Is murine gammaherpesvirus-68 (MHV-68) a suitable immunotoxicological model for examining immunomodulatory drug-associated viral recrudescence?

Abstract

Immunosuppressive agents are used for treatment of a variety of autoimmune diseases including rheumatoid arthritis (RA), systemic lupus erythematosis (SLE), and psoriasis, as well as for prevention of tissue rejection after organ transplantation. Recrudescence of herpesvirus infections, and increased risk of carcinogenesis from herpesvirus-associated tumors are related with immunosuppressive therapy in humans. Post-transplant lymphoproliferative disorder (PTLD), a condition characterized by development of Epstein Barr Virus (EBV)-associated B-lymphocyte lymphoma, and Kaposi’s Sarcoma (KS), a dermal tumor associated with Kaposi Sarcoma-associated virus (KSHV), may develop in solid organ transplant patients. KS also occurs in immunosuppressed Acquired Immunodeficiency (AIDS) patients. Kaposi Sarcoma-associated virus (KSHV) is a herpes virus genetically related to EBV. Murine gammaherpes-virus-68 (MHV-68) is proposed as a mouse model of gammaherpesvirus infection and recrudescence and may potentially have relevance for herpesvirus-associated neoplasia. The pathogenesis of MHV-68 infection in mice mimics EBV/KSHV infection in humans with acute lytic viral replication followed by dissemination and establishment of persistent latency. MHV-68-infected mice may develop lymphoproliferative disease that is accelerated by disruption of the immune system. This manuscript first presents an overview of gammaherpesvirus pathogenesis and immunology as well as factors involved in viral recrudescence. A description of different types of immunodeficiency then follows, with particular focus on viral association with lymphomagenesis after immunosuppression. Finally, this review discusses different gammaherpesvirus animal models and describes a proposed MHV-68 model to further examine the interplay of immunomodulatory agents and gammaherpesvirus-associated neoplasia.

• B-cell lymphoma
• immunosuppression
• immunotoxicology
• lymphomagenesis
• MHV-68
• PTLD
• viral recrudescence

Introduction

Immunosuppressive therapy is used for a variety of immune disorders as well as for prevention of tissue rejection after organ transplant. However, some of the side-effects of immunosuppressive therapy may lead to an increased prevalence of neoplasia or post-transplant lymphoproliferative disorder (PTLD). PTLD can be a concern after hematopoietic stem cell or solid organ transplant (SOT) and can lead to a variety of lymphoid cancers (Crespo-Leiro et al., 2008; Knowles, 1999; Zafar et al., 2008). PTLD occurs in the presence of immunosuppression, and may be associated with re-activation of EBV or KSHV from latency (Rezk & Weiss, 2007; Schwartz, 2001; Zafar et al., 2008). EBV⁻ individuals or those with primary or secondary immunodeficiencies are at greatest risk for PTLD (Gottschalk et al., 2005). EBV is known to latently infect up to 90% of adults and is frequently (∼80% of time) detected in post-transplant lymphomas (Tran et al., 2008). KSHV is the causative agent of Kaposi Sarcoma (KS), and is often found in HIV/AIDS-related malignancies (Sathy et al., 2008). The absence of direct small animal models for EBV and KSHV and the increasing prevalence of viral-induced neoplasia after immunosuppression for organ transplantation indicate a need for a better understanding of gammaherpesvirus pathogenesis and development of models to contribute to this understanding. An ideal mouse model for studying the role of immunosuppressive agents and their association with neoplasia would preferably be one that provides accurate predictions of safety/efficacy of the drug in human patients in an efficient study period. A good model would also proficiently help characterize novel immunosuppressive agents and provide meaningful clinical information to alleviate or cure human disease.

Gammaherpesviruses are large, double-stranded DNA viruses, characterized by a life cycle of lytic and latent infections (for a review of gammaherpesvirus molecular virology, the reader is encouraged to consult (Kieff & Rickinson, 2001; Roizman & Knipe, 2001). During a lytic phase of infection, infectious virus is produced and most of the viral genome is expressed. During latency, most viral genes are silenced and there is an absence of infectious virus production (Barton et al., 2011; Sunil-Chandra et al., 1992a,b, 1993). Latent infections permit lifelong viral persistence through evasion of the host immune response. Immune suppression can trigger the onset of lytic gene expression and re-activation from latency.

MHV-68 is genetically related to EBV and KSHV and was originally isolated from wood voles in Slovakia (Blaskovic et al., 1980; Nash et al., 2001). Infected mice with a primary or drug-induced immunodeficiency may develop lymphoma (Kulkarni et al., 1997; Sunil-Chandra et al., 1994; Tarakanova et al., 2005, 2008). Due to the fact that EBV and KSHV are restricted to human hosts, and since MHV-68 exhibits a similar disease course in the mouse, the purpose of this review is to demonstrate MHV-68 as a potential model to study viral re-activation in relation to human lymphomagenesis and PTLD. A murine gammaherpesvirus model is informative for a number of reasons. First, the gammaherpesvirus family is comprised of lymphotropic viruses with strong gene homology and function; all of these viruses are capable of playing a role in lymphomagenesis through similar mechanisms (Damania, 2004). Second, the innate and adaptive immune systems from mice to humans are strongly conserved, which argues for similar patterns of pathogenesis (Barton et al., 2011). Third, MHV-68 is capable of infecting laboratory mice and a variety of tissue culture cell lines, providing a valuable tool to study gammaherpesvirus pathogenesis in vitro and in vivo (Blaskovic et al., 1980; Virgin et al., 1997). In the sections that follow, an overview of gammaherpesvirus pathogenesis and relationship with lymphoma is presented, along with a description of the MHV-68 model and its potential for examining interactions of immunomodulatory agents and viral-associated neoplasia.

Gammaherpesvirus overview: MHV-68, EBV, and KSHV

Gene expression

All gammaherpesviruses have similarities in replication strategy (see Figure 1 for an overview of gammaherpesvirus pathogenesis) and follow a temporal pattern of gene expression of intermediate-early (IE), early (E), and late (L) genes (Roizman & Knipe, 2001). IE genes are expressed during acute viral replication and are expressed during the transition between latency and re-activation. IE genes are involved in regulation of gene expression or in transactivation that is important for initiation of viral replication. E genes are critical for setting up the enzymes needed for DNA replication, and these gene products are required for viral DNA replication to proceed. L genes express the structural proteins that make up the virus particle.

Figure 1. Overview of gammaherpesvirus pathogenesis. Primary gammaherpesvirus infection typically occurs in the respiratory tract and results in an initial lytic infection that is generally cleared within a few weeks via a robust CTL response in immune-competent individuals. The host is unable to clear all virally infected cells, and the virus down-regulates lytic genes and enters latency. Latency persists for the life of the infected host. Viral re-activation can periodically occur (shown by the bidirectional arrow) during immunosuppression that may be associated with lymphomagenesis in some patients.

Murine gammaherpesvirus-68 (MHV-68)

One of the major hurdles of studying EBV and KSHV is the lack of a robust small animal model to determine how these viruses may be associated with cancer. This problem is largely due to the restricted host range of the members of the gammaherpesvirus family (Roizman & Knipe, 2001). One current model of gammaherpesvirus pathogenesis is MHV-68, which is closely related to both EBV and KSHV. MHV-68 is thought transmitted via the respiratory route or through secretions and establishes a chronic infection that persists for the life of the infected mouse (Barton et al., 2011; Nash et al., 2001). Acute infection begins in the lungs and is resolved within ≈ 2 weeks (Flano et al., 2000; Speck & Ganem, 2010; Weck et al., 1999). Latency is then established in splenic B-lymphocytes (Simas et al., 1999; Sunil-Chandra et al., 1992b; Usherwood et al., 1996b). Similar to other gammaherpesviruses, re-activation from latency can be achieved via immunosuppression that can lead to lymphoproliferative disease (Nash et al., 2001; Sunil-Chandra et al., 1994).

Epstein Barr Virus

Epstein Barr Virus (EBV) infects ≈90–95% of adults and is the prototypical virus of the gammaherpesvirus sub-family of herpesviruses (for a comprehensive review of EBV molecular virology, see Kieff & Rickinson, 2001; Roizman & Knipe, 2001). EBV primary infection usually occurs at an early age of the host, and most cases are asymptomatic (Thorley-Lawson, 2005; Thorley-Lawson et al., 2008). Infections acquired in adolescence/adulthood can result in infectious mononucleosis (IM). For most of those infected, once an acute infection is controlled by the immune response, the virus remains latent for the lifetime of the host in B-lymphocytes (Kieff & Rickinson, 2001). However, infected patients who are immunocompromised are at a greater risk for re-activation of EBV from latency, which can contribute to the development of Burkitt’s lymphoma (BL), Hodgkin’s lymphoma, or other lymphoproliferative diseases.

Kaposi Sarcoma-associated herpesvirus (KSHV)

Kaposi Sarcoma-associated herpesvirus (KSHV or human herpesvirus 8/HHV-8) was identified as the etiologic agent of Kaposi’s Sarcoma (KS) which is a malignancy of the skin, oral cavity, and subcutaneous tissues (Moore et al., 1996; Ye et al., 2011). KSHV is also known to cause B-cell primary effusion lymphoma (PEL), body cavity-based B-cell lymphoma (BCBL), and Multicentric Castleman’s disease (MCD) (Chang et al., 1994; Ganem, 2005). KS is the most common cancer in untreated AIDS patients, but it is also known to occur during other states of immunosuppression such as during organ transplant (Mesri et al., 2010). For a comprehensive review of KSHV, see Speck & Ganem (2010), Wen & Damania (2010), and Barton et al. (2011).

Gammaherpesvirus molecular biology

Lytic gene products

All gammaherpesviruses express the replication and transcriptional activator gene (RTA/ORF-50) required for lytic expression and is necessary and sufficient for re-activation (Staudt & Dittmer, 2007). MHV-68 and KSHV share a homologous RTA gene (see Figure 2) whose expression activates the lytic cycle through DNA binding or protein–protein interactions (Speck & Ganem, 2010; Wu et al., 2000). EBV RTA is homologous to MHV-68/KSHV RTA and is encoded by BRLF1. EBV lytic gene expression is induced as a result of co-expression of RTA with the product of BZLF1 (also known as Z, ZTA, Zebra, or EB1) (Robinson et al., 2011). Expression of both of these genes is required for EBV lytic gene expression (Schwarzmann et al., 1998; Speck et al., 1997).

Figure 2. Gammaherpesvirus genomic organization and gene expression. MHV-68, EBV, and KSHV genomes are shown as indicated. Solid boxes refer to areas of similar genome conservation amongst gammaherpesviruses. 5′ and 3′ terminal repeats (TR) are not shown for simplicity. Arrows refer to direction of transcription. Unique genes or genes that play a known role in viral pathogenesis/recrudescence are labeled in the table and are discussed within the text.

MHV-68 and KSHV contain many cellular homologs that have been pirated from the infected host. MHV-68 and KSHV also express v-cyclin and v-GPCR, which share a similar function as their cellular counterparts. The M proteins are unique to MHV-68 from other EBV/KSHV proteins and share little to no homology with them (Virgin et al., 1997).

Latent gene products

Gammaherpesvirus also express a complement of gene products mostly expressed during latent infection and generally help the virus persist in an infected host. MHV-68 mLANA (murine latency-associated nuclear antigen) and KSHV LANA share mostly conserved functions, whereas a LANA homolog has not been identified for EBV. mLANA/LANA expression is required for establishment and maintenance of latent infection (see Figure 2) (Barton et al., 2011; Wen & Damania, 2010). KSHV LANA may also have other functions aside from latency maintenance, as its expression has been found in tumors from all three diseases known to be caused by KSHV: KS, MCD, and PEL (Dupin et al., 1999; Kellam et al., 1999; Parravicini et al., 2000). Although expression of mLANA is known to be required for MHV-68 latent infections, many of the other MHV-68 latency proteins (M proteins) have unidentified functions. KSHV latency proteins are also not well characterized, although it is known that KSHV latency is partly regulated through Kaposin proteins, microRNAs, and the viral homolog v-FLIP. These factors share important roles in modulation of signal transduction, cell cycle regulation, apoptosis inhibition, and immune function (Wen & Damania, 2010).

EBV latency proteins are better characterized than their other gammaherpesvirus counterparts. Epstein Barr Nuclear Antigens (EBNAs) are associated with latent EBV infections and have no direct KSHV or MHV-68 homologs (see Figure 2). There are six EBNAs: ENBA-1, -2, -3A, -3B, -3C, and -LP (Leader Protein) (Bornkamm & Hammerschmidt, 2001; Kieff & Rickinson, 2001). While the full complement of proteins required for immortalization of cells is not completely understood, all but EBNA-3B and -LP are thought to be required for immortalization of infected cells (Amon & Farrell, 2005; Bornkamm & Hammerschmidt, 2001). Latency-Associated Membrane Proteins (LMPs) are present only in EBV and consist of LMP-1, -2A, and -2B (see Figure 2). LMP-1 is expressed in a variety of EBV-associated cancers including PTLD, Hodgkin’s Disease, and nasopharyngeal carcinoma (Geiser et al., 2011).

Gammaherpesvirus immunology

Cell types acutely infected

The route of MHV-68 transmission is unclear, although it is thought to spread in the wild via the respiratory route, saliva, or in other secreted fluids. Acute infections occur in the lungs causing bronchiolitis and infection of alveolar epithelial cells (Nash et al., 2001; Sunil-Chandra et al., 1992a). Viral spread occurs from the lungs to the lymph node and extends to dendritic cells, macrophages, and B-lymphocytes (the eventual latency reservoir) (Barton et al., 2011; Nash et al., 2001; Sunil-Chandra et al., 1992a; Usherwood et al., 1996b; Weck et al., 1996, 1999).

EBV infections are tightly regulated and typically result in a latent infection of B-lymphocytes, although infection of epithelial cells and other lymphocytes have been reported (Bornkamm & Hammerschmidt, 2001; Kieff & Rickinson, 2001). EBV has been detected in oropharyngeal epithelial cells of patients with acute IM (Shannon-Lowe et al., 2009; Sixbey et al., 1984) or in the tongue epithelium of healthy EBV carriers (Walling et al., 2001). HIV-induced immunodeficiency can also be associated with oral hairy leukoplakia, a non-malignant lesion in the tongue that induces productive replication of EBV (Conant, 1987; Greenspan et al., 1985; Guerreiro-Cacais et al., 2004).

The primary tropism of KSHV is the B-lymphocyte (Ganem, 2010). Acute infections are mostly asymptomatic, but those who develop KS are usually immunosuppressed due to co-infection with HIV (Ganem, 2010; Kliche et al., 1998; Rappocciolo et al., 2008). Evidence exists for KSHV infection of endothelial cells, primarily from the isolation of KSHV DNA from KS spindle cells and in cells lining KS lesions (Ganem, 2010).

Immunologic control of acute infection

Control of acute gammaherpesvirus infection relies largely on T-lymphocyte immunity (see Table 1 for overview of gammaherpesvirus mechanisms of immune control). Following primary infection, MHV-68 induces a wave of inflammatory responses that induce macrophages by Day 3, a robust CD8⁺T-lymphocyte response after 1 week, and Type I interferon (IFNα/β) signaling (Nash et al., 2001; Sunil-Chandra et al., 1992a; Weck et al., 1997). The CD8⁺ T-lymphocyte response is critical for control of acute virus infection; depletion prior to infection is lethal for the infected animal (Ehtisham et al., 1993). CD4⁺ cells are required for maintenance of long-term CD8⁺ T-lymphocyte immune surveillance (Cardin et al., 1996), but they do not exert cytotoxic activity to resolve acute infection (Cardin et al., 1996; Ehtisham et al., 1993). Dendritic cells are also thought to play a role in the immunologic response since MHV-68 infection blocks their maturation. This suggests a role for MHV-68 in subversion of host adaptive immune surveillance (Flano et al., 2005; Hochreiter et al., 2007). Finally, the production of antibody takes ≈2 weeks for high titers to develop, and plays a role in long-term control of MHV-68 infection. However, control of acute infection does not rely on a robust antibody response (Simas & Efstathiou, 1998; Simas et al., 1999; Weck et al., 1997).

Table 1. Mechanisms of gammaherpesvirus immune control.

Cell type/effector	Impact on pathogenesis
CD8^+ T-lymphocytes	Primary mechanism to control acute lytic infection; shuts down lytic replication and down-regulates latency and re-activation
CD4^+ T-lymphocytes	Collaborate with CD8^+ T-lymphocytes to mediate lytic and latent infection; assist in long-term maintenance of CD8^+ T-lymphocytes response; other roles in immunity unclear
Interferon	IFN-α/β and -γ expression can induce an antiviral state and reduce lytic and latent replication as well as reduce re-activation
Antibody	Development of neutralizing antibody against virions occurs within a few weeks of infection and helps reduce re-activation and viral production
Apoptosis	Programmed cell-mediated death of infected cells which reduces lytic and latent replication and re-activation; decreases persistent viral replication

Primary infection with EBV normally occurs at an early age, with most of those infected showing no symptoms of disease. However, those who do not become infected until adolescence (or later) can develop IM. During IM, immunologically healthy individuals control acute infection mostly by cellular and humoral immunity (Thorley-Lawson & Gross, 2004). While large numbers of infected, resting B-lymphocytes circulate through the blood after infection, the production of antibody limits the spread of infectious virus while cytotoxic T-lymphocytes (CD8⁺ T-lymphocytes) destroy infected cells expressing EBV viral proteins (Kutok & Wang, 2006). In these cases, the percentage of T-lymphocytes specific to the host virus can reach as high as 50% (Chen, 2011; Kutok & Wang, 2006; Thorley-Lawson and Gross, 2004). While the host immune response is remarkably robust at limiting the spread of virus infection, elimination of the virus is not attained, resulting in a persistent latent infection of B-lymphocytes. Periodically, cells that do become re-activated are, in turn, killed by cytotoxic T-lymphocyte surveillance and/or by EBV-specific antibodies.

Infection with KSHV results in an asymptomatic, persistent infection in immunocompetent hosts. Unlike primary EBV infections, KSHV infection does not seem to be correlated with any specific illness. It is, therefore, very difficult to study the immune response in healthy individuals, and details of immune control in healthy carriers of KSHV remain enigmatic. In contrast, upon immunosuppression (such as what occurs in HIV-infected individuals and transplant recipients), the incidence of KS is >100-times more common (Boshoff & Weiss, 2002; Robey et al., 2010).

Despite a lack of understanding of the mechanisms of KSHV immune control, information from immunosuppressed individuals has highlighted the roles of the CD8⁺ T-lymphocyte response against the virus, which seems to play the biggest role in the T-lymphocyte mediated immune response (see Table 1). CD4⁺ T-lymphocytes are also important in controlling KSHV infection, but CD8⁺ T-lymphocyte epitopes are most frequently found in sera from HIV/KSHV-infected patients (Bihl et al., 2007; Brander et al., 2001; Robey et al., 2010).

Latency

Cell types latently infected

Gammaherpesviruses share conserved strategies for maintaining latency and also predominantly exploit B-lymphocyte biology to establish latency reservoirs. EBV and MHV-68 studies demonstrate latently infected cells in naïve, memory, or germinal center B-lymphocytes in the tonsils (Flano et al., 2002). MHV-68 experiments have also identified latent reservoirs in macrophages and dendritic cells (Flano et al., 2002; Willer & Speck, 2003). Studies have also shown establishment of latency in mice lacking mature B-lymphocytes and viral persistence in the lungs of B-lymphocyte-deficient mice (Usherwood et al., 1996b; Weck et al., 1996). KSHV studies have isolated latently infected B-lymphocytes in patients with PEL and MCD (Bechtel et al., 2003; Dupin et al., 1999) as well as in the endothelial tissue of spindle cells taken from KS-positive tumors (Bechtel et al., 2003; Boshoff et al., 1995; Staskus et al., 1997).

Phenotype of latency

MHV-68 and EBV share a latency strategy that utilizes the memory B-lymphocyte reservoir. MHV-68 infection also latently infects dendritic cells, macrophages, and a dynamic repertoire of splenic B-lymphocytes (Flano et al., 2000; Speck & Ganem, 2010; Weck et al., 1999). MHV-68 infects large numbers of splenocytes during the establishment of latency, some of which have been shown to spontaneously re-activate into the lytic cycle when co-cultured in vitro with permissive cell lines (Sunil-Chandra et al., 1992b; Weck et al., 1999).

MHV-68 infects a high proportion of naïve B-lymphocytes during establishment of latency, but the pattern and phenotype of latency changes over time. Naïve B-lymphocyte infection is not detectable at ≈3 months post-infection. The pools of latently infected splenocytes and non-B-lymphocyte populations drop sharply as well, and virus is detectable only in isotype-switched B-lymphocytes that reflect memory cells (Speck & Ganem, 2010; Willer and Speck, 2003). This latency pattern is hypothesized to be a hallmark of the gammaherpesviruses and serves as an efficient mechanism to avoid host immune surveillance.

EBV latency has been studied extensively and consists of four distinct growth patterns (Type 0, I, II, and III) characterized by differential expression of latent genes (typically EBNA1 and LMPs) in B and epithelial cell populations (Rowe et al., 1992). Type 0 latency occurs in peripheral blood within non-dividing B-lymphocytes and is characterized by no viral gene expression (Miyashita et al., 1995). Type I latency occurs in memory B-lymphocytes and has been observed in BL-derived cell lines. Latency II is the growth pattern observed in germinal center B-lymphocytes during which activated B-lymphocytes differentiate into memory B-lymphocytes. Latency III is the growth-promoting pattern that occurs after entry of the virus into resting B-lymphocytes and has been isolated in tonsils with acute infection of naïve B-lymphocytes (Thorley-Lawson et al., 2008).

Unlike EBV, the latency program of KSHV displays no obvious phenotype and the virus has neither been shown to immortalize nor transform primary B-lymphocytes (Ganem, 2010; Kliche et al., 1998; Rappocciolo et al., 2008). KSHV latency has been studied in cell culture and is similar to EBV in that latency is the default pathway for virus infection (Ganem, 2010). The mechanism of KSHV latency is still unclear, and it is still unknown whether the virus can infect germinal center B-lymphocytes or establish permanent latency in memory B-lymphocytes (Speck & Ganem, 2010).

Studies with MHV-68, EBV, and KSHV have all shown the strong preference of each virus for a latency program that is mostly specific for host B-lymphocytes. However, complete mechanisms of gammaherpesvirus latency establishment and re-activation remain unclear. The availability of molecular tools to study MHV-68 pathogenesis will likely be useful in elucidating these processes.

Immunologic control of latency

The innate and adaptive immune systems are integral in the immunological control of latency from gammaherpesvirus infection. Control of MHV-68 by the adaptive immune system occurs primarily through CD8⁺ T-lymphocytes (see Table 1 for an overview on immune mechanisms of gammaherpesvirus control). Virus-specific T-lymphocytes can be found at ≈6 days post-infection and peak at ≈ Day 10. During primary infection, an initial wave of CD8⁺ T-lymphocytes directed against lytic antigens develops, which decreases when the lytic infection is cleared. A second wave of CD8⁺ T-lymphocytes against different lytic antigens predominate during the peak of latency (Freeman et al., 2010; Gredmark-Russ et al., 2008; Stevenson et al., 1999). Most T-lymphocyte responses are directed against lytic antigens (Obar et al., 2004).

Very little is known about the mechanisms involved in CD4⁺ T-lymphocyte activation and maturation during MHV-68 infection, and very few CD4⁺ epitopes have been described (Flano et al., 2001). Nevertheless, infected CD4⁺ T-lymphocyte-deficient mice show evidence of viral re-activation and lytic infection over time, even though the initial acute infection is resolved (Cardin et al., 1996; Christensen et al., 1999; Sparks-Thissen et al., 2005). Control of acute and latent infection has also been shown to require IFNγ, whose secretion from CD4⁺ T-lymphocytes has been shown to be necessary for T-lymphocyte-mediated control of latent and productive infection (Sparks-Thissen et al., 2005).

CD8⁺ T-lymphocytes control latent infection mainly through the classic perforin, granzyme, and Fas-dependent cytotoxicity pathways (Loh et al., 2004; Tibbetts et al., 2002; Topham et al., 2001). During latency, CD8⁺ T-lymphocytes control the viral load in splenic B-lymphocytes, but not in peritoneal macrophages (Flano et al., 2000). This suggests different mechanisms of control for different cell types. However, in the absence of classical MHC Class I molecules, a large population of CD8⁺ T-lymphocytes develop that can effectively control all phases of MHV-68 infection. These T-lymphocytes are β2 microglobulin-dependent, suggesting they are restricted by one of the non-classical MHC Class I-like molecules (Braaten et al., 2006). Thus, control of latent and lytic MHV-68 infections seems to be a collaborative effort between B-lymphocytes, CD4⁺ and CD8⁺ T-lymphocytes, and IFNγ signaling (Stevenson et al., 1999).

EBV infection offers a unique viewpoint at cellular responses to gammaherpesvirus infection, since analysis of patients with IM can be tracked through the course of infection in humans. Studies from IM blood have shown that the immunodominant epitope responses in CD8⁺ T-lymphocytes are against lytic antigens that wane dramatically over time in the establishment of memory (Callan et al., 2000; Hislop et al., 2007; Steven et al., 1996, 1997). Latent antigen recognition has also been found in CD8⁺ T-lymphocytes, but this response is not as vigorous as that induced by lytic antigens. There have been reports of CD4⁺ T-lymphocyte responses against some latent and lytic antigens, but much less is known about these responses which are not very robust in peripheral blood during IM (Amyes et al., 2003; Hislop et al., 2007; Maini et al., 2000).

As mentioned above, infection with KSHV is asymptomatic in immunocompetent hosts, and its pathogenesis and mechanisms of immune control are not fully understood. KS development is >100-times more common in immunosuppressed patients (especially those co-infected with HIV); this argues for the role of a robust T-lymphocyte immune response in controlling KSHV infection and maintaining latency in immunocompetent hosts (Boshoff & Weiss, 2002). Concurrent with the host T-lymphocyte response, primary KSHV infection activates a number of immune pathways that drive the virus into a latent state. Binding of KSHV virions to host cells induces production of several cytokines, including IFN-α, -β, and -γ, as well as IFN-stimulated antiviral genes such as IRF-3 and IRF-7 (Ye et al., 2011). Induction of these cytokines shuts down the lytic program and favors the establishment of latency.

Although the pattern of latency establishment amongst MHV-68, EBV, and KSHV are not identical, they likely share conserved mechanisms. Models of MHV-68 infection may help explain the processes that lead to latent infections and the strategies gammaherpesviruses employ for immune evasion.

Activating factors and evasion of host defenses

Re-activation of both KSHV and MHV-68 requires expression of ORF-50, which encodes the transcriptional activator RTA (Replication and Transcription Activator). EBV encodes for RTA, but also encodes for Zta, a second transcriptional activator (see Figure 2 for genes involved in viral re-activation and Table 2 for activating factors). Expression of these genes is required for expression of lytic genes and viral re-activation by multiple activating factors (discussed below) (Lukac et al., 1998; Ragoczy et al., 1998; Sun et al., 1998; Zalani et al., 1996).

Table 2. Activating factors involved with gammaherpesvirus recrudescence.

Factors	Description and potential inducers
RAS/RAF	Phorbol esters, BCR cross-linkers have been shown to increase viral re-activation
Toll-like Receptors (TLRs)	Stimulators of TLR 3, 4, 5, and 9 result in increased viral reactivation
NF-κB	Inhibition of NF-κB results in increased viral replication High NF-κB expression promotes proliferation and survival of latently infected cells
Epigenetics	Histone deacetylase inhibitors increase viral re-activation and have been shown to reactivate latently infected MHV-68 cell lines
DNA Damage (DDR)	Unclear mechanism; may play a role in viral replication Chemicals such as etoposide, camptothecin, cisplatin, and staurosporine induce DDR and increase MHV-68 lytic gene expression in vitro

Chemical or physical stimulation can cause re-activation. Phorbol esters (e.g. PMA) or various BCR (B-cell receptor) cross-linking molecules such as anti-Ig/anti-CD40, induce re-activation (Speck et al., 1997). Other re-activation inducers include triggers of RAS/RAF signaling pathways (Yu et al., 2007). The expressions of factors that result in plasma cell differentiation also re-activate KSHV, EBV, and MHV-68. Thus, these chemicals that induce re-activation give clues to the mechanism of re-activation and may be informative for using MHV-68 murine models.

Gammaherpesviruses have also evolved several methods of re-activation (Table 2) that utilize host immune defenses to balance re-activation and latency in infected cells. This strategy maximizes viral persistence and allows spread to new hosts (see previous section and references within the text for immunological details). One mechanism of re-activation exploits the nuclear factor (NF)-κB signaling pathway. Inhibition of NF-κB correlates with increased gammaherpesvirus lytic replication, while high expression of NF-κB promotes proliferation and survival of latently infected cells (Barton et al., 2011; Brown et al., 2003; Krug et al., 2007, 2009). These results suggest roles for innate immunity in viral re-activation and persistence, and that compounds that affect NF-κB might be informative in the MHV-68 model.

Toll-like receptor (TLR) signaling also plays a role in gammaherpesvirus re-activation. TLR are integral components of innate immunity that recognize pathogen-associated molecular patterns (PAMPS) and initiate downstream expression of genes involved in host defense (Takeda et al., 2003). TLR can relay signals associated from host infection by bacteria, viruses, insects, and other microbes (Takeda et al., 2003). One MHV-68 study found that known stimulators of TLR-3, -4, -5, and -9 (i.e. dsRNA, LPS, flagellin, and unmethylated CpG DNA, respectively) resulted in enhanced viral re-activation (Gargano et al., 2009). This same study found that persistent latency was enhanced after periodic TLR stimulation, suggesting new pools for virus had been created that increased viral transmission to other cells (Gargano et al., 2009). These results suggest that compounds that impact upon TLR signaling might also be informative in the MHV-68 model.

Epigenetic factors also affect viral re-activation. Epigenetically-maintained viral genomes of latently infected cells are silenced largely through mechanisms that involve DNA methylation and histone deacetylation (Speck & Ganem, 2010). EBV and KSHV studies examining the role of DNA methylation of viral promoters found that inhibition of methylation causes viral re-activation (Chen et al., 2001; Countryman et al., 2008). MHV-68 studies likewise discovered that histone acetylation was sufficient to re-activate the lytic cycle as a result of demethylation of the RTA promoter (Ben-Sasson & Klein, 1981; Chang & Liu, 2000; Chen et al., 2001; Countryman et al., 2008). In vitro, histone deacetylase inhibitors (such as butyric acid [NaB]) have also been shown to re-activate MHV-68 latently infected cell lines that might also have biological relevance in vivo (Forrest & Speck, 2008; Moser et al., 2005; Yang et al., 2009).

The DNA Damage Response (DDR) pathway also plays a role in gammaherpesvirus re-activation. The DDR pathway is activated after damage to double-stranded DNA from a variety of stimuli (reviewed in Weitzman et al., 2010). Once DNA damage is detected in the cell, the cellular machinery attempts to repair it. Irreparable DNA damage may designate the cell for death via apoptosis (Weitzman et al., 2010). Although the mechanisms are unclear, the DDR is utilized by gammaherpesviruses for viral replication (Barton et al., 2011; Sinclair et al., 2006; Weitzman et al., 2010). Several chemicals have been shown to induce the DDR from latently infected cells, including etoposide, camptothecin, cisplatin, and staurosporine (Feng et al., 2002; Forrest & Speck, 2008). These results suggest that the MHV-68 model will likely be sensitive to compounds with mechanisms of action that affect these pathways.

Involvement of gammaherpesviruses in immunodeficiency and lymphomagenesis

Primary immunodeficiency

Primary immunodeficiencies (PID) are disorders that impact cellular, humoral, or non-specific host defenses. A PID can increase host susceptibility to infection and lead to malignancy (Lim & Elenitoba-Johnson, 2004). PID patients have an increased risk of developing lymphoproliferative disease (LPD), particularly after solid organ and bone marrow transplants. Of the PTLD cases, 60–70% are associated with EBV infection, although the contribution of other oncogenic viruses and autoimmune disorders cannot be ruled out (Evens et al., 2010). Examples of PID include Wiskott Aldrich Syndrome (WAS) and Ataxia Telangiectasia. WAS is an X-linked disorder characterized by autoimmunity, recurrent infections, and malignancy (Ochs & Thrasher, 2006). Ataxia Telangiectasia is an autosomal disease that can lead to neuronal degeneration, immunodeficiency, and lymphoid tumors (Biton et al., 2008).

The pathology of PTLD in PID patients is complex. Most of these patients are pediatric transplant recipients who are seronegative for EBV at the time of the transplant (Zafar et al., 2008). The majority of PTLD cases associated with these children occur within 6 months of the start of a primary EBV infection (Tran et al., 2008). These patients generally had received organs from EBV⁺ donors and their immune systems were unable to adequately block viral pathogenesis as a result of the immunosuppressive therapy used to prevent graft rejection.

Secondary immunodeficiency

Secondary immunodeficiencies occur in those born with normal immune systems that later become damaged from disease, radiation, or immunosuppressive drugs. One of the biggest contributors to secondary immunodeficiency is HIV infection and its subsequent progression to AIDS. Although AIDS patients may develop a variety of different cancers, they have a much higher risk of developing EBV- or KSHV-induced neoplasia, particularly during organ transplantation. Indeed, KSHV was first isolated from AIDS patients, is the causative agent of all KS types, and is the most common malignancy associated with HIV infection (Chang et al., 1994).

HIV infection can also lead to Non-Hodgkin’s Lymphoma (NHL) that is considered an ‘AIDS-defining condition’ (Knowles, 1999). NHL is associated with either EBV or KSHV infection and consists of morphologically diverse tumors of B-lymphocyte origin. AIDS-related NHLs can also be divided into different categories based on the site of origin of lymphoma-genesis, including nodal or extra-nodal sites, central nervous system (CNS), gastrointestinal tract, or in the body cavity (PEL) (Knowles, 1999). About 40% of these lymphomas are Burkitt lymphomas (BL) or Burkitt-like lymphomas (BLL), while the rest are divided between Large Cell Lymphomas (LCL) and Immunoblastic Lymphomas (IBL) (Knowles et al., 1988).

KSHV infection can also lead to Multicentric Castleman’s Disease (MCD) in AIDS patients, which is typically EBV⁻. Initial symptoms of MCD include fever, anemia, weight loss, and low white blood cell counts. MCD lesions occur at single (unicentric) or multiple (multicentric) nodes. MCD may occur in non HIV-infected patients, but is extremely common in those with AIDS (Cesarman, 2011). MCD patients typically have enhanced expression of pro-inflammatory cytokines (e.g. interleukin [IL]-6 and tumor necrosis factor [TNF]-α) that can enhance viral replication and increase symptoms (Ensoli et al., 2000; Wheat et al., 2005).

EBV and KSHV co-infection in AIDS patients can also lead to primary effusion lymphoma (PEL), a rare lymphoma where KSHV and EBV are often detected together (Fan et al., 2005). PEL occurs in body cavities like the pleural space, peritoneum, or pericardium. PEL symptoms are morphologically variable and typically involve a lack of expression of B-cell-associated antigens, even though the disease is of B-lymphocytic origin (Cesarman, 2011; Chen et al., 2007).

The MHV-68 model has some differences when compared to the pattern of AIDS, secondary immunodeficiencies, and development of gammaherpesvirus-induced lymphoma. It is well established that immunocompetent mice infected with MHV-68 can develop lymphoproliferative diseases in the face of immunosuppression that resembles lymphoma associated with EBV and KSHV infections (Sunil-Chandra et al., 1992a, 1994). However, mice infected with the retrovirus-induced immunodeficiency model defined as murine AIDS (mAIDS) exhibit strong lymphoproliferation and severe immunodeficiency, but are resistant to chronic infection with MHV-68 (Kulkarni et al., 1997).

Drug-induced immunosuppression

The use of immunosuppressive drugs is effective in the treatment of autoimmune diseases (such as rheumatoid arthritis [RA] and systemic lupus erythmatosis [SLE]) as well as in preventing complications in patients receiving solid organ or bone marrow transplantation procedures. Although the administration of these drugs has been shown to increase the incidence of PTLD, it is important to note that progression towards PTLD is a complex process. Multiple factors, including a mixture of pathogenic and genetic determinants, type of immunosuppressant used, dosage and duration of dosing, and type of organ transplant, all play a role in lymphomagenesis (Gutierrez-Dalmau & Campistol, 2007). However, once PTLD has been diagnosed, the course of action can range from reduction or withdrawal of the immunosuppressant, radiation therapy, anti-viral treatment, or a combination of these approaches. Here, we provide an overview of some of the most common drugs used for immunosuppression for organ transplantation and their relationship with viral-induced lymphomagenesis. We hypothesize that the MHV-68 model could be useful in the examination of immunomodulants used in both pre-clinical and clinical settings to help determine risk assessment associated with these compounds.

One of the earliest drugs used in preventing rejection as a result of SOT or HSCT was the anti-metabolite azathioprine that is incorporated into cellular DNA and inhibits purine nucleotide synthesis (Gutierrez-Dalmau & Campistol, 2007; Domhan et al., 2008). This drug also has defined effects on suppressing proliferation of B- and T-lymphocytes. Specifically, azathioprine is an inhibitor of inosine monophosphate dehydrogenase, an enzyme that is critical for B- and T-lymphocyte proliferation (Gutierrez-Dalmau & Campistol, 2007). Studies from the 1970s suggested a consistent risk of malignancy in renal transplant patients that were treated with azathioprine, although other cancer risks to patients include the development of skin tumors (including Kaposi’s Sarcoma) and lymphoma (McKhann, 1969; Weaver, 2012). Some newer anti-metabolite drugs, including leflunomide, methotrexate, and mycophenolate mofetil, were designed to be more effective and less toxic than azathioprine. While these drugs are associated with slightly lower risks of malignancy, some data from these studies are conflicting (Gutierrez-Dalmau & Campistol, 2007). Perhaps querying the MHV-68 model with these types of agents could elicit more insight into the DDR response impacted after treatment with these drugs.

Corticosteroids (e.g. prednisone) represent another class of immunosuppressants that are often used in conjunction with other immunosuppressant drugs. Corticosteroids are glucocorticoid receptor agonists that principally reduce overall inflammatory responses and the complement cascade. These drugs suppress cell-mediated and humoral immunity and can have some of the most severe side-effects since they have broad mechanisms of action. The contribution of these drugs to lymphomagenesis in PTLD patients could be their activities in promoting tumor cell resistance to apoptosis, particularly in solid tumors. These drugs can also inactivate B- and T-lymphocytes and decrease overall immune surveillance via NF-κB inhibition (Cidlowski et al., 1996; Herr et al., 2003; Jusko, 1995; Rutz, 2002). As noted earlier, NF-κB inhibition correlates with gammaherpesvirus re-activation, indicating that corticosteroid compounds may also be useful in studying the interplay between immunomodulation and gammaherpesvirus recrudescence in the MHV-68 model.

Calcineurin inhibitors (e.g. cyclosporine and tacrolimus) have also been used as part of immunosuppression regimens. These drugs, widely used in clinics, suppress the immune system by binding cell cyclophilin proteins and inhibiting IL-2 transcription. This, in turn, leads to an overall decrease in function of effector T-lymphocytes (Gutierrez-Dalmau & Campistol, 2007). One result of this then could be escape of virally-transformed cells (e.g. EBV-infected lymphocytes) to initiate re-activation of latent infection (Ryffel et al., 1992). Other reports have described the role of calcineurin inhibitors in activation of proto-oncogenic RAS (Datta et al., 2009). Indeed, these compounds may also be useful in the MHV-68 model and have already been tested with cyclosporine A (CsA). Some early animal studies found that ≈ 9% of mice infected with MHV-68 for ≥ 9 months developed lymphomas without drug treatment, and that the levels of mice with tumors increased to >50% following treatment with CsA (although the median time for tumor development did not change relative to mice not given the drug) (Nash et al., 2001; Sunil-Chandra et al., 1994). These studies indicate that an MHV-68 model to test immunomodulants will likely be an informative and amenable model to test both immunosuppression and tumor development.

Finally, some newer immunosuppressant drugs have recently been used in the treatment of certain cancers. The mammalian target of rapamycin (mTOR) is a serine/threonine kinase that is an important regulator of cell cycle progression, cell growth, protein synthesis, and cell signaling (Teachey et al., 2009). Drugs affecting this pathway include sirolimus (rapamycin) and everolimus that interfere with IL-2 function and stimulation of B- and T-lymphocytes. The drugs are relatively well-tolerated by patients when taken as the sole means of immunosuppression (Gutierrez-Dalmau & Campistol, 2007; Teachey et al., 2009). The strong effect of these drugs on halting cell cycle progression and cellular proliferation has indicated the potential utility of mTOR inhibitors in treatment of PTLD and skin cancers arising from gammaherpesvirus infection (Majewski et al., 2003; Nepomuceno et al., 2003).

The mechanisms underlying virus-induced lymphomagenesis are highly complex and may vary amongst patients and drug regimens employed, yet the MHV-68 model has a utility in studying the role of immunosuppressant drugs in gammaherpesvirus re-activation. While there may be differences in how infected mice respond to treatment with immunomodulants as compared to humans, the MHV-68 model will likely supplement knowledge about how/if they have an association with neoplasia.

Lymphomagenesis

Lymphomas in immunocompromised patients mostly occur after EBV or KSHV infection. Prior to discussion of lymphomagenesis, it is important to describe the process of gammaherpesviruses persistence in B-lymphocyte populations in immunocompetent hosts. Most lymphomagenesis models pertain to EBV and discuss the growth, default, latency, and lytic programs of EBV (Thorley-Lawson & Gross, 2004). During the growth program, EBV infects naïve B-lymphocytes and expresses a variety of latency-associated proteins. The expression of these proteins activates the naïve B-lymphocytes into proliferating blasts. This process is similar to B-lymphocyte activation by antigen. The growth program progresses into the default program, where B-lymphocytes differentiate into memory cells (Thorley-Lawson et al., 1982, 1985; Thorley-Lawson & Gross, 2004). Once the virus enters memory B-lymphocytes, the virus shuts down expression of lytic genes and enters a latent state. Latently infected cells enter the peripheral circulation and multiply by normal cell division. This process allows viral persistence without alerting the immune system of infection (Thorley-Lawson & Gross, 2004; Thorley-Lawson, 2005). This is the ideal path of infection for the virus, since it can persist in the host and escape immune surveillance. Periodically, latent virus re-activates that allows spread to new cells or hosts. The CTL response against cells expressing EBV antigens is quite robust, but the immune system is unable to completely eradicate the virus from the body (Laichalk et al., 2002). This ensures persistence and viral shed throughout the life of the host.

Since most people infected with gammaherpesviruses do not experience lymphomagenesis or symptoms of disease after virus infection, why do these viruses contribute to lymphoma? The normal EBV program of infection usurps naïve B-lymphocytes, causes them to differentiate out of the cell cycle and become resting memory cells. This process uses the normal pathways of B-lymphocyte development to maintain EBV infection in its host and spread to other hosts. However, for lymphomas to occur it is necessary for the EBV-infected cell to remain in the growth stage of the cell cycle and not become a memory cell. This scenario could occur if post-germinal center cells can no longer respond to signals that would cause differentiation into memory cells. This could happen if cells other than naïve B-lymphocytes were to get infected—the so-called ‘bystander cells’ (Kurth et al., 2000, 2003). This situation could also happen if the latently infected germinal center or memory cells were to receive a signal that resulted in inappropriate passage into the growth program, which would result in uncontrolled proliferation since these cells would not be able to exit the growth program. Under normal circumstances, most of these cells would be eliminated by the CTL response of the immune system. However, during immunosuppression, such as occurs with drug treatment during organ transplantation, the CTL response would be absent, resulting in uncontrolled cellular proliferation and potential lymphomagenesis.

In the laboratory, MHV-68 infected mice share a similar pattern of disease with EBV-infected individuals. As with EBV infection, it is thought that the CTL response is required not only for the control of acute MHV-68 infection, but to control a lifetime of persistence (Cardin et al., 1996; Ehtisham et al., 1993; Simas & Efstathiou, 1998). During establishment of latency, transitory lymph node enlargement and splenomegaly (similar to symptoms seen in IM) have been reported. Latently infected germinal center B-lymphocytes have also been detected in these animals (Simas & Efstathiou, 1998). Therefore, it is likely that MHV-68 will provide mechanistic details about lymphomagenesis that have some similarity to those occurring in humans after EBV infection and immunosuppression.

Existing models of gammaherpesvirus-induced lymphomagenesis

EBV re-activation is associated with PTLD, but the narrow host range of EBV makes it difficult to establish a small animal model. There has been considerable effort in establishment of an EBV murine model through the use of humanized mice. Humanized mice are typically immunocompromised and reconstituted with components of the human immune system (e.g. hematopoietic stem cells or peripheral blood lymphocytes) (Shultz et al., 2007). This model has been extensively used in examining the innate and adaptive immune responses to EBV infection and for B-cell lymphoma studies (Mauray et al., 2000; Melkus et al., 2006; Mosier et al., 1992; Rowe et al., 1991; Wagar et al., 2000; Yajima et al., 2008). However, humanized mice do have limitations, as their use precludes testing immunocompetent mice for modeling EBV pathogenesis. Different patterns of gene expression also exist between human and murine EBV infections (Haan et al., 2001; Shultz et al., 2007; Zychlinska et al., 2008). These differences in gene regulation could prove difficult for establishment of a small animal model of EBV infection and indicate the potential impact of the MHV-68 model.

Development of KSHV animal models has also been hampered by the narrow host range of the virus as well as technical problems in cultivating the virus in tissue culture. As a result, the details of primary infection are still poorly understood. Much of what is known about KSHV latency is derived from PEL-derived B-lymphocyte lines (Barton et al., 2011; Wen & Damania, 2010). Unfortunately, robust lytic infection in tissue culture has not been achieved and viral DNA cannot be stably maintained in KSHV cell lines (Mesri et al., 1996; Moore et al., 1996).

Humanized KSHV mouse models have been developed to study the pathogenesis of KSHV. When injected with PEL cell lines, NOD/SCID mice display malignancies that resemble PEL in humans (Wu et al., 2005). Treatment of these animals with the antiviral nucleoside analog azidothymidine (AZT) and IFNα induces apoptosis in the injected PEL cells, resulting in prolonged survival of the affected animals (Wu et al., 2005). This result shows the potential of the humanized mouse for studying gammaherpesviruses in pre-clinical settings to study KSHV involvement in lymphoproliferative disease. Humanized KSHV mouse models have also been used to show the virus’ ability to establish infection in B-lymphocytes, macrophages, NK cells, and dendritic cells of NOD/SCID mice (Adang et al., 2006; Parsons et al., 2006). It is not yet determined if KSHV humanized mice have similar limitations as the EBV mouse models.

An alternative to the humanized KSHV mouse model is infection of non-human primates (NHP), although experimentation with these animals is highly costly to maintain. Infection of rhesus macaques with the KSHV-related Rhesus Rhadinovirus (RRV) or Retroperitoneal Fibromatosis Herpesvirus (RFHV) (Giddens et al., 1985; Mansfield et al., 1999; Tsai et al., 1985, 1986; Wong et al., 1999). Pathological similarities between tissue isolated from RRV-infected macaques and lymphoid tissue found in MCD patients have helped to validate RRV as a model (Mansfield et al., 1999; O’Connor & Kedes, 2007; Wong et al., 1999). RRV is also capable of growing to high titers in tissue culture, which potentially facilitates its use as an in vitro model. The RFHV model is an alternative to RRV, although RFHV has not yet been isolated or cultured in vitro (Westmoreland & Mansfield, 2008). This difficulty, along with the cost associated with this system, may prevent the widespread use of RFHV as a model of KSHV pathogenesis.

MHV-68 as a predictive model of the role of immunomodulants in neoplasia

Experimental evidence

The MHV-68 model can provide important insight into gammaherpesvirus-associated re-activation and association with lymphomagenesis and may help predict risk for development of neoplasia after immunocompromise. Immunodeficient mice infected with MHV-68 develop a disease progression similar to immunocompromised individuals who develop EBV-driven LPD (Sunil-Chandra et al., 1994). Although the incidence of lymphomagenesis in wild-type MHV-68-infected mice is low (9%) and occurs after long incubation (≈3 years post-infection), treatment of infected mice with CsA increases the incidence of lymphoproliferative disease, of which 50% were mixed populations of T- and B-lymphocyte high-grade lymphomas (Sunil-Chandra et al., 1994). Further, some recent work has shown that MHV-68 growth transformed B-lymphocyte lines cause aggressive lymphomas in mice without intact immune systems, but not in mice that are immunocompetent (Liang et al., 2011). These reasons indicate that an MHV-68 model is likely to be informative, especially when studying immunomodulants.

The development of MHV-68 associated lymphoproliferative disease was also examined in β2-microglobulin (β2m)-deficient mice (Tarakanova et al., 2005, 2008). β2m is a component of MHC Class I molecules, and has been shown to be required for control of latent and persistent MHV68 infection (Coppola et al., 1999). Infection of β2m-deficient mice resulted in a high incidence and development of lymphoproliferative disease of B-lymphocyte origin classified as atypical lymphoid hyperplasia (ALH) (Tarakanova et al., 2005). Through an unclear mechanism, MHV-68 genes v-bcl2, v-cyclin, and M1 are thought to play a role in the onset of the lymphoproliferative disease (Tarakanova et al., 2008). Experiments with the MHV-68 model are therefore likely to elucidate some of the factors involved in some gammaherpesvirus pathologies.

The MHV-68 model has also been studied extensively in vitro, and a few murine B-lymphocyte lines have been developed from MHV-68 latently infected mice (S11 and A20-HE) (Forrest & Speck, 2008; Usherwood et al., 1996a). These cell lines are both responsive to stimulation with the phorbol ester PMA, which results in virus re-activation from latency (Table 2). An investigation into some of the chemicals known to induce viral re-activation was done recently in A20-HE cells, a latently infected murine B-lymphocyte line (Forrest & Speck, 2008). Treatment of these cells with PMA resulted in the most robust viral re-activation, but these cells were also shown to be responsive to other known re-activation stimuli (see Table 2) including etoposide, camptothecin, and staurosporine (DNA damage inducers), BCR cross-linkers, and NaB (histone deacetylase inhibitor) (Forrest & Speck, 2008). These in vitro results also suggest that an MHV-68 model is likely biologically relevant for examination of immunosuppressive agents.

Perspectives for an MHV-68 model

Ideally, an in vivo MHV-68 model would provide a model for human disease and a perspective on how to examine the clinical risk associated with novel immunosuppressive agents. Risk assessment of immunomodulants in pre-clinical settings remains problematic. The current standard for testing immunomodulants is a 2-year bioassay, which is not always predictive of human disease (Bugelski et al., 2010). Indeed, one report argues that ≈40% of all pharmaceuticals test positive in the 2-year assay (Brambilla & Martelli, 2009). Alternatively, due to differences in immune systems between humans and rodents, rodent assays may fail to predict drugs that would be toxic in humans. The 2-year bioassay also requires large numbers of animals (∼600) and long study durations, thus making a need for alternative screens for immunomodulatory compounds increasingly necessary. The MHV-68 model could provide another tool to meet this unmet need.

How then would an animal model of MHV-68 work? As summarized in Table 3, initial experiments would first need to confirm the route of infection most useful for the model. The literature describes both intranasal (IN) as well as intraperitoneal (IP) routes of infection. While IN infection is possibly the likely means by which a mouse is infected in the wild, an investigator would likely need to examine both the lungs and spleen for the presence of virus as acute lung infection occurs ≈2–3 weeks prior to splenic latency being established. In contrast, infection via the IP route might prove beneficial for research investigation, as virus could be delivered directly to the spleen. This would simplify the model so an investigator might only need to examine one organ for a presence of virus.

Table 3. Overview of MHV-68 model.

	Experimental summary	Description/comments
Characterize infection model and define molecular tools	1. Determine route of infection 2. Characterize infection 3. Characterize lytic and latent infection 4. Establish assays to detect virus	1. Intranasal (IN) vs Intraperitoneal (IP) 2. Characterize infection progression in WT and immnuosuppressed mice 3. Determine time-course of lytic infection and latency establishment 4. Determine most robust methods for viral detection (e.g. qPCR, viral titration, IHC, RNA ISH)
Validate model with known immunomodulants	1. Select immunomodulatory compound(s) likely to induce viral recrudescence 2. Determine dosing regimen	1. Choose compound(s) with well-known clinical history that are likely to induce recrudescence (e.g. CsA) as positive control for the assay 2. Establish dosing regimen and timelines of dosing after latency establishment to show proof of concept
Test model with novel immunomodulants	1. Treat animals with immunosuppressive compounds after establishment of latency to assay recrudescence 2. Determine utility of model using novel immunomodulatory agents	1. Viral recrudescence after treatment with immunomodulatory compounds provides an indication that the compound being tested has potential for association with neoplasia 2. MHV-68 model potentially provides a strong tool with which to predict how an immunomodulant might impact human disease 3. Timeline for recrudescence after immunosuppressive treatment may vary with agent used 4. Novel compounds may not induce viral reactivation but may still indicate increased risk of neoplasia via non-viral mechanisms 5. Novel compounds that induce viral reactivation may not necessarily be associated with increased risk of neoplasia 6. Need for validation of model via alternative assays

Concurrently with determining the route of infection for the model, appropriate tools for detecting viral presence need to be defined. Some of the most useful methods include analysis of viral load (viral titration), viral gene expression (quantitative PCR; qPCR), and assessment of infected tissue(s) for viral protein/nucleic acid (immunohistochemistry/IHC and RNA in situ hybridization [RNA ISH]). qPCR is the most straightforward and sensitive of these methods, and current technology is available to quickly design robust probes for multiple genes of interest. Protocols for these experiments can also be easily transferred to other laboratories to enhance assay reproducibility. qPCR would also allow high-throughput technology for large sample numbers. The other mentioned methods would certainly provide useful information but have some drawbacks. For example, viral titration is laborious and can suffer from irreproducibility as well as limit-of-detection restrictions. While IHC can provide excellent information as to specific infected cells/tissues harboring virus (which qPCR cannot provide), the method is also limited by detection problems if viral proteins are poorly expressed. IHC would also require development of a specific MHV-68 antibody. RNA ISH could alleviate some of these problems, but it is also very laborious and may still not permit detection of rare (e.g. latent) transcripts. Thus, qPCR is likely the method of choice for examining MHV-68 gene expression in novel assay development.

Once the molecular tools are better defined, the power of the mouse model can also be exploited to define gammaherpesvirus immunology. For example, MHV-68 infection cannot be controlled in animals with defective B-lymphocytes, both T-lymphocyte sub-sets, and IFNγ, indicating that immune mechanisms are required for keeping MHV-68 infection in check (Kulkarni et al., 1997). Further, numerous strains of immunodeficient mice exist, which could help further define the immunological pathways involved in the gammaherpesvirus life cycle. This information could then be useful in drug development for predicting the risk of drugs that directly impact those pathways.

A critical parameter in establishment of the MHV-68 model would also be establishment of study timelines that differentiate latency establishment from acute lytic infection. The acute phase of infection lasts ≈2 weeks in an immunocompetent mouse, followed by latency establishment. Defining latency in the course of a study is critical so that, when testing the model with immunosuppressive agents one can differentiate between lytic infection (as a part of acute infection) vs viral recrudescence (from latency). Thus, the goal would be that any viral recrudescence detected after treatment with a test compound would be solely due to effects of the agent and not to an unresolved acute infection. Nevertheless, as latency is established within ≈2 weeks post-infection, it is possible that dosing of immunomodulants could occur within a month of infection, thus dramatically reducing study timelines (especially as compared to the 2-year bioassay).

Once the infection model is better defined to include study timelines, tissues of interest, and end-points, the model should be validated with immunosuppressive compounds that have well-established clinical histories in an effort to determine the robustness of the model in predicting viral recrudescence from latency after immunosuppressive treatment. Perhaps the best drug to test this hypothesis is CsA, which has already been shown to result in a dramatic increase of lymphoma in MHV-68-infected mice (Nash et al., 2001; Sunil-Chandra et al., 1994). In a study to test this compound, latently infected mice could be dosed with CsA at various times during latency to determine if/when viral recrudescence arises. Once the experiment is conducted and validated, the investigator could then characterize other well-known immunosuppressive agents to determine if the MHV-68 model could be useful for querying about novel/unknown immunomodulants.

Suitability of MHV-68 as a model to assess immunomodulatory compounds

MHV-68 may be a useful tool for risk assessment of immunomodulatory therapeutics for recrudescence of herpes virus infections in humans. Table 4 summarizes some of the key aspects of the MHV-68 model and compares/contrasts it against what is known in humans about EBV and KSHV infection. All gammaherpesviruses share similar lifecycles and have conserved tropisms for B-lymphocyte latency establishment. Immunosuppression results in viral re-activation in all three of these viruses. Disease progression of MHV-68 in mice is similar to disease progression of EBV; in humans development of tumors is associated with all three viruses. In summary, these similarities demonstrate the potential relevance of MHV-68 infection of mice for herpes virus infections in humans.

Table 4. Utility of MHV-68 model for risk assessment of immunomodulatory compounds.

	Advantages	Disadvantages	Similarities to EBV/KSHV
Acute infection	1. Infection can be modeled in vivo 2. Infection can be modeled in vitro in various cell lines 3. Infects multiple cell types (macrophages, dendritic cells, and B-cells)	1. Route of transmission unclear 2. Only few cell types established become latently or persistently infected in vitro	1. Acute infection symptoms similar to EBV (pathogenic features of lung and lymphoid tissue) 2. CTL response to control acute infection 3. Acute infection usually resolved but virus never cleared 4. KSHV acute infection largely undefined
Latency	1. Established cell lines (A20, S11) to study latency in vitro 2. Virus can be used in vivo with default latency pathway 3. Immunosuppression can induce viral reactivation (lymphomagenesis)	1. Lymphomagenesis not as robust as EBV in vivo 2. Lymphomagenesis can take long periods to develop in immunocompetent or immunosuppressed animals	1. Latency primarily occurs in B-cells for all of these viruses 2. MHV-68 genes involved in latency are most homologous to KSHV
Pathology of PTLD	1. Use of immunosuppressive drugs can model virus reactivation and/or lymphomagenesis 2. Mouse models can be used to study a variety of viral reactivation mechanisms	1. MHV-68 may not fully represent PTLD progression 2. In vivo experiments (CsA) take a long time to cause LPD	1. Immunosuppression results in viral reactivation for all these viruses 2. Latency usually occurs in immunocompetent mice and/or humans
Lymphomagenesis	1. Established small animal model that is highly tractable 2. Potential for in vitro as well as in vivo modeling with latent cell lines and high titer viral stocks	1. Lymphomagenesis not as robust as EBV in vivo 2. Lymphomagenesis can take long periods of time to develop in in vivo models	1. MHV-68 disease progression similar to EBV 2. MHV-68 disease progression may have similarities to KSHV, but KSHV pathogenesis is not well-understood

However, there are important potential caveats to any animal model. Foremost, animal models are relevant for some aspects of the human disease, but rarely does a model fully recapitulate the full complexity of the human condition. Validation of the MHV-68 model with compounds of known human clinical significance is needed to more fully understand common mechanisms in both mice and humans. Further, the occurrence of viral recrudescence with immunomodulants in the MHV-68 model does not infer that that they will necessarily induce lymphoma in a clinical setting. One example of this phenomenon occurred in a mouse carcinogenicity study with the T-lymphocyte co-stimulation blocker Abatacept. Treatment of mice with Abatacept resulted in an increased frequency of malignant lymphomas and mammary gland tumors. Development of these tumors was possibly associated with decreased immune control of murine leukemia virus and mouse mammary tumor virus (Bristol-Meyers Squibb, 2007; Sibilia & Westhovens, 2007). Interestingly, Abatacept has not been shown to have these effects in clinical studies of human tumor viruses; thus, the relevance of these findings for human disease is currently unknown.

In conclusion, whether MHV-68 is a suitable human model for examination of immunomodulatory compounds is unclear absent further experimentation. Nevertheless, based on the biological similarities amongst gammaherpesviruses, MHV-68 appears to have potential as an alternative or a supplement to the current assays for risk assessment of immunomodulants. Taken together, the hope is that models such as these can be used to investigate and better characterize the association of immunosuppression and gammaherpesviruses with neoplasia.

Declaration of interest

This work was supported by Janssen Research and Development, LLC, a Division of Johnson and Johnson Pharmaceutical Research and Development, LLC. The authors are all stockholders with Johnson and Johnson. The authors report no conflicts of interest.

Acknowledgements

The authors would like to thank Dawn Baumgardner for review of this manuscript. The authors would also like to thank George Treacy for review of this manuscript and for his support of Biologics Toxicology. The authors also wish to thank Peter Bugelski as the inspiration for starting this work. Peter’s dedication and love of science will be sorely missed.