorescent detection and multiparameter algorithms. HCIS has been used to visualize and quantify the interaction of therapeutics in cell populations [
38]. Considering the high-throughput potential of image capture and analysis of HCIS, we developed a rapid and high-throughput method based on HCIS to determine neutralizing titers in B cells and epithelial cells. We validated this method in EBV infection of epithelial cell models (HNE1 epithelial cells infected with Akata-EBV-GFP virus) and B cell models (Akata B cells infected with CNE2-EBV-GFP virus). We compared the infection titers of CNE2-EBV-GFP and neutralizing titers of monoclonal antibodies determined by HCIS-based assays and FCM-based assays. A strong correlation was observed between CNE2-EBV-GFP viral titers defined by HCIS-based assay and FCM-based assay. The half maximal neutralizing concentration (NC
50) of monoclonal antibody 72A1 or CL55 was also similar in both assays. We evaluated the neutralizing titers of sera from healthy EBV carriers and sera from monkeys infected with rhesus lymphocryptovirus (rhLCV), a simian homolog of EBV [
39,
40]. Neutralizing titers in sera of healthy EBV carriers and infected monkeys determined by this HCIS-based assay in B cells and epithelial cells correlated highly with titers measured by the FCM-based assay. Finally, B cell neutralizing titers correlated with anti-gp350 IgG titers while anti-gHgL IgG titers correlated with epithelial cell neutralizing titers. This HCIS assay is a practical test with high-throughput potential, which will aid and facilitate further development of prophylactic vaccines and therapeutic treatments against EBV.
Methods
Human specimens
Sera were collected from age 40 to 60 EBV positive healthy carriers (VCA-IgM (−), VCA-IgG ( +), EBNA1-IgG ( +) and EBNA2-IgG (weak or -)) and their gender was documented by the investigators. This study was approved by the Institutional Ethics Committee of the Sun Yat-sen University Cancer Center, Guangdong, China. Written informed consent was obtained from all participants.
Cells lines
All cell lines were cultured at 37 °C in humidified air containing 5% CO
2. Akata cells (EBV negative, B cells) and HNE1 cells (EBV negative, epithelial cells) [
41] were cultured in RPMI 1640 (Invitrogen) with 10% fetal bovine serum (FBS; Invitrogen), and antibiotics (penicillin, 100 U/ml; streptomycin, 100 μg/ml; Invitrogen). CNE2-EBV cells (epithelial cells) [
42] and Akata-EBV cells (B cells) [
43], were propagated in RPMI 1640 (Invitrogen) with 10% FBS (Invitrogen) and antibiotics (penicillin, 100 U/ml; streptomycin, 100 μg/ml; Invitrogen), and maintained under G418 selection (700 μg/ml; MP Biomedicals).
Virus production
CNE2-EBV cells carrying the Akata-EBV-GFP genome were induced by 20 ng/ml 12-O-tetradecanoylphobol 13-acetate (TPA; Beyotime) and 2.5 mM sodium butyrate (NaB; Sigma Aldrich) for 12 h. After 72 h in culture, the supernatant was collected, centrifuged and then filtrated through a 0.45 μm filter to remove cell debris. The resulting virus, named CNE2-EBV-GFP, was concentrated 100 × by centrifugation at 50,000 g for 2.5 h and re-suspended by RPMI 1640 without FBS. The CNE2-EBV-GFP virions were stored at − 80 °C. A NIKON Eclipse Ti2-U microscope was used to capture images of non-induced CNE2-EBV cells as well as induced cells at 72 h post induction.
Akata-EBV cells carrying Akata-EBV-GFP were resuspended in RPMI 1640 without FBS and induced by 0.8% (v/v) goat anti-human IgG (Tianfun Xinqu Zhenglong Biochem. Lab). The medium was changed after 6 h induction. The Akata-EBV-GFP virus collection procedures and storage were the same as those used for the CNE2-EBV-GFP virus. A NIKON Eclipse Ti2-U microscope was used to capture images of non-induced Akata-EBV cells as well as induced cells at 72 h post induction.
Transmission electron microscopy
EBV virions were observed by negative staining electron microscopy. Briefly, viral samples were applied to 200-mesh carbon-coated copper grids for 5 min. The excess solution was removed, grids were washed twice with double distilled water and immediately stained for 30 s with freshly filtered 1.6% phosphotungstic acid (pH 6.5). Grids were examined using an FEI Tecnai T12 TEM (FEI, USA) at an accelerating voltage of 120 kV and photographed at a magnification of 150,000 and 250,000 fold.
CNE2-EBV-GFP virus titers definition by FCM
1 × 104 Akata cells were seeded in each well of a 96-well plate in 180 μl RPMI 1640 with 10% FBS and incubated with 20 μl of twofold serially diluted CNE2-EBV-GFP virus at 37 °C in a 5% CO2 humidified atmosphere. After 48 h incubation, cells were collected by centrifugation at 500 g for 5 min and washed once with PBS. Cells were resuspended in PBS without fixation for observation. The infection efficiency (percentage of GFP-positive cells) was determined using a CytoFLEX S (Beckman Coulter) and analyzed using FlowJo software X 10.0.7 (Tree Star). Half maximal infection dilution fold (ID50) was determined by GraphPad Prism 8.0.
CNE2-EBV-GFP virus titers definition by HCIS
1 × 104 Akata cells were seeded in a 96-well plate in 180 μl RPMI 1640 with 10% FBS and incubated with 20 μl twofold serially diluted CNE2-EBV-GFP virus at 37 °C in a 5% CO2 humidified atmosphere. After 48 h incubation, the plate was shaken to disperse the cells and let them be evenly distributed in the well. Images were captured and the total GFP positive spots of each well were calculated using the Operetta CLS high content imaging system (PerkinElmer).
Neutralizing titers evaluated by FCM
For B cell neutralization, 20 μl tenfold serially diluted monoclonal antibody 72A1 (starting from 100 μg/ml) or fivefold serially diluted sera from healthy EBV carriers or monkey (starting from 1:10) were mixed and incubated with 20 μl CNE2-EBV-GFP (a dose sufficient to infect 20% of cells) for 2 h at 37 °C in a 5% CO2 humidified atmosphere. The mixture was added to 1 × 104 Akata cells and incubated for 48 h. Uninfected cells were used as negative controls and cells incubated with CNE2-EBV-GFP in the absence of antibody or sera were used as positive controls. The infected cells were counted using a CytoFLEX S (Beckman Coulter) and analyzed using FlowJo software X 10.0.7 (Tree Star). The neutralizing activity of each sample was calculated as (%GFP positive cells of positive control–%GFP positive cells of samples with antibody or sera) × 100/ %GFP positive cells of positive control. Half maximal neutralizing concentrations (NC50) for monoclonal antibody or half maximal neutralizing dilution folds (ND50) for sera were determined by GraphPad Prism 8.0.
For epithelial cell neutralization, 20 μl twofold serially diluted sera from healthy EBV carriers (starting from 1:10) were mixed and incubated with 20 μl Akata-EBV-GFP for 2 h at 37 °C in a 5% CO2 humidified atmosphere. The mixture was added to 0.4 × 104 HNE1 cells and the medium was changed after 3 h. The following steps of data collection, analysis and calculation were the same as for the B cell neutralization.
Neutralizing titers evaluated by HCIS
For the B cell neutralization model, 20 μl tenfold serially diluted monoclonal antibody 72A1 (starting from 100 μg/ml) or fivefold serially diluted sera from healthy EBV carriers or monkey (starting from 1:10) were mixed and incubated with 20 μl CNE2-EBV-GFP virus for 2 h at 37 °C in a 5% CO2 humidified atmosphere. The mixture was added to 1 × 104 Akata cells and incubated for 48 h. Uninfected cells were used as negative controls and cells incubated with EBV in the absence of antibody or sera were used as positive controls. After 48 h incubation, the plate was shaken to disperse the cells to obtain even distribution in the wells. Images were captured and GFP positive spots were counted by Operetta CLS high content imaging system (PerkinElmer). The neutralizing rate of each sample was calculated as (number of total GFP positive spots of positive control–number of total GFP positive spots of samples with antibody or sera) × 100/ number of total GFP positive spots of positive control. Half maximal neutralizing concentrations (NC50) for monoclonal antibody or half maximal neutralizing dilution folds (ND50) for sera were determined by GraphPad Prism 8.0.
For the epithelial cell neutralizing model, 20 μl twofold serial diluted healthy EBV carriers sera (starting from 1:10) were mixed and incubated with 20 μl Akata-EBV-GFP for 2 h at 37 °C in a 5% CO2 humidified atmosphere. The mixture was added to 0.4 × 104 HNE1 cells and the medium was changed after 3 h. The following steps of data collection, analysis and calculation were the same as for the B cell neutralization model.
Enzyme-linked immunosorbent assay (ELISA)
Wells of 96-well ELISA plates (Corning) were coated with 100 ng/well gp350 or gHgL in PBS by incubation at 37 °C for 2 h. After washing with TBST (Tris Buffered Saline with Tween 20), blocking buffer (PBS containing 0.5% casein, 2% gelatin and 0.1% ProClin 300, pH 7.4) was used to block plates for 2 h at 37 °C. Five-fold serially diluted sera from monkeys or healthy EBV carriers (starting from 1:100) were added to each well, incubated for 1 h at 37 °C and then washed 5 times with TBST. Goat anti-human antibody conjugated with HRP (Promega) was added (1:5000 dilution) and incubated for 30 min at 37 °C. The colorimetric reaction was developed using the EL-TMB kit (Sangon Biotech). Absorbance was measured at 450 nm and 630 nm using a microplate reader (Molecular Devices).
Statistics
The Spearman correlation coefficient was used to evaluate the correlation between the results of different assays.
Discussion
Here we present a sensitive, high-throughput and robust HCIS-based approach to determine EBV infection titers, as well as neutralizing titers of sera or monoclonal antibodies against infection of B cells and epithelial cells. We validated this new HCIS-based assay by comparing its output to that of an established FCM-based assay [
35]. We observed consistent and constant agreement between the two assays in the determination of viral titers, monoclonal antibody neutralizing titers and sera neutralizing titers. We used HCIS to illustrate the significant correlation between gp350 IgG titers and B cell neutralizing titers across multiple sera from healthy EBV carriers and rhLCV-infected monkeys. Likewise, anti-gHgL IgG titers were correlated strongly with epithelial cell neutralizing titers in sera from multiple healthy EBV carriers. This HCIS-based assay can be applied more easily than FCM to high-throughput settings. This assay will be particularly efficient (i) to determine neutralizing titers of large-scale sera samples after vaccine inoculation in pre-clinical and clinical studies, (ii) to screen monoclonal antibodies and characterize their specific neutralization activity, and (iii) to assess the efficacy of EBV-specific antivirals to block B cell or epithelial cell infections. In addition, this HCIS-based assay will facilitate serological and epidemiological studies for large-scale samples to investigate the correlation between neutralizing titers and diseases outcome.
Previously reported methods to determine EBV-specific neutralizing titers include B cell transformation inhibition [
32], immunofluorescent-based assay [
33], competitive ELISA [
34], FCM-based neutralization assay [
35] and FIA-based neutralization assay [
36,
37]. Comparatively, FIA-based assays are more amenable to high-throughput settings. However, the published FIA assay relied on a CR2 overexpressing epithelial cell line to mimic B cell infection by EBV. Considering the dual tropism of EBV, it is necessary to consider a high-throughput approach applicable to B cells as well as epithelial cells. The HCIS-based assay described here has been validated in those two settings. First, the EBV-negative B cell line Akata was used in combination with the epithelial cell-derived CNE2-EBV-GFP virus. Second, the EBV negative epithelial cell line, HNE1 was used in combination with the B cell-derived Akata-EBV-GFP virus. These two models more realistically simulate the process of EBV natural infection. Neutralizing titers determined under these conditions are therefore more reliable. This assay also allowed accurate testing neutralization of antibodies against glycoproteins involved in infection of B cells (i.e. gp350) or epithelial cells (i.e. gHgL). However, the premise of an accurate analysis of GFP positive cells with HCIS is that the cells are single dispersed and evenly distributed in the wells. Therefore, epithelial cells need to be cultured in 96 well plates at a lower density for this assay. For B cells infection model, cells are clustered around the edges because of the edge effect of 96 well plates. It is necessary to shake the plate to disperse the cells and let them be evenly distributed in the well before imaging. Otherwise, the results will not be accurate.
Antibodies targeting different EBV proteins are raised with different peak times after infection [
45]. Importantly, high levels of neutralizing titers and high anti-gp350 IgG titers are considered low risk biomarkers for the development of NPC [
49]. It is known that gp350-specific neutralizing antibodies are the major contributors to B cell neutralization in healthy individuals [
17,
47]. Indeed, in this study B cell neutralizing titers of healthy EBV carriers determined by HCIS correlated strongly with anti-gp350 IgG titers determined by ELISA. A similar positive correlation was observed in sera from monkeys infected with rhLCV. Although no gp350-based vaccine has been approved yet, gp350 remains a major candidate for vaccine development when combined with a more efficient adjuvant such as AS01
B, Matrix-M and 3 M-052 [
50‐
52]. As for epithelial cell infection, gHgL specific neutralizing antibodies contributed to ~ 75% of the neutralizing activity [
17]. Five monoclonal antibodies targeting gHgL have been reported, which are AMMO1 (human) [
24], 6H2 (mouse) [
25], 1D8 (human) [
26], CL40 (mouse) [
27] and CL59 (mouse) [
27]. AMMO1 binds to gH domain I and II, 6H2 binds to gH domain IV and 1D8 binds to gH domain II. All three antibodies potently neutralize both B cell and epithelial cell infection. On the other hand, CL40 (domain II) and CL59 (domain I) only efficiently block epithelial cell infection. Interestingly, AMMO1, 1D8 and 6H2 antibodies protected humanized mice against EBV infection while 72A1 (against gp350) failed to reduce viral load in vivo [
25,
26,
53]. Here, using HCIS, we also demonstrated a strong correlation between anti-gHgL IgG titers and epithelial cell neutralizing titers in sera from multiple healthy EBV carriers. The gHgL complex participates in the infection process of epithelial and B cells as an activator of the membrane fusion effector gB. Consequently, gHgL needs to be taken into account for vaccine design. Indeed, antibodies induced by gHgL-ferritin nanoparticles were highly efficient at neutralizing infection of epithelial cells [
17]. gHgL-ferritin nanoparticles induced neutralizing antibodies in BALB/c mice and cynomolgus macaques and antibodies purified from immunized mice passively protected humanized mice from lethal EBV challenge [
17,
54].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.