Keywords

6.1 Introduction

As we begin to understand more about the human microbiome and its role in health and disease, attention has turned to understanding the virome. The virome includes not only viruses that infect human cells, but also endogenous retroviruses that have colonized the human genome and viruses that infect the bacteria that make up the microbiome. The focus of this Chapter will be on the viruses that persistently infect humans, becoming life long companions so to speak. These viruses include Epstein-Barr virus (EBV), Kaposi’s sarcoma herpes virus (KSHV), hepatitis C virus (HCV) and human T-cell leukemia virus (HTLV)-1. While all of these viruses can cause non-hematologic diseases, they are also associated with a number of hematologic malignancies (Table 6.1). The challenge with understanding the etiology of these malignancies is the relative ubiquity of their associated viruses which contrasts with the rarity of these malignancies. In this chapter, we will describe the biology of the persistent viruses that are part of our virome, the hematologic malignancies they are associated with and finally, potential mechanisms that drive persistent viral infections into disease.

Table 6.1 Viral hematologic malignancies
Full size table

6.2 Hematologic Malignancies

Hematologic malignancies derive from cells of the immune system and can be of either myeloid or lymphoid origin. While there are some malignancies that are derived from cells of myeloid origin, the cancers associated with viruses that are part of our virome derive primarily from cells with a lymphoid origin, e.g. B cells, T cells, and natural killer (NK) cells. These malignancies can be classified as either lymphomas or leukemias. Lymphomas are further classified as Hodgkin or non-Hodgkin lymphomas (NHL). According to the World Health Organization (WHO) criteria for classifying lymphomas, there are >60 subtypes of lymphomas [1]. The viral-associated lymphomas include the B cell derived malignancies such as Burkitt’s lymphoma, diffuse large B cell lymphoma, primary effusion lymphomas, plasmablastic lymphomas, T and NK cell lymphomas, and a spectrum of lymphomas arising in setting of immunosuppression (e.g. post-transplant lymphoproliferative disorders, HIV). While lymphomas typically are found in lymph nodes (but not always), leukemias are generally found as an expansion of lymphocytes or myeloid cells in the blood. Of the leukemias, only adult T-cell leukemia has a clear association with a viral infection, i.e. HTLV-1.

Because T and B cells have to undergo somatic gene rearrangement to generate T cell receptors and B cell receptors, respectively, as well as somatic hypermutation in the case of B cells, the machinery needed to alter the genome is activated in these cells. This is thought to increase their susceptibility to malignant transformation. These cells undergo repeated division throughout the life of the host, further increasing their vulnerability to additional genetic alterations. Finally, as we will describe below, the viruses that infect these cells encode oncogenes creating additional opportunities for transformation.

6.3 Viruses Associated with Hematologic Malignancies

6.3.1 EBV

EBV, also called human herpesvirus 4 (HHV-4), is a member of the gammaherpesvirus family and is a double stranded enveloped DNA virus. The viral genome is ~172 kilobase pairs (kbp) and encodes genes necessary for viral replication and for viral latency. There are two strains of EBV, EBV type 1 and type 2 that exhibit both genotypic and phenotypic differences [2, 3]. EBV type 1 is the predominant strain world-wide and is the most widely studied. EBV type 2 is more common in Africa and less frequently in Western and Asian populations [2, 4]. Greater than 90% of the global population is infected with EBV [5, 6] making it one of the most successful viruses and a prominent member of the human virome.

EBV is a strict human pathogen. Oral transmission through direct contact with infectious saliva is considered to be the primary route of transmission. There are two peaks of EBV infection as measured by seroconversion, age 2–4 years and 15 years [7]. In sub-Saharan Africa, most children are infected with EBV by 2 years of age [8, 9] with some infected at less than 6 months of age [10].

EBV can infect B cells, T cells, and NK cells along with epithelial cells. Life long persistence of the virus is thought to be in B cells [11], but recent data suggests that T cells might also serve as a reservoir for EBV type 2 [12]. EBV is unique among viruses in that, in contrast to most viruses that establish lytic infection a priori, primary infection of B cells results in establishment of a latent viral infection [13]. In culture, this leads to the immortalization of B cells and expression of all the latency genes [14, 15].

The study of EBV latency has led to a defining paradigm of EBV biology, e.g. the virus’ ability to establish different latency programs in normal and malignant B cells. EBV encodes nine latent proteins: latent membrane protein (LMP)-1, LMP-2a, LMP-2b, Epstein-Barr nuclear antigen (EBNA)-1, EBNA-2, EBNA3a, EBNA3b, EBNA3c, and EBNA-leader protein (LP). The latency program of EBV in health mirrors the latency program found in EBV-associated malignancies [16]. For example endemic Burkitt lymphoma (BL) expresses only EBNA-1, diffuse large B cell lymphoma of the elderly and Hodgkin’s disease expresses EBNA-1, LMP-1 and LMP-2 and immunoblastic lymphoma expresses all the latent proteins [17]. While the majority of the cells within the EBV-positive hematologic malignancies are typically latently infected, lytic transcripts and proteins are sometimes found [18]. The contribution of viral lytic cycle proteins to malignancy remains unknown but studies in humanized mouse model implicate at least the EBV immediate early protein, Zta, in lymphomagenesis [19]. In addition, small noncoding (nc) RNAs are also expressed during latency and in EBV-lymphomas including the EBV encoded small RNA (EBER) 1 and 2, and up to 50 microRNA’s [20]. Because the EBERs are highly expressed in infected cells, in situ hybridization to detect EBERs has been widely used clinically to detect EBV in lymphoma tissues [21]. Beyond their practical role in pathology, there is also indication that EBERs modulate host cell function and contribute to malignancy [22, 23].

EBV has been classified as Class I carcinogen by the International Agency for Cancer Research [24]. When you examine the list of EBV-associated hematologic malignancies, the variety is quite striking. EBV is associated with the B cell lymphoproliferative diseases found in immunodeficient hosts, as well as the following lymphomas: Burkitt lymphoma, Hodgkin lymphoma, diffuse large B cell lymphoma, plasmablastic lymphoma, and primary effusion lymphoma. In addition to B cell lymphomas, EBV is also found in extra-nodal NK/T lymphoma, angioimmunoblastic T-cell lymphoma, hydroa vacciniforme-like lymphoma and systemic T-cell lymphoproliferative disease of childhood [25]. A unique feature of EBV malignancies is the striking geographic prevalence of some types of malignancy. For example, endemic BL in sub-Saharan Africa has a clear link to P. falciparum malaria [26, 27] while T cell lymphomas are more prevalent in Asia [28].

The unique geographic and age prevalence of EBV-associated hematologic malignancies points towards the fact that EBV in most cases is likely necessary but not sufficient to drive lymphomagenesis. That said, extensive molecular and functional analysis of EBV latent proteins points towards clear roles for the viral proteins in driving lymphomagenesis. Of the nine EBV latent proteins, EBNA-1, EBNA-2, EBNA-3a, EBNA-3c, LMP-1 and LMP-2 have been shown to be essential for transformation of B cells [29]. While a discussion of the molecular studies of EBV latent proteins function is beyond the scope of this chapter, readers can refer to several recent comprehensive reviews [29,30,31].

6.4 KSHV

KSHV (also known as human herpesvirus-8), like EBV, is a human gammaherpesvirus and belongs to the subgroup gamma-2 herpesvirus. KSHV is a double-stranded enveloped DNA virus with a genome of ~160 kbp. The virus was discovered in 1994 by Chang and colleagues [32] in attempt to discover if there was an infectious cause of Kaposi’s sarcoma. KSHV shares many similarities with EBV including transmission through saliva [33] and life long latency reservoir in B cells [34]. However, KSHV has a much more limited worldwide distribution than EBV with geographic variability in it’s distribution. In Africa, there is the so-called “KSHV belt” with greater than 50% KSHV seroprevalence [35, 36], the Mediterranean region has between 10% and 30% seroprevalence, while in northern European and USA, the seroprevalence is less 10% [37].

Infection in endemic countries occurs in children with a peak age of seroconversion around 6 years [38]. Risk of infection in childhood increases if the mother is also infected [39]. Sexual transmission in the context of the HIV epidemic was thought to increase the prevalence of this infection but whether KSHV is transmitted through semen remains controversial [40]. The current consensus is that the primary mode of KSHV transmission is saliva [41].

KSHV establishes both a latent and lytic infection. During latency in B cells, several viral proteins are expressed including latency associated nuclear antigen (LANA), and K1 as well as three cellular gene homologues, viral(v) FLIP, vIL6 and vCyclin, along with viral microRNAs [42]. While it is clear that the virus establishes life-long latency in B cells [43, 44], early attempts to infect B cells ex vivo were not successful limiting the understanding of KSHV pathogenesis to infection of endothelial cells and by analogy to B cells. Subsequently, it was found that activation of B cells prior to infection resulted in susceptibility to KSHV infection [45] and that KSHV targets a subset of tonsillar B cells [46].

KSHV is the causative agent of two B cell diseases: primary effusion lymphoma (PEL) and multicentric Castleman’s disease (MCD), a B cell lymphoproliferative disease [47]. PEL is very rare and typically found only in those with underlying immunodeficiency primarily due to HIV [48]. PEL occurs in pericardial, pleural or peritoneal spaces. PEL cells are often co-infected with EBV [49] raising the question of whether these pathogens interact synergistically to promote lymphomagenesis [50]. MCD, while not a true malignancy, is a risk factor for development of plasmablastic lymphoma [51].

LANA is the only viral protein detected in all KSHV tumors [52] leading to an intense focus on LANA function. Several lines of evidence point to LANA’s oncogenic capacity including the multifunctional nature of the protein as demonstrated in numerous studies [53]. A compelling case for LANA’s oncogenic potential is from studies using transgenic mice; expression of LANA resulted in both B-cell hyperplasia and a slow onset of B cell lymphomas in a subset of older mice [54]. Two other proteins are also consistently detected in KSHV latently infected cells: vCYC and vFLIP [55]. Transgenic mice that express vFLIP generated tumors similar to PEL suggesting a role for this protein in lymphomagenesis [56]. Clues to the role of vCYC in lymphomagenesis come from studies where vCYC transgenic mice develop lymphomas [57]. Of note, this is only when the tumor suppressor protein p53 is deficient, highlighting the complex nature of oncogenesis and the requirement for multiple factors to drive lymphomagenesis.

6.5 Hepatitis C Virus

HCV was first described in 1989 [58] and is a member of the flavi-virus family. HCV is an enveloped single stranded positive RNA virus with a genome of only 9.6 kb. Following viral entry into hepatocytes, HCV replicates in the cytoplasm [59]. The virus encodes a large polyprotein that is cleaved to yield 3 structural proteins (core, E1, and E2) and 7 non-structural proteins (p7, NS2, NS3, NS4A, NS4B, NS5A and NS5B) [60]. There are no known oncogenes encoded by HCV.

Like many other small RNA viruses, HCV exhibits significant genetic heterogeneity. There are at least six major genotypes of HCV, with varying prevalence depending on geographic location [61]. There is no known disease association with a particular genotype. Upwards of 80 million people world-wide are chronically infected with HCV [62]. The geographic prevalence of HCV varies with China, Pakistan, Egypt and Nigeria having the highest prevalence and a significantly lower prevalence is observed in higher income countries [62]. HCV is transmitted to neonates through vertical transmission from infected mothers [63]. Transmission among adults is through sexual contact, sharing of contaminated needles, and iatrogenic [64]. Primary infection with HCV is generally asymptomatic. Following primary infection, the viral infection can be spontaneously cleared or establish a life-long chronic infection with ongoing viremia [65]. This is unlike the other viruses associated with hematologic malignancies which establish latent infections in their human host.

There is no doubt that HCV is a hepatotropic virus. Infection of lymphocytes has been more controversial. Both positive and negative strand HCV RNAs were detected in PBMC of chronically infected patients [66, 67]; the presence of the negative strand RNA suggests ongoing viral replication in lymphocytes. However, in follow-up studies, B cells were non-permissive for HCV infection in one study [68] and HCV infected B cells while not productively infected, promoted trans-infection of hepatocytes in a second study [69]. More recently, HCV variants were identified that had viral envelope glycoproteins with distinct lymphotropism as compared to other isolated variants from the same chronically infected patient that had hepatic tropism [70]. The presence of dual HCV variants within the same patient is intriguing and points to a dynamic interaction between lymphotrophic and hepatotropic HCV strains within the host. A recent study has found that CD86 (B7.2) is a co-receptor for lymphotropic variants of HCV on B cells [71] providing more credence that HCV infection of B cells is part of the biology of this virus.

Diffuse large B cell lymphoma (DLBCL) is the most common lymphoma subtype occurring with HCV infection in European cases [72]. Evaluation of a larger population cohort found that HCV is also associated with marginal zone lymphomas and lymphoplasmacytic lymphomas [73,74,75]. The incidence of HCV-associated NHL is higher in regions where the incidence of underlying HCV infection is high and likely represents up to 10% of NHL cases [76]. HCV is also associated with mixed cryoglobulinemia, a low grade B cell clonal lymphoproliferative disorder and a possible precursor to malignant B cells [74].

Beyond the epidemiologic associations of HCV with NHL, a stronger case for HCV as a cause of NHL came from a seminal study in 2002 [77]. Patients that had splenic lymphoma with or without concomitant HCV infection were given interferon therapy for treatment of HCV. Lymphoma regression occurred only in those patients that were HCV positive and had a sustained virologic response to the anti-viral treatment. Subsequent studies have observed lymphoma regression in HCV+ splenic marginal zone lymphoma patients using only anti-virals [78,79,80,81,82].

In the absence of an oncogene, three possible mechanisms have been proposed linking HCV infection to NHL [83]. The first possible mechanism is chronic antigen stimulation of B cells via binding of HCV proteins to cognate antigen receptors on B cells. A second mechanism would be through binding of the viral E2 glycoprotein to the CD81 molecule on the surface of B cells driving polyclonal activation of naïve B cells. CD81 has been shown to be a high affinity receptor for HCV-E2 [84]. Either of these two mechanisms would results in chronic B cell stimulation, a consequence of which could be down-stream accumulation of genetic changes in the B cells. A final possible mechanism is through direct infection of B cells. With the more recent studies showing a lymphotropic variant of HCV [71], this possibility is gaining greater credence. However, downstream effects of persistent HCV infection on B cells are unknown.

6.6 HTLV-1

HTLV-1 is a delta RNA retrovirus with a single stranded RNA genome of 9 kbp and was isolated in 1980 [85]. Similar to other retroviruses, HTLV-1 integrates as a provirus in the host genome. HTLV-1 is transmitted through exposure to bodily fluids including breast milk, semen and infected blood products ([86,87,88]. HTLV-1 establishes a life-long infection in CD4+ T-cells as well as CD8+ T cells and dendritic cells [89].

Carriers of HTLV-1 infection are found world-wide with an estimated ten million HTLV-1 infected people [90]. Several regions have high endemicity for HTLV-1 infection including Japan, the Caribbean and South America, and sub-Saharan Africa [91]. HTLV-1 infection causes adult-T-cell leukemia (ATL). The cancer was first described in Japan in the 1970s [92]. ATL, as the name implies, is a disease that occurs in adults, typically several decades following primary infection. Less than 8% of those infected with HTLV-1 will go on to develop ATL with men having a higher risk (4.5–7.3%) than women (2.1–3.8%) [93]. There are four clinical sub-types of ATL: acute, lymphoma, chronic and smoldering [94].

HTLV-1 encodes four proteins (e.g. gag, pro, pol and env) essential for viral replication. In addition, through complex splicing, several regulatory proteins are also generated from the relatively small genome. These include Tax, Rex, HBZ (also known as bZIP), p12, p13 and p30 [95]. Of these, Tax and HBZ are thought to be the key drivers of oncogenesis [96, 97]. A puzzle early on was that although the epidemiologic data was strong that HTLV-1 infection was linked to ATL [93], detection of Tax in leukemic cells was infrequent [96]. More recently, HBZ transcripts were detected at low levels in HTLV-1 infected cells suggesting a critical role for the HBZ protein in viral oncogenesis [97]. HBZ is a transcriptional transactivator and promotes cell proliferation [97]. Tax binds to DNA and is also a transcriptional transactivator [98]. Tax is an oncoprotein based on classic definitions, e.g. immortalizes cells in vitro, can stimulate colony formation in soft agar and Tax expressing cells can generate tumors following xenoengraftment in immunodeficient mice [99, 100]. In regards to ATL etiology, Tax is thought to induce genomic instability resulting in accumulation of mutations [95].

6.7 Mechanisms of Oncogenesis

Persistent viruses encode well-characterized oncogenes but are rarely directly oncogenic. Rather, life-long infection by these members of the human virome is typically only the first step of many that lead to carcinogenesis. Unraveling the role of these viruses in hematologic malignancies has been a challenge for many scientists over the last 50 years. Through that research, some common mechanisms have been elucidated.

The age at which the persistent viral infection is acquired impacts subsequent cancer risk. For example, while early age at infection with EBV is asymptomatic [8, 10, 101], this also increases the risk for endemic Burkitt’s lymphoma [102]. In contrast, delayed infection with EBV until young adulthood leads to infectious mononucleosis, a self-limiting lymphoproliferative disease but it also is associated with an increased risk for Hodgkin’s lymphoma [103, 104]. Similarly, infection with HTLV-1 through breastfeeding increases the risk for ATL [105, 106], while delay of infection to later in life results in increased risk of tropical spastic paraparesis.

Why the age of infection leads to differential risk for hematologic malignancy is not well understood. In regard to BL, one possible mechanism would be through the increased number of latently infected circulating B cells [10]. Although these cells are not malignant, the expanded pool of latently infected cells would drive a stochastic balance whereby the chance for a secondary oncogenic hit increases. P. falciparum induces an enzyme, activation induced deaminase (AID), that has been shown in mouse models to drive the c-myc translocation characteristic of BL [107]. AID is elevated in circulating B cells in children living in areas where malaria is endemic and in tonsils of children from malaria endemic regions [108, 109]. AID is also elevated in peripheral blood prior to emergence of NHL in HIV/AIDs patients [110] suggesting that sustained activation of this enzyme in B cells is a common risk factor for B-cell lymphomagenesis. High HTLV-1 viral loads are also seen as a risk factor for ATL [111].

In all of these viral infections, there is either continual virus production as is the case with HCV or reactivation of the virus from latency as with HTLV-1, EBV and KSHV. This can lead to chronic antigen exposure throughout the life of the host and potentially driving exhaustion of the adaptive CD8+ T cell response to these pathogens [112,113,114]. The loss of the CD8+ T cell response is thought to result in failure to clear pre-malignant cells that then are driven to malignancy through expression of viral oncogenes. That many of these malignancies only occur long after the primary infection and patients with these lymphomas have exhausted viral specific T cells [115, 116] supports this model. In addition, studies in both EBV [117] and HTLV-1 [118] infected lymphocytes reveal transient expression of EBNA-1 or Tax, respectively suggesting an additional escape mechanism from CTL responses.

The above speaks to the immune cost for containing these members of our virome. With the loss of immunity due either to iatrogenic effects as a consequence of allogeneic stem cell and solid organ transplantation or due to HIV infection, the risk for emergence of lymphoproliferative diseases and lymphomas associated with these viruses is elevated [119,120,121]. Moreover, if immune function is not restored, the lymphoproliferative diseases can lead to lymphomas. This has been shown for patients with KSHV and MCD [51], HCV and mixed cryoglobulinemia [122], and EBV and LPD [123].

Many of the viral-associated hematologic malignancies require additional exogenous co-factors. For example, endemic Burkitt’s lymphoma, a common pediatric cancer in sub-Saharan Africa, is etiologically linked to both EBV infection as well as Plasmodium falciparum malaria [126]. Primary effusion lymphoma is primarily found in patients that are co-infected with HIV and KSHV [50]. The EBV EBNA-1 protein was found to enhance HCV replication suggesting a potential interaction between these two members of the virome [124].

While the oncogenic capacity of these viruses is clear, how the viral encoded oncogenes contribute to the emergence of malignancy is a bit of a conundrum. This is because it is rarely possible in healthy infected individuals to detect the expression of the viral oncoproteins in infected cells. For example, while HTLV-1’s Tax protein has oncogenic capacity, less than 5% of HTLV-1 infected cells isolated from peripheral blood express Tax and this can only be detected by sensitive PCR [125]. In EBV latently infected memory B cells—the site of long-term latency—only the EBNA-1 protein is detected and only then in memory B cells that have entered the cell cycle [117]. Moreover, it would be detrimental to long term persistence for these viruses to continuously express viral genes as the immune system would be able to eliminate those cells. One possible mechanism that would account for this paradox is the transient re-expression of viral oncoproteins that then can act as an initiator of oncogenesis by dysregulating key cellular pathways. During the transient activation, these viral oncoproteins could modulate cellular pathways including suppression of apoptosis and promotion of cell cycle.

A final thought—as we gain a better understanding of the role of microbiome in human health and disease, it seems likely that the microbiome will also have a role in leading to hematologic malignancy. The nature of that interaction is for future scientists to discover.