Infants Younger Than 24 Months With Severe Malaria Have a Serologic Profile of Active Infection With Epstein-Barr Virus Authors: Wayne T Cheng, Balotin Fogang, Aarti Jain, D Huw Davies, Philip L Felgner, Carole Eboumbou, Paul N Koki, Samuel H Speck, Chester J Joyner, Lawrence S Ayong, Tracey J Lamb Published: 2026 ABSTRACT AbstractBackgroundPrimary Epstein-Barr virus (EBV) infection occurs during late adolescence and is characterized by the symptomatic manifestation of infectious mononucleosis. Asymptomatic primary EBV infection in malaria-endemic areas often occurs in young children by the age of 2 years. Resulting in humoral immune suppression to unrelated antigenic challenges for approximately 4 weeks.MethodsWe undertook a cross-sectional study of 195 infants aged 6 to 24 months in Cameroon. The EBV infection status of each child was determined by a standard serologic classification system, and the magnitude, breadth, and invasion-blocking capacity of the anti–Plasmodium falciparum antibody response were quantified.ResultsAn overall 24% of children were serologically positive for active EBV infection, and the highest proportion of severe malaria cases was in children with active EBV. An elevated magnitude and breadth of the antibody response with increased in vitro invasion-blocking capacity were observed in children with active EBV, but circulating parasitemia in vivo was similar.ConclusionsPrimary EBV infection may be a risk factor for developing severe malaria in children aged 6 to 24 months. Targeting EBV infection in young children may be beneficial in protecting against the development of severe falciparum malaria in children living in malaria-endemic areas. Every year approximately 90% of the 250 million clinical cases of malaria are confined to sub-Saharan Africa, resulting in >600 000 deaths with children aged <5 years and pregnant women being the most affected. Plasmodium falciparum (Pf), the most virulent malaria parasite species, is responsible for most of the infections and fatalities, as attributed to untreated severe syndromes such as severe malarial anemia, acute respiratory distress syndrome, and cerebral malaria [1]. Exposure to endemic malaria and primary Epstein-Barr virus (EBV) infection occurs in early childhood via exposure to maternal saliva and breast milk [2]. In sub-Saharan Africa, up to 97% of children have experienced primary EBV infection by 12 months of age [3] after maternal antibodies have waned [4]. Immunity that controls Plasmodium infection levels and clinical symptoms is thought to build over time with repeated infections. Immunity to severe malaria, in particular noncerebral severe malaria, develops after just 1 or 2 Plasmodium infections in areas of high transmission [5]. Not all children develop severe disease [6]. It is not fully understood why some infants cannot acquire immunity to severe, lethal disease once protective maternal antibodies wane. Primary EBV infection is generally asymptomatic in young children [7] but can lead to suppression of humoral immunity to unrelated antigens in young children and adults [8, 9]. Several studies have shown impaired anti-EBV cytotoxic T-cell responses in children who are coinfected [10, 11], and this has been associated with higher EBV viremia [12, 13] and reactivation [14]. In Gabon, increased EBV viremia has been associated with increased rates of reinfection with Pf in children who are coinfected [14]. The highest Pf parasitemia levels are observed in children aged 6 to 12 months and coincide with the age range where most children in sub-Saharan Africa become infected with EBV [15]. Interactions between EBV and Plasmodium infections can lead to the loss of immune mechanisms preserving viral latency and the development of an EBV-related malignancy known as endemic Burkitt lymphoma [16]. However, few details are known regarding the potential impact of primary EBV infection on the severity of Pf infections. Humoral immunity is correlated with protection against the uncontrolled growth of Pf in infected red blood cells (iRBCs) [17]. Published data argue that the generation of antibodies reacting to parasite proteins, such as the variant antigen Pf erythrocyte membrane protein 1 on the iRBC surface [18], is a critical mechanism of antimalarial immunity [19]. In some studies, passive transfer of antibodies from adults immune to malaria has been shown to reduce parasitemia in children with severe malaria [19]. Antibodies contribute to the control of blood stages by several mechanisms, including the promotion of antibody-dependent phagocytosis of iRBCs [20] and blocking invasion of RBCs [21]. This nonsterilizing immunity to malaria is short-lived and requires constant reinfection to be maintained [22]. In children, short-lived antibody responses are not always boosted upon reinfection with Plasmodium [23], suggesting that the establishment of memory may be defective. However, no mechanism has been defined. Because immunity to malaria infection is generated when EBV exposure and infection occur, understanding the role of EBV in developing antimalarial immunity to prevent severe disease in early life is essential. Here, we show that active EBV infection increases the likelihood of developing severe malaria in 6- to 24-month-old children. This effect was not seen in children who were serologically negative for EBV or had serologic evidence of latent EBV infection. Although this was associated with an increase in circulating parasite density, active EBV infection appeared to increase the magnitude and breadth of the anti-Pf humoral response that could better block invasion of Pf into new RBCs, suggesting that primary EBV infection did not impair anti-Plasmodium humoral responses. Thus, active EBV must interfere with another component of antimalarial immunity essential for preventing severe malaria. Nevertheless, this study shows that active EBV infection in young infants is a risk factor for developing severe malaria and supports that an efficacious EBV vaccine may protect infants from the consequences of endemic Burkitt lymphoma while reducing their risk for severe malaria. METHODS Ethical Approval The study was approved by the Cameroon National Ethics Committee on Human Health Research (ethical clearance 2016/01/683/CE/CNERSH/SP), the Emory University Institutional Review Board (IRB00084993), and the University of Utah Institutional Review Board (IRB_00099332). Administrative research authorization was obtained from the Ministry of Public Health, Cameroon (6310716). Study Design This cross-sectional study was carried out between 2015 and 2017 in 4 district hospitals (Efok, Obala, Efoulan, and Nkoleton) and 1 pediatric hospital (Centre Mère et Enfant Chantal Biya) in the central region of Cameroon, where malaria transmission is perennial and hyperendemic, with Pf responsible for >99% of malaria cases [24]. This study included children aged ≤24 months who presented at the emergency units for consultation with at least 1 symptom of malaria as diagnosed by the clinical officer. Not all children with fever were found positive for Pf, and these cases were classified as malaria negative with fever attributed to other causes. Children with evidence or history of meningitis, encephalitis, developmental delay, or other neurologic conditions were excluded. The demographics of the study population are shown in Supplementary Table 1. Plasmodium infections were diagnosed by Pf/panantigen rapid diagnostic testing (Standard Q Rapid Test; SD Biosensor) and microscopy according to standard World Health Organization procedures. Severe malaria was defined according to World Health Organization criteria—for our study, the presence of Pf infection in circulation with 1 or more of the following: severe anemia (<5 g/dL of hemoglobin), <15% hematocrit, neurologic attacks, convulsions, coma, respiratory distress, and prostration. Uncomplicated malaria was defined as the presence of fever or a history of fever in the last 48 hours with Plasmodium-positive thick blood smears but without signs of severe malaria. Children with Plasmodium infection (severe or uncomplicated) were treated according to national guidelines. Those who were negative on thick smears and rapid diagnostic tests were used as uninfected/subpatent controls. Samples from adults aged >20 years living in Esse Health District, Centre Region, Cameroon [25], were used as positive controls for studying the acquired immune response. Venous blood samples were collected in Ethylenediaminetetraacetic acid (EDTA) tubes before antimalaria treatment and transported in refrigerated boxes to the Molecular Parasitology Laboratory of Centre Pasteur Cameroon. Plasma was obtained by centrifugation at 3000 rpm for 5 minutes and immediately stored at −80 °C. Anti-Pf Enzyme-Linked Immunosorbent Assay Plasma antibody levels were determined by enzyme-linked immunosorbent assay (ELISA) using Pf protein soluble extracts as previously described [25]. Total soluble antigens were extracted from the Pf 3D7 following freeze-thaw fractionation procedures [25]. EBV Diagnostic ELISAs ELISAs to EBV viral capsid antigen (VCA) and EBV nuclear antigen 1 (EBNA-1) were quantified by commercial ELISA kits according to the manufacturer's recommendations (GenWay Biotech). The applied cutoff was 10 U/mL for all 3 ELISAs (VCA IgM, VCA IgG, and EBNA-1 IgG). Infection status (negative, active, or latent infection) was defined as previously described [26] (Supplementary Table 2). Pf Protein Microarray Analysis All sequences for this array were derived from a single reference genome (3D7). Microarrays (Antigen Discovery Inc) were printed and probed following previously published methods [27]. Plasma samples were diluted 1:200 in Protein Array Blocking Buffer (GVS Life Sciences) supplemented with 10% Escherichia coli lysate (Antigen Discovery Inc) to block any reactivity against E coli in the arrayed in vitro transcribed/translated (IVTT) proteins. Bound IgG and IgM were then visualized with appropriate secondary antibodies. Slides were scanned on a GenePix 4200AL (Molecular Devices) and spot intensities quantified in ScanArray Express software (Perkin Elmer). The pixel intensity for each antigen was corrected by subtracting background based on the average of “control spots” containing the components of IVTT reactions lacking plasmid template. The minimum pixel intensity value +1 was then added to the entire dataset to ensure that all intensities were positive. These intensities were used to calculate the estimated area under the curve using the trapezoidal method for each sample. Mean intensities were considered a proxy for the titer of IgG and IgM present against Pf antigen, and the estimated area under the curve was used as a proxy for the total IgG and IgM against all Pf antigens in the array. The seropositivity rate for each sample of antigens was determined by using the corrected intensity data to calculate the foldover IVTT control (FOC). FOC normalization provides a relative measure of the specific antibody binding to the nonspecific antibody binding to the IVTT controls. Antigens with an FOC value >2 were considered seropositive for IgG and IgM. Growth Inhibition Assay Asexual stages of Pf NF54 (Patient E isolate) were cultured by standard methods. Trophozoites and early schizonts were magnetically enriched and resuspended to 1% parasitemia in human blood. The samples were transferred to 96 well-round bottom plates (Greiner Bio-One), and the hematocrit was adjusted to 5% using media containing plasma from the available pediatric samples at a 1:200 dilution. Cytochalasin D at 10 µM and malaria-naive serum from individuals in the United States were used as positive and negative controls, respectively. The initial parasitemia at plating and after invasion about 24 hours after plating was quantified with flow cytometry by staining cultures with 10-µg/mL Hoechst 33342 before acquisition. Percentage inhibition was calculated by subtracting background events from an uninfected/subpatent sample and dividing the percentage of infected erythrocytes with normal human serum from the negative control. This number was then converted to a percentage and subtracted from 100 to determine the growth inhibitory capacity for each sample. Statistical Analyses Data were log10 or arcsine transformed to improve model fit for statistical analyses. Contingency tables followed by chi-square tests were used to assess the relationship between EBV status and malaria infectious status. To evaluate the impact of EBV status on antibody responses to Pf infection, a 1-way analysis of variance, followed by a Fisher least significant difference post hoc test, was used to determine statistical significance between groups. Quantile regressions were performed to estimate the effect of malaria infection status on Pf antibody responses within EBV groups. A generalized regression was used to assess the efficacy of plasma samples on parasite invasion blocking. Comparisons were considered significant when least significant difference– or false discovery rate–adjusted P values for multiple-hypothesis correction were <.05. Statistical tests were performed with JMP 17 Pro. RESULTS Children With Active EBV Infection Are More Likely to Have Severe Malaria Participants were grouped by serologic responses to VCA and the latency-associated protein EBNA-1 [26], where it is generally easy to distinguish active and past infections in individuals who are immunocompetent. Children negative for EBV were negative for both antigens. A serologic response to either VCA or EBNA-1 was classified as EBV exposed. In this study, active EBV infection was defined as a positive serologic response to VCA but not EBNA-1. Latent infection was defined as lacking an IgM response to VCA and/or IgG response to EBNA-1. Within latent infection, we identified individuals who also had a response to VCA but not EBNA-1 (early latent or lack of conversion to latent) or were positive for IgM VCA, IgG VCA, and IgG EBNA-1, which may have resulted from reactivation of a latent infection. For classification purposes, samples were removed from the analysis in cases of ambiguity, as follows: (1) VCA IgG in the absence of VCA IgM and EBNA-1 IgG; (2) the simultaneous presence of VCA IgG, VCA IgM, and EBNA-1 IgG; and (3) the presence of EBNA-1 IgG in the absence of VCA IgG and IgM. These classifications are summarized in Supplementary Table 2. In the cohort of individuals, 50.7% were serologically positive for EBV infection, with 24% having an active EBV infection (Figure 1A). The breakdown of EBV categorization according to malaria status is shown in Table 1. Similar to mouse models of primary gammaherpesvirus-Plasmodium coinfections [28], the presence of active EBV infection was associated with severe malaria in our cohort (Figure 1B). Although children who were serologically negative for EBV tended to be younger (Figure 1C), the circulating parasitemia in individuals who were microscopically positive did not differ from older children who had latent EBV infection (Figure 1D). However, active EBV infection was associated with significantly higher circulating Pf parasitemia as compared with negative or latent cases, and as expected, adults had lower parasitemia than infants. Active Epstein-Barr virus (EBV) infection is associated with a greater probability of a severe malaria episode in infants <24 months old. A, Distribution of EBV status groupings in children aged 6 to 24 months. B, Mosaic plot of malaria status within EBV groups. C, Distribution of ages by EBV status. D, Distribution of circulating Plasmodium falciparum–infected red blood cells in children by EBV status. Children were positive for malaria if P falciparum infection was detected by either rapid diagnostic testing or microscopy. Parasitemia values for children who were rapid diagnostic test positive but microscopy negative were considered below the limit of detection and given an arbitrary value of 5. Adults served as positive controls for all experiments. Error bars represent SEM. *P < .05 by analysis of variance with least significant difference pairwise comparisons. Table 1.Cohort Characterization of EBV on Plasmodium Infection in Young ChildrenCharacteristicChildren (n = 150) aAdults (n = 34)Malaria Malaria uninfected86 (56.6)10 (29.4) Malaria infected66 (43.4)24 (70.6)EBV Negative74 (46.7)… Past40 (26.3)13 (38.2) Recent or current38 (25.0)21 (61.8)Coinfection Malaria/negative EBV22 (14.7)… Malaria/latent EBV16 (10.7)10 (29.4) Malaria/active EBV27 (18.0)14 (41.2) Malaria (−) or subpatent/negative EBV52 (34.7)… Malaria (−) or subpatent/latent EBV24 (16.0)3 (8.8) Malaria (−) or subpatent/active EBV9 (6.0)7 (20.6)Data are presented as No. (%).Abbreviation: EBV, Epstein-Barr virus. aAged 6 to 24 months. Cohort Characterization of EBV on Plasmodium Infection in Young Children Data are presented as No. (%). Abbreviation: EBV, Epstein-Barr virus. aAged 6 to 24 months. Children With Active Infection Had the Highest Anti-Pf Response Antibody Response A previous study in a mouse model of gammaherpesvirus-Plasmodium infection [28] turned a nonlethal mouse model of Plasmodium infection into a lethal infection due to suppression of humoral responses. Thus, we hypothesized that children with an active EBV infection would have an inferior antibody response to Pf antigens. Contrary to the hypothesis, the antibody responses against Pf were increased in children with active EBV (Supplementary Figure 1). Because the ELISA measured IgM, IgG, and IgA simultaneously against lysates, the samples were then probed against protein microarrays, which enabled the assessment of the magnitude and breadth of the anti-Pf response for IgG and IgM. Similar to the ELISA, children with active EBV infection had a greater magnitude of Pf-reactive IgM (Figure 2A) and IgG (Figure 2B). The breadth of the IgM (Figure 2C) and IgG (Figure 2D) response was also greater in children with active EBV infection. As expected, adults had the highest antibody measurements as compared with children. Infants with active Epstein-Barr virus (EBV) infection had more anti–Plasmodium falciparum (anti-Pf) reactive antibodies relative to those with latent EBV or those who had no serologic evidence of EBV infection. Comparison of intensity by estimated area under the curve (eAUC) among EBV status groups: A, IgM; B, IgG. Comparison of breadth by EBV status: C, IgM; D, IgG. Adults served as positive controls for all experiments. Error bars represent SEM. *P < .05 by analysis of variance with least significant difference pairwise comparisons. Breadth of the Anti-Pf Response Is Associated With EBV Infection Independent of Age When EBV status was subsetted by malaria status and severity, latent and active EBV infection were both associated with an increased magnitude of the antibody response against Pf lysates (Figure 3A), and active EBV infection was associated with greater breadth of the IgM and IgG responses (Figure 3B and 3C, Table 2). Age did not contribute to anti-Pf humoral responses measured within the cohort of this study. This observation is not surprising given the age bracket of children included in the study, and it is in agreement with published studies [29]. Thus, the intensity and breadth of the anti-Plasmodium reactive humoral response were driven by Plasmodium infections in latent and active EBV infections. Symptoms of malaria are associated with greater humoral anti–Plasmodium falciparum (anti-Pf) responses in infants infected with Epstein-Barr virus (EBV). A, Enzyme-linked immunosorbent assay measurement of total anti-falciparum IgG in plasma from adults or children by malaria status within EBV groups. Comparison of breadth by malaria status within EBV groups: B, IgM; C, IgG. Adults served as positive controls for all experiments. *P < .05 by quantile regression with least significant difference– or false discovery rate–adjusted P values for multiple-hypothesis correction. Error bars represent SEM. OD, optical density. Table 2.Effects Tests of Anti-Pf Antibody Responses With Epidemiologic Characteristics in Young Children With EBV InfectionEffect TestSum of Squares F RatioProb > F aIgM intensity Age, mo0.00010.1323.7166 Weight0.00081.9135.1690 Hb, g/dL0.00040.9259.3378 HCT, v/v %0.00030.6825.4103 Body temperature, °C<.00010.0004.9849 EBV status0.012315.2896<.0001IgM breadth Age, mo0.59521.9314.1671 Weight0.08370.2716.6032 Hb, g/dL0.21780.7067.4022 HCT, v/v %0.12430.4033.5266 Body temperature, °C0.12930.4197.5183 EBV status16.379926.5787<.0001IgG intensity Age, mo0.00020.1596.6902 Weight0.00090.8262.3651 Hb, g/dL0.00131.1095.2942 HCT, v/v %0.00151.3511.2473 Body temperature, °C0.00161.4555.2299 EBV status0.01627.1725.0011IgG breadth Age, mo0.33370.7876.3765 Weight0.02940.0693.7927 Hb, g/dL0.11550.2726.6025 HCT, v/v %0.16150.3812.5381 Body temperature, °C0.87162.0572.1540 EBV status4.10014.8388.0095Abbreviations: EBV, Epstein-Barr virus; Hb, hemoglobin; HCT, hematocrit. aStandard least squares regression to identify active predictors in antibody breadth and intensity. Effects Tests of Anti-Pf Antibody Responses With Epidemiologic Characteristics in Young Children With EBV Infection Abbreviations: EBV, Epstein-Barr virus; Hb, hemoglobin; HCT, hematocrit. aStandard least squares regression to identify active predictors in antibody breadth and intensity. Plasma From Infants With Active EBV Who Were Parasitemic Exhibits Better Invasion-Blocking Capacity Since antibody magnitude and breadth do not necessarily predict neutralization, a growth inhibition assay was performed with samples where enough plasma was available. The greater magnitude and breadth of humoral response in children with active EBV infection were associated with a better invasion inhibition capacity in vitro and were similar to adults (Figure 4A). This was specific to Pf infection because individuals with active EBV infection who were negative for Pf per rapid diagnostic test and microscopy did not inhibit parasite growth in vitro (Figure 4B). Surprisingly, the best correlate of invasion-blocking capacity was IgM magnitude and breadth (Table 3), particularly in children who were positive for Plasmodium infection (Figure 4C). Active Epstein-Barr virus (EBV) infection is associated with greater in vitro invasion inhibition in children positive for Plasmodium falciparum infection. A, P falciparum NF54 growth inhibition by plasma among EBV status groups. Data were averaged between 2 experimental replicates. Error bars represent SEM. *P < .05 by 1-way analysis of variance with a Fisher least significant difference post hoc test. Comparison within EBV groups: B, Plasmodium infection status;C, malaria disease status. Adults served as positive controls for all experiments. *P < .05 by a generalized regression with a false discovery rate–adjusted post hoc test. Table 3.Correlation of Plasmodium falciparum Invasion Blocking With Antibody Responses in Young Children With Active EBV InfectionInfection StatusSpearman ρProb > |ρ|FDR-Adjusted P ValuePrimary EBVIgM Intensity0.59120.0001.0006 Breadth0.47180.0037.0074IgG Intensity0.28490.0921.1228 Breadth0.20170.2382.2382Malaria statusPositive IgM breadth0.22810.2526.2887 IgG breadth−0.32970.0931.2470Negative IgM breadth0.54390.1301.2470 IgG breadth0.65000.0581.2323Abbreviations: EBV, Epstein-Barr virus; FDR, false discovery rate. Correlation of Plasmodium falciparum Invasion Blocking With Antibody Responses in Young Children With Active EBV Infection Abbreviations: EBV, Epstein-Barr virus; FDR, false discovery rate. DISCUSSION The overlap in endemicity between Pf infections and EBV in children is associated with the development of endemic Burkitt lymphoma [16], a nonaggressive non-Hodgkin lymphoma that manifests in 5- to 9-year-old children in sub-Saharan Africa [30]. However, the effect of active EBV infection on Pf infections has not been reported despite the possibility of immunologic interactions when young children are coinfected [31]. As shown in a mouse model of gammaherpesvirus-Plasmodium infection [28], the severity of malaria can be increased in infants with active EBV infection. The hypothesis that EBV may suppress antibody responses to an incoming infection rests on the fact that EBV largely resides in germinal center B cells [32]. Although there was some evidence suggesting that Plasmodium vivax responses were impaired in a study involving 56 adults living in a P vivax–endemic area of the Amazon, there has never been a thorough analysis undertaken when infants first become infected with EBV [33]. Contrary to mouse MHV68–Plasmodium yoelii XNL infections [28], the mechanism does not appear to be due to EBV-induced suppression of anti-Pf humoral responses. In contrast, our data suggest that EBV infection augments anti-Pf humoral responses, in particular the magnitude and breadth of the IgM response, which can exert substantial invasion-blocking capacity in vitro. Despite an augmented IgM response that exhibits increased invasion-blocking capacity in vitro (Figure 4), circulating parasitemia was higher in infants presenting with active EBV infection. There are several possible explanations underlying this finding. EBV infection may induce suppression of immune mechanisms other than antibody production, which may also contribute to the control of Pf iRBCs. These could include modulation of innate responses [34]. For example, EBV is known to inhibit interferon regulatory factors to prevent induction of type I interferons [35]. Although type I interferons have been shown in mouse models to suppress priming of Tfh by B cells, resulting in reduced IgM and IgG responses [36], type I interferon is needed to augment CD4 T-cell–derived IFN-γ [37], which activates several parasite clearance mechanisms but is also correlated with pathogenesis [38]. Additionally, EBV transcribes a homolog of IL-10 [39] that likely serves to increase B-cell proliferation and expand the latency reservoir but can suppress Th1 responses. IL-10 helps protect against the pathogenesis of malaria but may impair control of blood stage parasitemia. The serum concentration of IL-10 during infection with Pf is associated with higher parasite densities and less effective parasite clearance in Tanzanian children [40]. This study departs from the current thinking that the main consequence of infant primary EBV infection is simply the eventual development of endemic Burkitt lymphoma in some children and suggests that it could be a risk factor for the development of severe malaria. The outcome of this work may extend to other childhood infections prevalent in sub-Saharan Africa, particularly related herpesviruses such as Kaposi sarcoma–associated herpesvirus and cytomegalovirus, but further studies are required. Supplementary Material