Gammaherpesviruses and B Cells: A Relationship That Lasts a Lifetime

Abstract

Gammaherpesviruses are highly prevalent pathogens that establish life-long infection and are associated with diverse malignancies, including lymphoproliferative diseases and B cell lymphomas. Unlike other viruses that either do not infect B cells or infect B cells transiently, gammaherpesviruses manipulate physiological B cell differentiation to establish life-long infection in memory B cells. Disruption of such viral manipulation by genetic or environmental causes is likely to seed viral lymphomagenesis. In this review, we discuss physiological and unique host and viral mechanisms usurped by gammaherpesviruses to fine tune host B cell biology for optimal infection establishment and maintenance.

Gammaherpesviruses as Manipulators of B Cell Differentiation

Gammaherpesviruses are ancient double-stranded DNA (dsDNA) viruses that have coevolved with their respective host and, similar to other herpesviruses, establish lifelong infection in a significant proportion of humans worldwide. All herpesviruses execute two distinct life cycles: lytic replication and latent infection. Lytic cycle is a highly immunogenic process accompanied by expression of dozens of viral genes, viral genome replication by the viral DNA synthesis machinery, and production of infectious virions. Due to its immunogenicity, lytic replication in a chronically infected host occurs at very low levels and only at limited anatomic locations.

In contrast, latency, the stealth mechanism that allows herpesviruses to maintain a lifetime infection, is the predominant herpesvirus life cycle in an immunocompetent host. Latently infected cells maintain the viral genome as an episome, which is replicated by the cellular DNA synthesis machinery during cell division. During latency, only a few, if any viral genes are expressed, with the viral transcripts acting at the RNA level or, if translated, possessing functions that attenuate proteasome-dependent degradation. Switch from latent to lytic life cycle, termed reactivation, is an important driver of herpesvirus transmission and pathogenesis.

Unlike other members of Herpesviridae family, gammaherpesviruses of many species, including the two known human gammaherpesviruses, Epstein-Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus (KSHV), are associated with diverse cancers and lymphoproliferative diseases, including B cell lymphomas (14). In addition to cancer, EBV is the causative agent of infectious mononucleosis, a clinical entity associated with primary EBV infection and characterized by malaise, lymphadenopathy, splenomegaly and hepatomegaly, and fever (32). Intriguingly, infectious mononucleosis only manifests upon acquisition of EBV in adolescence or beyond, via saliva exchange. EBV acquisition before adolescence does not produce unique symptoms and the route of transmission is not well established.

While transmission of EBV is believed to occur, at least in adolescents, via oral shedding of the virus and subsequent exchange of saliva, prospective study of oral washes collected from initially EBV naive college students failed to detect EBV genomes above 100 copies/mL before the onset of infectious mononucleosis symptoms (31). Furthermore, in many cases, presence of EBV in oral washes coincided with the detection of the virus in blood, highlighting the fact that a better understanding of initial phases of EBV infection is needed. In contrast, chronic EBV infection can be detected at low levels (1 in 10⁴–10⁶ circulating memory B cells) in most healthy blood donors (81,132).

Unlike EBV that infects >95% of adults, the seroprevalence of KSHV varies from 5% to >90% depending on the geographical location, age, and HIV status; transmission occurs by both sexual and nonsexual routes (80). While KSHV tropism for B cells is clear (79,82), circulating B cells of healthy donors have undetectable levels of KSHV genome (compared to readily detectable EBV presence) (45), and manipulation of B cells during natural KSHV infection is still poorly understood.

In contrast to the transient infection of B cells by other viruses, such as lymphocytic choriomeningitis virus (LCMV) (126), Dengue virus (65), and Chandipura virus (102), latent EBV (and, likely, KSHV) is predominantly hosted by memory B cells. This presents a logistical challenge for the primary infection and the establishment of viral latency, as the memory B cell compartment represents a very small proportion of total B cells and is widely distributed across secondary lymphoid organs and tissues of the host. Therefore, several hypotheses were originally proposed to explain how EBV managed to gain access to the memory B cell compartment.

One such hypothesis put forth by David Thorley-Lawson postulated that naive B cells, that are far more abundant than memory B cells, are infected by EBV and, via expression of select latency-associated EBV genes, are driven to proliferate to eventually become long lived memory B cells (120). Thorley-Lawson proposed that a portion of EBV-infected activated B cells enters the germinal center, a site of robust B cell proliferation, somatic hypermutation, and class switching within secondary lymphoid organs. While the majority of B cells are selected against in this process, it was proposed that EBV-infected B cells expressed the two latent proteins LMP1 and LMP2A, which functionally mimic CD40 and B cell receptor (BCR) signaling to stimulate cell survival (120). This hypothesis was based on the parallels between EBV infection that ultimately leads to latent infection of memory B cells, and the physiological process of germinal center B cell differentiation whereby B cells become activated, undergo proliferation and selection, and become long-lived memory B cells.

A compelling evidence to support the germinal center hypothesis was offered in the subsequent study from the Thorley-Lawson's group showing that EBV-infected B cells in human tonsils have functional and phenotypic markers of germinal center B cells, including expression of two key germinal center proteins, BCL6 and activation-induced cytidine deaminase (AID), and were located within the germinal center follicle (100). Similarly, before the observation made by the Thorley-Lawson's group for EBV, the Efstathiou group reported expression of murine gammaherpesvirus 68 (MHV68) genes detected by in situ hybridization in germinal centers of latently infected mice (11). These observations, and the fact that many EBV-positive B cell lymphomas display phenotypic and genetic markers of germinal center progression (121), offer a strong support for the germinal center-based establishment of gammaherpesvirus latency and the germinal center origin of viral lymphomagenesis.

However, evidence supporting alternative models of EBV infection (60) questioned the importance of physiological germinal center response for the establishment of life-long latency. Unfortunately, further resolution of the competing models remains challenging due to the exquisite species specificity of human gammaherpesviruses and the inability to faithfully model germinal center responses in culture.

Recently, humanized mouse models have made it possible to study some aspects of EBV infection and lymphomagenesis in an intact host. These recent studies have offered an insight into antiviral T cell responses and have produced paradigm shifting conclusions regarding the “oncogenicity” of latency-associated EBV genes, including dispensable nature of classic viral oncogenes and important role of lytic EBV cycle during viral lymphomagenesis (70,86,99). However, the humanized mouse models do not faithfully recapitulate either the innate immune responses, due to species restriction of EBV or KSHV infection, or the physiological B cell differentiation. Furthermore, genetic manipulation of the human host is still highly challenging in the humanized mouse models.

To overcome the limitations associated with the studies of EBV and KSHV natural infection, we and other groups have utilized the MHV68 experimental system. MHV68 is a natural rodent gammaherpesvirus that is genetically and biologically related to EBV and KSHV (10,33,131), including induction of B cell lymphomas in immunocompromised mice (117). Over the past decades, the MHV68 experimental system has offered significant insights into the immunology of gammaherpesvirus infection of a natural host, insights that were possible due to the ability to genetically manipulate both the virus and the host and the power of mouse immunological tools. In this review, we will discuss current understanding of how gammaherpesvirus infection interplays with and alters physiological B cell responses and highlight key unanswered questions in the field.

B Cell Differentiation: Under the (Gammaherpesvirus) Influence

During infection by an invading pathogen, a subset of naive B cells undergoes differentiation through the germinal center response in secondary lymphoid organs to become highly antigen-specific B cells. The two resulting populations of B cells, plasma cells, and memory B cells secrete antibodies targeting the current infection or recognize future infection by the same pathogen and rapidly respond, respectively. EBV and MHV68 manipulate the physiological B cell differentiation for their own benefit.

To initiate a physiological germinal center response, B cells specific to an invading pathogen will become activated and present antigen to CD4⁺ T cells within the intrafollicular zone, or the border between the B and T cell zone (49). Upon establishing this antigen-specific interaction, both cell populations will express BCL6 (49,54), a key transcription factor for germinal center B cell and T follicular helper cell (Tfh) development and survival (13,27,140). Tfh will then upregulate CXCR5, PD-1, ICOS, and GL7, allowing migration into the germinal center follicle (49). B cells will also migrate into the follicle and express GL7 and CD95, becoming germinal center B cells.

The germinal center is anatomically subdivided into the dark and light zones, originally based on classic histology studies that used morphological criteria, such as size and nuclear contour along with the relative abundance of T cells to show distinct zones of the germinal center (109). Germinal center B cells undergo rapid proliferation in the dark zone of the germinal center (1), with the upregulation of antiapoptotic factors such as Bcl-2, and reduced expression of cell cycle checkpoint p21 (57) and the p53 tumor suppressor (92). This rapid proliferation is coupled to mutagenesis that generates somatic hypermutation and class switching, processes that aim to increase antibody affinity and avidity. Somatic hypermutation and class switch recombination is mediated by AID (28,87). Due to the random nature of AID-mediated mutagenesis, continuous selection of germinal center B cells with antigen-specific BCR occurs via BCR-antigen interaction and costimulatory signals from Tfh in the light zone; the lack of interaction with either leads to cell death (56). Following selection, B cells can recirculate through the dark zone or differentiate into either plasma cells or memory B cells. Memory B cell differentiation is driven by continued PAX5 expression and a downregulation of BCL6 (5,16,59). To facilitate plasma cell differentiation, PAX5 must be silenced (16,89) in addition to an upregulation of BLIMP1, XBP1, and IRF-4 (57,103).

This differentiation process is usurped during gammaherpesvirus infection (Fig. 1). EBV and MHV68 infect naive B cells, with subsequent entry of both infected and uninfected B cells into the germinal center. EBV and MHV68 infection is latent in germinal center B cells (64,100), with up to 20% of germinal center B cells harboring MHV68 genome at the peak of viral latency (22,75). Due to the latent nature of infection and rapid rate of cellular proliferation, gammaherpesviruses usurp proliferation of germinal center B cells to achieve exponential increase in the viral reservoir without the need to undergo highly immunogenic lytic replication. Following further differentiation, infected B cells emerge as memory B cells that host life-long latent infection (125), or plasma cells, where EBV and MHV68 reactivate, switching from latent to lytic life cycle. Production of infectious virions by plasma cells leads to infection of naive B cells, creating a positive feedback loop that ensures the establishment of gammaherpesvirus latency.

FIG. 1.

Germinal center model of gammaherpesvirus infection. Gammaherpesviruses infect naive B cells, followed by activation and entry of both infected and bystander B cells into the germinal center reaction. Latently infected B cells (indicated by red viral episome in the nucleus) undergo cycles of proliferation, somatic hypermutation, and selection to emerge as latently infected memory B cells or plasma cells. Differentiation into plasma cells triggers viral reactivation, with subsequent production of infectious virions.

In contrast to physiological B cell differentiation, gammaherpesvirus infection elicits a rapid and robust increase in the titers of class-switched antibodies reactive against self and foreign species antigens. In contrast, titers of class-switched virus-specific antibodies arise with much slower kinetics. MHV68 induces peak titers of total class switched antibodies by about 10 days postinfection; however, MHV68-directed class switched antibodies do not peak until 3 weeks postinfection; kinetics of antiviral antibodies may be further modified by the original inoculum dose (104,111). Similarly, class-switched EBV-specific antibody responses do not peak until 6–12 weeks after EBV exposure (108).

The difference in the kinetics of overall versus virus-specific class-switched antibody induction is accounted for by the rapid increase in self-reactive antibodies. As such, levels of class-switched, self-directed antibodies, including anti-dsDNA antibodies, peak by 14 days post-MHV68 infection (40,104). Furthermore, the titers of MHV68-driven self-directed antibodies are much higher than those induced by LCMV (24), an unrelated RNA virus that infects multiple cell types, including B cells (126). Similarly, EBV infection generates robust responses against self and foreign antigens. In fact, the presence of high titer antibodies against horse antigens is diagnostic of a recent EBV infection in humans (36). Unfortunately, the mechanisms underlying this selective increase in self-directed and irrelevant antibodies driven by gammaherpesvirus infection are still poorly understood, but could potentially offer insights into the genesis of B cell-driven autoimmune disease and virus-driven lymphomas.

Because mechanisms underlying the gammaherpesvirus-driven self-directed humoral response remain enigmatic, it is difficult to test the role of nonspecific B cell differentiation in chronic infection. However, it is clear that EBV and MHV68 selectively establish infection in B cells that do not encode a virus-specific BCR (25,125). As EBV and MHV68 reactivation occurs in plasma cells (61,64) that also produce high levels of antibodies, it is tempting to speculate that intracellular virus-directed antibodies could attenuate assembly or infectivity of reactivating virus. Therefore, by preferentially establishing latency in virus-nonspecific B cells, gammaherpesviruses would ensure optimal reactivation kinetics, a feature that is particularly important in an immunocompetent, chronically infected host.

Robust, MHV68-driven germinal center response is transient in nature. Despite mild follicular hyperplasia maintained in long-term-infected animals, the magnitude of the germinal center response decreases by 42 days postinfection from its peak levels observed at 16 days postinfection (76) along with the decrease and stabilization of the frequency of latently infected splenocytes. Similarly, exuberant immune activation associated with EBV-driven infectious mononucleosis resolves over a period of weeks to months.

Interestingly, latent MHV68 genome is also present in developing and transitional B cells (17,18,116). The elegant study from the Tibbetts group demonstrated that transient treatment of long-term-infected animals with anti-IL-7, which led to depletion of transitional, but not naive or germinal center B cells, decreased the frequency of MHV68-positive naive and germinal center B cells (17). In contrast, transient depletion of IL-7 before and during early stages of MHV68 infection had no effect on the latent viral reservoir in germinal center B cells, indicating that, as the chronic infection transitions to its long-term stage, gammaherpesviruses may become more reliant on the infection of developing B cells to maintain the latent reservoir. While EBV is associated with fibrin-ring granuloma and hemophagocytic histiocytosis (15), diseases that manifest in the bone marrow, EBV infection of developing B cells in vivo has not been examined.

Living with the Enemy: Host Factors That Affect Gammaherpesvirus-Driven B Cell Differentiation

Given the limitations that are associated with the exquisite species specificity of human gammaherpesviruses, host factors that are involved in the interaction between gammaherpesviruses and B cell differentiation in vivo have been primarily defined using the MHV68 system. Intranasal inoculation of mice that lack peripheral B cells (μMT^−/− mice) results in significantly attenuated establishment of MHV68 latency in the spleen (113,130,135). In contrast, lack of peripheral B cells, does not preclude long-term latent MHV68 infection in the spleen (136), likely due to ability of gammaherpesviruses to infect and establish latency in myeloid cells (41,63,95,97,105,134,137). However, long-term MHV68 reactivation is poorly controlled in μMT^−/− mice, particularly in the peritoneal cavity, and μMT^−/− mice receiving high-dose intraperitoneal MHV68 inoculum eventually succumb to the infection (136).

Interestingly, lack of MHV68 antibody in μMT^−/− mice is not likely to be critical for the control of infection, as chronic MHV68 infection was well-controlled in mice that could only generate anti-hen egg lysozyme antibodies (77). It was proposed that, following infection, these transgenic B cells instead served as antigen presenting cells (77), facilitating anti-MHV68 T cell responses. Similarly, EBV-infected B cells can present viral antigen to T cells in culture, with subsequent inhibition of EBV-driven transformation (6). Despite well-controlled MHV68 infection in immunocompetent mice that cannot produce MHV68-specific antibodies, MHV68-specific antibody is important to attenuate increased viral reactivation and persistent replication in immunocompromised mouse models (38,53) and preexisting antiviral antibody attenuates acute replication and early latent MHV68 infection (3,123).

In addition to the mere presence of peripheral B cells, it is clear that the induction of germinal center response is critical for the peak levels of MHV68 latency. The first studies to offer evidence supporting the importance of T cell-dependent B cell differentiation during MHV68 infection demonstrated decreased splenomegaly and MHV68 reactivation in CD4 T cell depleted mice and an important role of CD4 T cells in MHV68-driven nonspecific B cell differentiation (112,129). Subsequently, it was demonstrated that the cognate T-B cell interaction via CD40/CD40L and MHC class II expression on B cells was required for the stimulation of non-MHV68-specific antibodies (104). These early studies led to elegant publications from the Speck group cementing the critical importance of Tfh and IL-21 in MHV68-driven germinal center response and the establishment of viral latency (20,21).

B cell activation is important before the entry into the germinal center response. Not surprisingly, infected cell-intrinsic NF-κB signaling supported the efficient establishment and maintenance of MHV68 latent infection, as demonstrated using an MHV68 mutant expressing dominant negative IκBα (58). The importance of NF-κB has been further demonstrated by the fact that EBV and KSHV encode viral proteins that modulate NF-κB activity. For example, EBV encodes LMP1, which activates NF-κB signaling by acting as functional mimic of CD40 receptor (55,122,127). KSHV encodes vFLIP, which drives NF-κB activation via TRAF2 and TRAF3 interactions (43,74). Importantly, both of these viral proteins facilitate cellular transformation (42,48).

Consistent with the importance of B cell activation for the establishment of gammaherpesvirus latency, MyD88^−/− mice demonstrate decreased B cell activation, MHV68-driven germinal center response, and establishment and maintenance of latent infection, with the phenotypes likely driven by a B cell-intrinsic MyD88 deficiency (39). MyD88 is involved in a number of signaling pathways, including IL-1 signaling. We have recently shown that IL-1R1^−/− mice demonstrate attenuated germinal center response following MHV68 infection, with attenuated induction of virus-specific and self-directed antibodies (24). However, frequency of MHV68 DNA-positive splenocytes was comparable in control and IL-1R1^−/− mice, with most of infected cells representing germinal center B cells ((24) and unpublished observations), suggesting that several MyD88-dependent pathways function to support the establishment of MHV68 latency.

Given the role of B cell activation in the gammaherpesvirus-driven germinal center response and the establishment of latent infection, we hypothesized that host factors restricting B cell activation would be deleterious for the virus. SHP1 (encoded by PTPN6), a tyrosine phosphatase expressed in hematopoietic cells, is a negative regulator of immune cell activation (50,66). In B cells, SHP1 is a cytoplasmic protein that localizes to the BCR to dephosphorylate several substrates, including Igα-Igβ subunits, Syk, and BLNK, ultimately attenuating BCR-proximal signaling [reviewed in Tamir et al. (115)]. Correspondingly, SHP1 expression is significantly decreased in EBV or KSHV positive B cell lymphomas, including Burkitt's, Diffuse Large B cell, Monomorphic Posttransplant Lymphoproliferative Disease, and Primary Effusion lymphomas (26,85,91,133). Surprisingly, we found that B cell-, but not T cell-intrinsic SHP1 expression supported efficient germinal center B cell expansion and latency establishment during MHV68 infection (46). These findings are consistent with the role of SHP1 expression in germinal center B cells during the physiological germinal center response (51), supporting the concept that gammaherpesviruses usurp some physiological B cell differentiation mechanisms to achieve their goal of establishing life-long infection.

However, it is also clear that the mechanisms governing gammaherpesvirus-driven germinal center response do not fully overlap with those guiding physiological germinal center reactions. In addition to the robust gammaherpesvirus-driven induction of self- but not virus-directed class-switched antibodies, specific host factors seem to be selectively important or dispensable for the gammaherpesvirus-induced B cell differentiation. We showed that MHV68-driven germinal center response was significantly exaggerated and failed to contract in IRF-1^−/− mice (75) along with a significant elevation in the frequency of MHV68 DNA-positive splenocytes. In contrast, IRF-1^−/− mice immunized with sheep red blood cells or infected with LCMV, displayed germinal center responses equivalent to those observed in wild-type mice (75). Thus, IRF-1 selectively attenuates gammaherpesvirus-driven germinal center response, with little effect on B cell differentiation induced by immunization or an RNA virus infection.

Similarly, while B cell intrinsic STAT3 expression is required for germinal center response induction and maintenance following several physiological stimuli, it is entirely dispensable for the MHV68-driven germinal center response (29,96). It is likely that many more mechanisms unique to gammaherpesvirus-driven B cell differentiation await to be discovered in the future studies.

Beyond the germinal center response, expression of BLIMP1, XBP1, and IRF-4 facilitate differentiation of germinal center B cells into plasma cells (57,103). This differentiation triggers reactivation, a switch from latent to lytic viral cycle. The question of how these viruses “know” that the B cell has differentiated has been approached from the perspective of gene expression regulation. Specifically, gammaherpesvirus reactivation is initiated by the expression of viral transcription factor(s) that are critical for the induction of downstream lytic viral genes. Promoters of such lytic cycle masterminds in EBV, KSHV, and MHV68 include elements bound by XBP1, such that expression of XBP1 in vitro induces viral reactivation (9,23,73,114,139,141). Intriguingly, B cell-specific XBP-1 deficiency had no effect on MHV68 reactivation in vivo (73), indicating that there is a redundancy among plasma cell transcription factors in mediating gammaherpesvirus reactivation. In contrast, B cell-intrinsic deficiency of IRF-4 resulted in attenuated generation of plasma cells following MHV68 infection along with a dramatic decrease in viral reactivation (73).

Gammaherpesvirus Proteins and B Cell Differentiation: When Extra Convincing of the Host Is Needed

In contrast to host factors that are likely to be expressed in a majority of cells within the relevant B cell subpopulation, only a small proportion of any B cell population will be infected. For example, up to 20% of germinal center B cells are MHV68 positive at the peak of viral latency, with the frequency of virus-positive cells decreasing as long-term infection is established; the virus is even less abundant in other B cell subpopulations (19,35). Thus, gammaherpesviruses are likely to trigger a combination of infected cell intrinsic and extrinsic mechanisms to drive germinal center response and differentiation of both infected and uninfected B cells.

MHV68 latency-associated nuclear antigen (LANA, encoded by orf73) is a latent gammaherpesvirus protein, with genetic and functional homologs present in KSHV and EBV (37,84,128). MHV68 LANA, while mostly dispensable for acute infection in vivo, is critical for the efficient establishment of latency and viral reactivation, a devastating defect that cannot be rescued by increasing the infectious dose of LANA null MHV68 mutant (37,84). LANA is expressed during MHV68 latent infection (88) and, similar to its counterparts in human viruses, is responsible for tethering the viral episome to the cellular chromosomes upon division of the latently infected cell, ensuring the “inheritance” of the viral plasmid by the daughter cells.

Despite similar acute infection parameters, mice infected with LANA-deficient MHV68 mutant display decreased expansion of Tfh population and complete lack of induction of anti-dsDNA antibodies (40), suggesting that efficient establishment of latency and/or other LANA functions are critical to drive self-reactive B cell response. Interestingly, Sh2d1a deficient mice that have significantly attenuated Tfh responses also fail to increase levels of anti-dsDNA antibodies following MHV68 infection (40), highlighting the important role of germinal center response in the gammaherpesvirus-driven differentiation of self-reactive B cells.

Consistent with observations during the MHV68 infection, KSHV LANA transgene driven by the endogenous viral promoter induces follicular hyperplasia, germinal center responses, and B cell lymphomas in transgenic mice (34). Thus, the manipulation of B cell differentiation by gammaherpesvirus LANA proteins is conserved across host species and goes beyond the maintenance of the viral episome, as the B cell phenotypes in LANA transgenic mice occur in the absence of the viral genome.

In addition to LANA and its functional homologs, several unique (at least sequence-wise) EBV proteins expressed in latently infected B cells are thought to directly affect the germinal center stage of infection. EBV LMP1 is expressed in infected germinal center B cells (4) and is a powerful, constitutively active functional homolog of CD40 that is postulated to support proliferation and survival of EBV-infected germinal center B cells [reviewed in Kieser and Sterz (52)]. Interestingly, EBV-driven lymphomagenesis in humanized mice is independent of LMP1 in the presence of CD4 T cells. In contrast, depletion of CD4 or blocking CD40 signaling completely abolishes lymphomagenic ability of LMP1-deficient EBV mutant, highlighting the functional redundancy of LMP1 and CD4 T cell help/CD40 signaling (70).

LMP1 is coexpressed in infected germinal center B cells with LMP2A, with the latter being a functional mimic of BCR signaling. EBV infection of “crippled” germinal center B cells that no longer express BCR due to debilitating heavy-chain mutations can rescue the survival and proliferation of these cells in vitro and yield clonal lymphoblastoid cell lines, with LMP2A presumably playing a role in this process (71). LMP2A expression driven by highly active immunoglobulin heavy-chain promoter and intronic enhancer in transgenic mice resulted in attenuated heavy-chain rearrangement and increased numbers of IgM-negative circulating B cells (12).

Similar to LMP1, deletion of LMP2A has minimal effect on EBV-induced lymphomagenesis in humanized mice (69). However, simultaneous deficiency of LMP1 and LMP2A results in fewer tumors with delayed onset, highlighting biologically redundant functions of these two EBV proteins with respect to viral lymphomagenesis, at least in the context of humanized mouse model (69). The role of LMP1 and LMP2 in the EBV-driven germinal center response remains unclear due to the limitations of humanized mouse models. However, expression of these two EBV genes gives rise to CD8 T cell epitopes that may be important for the contraction of the EBV-driven germinal center response and control of latent viral reservoir (78).

While genes homologous to EBV LMP1 and LMP2A are not present in other gammaherpesviruses, functional homologs certainly exist. Expression of KSHV-encoded K1 in place of LMP2A prevents apoptosis of primary BCR negative human B cells infected with a chimeric EBV mutant in vitro (110). High levels of K1 expression as a SV40 promoter-driven transgene in mice led to lymphoid hyperplasia with a small proportion of transgenic mice developing frank B cell lymphomas at 18 months of age (8). The role of endogenous levels of KSHV K1 expression in B cell differentiation during chronic infection remains unclear.

MHV68 M2 is a multifunctional latent protein that is expressed in infected germinal center B cells (118). M2 has several B cell-specific functions: at low doses of infection, it facilitates MHV68 infection of germinal center B cells (44), at high doses of infection M2 promotes differentiation of MHV68-infected germinal center B cells to plasma cells and subsequent viral reactivation (118). Importantly, M2 expression by B cells, in the absence of infection, was sufficient to increase expression of IL-2, IL-6, MIP1α, and IL-10, with IL-10 supporting the expansion of M2 expressing B cells (106). The induction of IL-10 by M2 was later shown to be IRF-4-dependent (94). The ability of MHV68 M2 to stimulate cytokine production by B cells is particularly interesting, because KSHV encodes viral IL-6 and MIP-1α (90), while EBV encodes viral IL-10 (83,101), indicating overlapping functions between M2 and human gammaherpesvirus proteins. Finally, adoptive transfer of naive B cells that express MHV68 M2 drives their differentiation into germinal center B cells or plasma cells without any additional stimulation of the recipient host (118), phenotypes that mimic those attributed to LMP1 and/or LMP2A expression.

All gammaherpesviruses encode a single conserved protein kinase. Despite the conserved nature of this viral protein, the role of the gammaherpesvirus kinase during chronic infection remains poorly understood. Our group was the first to demonstrate that both expression and enzymatic activity of the MHV68 protein kinase (orf36) support the establishment of chronic infection, especially under conditions of low infectious dose (24,116), however, the underlying mechanism remained unclear.

Recently, we demonstrated that expression and enzymatic activity of MHV68 orf36 promotes MHV68-driven germinal center response and generation of self-reactive, but not MHV68-specific antibodies, for the first time genetically separating physiological, virus-directed, and abnormal, self-reactive B cell differentiation induced by gammaherpesvirus infection (24). This observation was particularly unexpected, as known gammaherpesvirus “manipulators” of germinal center response (discussed above) are expressed in latently infected cells, whereas gammaherpesvirus protein kinases are classically associated with the lytic viral cycle. Recently, we have detected MHV68 orf36 expression in splenocytes harvested from latently infected mice (Paul Sylvester and Vera Tarakanova, unpublished observations), with ongoing studies defining the subsets of B cells expressing MHV68 protein kinase during chronic infection.

Similar to the functional conservation of latent gammaherpesvirus proteins observed across species, expression of KSHV protein kinase orf36 as a transgene (under the control of endogenous viral promoter) resulted in expression of viral kinase in B cells, increased germinal center response and class-switching, and eventual generation of B cell lymphomas that continued to express the KSHV protein kinase (2). This important study was not only the first to demonstrate the oncogenic role of the gammaherpesvirus gene classically associated with the lytic infection but also highlighted the functional conservation of gammaherpesvirus protein kinases across different host species. Because gammaherpesvirus protein kinases are multifunctional proteins that interact with numerous systems of the host, future studies need to define the specific viral kinase–host interactions that are responsible for the manipulation of B cell differentiation and lymphomagenesis.

Gammaherpesvirus Pathogenesis: The Case of Deranged B Cells

Similar to other oncogenic viruses, gammaherpesvirus-driven lymphomagenesis is almost certainly an accidental outcome of well-evolved virus–host interactions, an accident that is driven by changes in host genetics and/or environmental influence. The accidental nature of gammaherpesvirus lymphomagenesis is best illustrated by the challenges that have to be overcome to faithfully model such lymphomagenesis in vivo. With a few exceptions (i.e., KSHV orf36 transgenic mouse model that generated lymphomas in 66% of immunocompetent mice) (2), expression of a single gammaherpesvirus gene or even the entire KSHV latency locus under the endogenous transcriptional control produces lymphomas in only a limited number of immunocompetent animals (11–16%) (34,107), consistent with the low frequency of gammaherpesvirus lymphomagenesis in the human population. The incidence of lymphomas typically can be increased by placing the viral gene under the control of a robust constitutively active promoter, introduction of other oncogenic host mutations, or using severely immunocompromised animals; in other words, creating conditions that disrupt the physiological virus–host interactions.

Unfortunately, because the risk factors underlying gammaherpesvirus lymphomagenesis remain poorly defined and the low frequency of infected individuals that develop cancer, it is currently highly challenging to gain any insight into the mechanism of “natural” gammaherpesvirus lymphomagenesis in humans. Signature genetic mutations have not been reported in gammaherpesvirus-driven lymphomas, with an exception of EBV-positive Burkitt's lymphomas that are characterized by the chromosomal translocation that stimulates expression of c-myc. What is clear is that many gammaherpesvirus-positive lymphomas have evidence of germinal center origin, whether by retaining phenotypic markers of germinal center B cells and/or genetic signatures consistent with somatic hypermutation (121).

In support of the germinal center origin of viral lymphomagenesis, germinal center stage of B cell differentiation is, arguably, the most prone to transformation due to rapid cellular proliferation, downregulation of tumor suppressors, and expression of mutagenic AID. Finally, expression of EBV or KSHV proteins that are expressed in latently infected B cells, including germinal center B cells, is in many cases sufficient, especially at high levels of expression, to generate lymphomas or lymphoproliferative disease in transgenic animals; the role of these viral proteins during lymphomagenesis in humans is difficult to define.

Intriguingly, an emerging area of research has established an important role for lytic gammaherpesvirus cycle and/or proteins in viral lymphomagenesis. Infection of humanized mice with lytic replication deficient EBV mutant resulted in a five-fold decrease in the incidence of lymphomas compared to wild-type EBV, despite equivalent establishment of latent infection (68). Furthermore, infection with a “superlytic” EBV mutant had given rise to the same incidence of lymphomas in humanized mice as infection with wild-type EBV (67). Similarly, expression of KSHV protein kinase, a classical lytic viral gene, was sufficient to drive B cell lymphomas in a majority of transgenic mice (2). Given this emerging connection, it is critical to revisit the traditional dogmas sorting gammaherpesvirus genes into latent and lytic “boxes” and comprehensively redefine viral gene expression in the relevant B cell subpopulations, including germinal center B cells in vivo, during natural infection.

In a counterargument of the germinal center origin of gammaherpesvirus positive lymphomas, AID expression in B cells can be induced by EBV and KSHV infection (7,47), potentially stimulating genetic changes associated with germinal center response without the actual experience of progressing through this stage of differentiation. Furthermore, KSHV infection of mature B cells also activates expression of RAG, a mutagenic V(D)J recombinase normally limited in expression to early stages of B cell development (124). This KSHV-driven reexpression of RAG induces further V(D)J recombination, offering an explanation for a well-known bias of KSHV infection for Igλ B cells.

It is possible that virus-induced expression of mutagenic enzymes cooperates with physiological and environmental stimuli of AID expression to promote lymphomagenesis. One such environmental stimulus is Plasmodium infection, which leads to prolonged expansion of germinal center response along with sustained expression of AID (98), including in nongerminal center B cells (138), and is a well-established risk factor for endemic EBV-driven Burkitt's lymphoma. Thus, it is important to continue questioning and refining the germinal center model to gain insight into the mechanisms underlying gammaherpesvirus lymphomagenesis.

While the association of gammaherpesvirus infection with cancer is well-established, there is an additional, although much more controversial, association between EBV infection and autoimmune disease. Despite controversy, this association seems intuitively plausible, given the unique ability of gammaherpesviruses to induce robust nonvirus-specific B cell differentiation, especially during the early stages of chronic infection, a process that could synergize with host genetic or environmental susceptibility to autoimmune disease. However, causative association between autoimmune disease has been challenging to demonstrate given the high prevalence and life-long nature of EBV infection.

EBV infectious mononucleosis is the least controversial risk factor for multiple sclerosis (MS), with infectious mononucleosis occurrence shown to temporally precede and greatly increase the risk of the subsequent MS development (72,119). While the causative link between EBV infection and the development of systemic lupus erythematosus (SLE) is more debatable, buffy coat EBV viral loads are significantly increased in SLE patients with active compared to inactive disease (93) along with impaired EBV-, but not HCMV-specific CD8 T cell responses (62). This increased EBV activity is thought to exacerbate severity of already established SLE, including nephritis [discussed in Draborg et al. (30)]. The role of EBV in other autoimmune diseases remains a subject of debate and it is likely that, similar to other disease models, the infection synergizes with other host susceptibilities (genetic, environmental, etc.) to promote disease.

Conclusion and Future Directions

The fascinating relationship between gammaherpesviruses and B cell differentiation lies at the heart of chronic infection and viral pathogenesis. While a number of host and viral factors involved in this relationship have been identified, we still know surprisingly little regarding this relationship. This is an unacceptable knowledge gap considering the fact that there are currently no approaches to generate sterilizing immunity, clear existing gammaherpesvirus infection, or to precisely identify and manage susceptibility to gammaherpesvirus-driven disease in humans. Because many aspects of B cell differentiation only occur in the context of an intact host, animal models are critical to define the relevant molecular and cellular mechanisms that selectively govern the intricate dance between the gammaherpesvirus infection and B cell differentiation, especially for human gammaherpesviruses. Defining unique mechanisms that regulate gammaherpesvirus-driven B cell differentiation will lay the foundation for targeted preventative and therapeutic approaches against gammaherpesvirus-driven disease.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by 1F31CA243364 (K.E.J.), CA183593, and CA203923 (V.L.T.).