Malaria – how this parasitic infection aids and abets EBV-associated Burkitt lymphomagenesis

Epstein-Barr virus and endemic Burkitt lymphoma

Since the discovery of Epstein-Barr virus (EBV) within a B cell tumor in 1964 [1], how EBV promotes B cell tumorigenesis has been extensively studied and are reviewed in detail elsewhere (reviewed in [2]). As a scientific community we are generally in agreement that EBV is a causative agent for several cancers, in that viral proteins, mircoRNA and epigenetics play a role in driving cell proliferation and rescue from apoptosis (reviewed in [3,4]). It also appears that EBV takes advantage of rare compensatory host cell mutations that disable apoptotic pathways in conjunction with the oncogenic c-myc translocation that drives unrestrained cell growth and proliferation [5••]. In immunocompetent individuals, EBV alone is insufficient for malignant transformation since this gamma-herpesvirus is a life-long and typically asymptomatic infection in most adults around the world [6]. How EBV promotes cancer is not the focus of this review, however, in the interest of the ensuing discussion it should be noted that healthy children residing in areas with a high incidence of endemic Burkitt lymphoma (eBL) experience their primary EBV infection, asymptomatically before 3 years of age if not within the first year of life [7••,8, 9], and holoendemic Plasmodium falciparum malaria exposure has been associated with higher frequencies of EBV reactivation and more episodes of viremia compared to early-age EBV-infected children not repeatedly co-infected with malaria [7,10–13•].

The postulated roles for falciparum malaria in the etiology of endemic Burkitt lymphoma

There are two synergistic mechanisms by which malaria is hypothesized to contribute to eBL etiology: 1) by inducing defects in immune surveillance to EBV antigens and thereby failing to limit the number of latently EBV-infected B cells during episodes of lytic reactivation; and 2) by inducing B cell activation that leads to EBV lytic reactivation as well as aberrant expression of activation-induced cytidine deaminase (AID) and thereby increasing the likelihood of a successful c-myc translocation triggering oncogenesis. Given this background, we will review the evidence gathered to date to address the long-standing question, ‘How precisely does malaria contribute to eBL tumorigenesis?” and highlight as yet unresolved questions.

Epidemiologic overlay of endemic Burkitt lymphoma on Plasmodium falciparum malaria

Plasmodium falciparum malaria being named as a co-factor in eBL etiology was suggested after Denis Burkitt’s famous ‘tumor safari’ which first described rainfall and altitude as being geographically associated with eBL incidence in Africa [14]. The suggestion that malaria played a role in eBL pathogenesis [15] was also made during the first global push for malaria eradication (1950-1970) and when hemoglobin (Hb) AS heterozygosity was discovered to protect individuals against severe malaria but resulted in sickle cell disease when people where HbAA homozygous [16]. It was subsequently postulated that severe malaria equated to a higher risk of eBL and an initial study appeared to show that HbAS may protect against eBL [17]. However, a series replication studies conducted over the course of 30 years and in three different African countries concluded that children diagnosed with eBL had the same frequency of HbAS heterozygosity as appropriately matched population-based controls [18–20]. In addition, the lack of documented family clusters of eBL, especially in families with many siblings, also argues against a strong, simple host-genetic predisposition. This lack of a link between severe malaria and eBL is also supported by considering the age-dependent epidemiology of both of these diseases. Age-structured modeling consistently demonstrates that severe malaria susceptibility decreases as a child reaches 5 years of age when residing in malaria holoendemic areas (reviewed in [21]). Of note, this is the age at which the incidence of eBL precipitously increases until it tapers off by 9 years of age (reviewed in [22]). Thus, the clinical and epidemiologic evidence points toward syndemic mechanisms that require prolonged malaria-induced perturbations of EBV homeostasis and immune surveillance in order to culminate in eBL tumorigenesis.

Early-age primary Epstein-Barr virus infection: implications for immune surveillance within the context of malaria co-infections

The prolonged time interval between primary EBV infection in African children and the induction of pediatric eBL could in part, be explained by the natural progression of acquired immunity to malaria and how this changes malaria from an acute to chronic infection. Anti-malarial immunity is intrinsically complex and depends on age at time of infection, cumulative exposure and genetic variation of the parasite [23]. Therefore, a simple Th1 versus Th2 immune response dichotomy does not necessarily apply to human malaria (reviewed in [24]). In addition, infants born to mothers residing in malaria endemic areas benefit from some degree of transplacentally transferred antibody-mediated immune protection (reviewed in [25]). As maternal antibodies naturally wane by around 6 months of age, the majority of infants and young children experience repeated, acute uncomplicated malaria infections, as opposed to manifestations of severe malaria (reviewed in [26]). With cumulative malaria exposure and immunologic maturation, that tends to occur around the age of 5 years, children develop premunition. This is also known as anti-disease immunity which allows children to tolerate chronic, asymptomatic parasitemias [21].

In keeping with the concept of dynamic human immune heterogeneity, recent studies demonstrate a reduced transfer of maternal antibodies against EBV and signs of increased viral reactivation when some mothers are infected with malaria during pregnancy [27,28]; which could result in earlier-age, higher viral load infections during infancy [10]. Acute, uncomplicated malaria has been associated with EBV lytic reactivation and a higher frequency of episodes of measurable viremia [7,10,12]. The apparent immune defects determined from infant cohort studies conducted in Kenya include decreased IFN-γ responses to EBV lytic and EBV latent antigens, in addition to skewed EBV-specific CD8+ T cell immune profiles that may not be as efficient at limiting expansion of EBV latency [13,29–31••]. The role of malaria in altering EBV-specific immunity appears to be an indirect effect whereby high EBV antigen load over time results in a degradation of EBV-specific immune surveillance and signals that prevent the development of immunologic effector-memory.

Effective control over EBV has been shown to be mediated by both innate and adaptive immunity when this infection occurs during adolescence or adulthood and has been shown to undergo temporal changes during the course of infection (reviewed in [32]). Preliminary studies in the Moormann lab suggest that natural killer (NK) cell subsets [33–35•] of children co-infected with malaria are significantly different in phenotype and function compared to age-matched, non-malaria exposed children, with additional defects apparent in children who are diagnosed with eBL (unpublished, Moormann). In addition, longitudinal infant cohort studies suggest that cytotoxic T cell mediators [32] differ in phenotype and function after cumulative, high burden exposure to malaria which may also prevent proper immune control over EBV (unpublished, Moormann). The delayed age of onset for eBL could possibly be explained by NK cells being more important in controlling EBV [34•,36] primary EBV infection during infancy, prior to the maturation of antigen presentation signaling pathways that promote the development of effector-memory T cells. Transcriptome and computational analyses of human immune cell subsets are starting to shed new light into which transcription factors regulate the selection and depletion of T cell subsets (reviewed in [37]) and can be applied within the context of EBV and malaria co-infections in children.

EBV-infected memory B cell susceptibility to malaria-induced aberrant activation-induced cytidine deaminase

If EBV-specific immune surveillance is sufficiently impaired and malaria has become a chronic infection, the stage is set for a prolonged assault on EBV-infected B cells. At this point, EBV latency within memory B cells has been established (reviewed in [4]) and viral proteins, LMP1 and LMP2 are mimicking host CD40 and B cell receptors (BCR), respectively providing signals for cell growth and down regulation of pro-apoptotic signals (reviewed in [3]). In addition, there is evidence that AID, which is the key to somatic hypermutation and class switch recombination for antibody generation in memory B cells (reviewed in [38]), can be expressed outside the germinal center environment as demonstrated by peripheral blood AID expression of children co-infected with malaria [39••] and EBNA3C directly inducing AID in B cells [40]. Malaria has also been shown to induce AID in a p53 deleted Eμ-mouse model and in vitro studies using human tonsillar cells [41•,42•]. A caveat to the Robbiani et al. study is that Plasmodium chabaudi mouse malaria appears not have a functional homologue to Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) which is also involved in parasite sequestration [24,43]. The cysteine-rich interdomain region 1 alpha (CIDR1a) of PfEMP1 has been identified as the T cell-independent polyclonal B cell activator, Ig binding protein, and inducer of EBV lytic reactivation [44••,45••]. Therefore, induction of AID within this model system could be mediated by inflammatory signals not specific to malaria [46,47]. Along this line, the more benign forms of human malaria P. vivax, P. ovale, P. malariae and P. knowlesi have not been implicated in eBL etiology. Further studies of AID induction across plasmodia species would be required to determine the antigen-specificity of malaria-derived red blood cell variant surface proteins [43]. In the study by Torbor et al, it appears that malaria induces AID expression by also engaging the innate toll-like receptor, TLR9 [48] and CD40 receptors as would be provided by cognate antigen-specific CD4+ follicular helper T cell signaling pathways. Malaria has been recently described to induce less functional Th1-polarized CXCR3+ follicular helper CD3 T cells in children [49]. If this defect influences the fate of T cell help for EBV-infected B cells within children remains speculative. In addition, the impact of activation dynamics or AID thresholds needed to achieve somatic mutations within a co-infection model remain to be explored, leaving these biologically relevant questions open to further investigation.

It remains controversial if EBV is able to establish latency in naïve memory B cells by bypassing the germinal center (GC) reaction (antigen-independent) or if eBL pathogenesis requires GC transit and is antigen-dependent [50,51]. More recent studies of human B cell subsets altered by malaria [52,53] and the identification of two possible routes by which memory B cells can undergo class switch recombination, GC-dependent or independent [54], compel us to revisit the question of which B cell compartments harbor EBV in children who develop eBL. Therefore, the relevance of BCR antigen-specificity in eBL pathogenesis remains unanswered. However, it is tempting to speculate that if the surface IgM of EBV-infected memory B cells encounter their cognate antigen, triggering a ‘secondary’ GC reaction and affinity maturation, this could provide malaria another point at which to influence B cell activation. Combined with the non-specific mechanisms described above to induce AID this dual stimulation by both EBV and malaria could thus create a supercharged environment for genetic instability and oncogenesis.

Is malaria the only parasite that could modulate immune surveillance that abets EBV-infected B cell tumorigenesis?

No. Schistosomiasis is an equally common chronic childhood infection in lake regions of Africa [55,56] where we find eBL, and it induces Th1-cytokines during early infection that dramatically shifts to Th2-cytokines during chronic infection [57]. S. mansoni antigens are known to induce a robust type 2 cytokine response, including production of IL-4, IL-5, IL-10, and IL-13 [58]. A pre-existing polarizing cytokine milieu is likely to affect NK and CD8+ T cell function prior to or during their activation by a viral infection [59–61]. Strong type I interferon signaling drives NK cell maturation to the terminal effector stage, producing an NK cell population with an impaired ability to respond to herpesvirus infection [60,62], whereas strong IL-4 signaling by CD8+ T cells is characterized by poor cytotoxicity and reduced secretion of IFN-γ [59]. Studies of NK cell education and tolerance during chronic infections reveal a broad range of pathogen-specific ligand interactions and highlight differences between early and late differentiated NK cell subsets in protection from EBV which contrast those engaged to respond to CMV [34•,35•,63]. It remains to be determined if loss of immune control over EBV early in life is mediated by NK cell tolerance or T cell clonal deletion or exhaustion and if the cytokine milieu induced by other parasitic co-infections plays a role in eBL pathogenesis. However, the possibility that other chronic parasitic infections engender immune dysfunction or exhaustion is open for debate [64].

Clarifying malaria’s role in endemic Burkitt lymphoma pathogenesis and questions remaining

In summary, malaria in not generally immunosuppressive but is a powerful driver of EBV-associated eBL pathogenesis. This review has gathered evidence in hopes of dispelling this lingering dogma. In fact, malaria is highly immunogenic (reviewed in [23]) and children diagnosed with eBL display robust immune responses to malaria antigens while deficient in EBV-specific immunity [65••]. The impact of chronic malaria on EBV-specific immune surveillance appears to be due to its ability to drive EBV lytic reactivation in B cells that repeatedly triggers the cascade of EBV lytic and latent antigen expression that eventually leads to impaired immunosurveillance and subsequent, unhindered interactions between malaria-derived ligands/proteins and EBV-infected B cells that are already primed and poised for oncogenesis. Figure 1 illustrates the synergistic mechanisms by which malaria contributes to impaired EBV immune surveillance and aberrant AID expression within an EBV-infected B cell.

In vitro cell culture and animal models have been instrumental toward increasing our understanding of EBV biology and mechanistic pathways responsible for B cell transformation. Yet these models systems are limited in their ability to resolve many remaining question as to how malaria contributes to eBL pathology in children. The use of humanized mouse models show promise in addressing questions about human infectious diseases and cancer etiology [66,67]. Combined with descriptive natural co-infection studies of humans and computational approaches to human immunology and virology [68], the interwoven array of immune mediators are being identified and can now be interrogated within malaria and EBV co-infection studies. This new appreciation of human immunology will lead to designing appropriate interventions to prevent malaria and/or EBV infections in infants. Learning how to silence or divert signaling pathways that interfere with normal EBV immune surveillance and improving our understanding of how chronic malaria contributes to eBL etiology has implications for preventing this pediatric cancer in Africa.