THE ROLE OF OTARINE HERPESVIRUS 1 IN CALIFORNIA SEA LION (ZALOPHUS CALIFORNIANUS) UROGENITAL CARCINOMA

.

By

ALISSA C. DEMING

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2018

© 2018 Alissa C. Deming

To my family

ACKNOWLEDGMENTS

I thank my parents for always supporting my educational endeavors and instilling in me a love, respect and curiosity of nature which has driven me along my path. I want to thank my mother, father, brother, sister-in-law, nanny, aunt, cousin and friends for encouraging me to follow my dreams. I also thank my primary advisors, James

Wellehan and Frances Gulland, for their constant support throughout my career.

Particular thanks go to Katie Colegrove for her constant support through this process. I am additionally grateful to the members of my committee: Rowan Milner, Rolf Renne, and Nancy Denslow, for their invaluable support and advice. Also, I want to thank Linda

Archer and April Childress for their fearless leadership in the lab. I am also thankful to my friends and coworkers including Ashley Barratclough, Cara Field, Padraig Duignan,

Christine Fontaine, Barbie Halaska, Tenaya Norris, Sophie Whoriskey, Sophie

Guarasci, Jennifer Luff, Galaxia Cortes, and Jessica Jacobs for their personal and academic support. I am also indebted to The Marine Mammal Center (TMMC) for awarding me the Geoffrey Hughes Veterinary Research Fellowship, which provided financial support for my PhD research, without which this project would have never been possible. A huge thanks to all the incredible volunteers, veterinary technicians, research biologists, and staff that make the rehabilitation and research that is done at

The Marine Mammal Center possible. And most importantly, a special thanks to all the

California sea lions with and without that made this study possible- even in your death you shed light.

4

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 7

LIST OF FIGURES ...... 8

LIST OF ABBREVIATIONS ...... 10

ABSTRACT ...... 12

CHAPTER

1 SEA LIONS, CANCER, AND ...... 14

Urogenital Carcinoma in California Sea Lions ...... 14 Cancer ...... 18 Oncogenic Viruses ...... 20 Herpesviruses ...... 22 Latently Expressed Viral Proteins: LANA/EBNA1 and vFLIP ...... 25 Host-derived Viral Genes: vBCL2 and vCDK4 ...... 26 Endogenous Viral Elements ...... 27

2 PREVALENCE OF UROGENITAL CARCINOMA IN STRANDED CALIFORNIA SEA LIONS (ZALOPHUS CALIFORNIANUS) FROM 2005-2015 ...... 32

Background ...... 32 Materials and Methods...... 33 Retrospective Data Collection ...... 33 Statistics ...... 34 Results ...... 34 Prevalence of Urogenital Carcinoma ...... 34 Urogenital Carcinoma in Adults versus Subadults and Juviniles ...... 34 Histologic Evaluation ...... 35 Metastatic Lesions in Urogenital Carcinoma ...... 35 Temporal Trends in Prevalence of Urogenital Carcinoma ...... 36 Discussion ...... 36

3 OTARINE HERPESVIRUS 1 GENOME AND ONCOGENES ...... 43

Background ...... 43 Materials and Methods...... 44 Sample Collection ...... 44 DNA Extraction and PCR ...... 45 OtHV1 and OtHV4 qPCR for Viral Load Assessment ...... 46

5

OtHV1 Genome Sequencing ...... 48 OtHV4 Genome Sequencing ...... 50 Bioinformatics and De Novo Assemblies ...... 51 Gene Comparisons: Percent Identity and Differences of Select Genes ...... 52 Phylogenic Analysis ...... 52 Results ...... 54 Histological Classification ...... 54 PCR and qPCR Results ...... 54 OtHV1 Genome Assembly ...... 55 OtHV4 Genome Assembly ...... 56 OtHV1 and OtHV4 Annotation and Genome Comparison ...... 58 OtHV1 and OtHV4 Percent Identity and Distance Comparisons ...... 61 OtHV1 and OtHV4 Represent New Oceanic Gammaherpesvirus Genus ...... 62 Discussion ...... 62

4 OTHV1 GENE EXPRESSION IN UROGENITAL CARCINOMA ...... 77

Background ...... 77 Materials and Methods...... 81 Cases and Tissue Preparation ...... 81 Basescope in situ Hybridization ...... 83 Image Analysis ...... 85 Data Analysis ...... 86 Results ...... 86 Gross Findings ...... 86 Histological Classification ...... 87 OtHV1 Gene Expression in Sea Lion Urogenital Carcinoma ...... 88 Discussion ...... 90

5 CONCLUSION ...... 105

LIST OF REFERENCES ...... 107

BIOGRAPHICAL SKETCH ...... 116

6

LIST OF TABLES

Table page

2-1 Prevalence of urogenital carcinoma in California sea lions (Zalophus californianus) of different age and sex classes ...... 38

2-2 Non-urogenital carcinoma cancer in California sea lions (Zalophus californianus) ...... 39

2-3 Number and prevalence of various stages of urogenital carcinoma diagnosed histologically in juvenile, subadult and adult California sea lions ...... 40

3-1 Herpesvirus species list for concatenated phylogenetic tree...... 66

3-2 Herpesvirus gene list for concatenated phylogenetic tree...... 67

4-1 Normal cervical or vaginal cases from California sea lions ...... 96

4-2 Urogenital carcinoma cases from California sea lions ...... 97

4-3 Basescope custom probes targeting 5 viral genes, 2 positive controls and a negative control ...... 98

4-4 Quartile results for percent positive labeling in control cases for all OtHV1 RISH probes and positive controls ...... 100

4-5 Quartile results for percent positive labeling in urogenital carcinoma intraepithelial neoplasia lesions for all OtHV1 RISH probes and positive controls ...... 100

4-6 Quartile results for percent positive labeling of BaseScope probes in invasive urogenital carcinoma lesions for all OtHV1 RISH probes and positive controls 100

7

LIST OF FIGURES

Figure page

1-1 Large, in situ abdominal in a female California sea lion with disseminated urogenital carcinoma ...... 29

1-2 Ex situ urogenital tract in a female California sea lion with disseminated urogenital carcinoma ...... 30

1-3 Ex situ urogenital tract of a female California sea lion with disseminated urogenital carcinoma ...... 31

2-1 Total number of California sea lions (Zalophus californianus) necropsied and diagnosed histologically with urogenital carcinoma and without urogenital carcinoma ...... 42

3-1 Annotated draft genome of OtHV1 ...... 68

3-2 OtHV4 MiSeq contig assembly to OtHV1 draft genome with OtHV4 contig map ...... 69

3-3 Amino acid alignment and comparison for herpesvirus core gene Terminase (ORF7)...... 70

3-4 Amino acid alignment and comparison for herpesvirus core gene Glycoprotein B (ORF8) ...... 71

3-5 Amino acid alignment and comparison for herpesvirus core gene DNA Polymerase (ORF9) ...... 72

3-6 Amino acid alignment and comparison for herpesvirus core gene Major Capsid Protein (ORF25) ...... 73

3-7 Amino acid alignment and comparison for various viral vFLIPs and host cFLIPs ...... 74

3-8 Concatenated Bayesian/Maximum Likelihood analysis phylogram based on 4 herpesvirus core genes (polymerase, terminase, glycoprotein B, and major capsid protein) using predicted amino acid sequences ...... 75

3-9 Concatenated Bayesian/Maximum Likelihood analysis phylogram based on host FLIP/FADD-like amino acid sequences and vFLIP host derived amino acid sequences ...... 76

4-1 H&E. sections of normal cervix from two adult California sea lions at different times of year ...... 95

8

4-2 H&E. Histology sections of cervical intraepithelial neoplasia and invasive urogenital carcinoma lesions from the cervix of an adult California sea lion...... 95

4-3 Histology sections of cervix from California sea lions labeled with RISH positive control probe polR2A ...... 99

4-4 Representative series of H&E and RISH labeled tissue from control cervix, CIN lesion, and invasive lesion with OtHV1 RISH probes (vBCL2, vCDK4, EBNA1, vFLIP, and vEVE), positive controls (PPIB and polR2) and negative control (dapB)...... 101

4-5 Series of early, non-invasive urogenital carcinoma of the cervix with normal transitioning into cervical intraepithelial carcinoma (CIN) ...... 103

4-6 Dot plot showing mean percent positive staining for all healthy cervix, CIN and invasive tumors ...... 104

9

LIST OF ABBREVIATIONS aa Amino acids bp Base pairs

CIN Cervical intraepithelial neoplasia

CSL California sea lions (Zalophus californianus) dapB Bacillus subtilis dihydrdipicolinate reductase

DNA Deoxyribonucleic acid ds Double stranded

EBNA1 Epstein-Barr Nuclear Antigen 1

EBV Epstein Barr Virus

HHV4 Human herpesvirus 4

HHV8 Human herpesvirus 8

Kb Kilobase

KSHV Kaposi’s herpesvirus

LANA Latency-associated nuclear antigen

ML Maximum Likelihood

NFS Northern fur seal (Callorhinus ursinus) nt Nucleotide

OCIAD1 Ovarian cancer immunoreactive antigen domain-containing protein 1

ORF Open reading frame

OtHV1 Otarine herpesvirus 1

OtHV4 Otarine herpesvirus 4

PCR Polymerase chain reaction

POL2A DNA-dependent RNA polymerase II

10

PPIB Peptidyl prolyl isomerase B qPCR Quantitative polymerase chain reaction

SS Single stranded

TMMC The Marine Mammal Center

UGC Urogenital carcinoma

µL Microliter vBCL2 Viral B-cell 2 vCDK4 Viral cyclin-dependent kinase 4 vFLIP Viral Fas-associated death domain-like interleukin-1β-converting enzyme-inhibitory protein vEVE Viral endogenous viral elements

Zc-POL2A Zalophus californianus DNA-dependent RNA polymerase II

Zc-PPIB Zalophus californianus peptidyl prolyl isomerase B

11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

THE ROLE OF OTARINE HERPESVIRUS 1 IN CALIFORNIA SEA LION (ZALOPHUS CALIFORNIANUS) UROGENITAL CARCINOMA

By

Alissa C. Deming

August 2018

Chair: James X. Wellehan Major: Veterinary Medical Sciences

Urogenital carcinoma is a common cancer in California sea lions (Zalophus californianus). This disease has been strongly associated with a herpesvirus, Otarine herpesvirus 1 (OtHV1), but a causative link has yet to be established. Here we assess the prevalence of the disease over the past decade and explore the possible oncogenic role of OtHV1 in this cancer. From 2005 to 2015, The Marine Mammal Center necropsied 932 adult California sea lions, 23% (n=211) of which had urogenital carcinoma. To better understand the oncogenic potential of OtHV1, next generation sequencing was used to sequence the OtHV1 genome. Phylogenetic analyses group

OtHV1 into the gammaherpesvirus subfamily but indicate an early branching consistent with a novel genus, with a proposed name of Marmamvirus, also containing herpesviruses identified from other oceanic species (Callorhinus ursinus- Northern fur seal and Tursiops truncatus-common bottlenose dolphin). Within the OtHV1 genome several viral genes associated with oncogenesis by other herpesviruses were identified.

These potential viral oncogenes include: vFLIP, vBCL2, LANA-like, CDK4-like and vEVE. We used RNA in situ hybridization to determine whether OtHV1 is localized and

12

transcriptionally active in primary urogenital carcinoma lesions. There was no expression of OtHV1 genes in the cervical epithelium of California sea lions without urogenital carcinoma. However, there was very high expression of all OtHV1 genes examined in cervical epithelium of California sea lions with urogenital carcinoma. This indicates that OtHV1 expression is associated with tumor tissue rather than cervical epithelium of animals without urogenital carcinoma. The homologous viral oncogenes and localization of viral expression in cervical urogenital carcinoma lesions strongly support the hypothesis that OtHV1 plays a significant role in the development of this persistently common cancer observed in California sea lions.

13

CHAPTER 1 SEA LIONS, CANCER, AND VIRUSES

Urogenital Carcinoma in California Sea Lions

California sea lions (Zalophus californianus) are marine mammals that inhabit the

Pacific coastline from southern Mexico to Alaska. Five distinct geographic populations have been established using mitochondrial DNA analysis: Pacific Temperate, Pacific

Subtropical, Southern Gulf of California, Central Gulf of California and Northern Gulf of

California (Schramm et al. 2009). The primary rookeries for the US stock are located on the California Channel Islands (San Miguel, San Nicolas, Santa Barbara, and San

Clemente) where over 99% of breeding takes place from late June to early August

(Laake et al. 2018). Animals in the Pacific Temperate stock move into both Canadian and Mexican waters and some males from Pacific Subtropical Mexican stock may spend the majority of the non-breeding season in US waters (Carretta et al. 2015). As a result of The Marine Mammal Protection Act of 1972, the Pacific Temperate population has grown from under 100,000 animals in the 1970s to an estimated 250,000 individuals as of 2014 (Carretta et al. 2015). These long-lived mammals share physiologic characteristics, coastal environments and marine food sources with people, thus act as a good sentinel species for understanding environmental and human health.

Over the past few decades, there has been a high prevalence of urogenital carcinoma in stranded California sea lions along the west coast of the US. However, there have been no documented cases of urogenital carcinoma in the Mexican population. Although the Mexican population may be more resilient to the development of this disease, the lack of documented urogenital carcinoma in this population may also be due to decreased surveillance due to the lack of rehabilitation and minimal necropsy

14

of stranded pinnipeds in Mexico (Barragán-Vargas et al. 2016) or lower chlorinated hydrocarbon levels in the Mexican population of sea lions (Del Toro et al. 2006).

In the mid-1990’s, it was reported that 18% of the subadult and adult California sea lions admitted to The Marine Mammal Center (TMMC) over the previous 15-year period had disseminated carcinoma of urogenital origin (Gulland et al. 1996). The primary originates in the genital tract (cervix, vagina, penis or prepuce) and aggressively spreads to the surrounding lymph nodes, liver, lungs, spleen, and/or kidneys (Figure 1-1, 1-2, and 1-3). Weight loss and cachexia ensue, and the sequelae of metastatic disease include vaginal/penial prolapse, hind-limb edema, hind-limb paresis/paralysis, hydroureter, hydronephrosis, and secondary bacterial leading to the death or stranding of the animal (Deming et al. 2018). Most often, sea lions with urogenital carcinoma strand with end stage disease (widespread metastases) and humane euthanasia is elected. Complete necropsies are performed on all animals that die or are euthanized at TMMC. Histological evaluation is performed on all major organ systems of select cases and on all animals with unexplained abnormalities identified during gross necropsy. Occasionally non-invasive cervical intraepithelial neoplasia (CIN) or locally invasive dysplastic epithelial lesions of the cervix or penis are incidentally observed on histology from sea lions that died of non-cancer related causes. These early genital lesions are the primary site of tumor development in urogenital carcinoma.

Herpes-like intranuclear inclusion bodies have been observed using electron microscopy and histological examination of primary urogenital tumors and metastases; immunohistochemistry using a cross-reactive Epstein-Barr virus (EBV) monoclonal

15

antibody and nucleotide sequencing classify this virus as a gammaherpesvirus

(Lipscomb et al. 2000). This virus, initially named California sea lion herpesvirus

(GenBank accession # AF193617.1), is referred to as Otarine herpesvirus 1 (OtHV1) in more recent literature. Using PCR to test for the virus in tumor tissue, OtHV1 was found to be present in all urogenital tumors examined and it was speculated that OtHV1 may be oncogenic (King et al. 2002). This study also found that OtHV1 was detectable in both muscle and brain of animals of unknown tumor status, and additional studies have identified the virus in genital swabs of animals without cancer (Buckles et al. 2007). A limited survey for OtHV1 in free-ranging animals suggested that with OtHV1 is relatively common in stranded California sea lions from the Pacific Temperate population and not exclusively associated with cancer (Buckles et al. 2007). Speculation that OtHV1 is an incidental finding in cancer animals, versus a causative agent, has been raised (Barragán-Vargas et al. 2016) and limitations of working with marine mammals, as well as the suspected multifactorial etiology of this disease, prevent studies to prove Koch’s postulates with this virus.

Additional cofactors have been associated with urogenital carcinoma in sea lions, including contaminant exposure (Ylitalo et al. 2005, Randhawa et al. 2015) and genetic predisposition (Acevedo-Whitehouse et al. 2003; Bowen et al. 2005; Browning et al.

2014). Legacy contaminant levels, such as polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT), in stranded California sea lions along the

California coast are among the highest levels reported in any marine mammal population studied (Le Boeuf and Bonnell 1971; DeLong et al. 1973; Kannan et al.

2004). California sea lions with urogenital carcinoma were found to have eight and six

16

times higher blubber levels of PCBs and DDTs, respectively compared to animals without urogenital carcinoma (Randhawa et al. 2015). An IHC study of expression of hormone receptors in California sea lions, suggests that environmental contaminants that interact with steroid hormone receptors and endogenous hormones may play a role in the initiation and/or progression of this cancer (Colegrove et al. 2009a). Loss of estrogen receptor expression with continued progesterone receptor expression in intraepithelial urogenital carcinoma lesions in sea lions is a pattern similar to that observed in in women. Thus, endogenous hormones or exogenous hormone mimics may contribute to sea lion urogenital carcinoma in a similar manner to how hormone cofactors play a role in cervical carcinogenesis in women.

Pinnipeds have a synchronized estrus, embryonic development, diapause, pregnancy, lactation, and parturition cycle. Studying potential hormone-associated in this species is eased by this predictable temporal pattern of their reproductive cycle. California sea lions typically give birth in June. Breeding, parturition and nursing occur in rookeries on the Channel Islands of the southern California coast and on islands off Baja California, Mexico (Odell, 1975). Estrus, breeding and conception occur 3-4 weeks following partition, with a 2-month embryonic diapause and implantation occurring in early October. The average gestation period is 9 months, with parturition in June (Reeves et al. 1992). Typically, increased serum progesterone occurs in all adult female in the fall, with persistent elevation in pregnant animals in the spring compared to non-pregnant animals (Greig et al. 2007). Estrogen levels, however, have not been documented to follow a predictable pattern. Histologic examination of the healthy cervix and proximal vagina of the California sea lions has shown variability

17

throughout the year, ranging from pseudostratified columnar to stratified squamous

(Colegrove et al. 2009b). During pupping and estrus, the endo- and exo- cervix have increased number of invaginations and glands with squamous and pseudostratified columnar epithelium. This is followed by a thinner pseudostratified columnar cervical epithelium in diapause (August to early October) that persists through fall, winter and spring. Cervical and vaginal thickness is typically thicker- three to eight cell layers thick during pupping and estrus, with some areas of squamous epithelium reaching 10-12 cell layers thick in some invaginations during this period (Colegrove et al. 2009b). These thick layers observed during pupping and estrus can sometimes be difficult to distinguish between the early stages of dysplasia and metaplasia characteristic of urogenital carcinoma.

The primary goal of this work is:

1. Determine the current prevalence of urogenital carcinoma in stranded

California sea lions examined at TMMC;

2. Determine the genomic sequence of OtHV1 and identify potential viral oncogenes; and

3. Assess OtHV1 oncogene expression in primary cervical tumors using RNA in situ hybridization.

The purpose of this work is to help clarify whether OtHV1 plays a role in the development of urogenital carcinoma and determine if California sea lions would be a good model for studying naturally occurring, virally-induced .

Cancer

The Merriam-Webster dictionary defines cancers as: “a malignant tumor of potentially unlimited growth that expands locally by invasion and systemically by

18

metastasis”. The characteristics of unlimited growth, invasion and metastasis occurs when there is a malfunction of cellular regulatory mechanisms. When cells behave appropriately, they follow a strict set of rules that control cellular replication, localization and cell death, thus maintain homeostasis within the body. Cancer cells have mutations or malfunctions in these cell regulatory pathways that initiate increased cellular proliferation, immortality, and metastasis, disrupt homeostasis and result in .

Cancer most often arises from an accumulation of genetic mutations in oncogenes, including tumor suppressor genes or genes involved in DNA replication and repair (Damania et al. 2004). The significant redundancy of cellular checkpoints exemplifies the importance of these pathways. Breakdown of these checkpoints allow a cell to transform from a normal cell into a malignant cell. This malignant transformation results in a self-sufficiency in growth signaling, insensitivity to growth-inhibitory signals, evasion of , limitless replication, sustained angiogenesis, invasive characteristics and the ability to spread to distant sites (metastasis). These mutations can be the result of genetic predisposition, random mutation events, environmental carcinogens or infectious agents, and are often a combination of these factors.

Regardless of the inciting cause, neoplastic cells often have disruption in similar cellular pathways. Established in the early evolution of multicellular life, these fundamental pathways are shared across species. Studying and understanding the etiology of cancer in all species opens the door for potential prevention, screening, and immunization.

19

Oncogenic Viruses

It is estimated that viral infections contribute to 15-20% of cancers in humans and animals (Parkin, 2006). Oncogenic viruses include both RNA (Retrovirdae and

Flaviviridae) and DNA virus families (Hepadnaviridae, , and

Papillomaviridae). Human oncogenic viruses associated with include:

Human papillomaviruses (HPV; cervical cancer, anogenital cancer, skin cancer, ), Epstein-Barr virus (EBV; Burkitt’s lymphoma, Hodgkin’s disease, , post-transplant ), Kaposi’s sarcoma-associated herpesvirus (KSHV; Kaposi’s sarcoma, Multicentric Castleman’s disease, and primary effusive lymphoma), and C (), and human T- lymphotropic virus 1 (HTLV-1; adult T-cell leukemia). In non-human species, similar viruses are associated with neoplasms, from primates to chickens to frogs (Klein 1972).

Historically, studying viruses in animals have helped scientists uncover the conserved cellular pathways that drive cell replication and apoptosis at the most fundamental cellular level. The first oncogenic retrovirus was described by

Peyton Rous in 1911 while studying the etiology of in domestic fowl (Rous

1911). Coined Rous Sarcoma Virus (RSV), this positive sense RNA virus is composed of four genes: gag- encoding capsid protein, pol- encoding reverse transcriptase, env- encoding envelope gene, and src- a host acquired gene that was found to phosphoralate tyrosine in host proteins involved in the regulation of mitosis. Nearly 50 years after RSV was discovered, researchers Harry Rubin and Howard Temin established that it was the src gene that gave RSV it’s oncogenic characteristics- driving uncontrolled cell growth in the domestic fowl sarcomas. The src gene in RSV was the first retroviral oncogene discovered. Since then, various naturally occurring viral

20

infections in a variety of host species have served as models for studying viral oncogenesis and have helped reveal the basic, highly conserved cellular processes that drive cellular division and apoptosis in cells across species.

In order to replicate as obligate intracellular parasites, viruses have established mechanisms to exploit cellular proliferation pathways, prevent apoptosis, and evade the . The delicate balance between immune surveillance and viral latency can be interrupted in an immune suppressed host, resulting an increased risk of cellular transformation. This is particularly evident in AIDS patients, who have a significantly higher risk of developing virally induced cancer (Grulich et al. 2007). Recent research shows AIDS patients are 500 times more likely to be diagnosed with Kaposi’s sarcoma,

19 times more likely to be diagnosed with anal cancer, 12 times more likely to be diagnosed with B-cell non-Hodgkin’s lymphoma, 3 times more likely to be diagnosed with cervical cancer, and 3 times more likely to be diagnosed with

(Hernández-Ramírez et al. 2017). This has also been seen in certain animal populations, where cancer associated with viral infections can be correlated to immune compromise, environmental contaminants, genetic predispositions, and/or co-infections with other pathogens (Vogt 1965, Lancaster and C Olson1982, Ylitalo et al. 2005,

Randhawa et al. 2015, Acevedo-Whitehouse et al. 2003; Bowen et al. 2005). This strengthens the evidence that there is a complex relationship between immunity, infectious agents and cancer, adding value to studying the pathophysiology of virally induced cancers in a natural model with an intact immune system and representative microenvironment.

21

Herpesviruses

Herpesviruses are large, double stranded, linear DNA viruses that encode 100 to

200 genes. They are encased within an icosahedral capsids and wrapped in tegument layer and lipid bilayer envelope. Herpesviruses exhibit a pattern of latent and lytic infections, and most often are subclinical, persistent infections that do not produce clinical signs in their well-adapted host following initial infection. Immune compromise, the effects of latency protein or other inciting events can induce recrudescence or disease processes ranging from rashes to encephalitis to neoplasia. There are well over

200 herpesviruses known to infect humans and animals. All herpesviruses share a common evolutionary origin and have coevolved with their particular host. When herpesviruses manage to infect aberrant hosts, it is often fatal.

The order contains viruses classified into three families:

Herpesviridae (infecting reptiles and mammals), (infecting amphibia and fish) and (infecting molluscs) (Mettenleiter, 2009). The family

Herpesviridae is further divided into 3 subfamilies: ,

Betaherpesvirinae, and . Alphaherpesviruses are characterized by a short reproductive cycle, rapid destruction of the host cell, and have the ability to replicate in a wide variety of host tissues. Typically, viruses in the alpha-subfamily infect mucoepithelial cells and spend their latency located in the neuron. Some examples of alphaherpesviruses include virus 1 and 2 (oral and genital cold sores), varicella–zoster virus (chickenpox), and virus (PRV). have a tissue tropism for epithelial mucosa, neurons, leukocytes or salivary glands, and often establish latency in leukocytes. Some viruses in the beta-subfamily include human (HHV5), (HHV6A and B), HHV7, Elephantid herpesvirus

22

1 (aka Elephant endotheliotropic herpesvirus), and (infects rodents).

Gammaherpesviruses are further subdivided into four genera: ,

Rhadinovirus, , and . Viruses in the Gamma subfamily are lymphotropic, establishing lifelong latency in host B- and/or T- cells, with intermittent lytic replication in epithelial and/or fibroblast cells. Examples of gammaherpesviruses include: Epstein-Barr virus (HHV4), Kaposi’s sarcoma-associated herpesvirus (KSHV),

Rhesus monkey (VRP), Equid herpesvirus 2, murid herpesvirus 68, common bottlenose dolphin gammaherpesvirus 1 and Otarine herpesvirus 1 and 4. A clinically important feature shared by some of the gammaherpesviruses is their ability to induce neoplastic disease in the host (Damania, 2004).

Herpesviruses exhibit co-speciation with host lineages, making them fairly host specific. Herpesviruses share high amino acid sequence similarities among many of their core gene products, supporting a common evolutionary origin. Additional accessory genes can be acquired from the host genome, giving a virus permissiveness in a particular host as well as potential pathogenic properties. Evolutionarily, at least

20% of herpesvirus genes are thought to be host in origin, supporting the theory that some viruses are capable of stealing host genes and incorporating them into their own genome (Alcami 2003, Elde and Malik 2012). This results in a virus with the ability to express host-like proteins. Phylogenic analysis shows that independent, parallel acquisition of some accessory genes has occurred- meaning that viruses have stolen the same host gene independently at different points in their evolutionary history and these particular genes are more closely related to their host genes than to other herpesviruses with similar genes. This convergent evolution demonstrates the important

23

role these genes play in the virus’ survival strategy. In particular, certain host-acquired accessory genes have been shown to play fundamental roles in signal transduction pathways, inducing cellular proliferation, blocking apoptosis, inactivating tumor suppressor pathways and avoiding a host’s immune surveillance (Elde and Malik 2012,

Rappoport and Linial 2012). In addition to host genes, herpesviruses have also been shown to acquire genes through horizontal transfer of retroviruses (Aswad 2015).

Both gammaherpesviruses endemic in humans have been linked to human malignancies: Epstein-Barr virus (EBV), which is the etiological agent of nasopharyngeal carcinoma, African Burkitt’s lymphoma, post-transplant lymphomas,

Hodgkin’s disease and some gastric cancers (Damania et al., 2004; Pagano et al.,

2004), and Kaposi’s sarcoma-associated herpesvirus (KSHV), linked to Kaposi sarcoma

(KS) and two lymphoproloiferative diseases (primary effusion lymphoma and multicentric Castleman’s disease) (Ganem 1997, Damania et al., 2004; Pagano et al.,

2004). In vitro studies show EBV and KSHV have various genes that play roles in modulating host cellular pathways including transcription, cell cycle progression, signal transduction, and apoptosis. However, their role in tumorigenesis in vivo is limited due to the lack of appropriate lab animal models (Grinde, 2013). One way to fill the lack of traditional laboratory models is to study naturally occurring viral-associated cancers in other species.

Interference of host cellular pathways can result in cellular transformation, causing cell immortality, uncontrolled proliferation, and predisposing a cell to malignancy. Cellular mechanisms controlling these fundamental pathways are highly conserved across species and viruses have independently developed ways of

24

interfering with similar pathways. Below are some examples of well-established viral oncogenes or tumorigenic characteristics, and their roles in malignant transformation of a host cell. These categories include oncogenic latent viral proteins, host-derived viral oncogenes, and endogenous retroviral elements.

Latently Expressed Viral Proteins: LANA/EBNA1 and vFLIP

Following initial lytic infection of a herpesvirus, the host’s immune response causes the virus to retreat into a latent stage which results in persistent, lifelong infection. In latency, herpesviruses are circularized into an episomal form of DNA and express a limited set of viral genes. These genes provide core functions that ensure duplication of the viral genome during host cell division but does not produce large amounts of viral proteins as seen in the lytic stage (Ballestas et al. 1999, Cotter and

Robertson 1999). These latently expressed genes also help in blocking apoptosis and altering immune signaling pathways to ensure viral persistence, survival of the infected host cell, and helps the virus evade detection by the host immune system. A common finding in KSHV- and EBV-related cancers is the majority of tumor cells are latently infected with these viruses (Boshoff et al. 1995, Puglielli et al. 1997). KSHV expresses

Latency-Activating Nuclear Antigen (LANA), vCyclin and vFLIP in latency, and there are patterns of differential expression of these genes during different stages of latency.

LANA has been found to be the only protein expressed in all stages (Renne et al. 2001).

During latency, EBV protein expression is limited to EBNA1, EBNA2, EBNA3 and LMP.

The formal analog of LANA in EBV is EBNA1, and like LANA, EBNA1 is expressed during all stages in latency. Studies focusing on LANA and EBNA1 have provided evidence that they have the potential to be major drivers of cellular transformation by interacting with viral promoters and host transcriptional factors which have effects on

25

host cellular pathways including cell cycle, apoptosis, immune modulation, epigenetic modification, and altered signal transduction pathways (Renne et al. 2001).

Host-derived Viral Genes: vBCL2 and vCDK4

Cellular B-Cell Lymphoma 2 (cBCL2) is a family of proteins that play critical roles in regulating programmed cell death pathways by blocking or inducing the release of proapoptotic signaling molecules from mitochondria (Newmeyer and Ferguson-Miller

2003). Various proteins in this family can be pro-death (Bax, Bak) or anti-death (BCL2,

BCL-XL). Interactions and neutralization between these pro- and anti- apoptotic proteins suggest that the ratio of these proteins within a cell determines the susceptibility of a cell to death signals (Oltvai et al. 1993). Homologues of BCL2 have been identified in a huge range of species, from nematodes to humans (Gross et al. 1999), and overexpression of host specific BCL2 has been found to universally protect against cell death. The highly conserved nature of this family of proteins indicates a fundamental role in controlling apoptosis. Dysregulation of this pathway has been identified to be at the root of cellular transformation in numerous types of cancers.

Several large DNA viruses have acquired a viral BCL-2 (vBCL2) homologue from their host as a means to obstruct cell death pathways in infected cells, thus ensuring cell survival and viral propagation (Polster et al. 2004). DNA virus families found to have a functional BCL2 homologue include the Herpesviridae, Poxviridae, and

(Polster et al 2004; Kvansakul and Hinds 2013). All gammaherpesviruses sequenced to date encode at least one BCL2 homologue derived from their respective hosts. EBV has two BCL2 homologues, BHRF1 and BALF1 (Henderson et al 1993; Marshall et al 1999), and KSHV has one, viral Bcl-2 (vBcl-2) (Sarid et al 1997; Cheng et al 1997).

Interestingly, host derived vBCL2 often include a portion of the gene responsible for

26

blocking apoptosis but lack the regulatory mechanisms that control the activity of their cellular counterparts as well as the membrane-anchor domain. In these instances, expression of vBCL2 results in inhibition of apoptosis, but is not subject to the regulatory mechanisms of its cellular counterpart. In addition, vBCL2 has also been found to play an important role in inhibiting autophagy, thus inhibiting intracellular pathogen processing via the CD4+ antigen presenting pathway, aiding in evasion of the host’s innate immune system by virally infected cells (Feng et al. 2013).

In order for a virus to survive long term and replicate effectively in a host, it must avoid detection by the host’s immune system. The ability to avoid immune surveillance allows viruses to persist latently in hosts and move host cells towards malignancy

(Michaelis et al., 2009). Thus, virally induced oncogenic transformation can occur directly by effecting the function of major cellular growth control proteins; or more indirectly through immune suppression, chronic , suppression of apoptosis and induction of genetic instability (zur Hausen 2009). This is especially true in viruses that are known to develop latent infections that periodically reactivate. Herpesviruses, in particular, have developed a number of ways to avoid host immune detection, control the cell cycle and prevent apoptosis in host cells, and as a result, several cancers in humans and animals have been associated with herpesvirus infections.

Endogenous Viral Elements

As discussed above, viruses have the capacity to acquire genes from their host potentially conferring a selective advantage to the virus (Ploegh 1998). More recently, the opposite has also been found to be true; viral elements have been found to be incorporated into mammalian genomes (Parrish and Tomonaga 2016). Both retroviruses and DNA viruses are capable of integrating their viral genes or entire viral

27

genomes into the host genome, and if this integration occurs in germ line cells these viral elements can be passed to successive generations. Known as endogenous viral elements (EVE), these relics of ancient infections account for at least 8% of the human genome. Most often these integrations result in non-coding introns and “junk” DNA after the accumulation of mutations over time. However, some integrations do still maintain intact open reading frames (Mayer et al. 1999). When this occurs, it can be because the

EVE has become adapted to serve important functions in the host, such as an EVE that has been identified as a key factor driving the evolution from egg-bearing to placental mammals (Mi et al. 2000).

28

Figure 1-1. Photo courtesy of Dr. Padraig Duignan. Large, in situ abdominal metastasis (encircled by dashed line) in a female California sea lion with disseminated urogenital carcinoma. The right and left ureters have become obstructed by the metastatic mass, resulting in hydronephrosis, a common sequelae of this disease.

29

Figure 1-2. Photo courtesy of Dr. Padraig Duignan. Ex situ urogenital tract in a female California sea lion with disseminated urogenital carcinoma. A large, primary mass in the cervix (encircled by dashed line) and a large, abdominal metastatic mass associated with the sublumbar lymph nodes (white arrow) are shown. Sagittal cross sections of the right and left kidneys have been made to reveal the moderate to severe hydronephrosis of the left kidney (white arrow) and a lesser affected but enlarged right kidney.

30

Figure 1-3. Photo courtesy of Dr. Padraig Duignan. Ex situ urogenital tract of a female California sea lion with disseminated urogenital carcinoma. A sagittal cross section has been made in the large, primary mass of the cervix (white arrow) showing a locally extensive, white to beige mass extending throughout the cervix, extending from the vaginal vault to the distal uterus.

31

CHAPTER 2 PREVALENCE OF UROGENITAL CARCINOMA IN STRANDED CALIFORNIA SEA LIONS (ZALOPHUS CALIFORNIANUS) FROM 2005-2015

Background

Cancer is rare in free-ranging marine mammals, with the exception of two populations: beluga whales (Delphinapterus leucas) in the St. Lawrence River (De

Guise et al. 1994), and California sea lions (Zalophus californianus) on the west coast of the United States (Gulland et al. 1996). Urogenital carcinomas comprise the majority of cancers documented in California sea lions (Gulland et al. 1996), whereas a variety of cancers (, carcinomas, and ) have been identified in belugas that have high levels of contaminants in their blubber (De Guise et al. 1995). In recent years, the annual prevalence of cancer reported in necropsied belugas from this population has declined (Lair et al. 2016); however, the current trend in prevalence of urogenital carcinoma in California sea lions is unknown. Environmental contaminants are also common in California sea lions, with cancer being eight- and six-times more likely in animals with higher blubber levels of PCBs and DDTs, respectively (Randhawa et al. 2015). Thus, contaminant exposure may play a role in initiation or promotion of urogenital carcinoma, which likely has a multifactorial etiology. Other cofactors associated with urogenital carcinoma in California sea lions include Otarine Herpesvirus

1 (OtHV1) and genetics (King et al. 2002; Bowen et al. 2005; Browning et al. 2014).

Urogenital carcinoma is most frequent in adults, and sexual transmission of OtHV1 has been suggested (Buckles et al. 2007). There is no evidence for direct transmission of a clonal cancer (Leathlobhair et al. 2017).

From 1979 to 1994, metastatic carcinoma was diagnosed at post-mortem examination in 18% (66/370) of wild subadult and adult California sea lions (Gulland et

32

al. 1996). Urogenital carcinoma originates in the genital epithelium, and local invasion and widespread metastasis are common (Lipscomb et al. 2000). From 1991 to 2000,

15% (88/568) of adult California sea lions, most of which were younger adults, at TMMC were diagnosed with urogenital carcinoma (Greig et al. 2005). Cancer continues to be observed in stranded California sea lions along the California coast, but the current prevalence has not recently been evaluated. Our objective was to assess the occurrence of urogenital carcinoma in California sea lions from 2005 to 2015 and to determine if temporal changes in prevalence have occurred.

Materials and Methods

Retrospective Data Collection

The Marine Mammal Center (TMMC) is a rehabilitation facility that examines stranded marine mammals from 700 miles of California coastline (Grieg et al. 2005).

Necropsy records were reviewed from juvenile, subadult, and adult California sea lions examined at TMMC from January 1, 2005 to December 31, 2015. Records included gross findings, sex, weight, standard length, nutritional status (score 1-5), evidence of pregnancy (lactation, placental scarring, fetus), stage of sagittal crest development, and tooth development and wear. Standard length, tooth wear and sagittal crest development were used to determine age class as defined in Grieg et al (2005): pup (0-

1 yr), yearling (1-2 yr), juvenile male (2-4 yr), subadult male (4-8 yr), juvenile or subadult female (2-5 yr), adult female (5+ yr), and adult male (8+ yr). Juvenile and subadult age classes were pooled and considered separately from adults in some data analyses, as the prevalence of cancer is known to be significantly higher in sexually mature animals

(Grieg et al. 2005). Samples from all major organs were examined histologically from all fresh dead or euthanized animals. Any patient with a diagnosis of cancer was included

33

in this study. Cases histologically characterized as carcinoma in situ or locally invasive carcinoma with no metastasis (Colegrove et al. 2009b) were defined as early stage disease. Advanced stage disease was defined by gross and histological evidence of metastasis.

Statistics

Prevalences were compared using chi square analysis, and a beta regression model was used to examine differences in proportions using the R package,© 2016 The

R Foundation (accessed March 2017).

Results

Prevalence of Urogenital Carcinoma

Necropsies were performed on 985 juveniles and subadults, and 932 adults

California sea lions. There were 1,003 females and 914 males (Table 2-1). Fourteen percent (263/1917) all sea lions necropsied had cancers (Table 2-1), of which 90%

(237/263) were grossly or histologically identified as urogenital carcinoma, while 10%

(26/263) were other cancers including uterine leiomyomas, lymphomas, and adrenal cortical carcinomas (Table 2-2).

Urogenital Carcinoma in Adults Versus Subadults and Juviniles

Of the juveniles and subadults, 2.6% (26/985) had urogenital carcinoma, whereas adults had a significantly higher prevalence of 23% (211/932; P<0.001). In juveniles and subadults, metastasis was documented in 77% (20/26) of cases, and urogenital carcinoma was the cause of death in all but one of these animals. There were

13 to 30 (mean 19) cases of urogenital carcinoma in adult California sea lions each year

(Fig. 2-1). Annual prevalence varied among years (range 11%–51%) but when year was included in the model to test for a trend in the proportions (i.e., a decline or increase

34

over time), there was no statistically significant change in prevalence over time (type II chi squared test P=0.560).

Histologic Evaluation

Histology was performed on 232 of the 237 urogenital carcinoma cases identified on gross necropsy. Of these cases, 22% (50/232) were early stage and 78% (182/232) were advanced stage (Table 2-3). Carcinoma in situ of the cervix, vagina, penis, or prepuce with no evidence of metastasis was diagnosed in 19% (45/232) of urogenital carcinoma cases, and locally invasive carcinoma with no evidence of metastasis in 2%

(5/232). All early stage cases were considered incidental findings, and the primary cause of stranding was most often associated with domoic acid intoxication, trauma, leptospirosis, or pneumonia. Carcinoma in situ of the cervix, vagina, penis, or prepuce with metastasis was diagnosed in 17% (39/232) of urogenital carcinoma cases; metastases but no primary lesion were seen in 12% (27/232) of urogenital carcinoma cases; and invasive carcinoma in the cervix or vagina or in the penis or prepuce with metastases was present in 50% (116/232) of urogenital carcinoma cases. Metastatic disease was considered the primary cause of death in 95% (172/182) of California sea lions with advanced stage urogenital carcinoma.

Metastatic Lesions in Urogenital Carcinoma

Metastatic lesions were identified in a number of organs (Table 2-4), especially the sublumbar lymph node. Nodal enlargement secondary to metastases frequently resulted in obstruction of ureters with sequela of hydroureter and hydronephrosis.

Hydronephrosis was present in 62% (114/185; male=19/42, female=95/143) of advanced stage cases. Clinical signs in advanced stage cases included hind limb swelling, paresis, or paralysis (49%, 90/185; male=21/42, female=69/143) and genital

35

swelling or penile or vaginal prolapse (62%; 114/185; males=29/42, female=85/143).

None of these findings were present in early stage cases. State of nutrition (SON) was significantly lower (two-tailed t-test, P=0.008) in animals with advanced stage disease

(SON mean=1.98 SD=0.68, n=182) compared to animals diagnosed with early stage disease (SON mean=2.34 SD=0.72, n=50).

Temporal Trends in Prevalence of Urogenital Carcinoma

A high prevalence of cancer in stranded California sea lions was confirmed, with

263 cancer cases observed over a decade and urogenital carcinoma accounting for

90% of cases. The number of urogenital carcinoma cases observed from 2005 to 2015 were more than triple the number described for this species between 1979 and 1994

(Gulland et al. 1996), and more than twice the number observed from 1990 to 2001

(Greig et al. 2005). Although the prevalence in necropsied animals was highly variable among years with no significant temporal trend, the increase in numbers of cases is likely a reflection of the increasing California sea lions population which has expanded from approximately 161.000 in the mid-90s to over 250,000 today (Carretta et al. 2015).

Discussion

Most cancers occurred in adult California sea lions in our study, with few cases in juveniles, a distribution common in humans and animals (Howlader et al. 2017). The prevalence of carcinoma was highest in adult males, which is of interest because they also have the highest levels of organochlorine contaminants (Randhawa et al. 2015) and a higher prevalence of OtHV1 infection (Buckles et al. 2007). We did not diagnose

UGC in yearlings or pups, however, there are rare reports of UGC in yearling California sea lions (Greig et al 2005). The virus associated with urogenital carcinoma, OtHV1,

36

has been identified in California sea lion pups (Buckles et al. 2007), although no known cases of urogenital carcinoma have been documented in this age class.

In free-range California sea lions, diagnosis of urogenital carcinoma in live animals is difficult. The number of cases with advanced stage metastatic disease, but with no observed primary genital lesion, identified the risk of false negative genital when testing live animals for carcinoma. Unbiased sampling of free-ranging

California sea lions is challenging, making its true prevalence hard to establish. There is an inherent sampling bias associated with assessing prevalence of any disease in a stranded group of animals, as the results of sampling those animals reflects the factors influencing their stranding. Reasons other than cancer, such as changes in prey distribution, toxic algal blooms, and disease outbreaks can increase strandings, altering the prevalence of cancer in the sample (Greig et al. 2005). However, the long-term data

(1997 to present) collected on sea lions at TMMC, minimized the effect of short term influences on the mortality patterns that we observed.

The high number of animals that were diagnosed with this aggressive cancer demonstrated that urogenital continues to be common in stranded California sea lions.

Thus, urogenital carcinoma in sea lions can serve as a spontaneous disease model to help elucidate the complex cellular derangements that accrue in cellular transformation, malignancy, and metastasis (Browning et al. 2015).

37

Table 2-1. Prevalence of urogenital carcinoma in California sea lions (Zalophus californianus) of different age and sex classes necropsied at The Marine Mammal Center, Sausalito, California, from 2005-2015. Juvenile and subadult age classes were pooled because juvenile and subadult females cannot be easily distinguished and considered separately from adults Number of sea Number of sea Age class Sex lions with Prevalence (%) lions necropsied carcinoma Adult Females 821 174 21.2

Males 111 37 33.3

Total 932 211 22.6

Juvenile–subadult Females 182 5 2.7

Males 803 21 2.6

Total 985 26 2.6

Total Females 1003 179 17.8

Males 914 58 6.3

All 1917 237 12.4

38

Table 2-2. Non-urogenital carcinoma cancer in California sea lions (Zalophus californianus) that were necropsied and evaluated histologically at The Marine Mammal Center, Sausalito, California, from 2005-2015. Age, sex, tumor type, location and metastasis are listed. The presence of urogenital carcinoma is also noted, as several animals had both urogenital carcinoma and an unrelated cancer concurrently. (A=adult, J=juvenile, F=female, M=male). S Age Metastasis Concurrent urogenital e Primary tumor class location carcinoma x A F Adrenal adenoma None Cervix- carcinoma in situ A F Adrenal cortical adenoma None Cervix- carcinoma in situ A F Adrenal cortical carcinoma Lymph node* Vagina and cervix- carcinoma in situ A F Leiomyoma- cervix None Vagina and cervix- carcinoma in situ A F Leiomyoma- uterus; None Vagina and cervix- - tongue carcinoma in situ A F Leiomyoma- uterus None No A F Leiomyoma- uterus Oropharynx Vagina and cervix- invasive carcinoma A F Leiomyoma- uterus None No A F Leoimyoma-uterus None No A F Leiomyosarcoma- uterus None No A F Lymphoma- None No A F Lymphoma- multiple LNs Multiple lymph No nodes A F Lymphoma- T-cell Duodenum, Vagina and cervix- invasive ileum, liver, carcinoma lymph nodes A F Ovarian papillary Uterus and ovary Vagina and cervix- invasive * carcinoma A F Ovarian papillary adenoma None No A F Pancreatic adenoma None No A F Poorly differentiated sarcoma-eye; None No Hibernoma- free abdominal mass A F Poorly differentiated sarcoma-eye Lungs No A F Hemangioma- uterus None Vagina and cervix- invasive carcinoma A F Spindle cell carcinoma- abdominal Unknown (patient Unknown (patient released) mass () released) A F - tonsil Lymph nodes, No lung, liver A F Lipoma- vagina None No A M Seminoma None No J M Hemangioma- liver None No J M Lymphoma- thoracic masses and Lungs, heart No gastric mass

39

Table 2-3. Number (n) and prevalence (%) of various stages of urogenital carcinoma diagnosed histologically in juvenile, subadult and adult California sea lions (n=232; (Zalophus californianus) necropsied at The Marine Mammal Center, Sausalito, California, from 2005-2015. Early stage cases had no evidence of metastasis, and included carcinoma in situ and locally invasive carcinoma; advanced stage cases had metastatic lesions. Females and males are considered separately, as well as together, to show prevalence of each stage of urogenital carcinoma in all cases and between sexes.

Stage of urogenital Number of urogenital carcinoma Histologic description carcinoma cases (%)

Females Males All

16/57 45/232 Early stage Carcinoma in situ 29/175 (17) (28) (19)

Locally invasive carcinoma 5/175 (3) 0/57 (0) 5/232 (2)

12/57 39/232 Advanced stage Carcinoma in situ with metastases 27/175 (15) (21) (17)

100/175 16/57 116/232 Invasive carcinoma with metastases (57) (28) (50)

No primary identified but 13/57 27/232 14/175 (8) metastases present (23) (12)

40

Table 2-4. Distribution of metastases detected via histologic evaluation of California sea lions (Zalophus californianus) with advanced urogenital carcinoma (n=182). Location of metastatic lesion Number (%) of animals

Lymph node 171 (94) Lung 127 (70) Liver 87 (48) Uterus 78 (55) Ovary 66 (37) Kidney 66 (36) Urinary bladder 52 (29) Adrenal gland 47 (26) Mesentery 33 (18) Diaphragm 26 (14) Spleen 22 (12) Urethra/ureter 11 (6) Skeletal muscle 11 (6) Small intestine 11 (6) Aorta 8 (4) Pericardium 6 (3) Prostate 5 (12) Vertebrae 5 (3) Large intestine 5 (3) 4 (2) Heart 4 (2) Brain 4 (2) Mediastinum 4 (2) Renal capsule 2 (1) Testicle 3 (8) Tonsils 3 (2) Pituitary 1 (1)

41

d

e 250

i

s

p

o non-UGC cases

r

c

e

n

200 11%

s UGC cases

n 11% 21

o

i

l 20

a

e

s 150

t

l

u

d 16%

a

f 19

o

100

r

e b 22% 25% 46% m 34% 29% u 40% 17 31%

N 50 13 51% 30 21 19 15 15 21

0

5

6

7

8

*

0

1

2

3

*

*

0

0

0

0

9

1

1

1

1

4

5

0

0

0

0

0

0

0

0

0

1

1

2

2

2

2

0

2

2

2

2

0

0

2

2

2 Year

Figure 2-1. Total number of California sea lions (Zalophus californianus) necropsied (grey and black combined) and diagnosed histologically with urogenital carcinoma (black) and without urogenital carcinoma (grey) at The Marine Mammal Center by year from 2005 to 2015. Each year’s prevalence is indicated by percentage above bar. *=2009 was an El Niño year

42

CHAPTER 3 OTARINE HERPESVIRUS 1 GENOME AND ONCOGENES

Background

A suspicion of viral etiology in California sea lion urogenital carcinoma was raised when herpes-like intranuclear inclusions were observed on histology and electron microscopy of the cervical lesions (Gulland et al. 1996; Lipscomb et al. 2000). Further identification of the virus was performed via immunohistochemistry, using an Epstein-

Barr Virus monoclonal antibody, and PCR of the DNA-dependent DNA polymerase

(dpol) which classified OtHV1 as a gammaherpesvirus (Lipscomb et al. 2000; King et al.

2002) Two gammaherpesviruses that infect humans, Kaposi’s sarcoma-associated herpesvirus (KSHV) and Epstein-Barr virus (EBV), have been associated with cancer, as well as several others in non-human gammaherpesvirus species. Because of the potential oncogenic characteristics of gammaherpesviruses and the strong correlation between OtHV1 and urogenital carcinoma in California sea lions, it has been hypothesized that OtHV1 plays a role in the development of urogenital carcinoma in sea lions in a similar way as KSHV and EBV (King et al. 2002). An age-prevalence study showed genital swabs were PCR positive for OtHV1 in 46% of adult males (n=52), 22% of adult female (n=72) and 6% in juvenile sea lions (n=120), and OtHV1 was most commonly detected in swabs from the lower genital tract compared to pharyngeal swabs and peripheral mononuclear cells (Buckles et al. 2007). These findings indicate sexual transmission as a primary route of infection, but one positive premature pup in the study indicated the potential for vertical transmission as well.

Currently only two genes from OtHV1 have been partially sequenced, 448 bp of the terminase gene (GenBank accession number AF236051.1) and 1,345 bp of the

43

DNA polymerase gene (GenBank accession number AF236050.1) listed under

California sea lion herpesvirus in GeneBank (King et al. 2002). To date, there has not been any attempt to identify potential viral oncogenes in the OtHV1 genome. Recent advances in DNA sequence, specifically next generation sequencing platforms, have made it possible to generate novel viral genomes from clinical samples (Radford et al.

2012).

Here we have 3 specific aims:

1. Design and validate novel OtHV1 specific PCR and qPCR.

2. Generate and annotate the OtHV1 and OtHV4 genomes using next generation

sequencing.

3. Perform comparative analysis and phylogenetic relationship of the OtHV1 and

OtHV4 to other herpesviruses to determine if any previously identified

herpesviruses oncogenes are present in OtHV1.

Materials and Methods

Sample Collection

The cervical tumor tissue used for sequencing OtHV1 was collected on necropsy from a stranded, adult California sea lion (Zalophus californianus) that presented emaciated and lethargic to The Marine Mammal Center in 2012. Humane euthanasia was performed following diagnosis of metastatic urogenital carcinoma during admission examination (Marine Mammal Protection Act permit number 18786). At necropsy, urogenital carcinoma with widespread metastasis was confirmed. Cervical tumor tissue was collected and frozen (-80°C). Samples from all gross lesions and major organs were fixed in 10% neutral buffered formalin. Efforts to cultivate OtHV1 in cell cultures

44

from fresh cervical tumor tissue was unsuccessful, therefore OtHV1 genome was sequenced directly from frozen cervical tumor tissue.

A vaginal swab collected from a Northern fur seal (Callorhinus ursinus) in Pribilof

Islands, Alaska, US in 2011 (Marine Mammal Protection Act permit number 932-

1905iMA-009526), previously identified to by our lab to be positive for OtHV4 (Cortés-

Hinojosa et al. 2016), was used for OtHV4 genomic sequencing. This animal was a wild

Northern fur seal sampled during routine health assessments of the population in determined to be clinically healthy on routine examination; however, histology was not performed on the cervix as pre-mortem tissue biopsies are not collected during routine health assessments of this species. Urogenital carcinoma has not been documented in

Northern fur seals, and to date, no disease process has been identified in association with OtHV4 infection.

DNA Extraction and PCR Screening

DNA was isolated from the California sea lion cervical tumor using a commercial kit (DNeasy Blood and Tissue, Qiagen Inc., Valencia, California, USA). For the Northern fur seal vaginal swab DNA was isolated using a Maxwell automated extractor and

Maxwell 16 Buccal Swab Purification kit (Promega, Madison, Wisconsin, USA). DNA extracts from California sea lion cervical tumor and Northern fur seal vaginal swab were screened using herpesvirus conventional consensus PCR as describe in VanDevanter et al. (1996) targeting DNA-dependent DNA polymerase gene with Platinum Taq DNA

Polymerase (Invitrogen, Carlsbad, CA, USA). Briefly, round 1 primers DFA, IKL, KG1 were used in a semi-nested PCR followed by round 2 primers TGV and IYG. PCR conditions for round 1 and round 2 were as follows: denaturation for 5 minutes at 94˚C followed by 40 cycles of denaturation at 94˚C for 30 seconds, annealing at 46˚C for 60

45

seconds and extension at 72˚C for 60 seconds, followed by an elongation step at 72˚C for 7 minutes. PCR products were electrophoresed in 1% agrose gels and all products of expected band size (round 1: 450 bp, round 2: 220bp) were extracted using QIAquick

Gel Extraction kit (Cat No. 28706; Qiagen Inc.) and sequenced at the University of

Florida Interdisciplinary Center for Biotechnology Research using ABI 3130 DNA sequencer (Life Technologies, Carlsbad, California, USA).

OtHV1 and OtHV4 dpol specific PCRs were then performed on both samples. A novel OtHV1 dpol primer set was designed, and previously published PCR primers specific for OtHV4 was used (Corté s-Hinojosa et al. 2016). For the novel OtHV1 PCR primers set we designed primers using Primer3 in unique nucleotide areas of OtHV1 to ensure specificity between OtHV1 and OtHV4 dpol. OtHV1 specific primers were:

OtHV1_polF: 5’-CTTCGCATGGGTGGACTACT-3’, OTHV1_polR: 5’-

TCATGCCTACTAGCAGCAGC-3’ with an expected band size of 344bp. We used platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA, USA) and conditions were as follows: denaturation for 5 minutes at 94˚C followed by 45 cycles of denaturation at

94˚C for 1 minute, annealing at 58 ˚C and extension at 72˚C for 60 seconds, and an elongations step at 72˚C for 7 minutes. PCR products were electrophoresed in 1% agrose gels, and positive bands of expected size were extracted using QIAquick Gel

Extraction kit (Cat No. 28706; Qiagen Inc.). All PCR positive products were sequenced at the University of Florida Interdisciplinary Center for Biotechnology Research using

ABI 3130 DNA sequencer (Life Technologies, Carlsbad, California, USA).

OtHV1 and OtHV4 qPCR for Viral Load Assessment

OtHV1 and OtHV4 viral loads, in California sea lion cervical tumor and Northern fur seal vaginal swab respectively, were measured using hydrolysis probe real-time

46

PCR (qPCR) with primers and probe targeting unique regions of each virus. qPCR was conducted using 7500 Fast Real-Time PCR System (Applied Biosystems) with a standard fast protocol. For OtHV4 primers, probe and qPCR conditions were followed as described in Corté s-Hinojosa et al. (2016). For OtHV1, novel qPCR primers and probe were designed using Primer express version 3.0 (Applied Biosystems, Foster

City, California, USA) to target unique areas of the OtHV1 DNA-dependent DNA polymerase gene (dpol; forward primer: OtHV1qPCRF 5’-TCCCACGCTGTTTCGAATG-

3’, reverse primer: OtHV1qPCRR 5’- AGCTCCGAGTCGTGTACACAGTAT -3’, probe:

OtHV1 Probe 5’-{FAM}-TCGCGCTCGCATCGGCA-{BNQ]-3’) with Black Hole Quencher probe and FAM reporter dye. PCR primers and conditions in above section

(OTHV1_polF, OTHV1_polR) were used to generate 344 bp fragment of OtHV1 dpol providing template for standard curve generation. PCR product was run on 1% agrose gel, extracted with QIAquick Gel extraction kit and quantified using NanoDrop 8000 spectrophotometer (Therma Scientific, Wilmington, Delware, USE). Ten-fold dilution series ranging from 1 to 10^7 copies per well were made with PCR product and diluted with Tris-EDTA buffer. A 10^-1/slope was calculated for efficiency (Bustin et al. 2009).

Each 20µL reaction was composed of 4uL DNA extract, 10µL qPCR Master mix

(TaqMan® Fast Universal PCR Master Mix 2X, Applied Biosystems), 3µL of molecular grade water, and 1µL of each primer at a dilution of 18µM. Sensitivity and specificity was performed using 10-fold dilution series of OtHV1 dpol PCR product discussed above to estimate the analytical sensitivity (lower limit of detection for the assay). Ten diagnostic samples from California sea lion cervical tumors previously submitted to the authors’ laboratory, which tested positive for OtHV1 on PCR and were confirmed to be

47

OtHV1 by Sanger sequencing. An additional 10 OtHV1 negative California sea lion cervix samples were included to assess species-specific negative controls. To test for viral specificity, four OtHV4 positive samples from Northern fur seal vaginal swabs, one

OtHV3 positive sample from California sea lion esophageal ulcer, and one herpesvirus positive sample from sea turtle cutaneous fibropapilloma were used to assess primer and probe specificity. All samples were run in duplicate with an internal positive control of 18S ribosomal universal eukaryote DNA primer/probe (VIC Probe, Applied

Biosystems) on a 7500 Fast Real-Time PCR System (Applied Biosystems) using standard Fast protocol with thermocycling conditions: 94°C for 20 seconds once, followed by 45 cycles at 94°C for 3 seconds and 60°C for 30 seconds. All reactions were ran on 96-well polypropylene plates (Olympus Plastics, Genesee Scientific) and had 3 no template (molecular grade water) for negative controls and 1 to 10^7 standard curve in triplicate. Data was analyzed using 7500 Fast Real-Time PCR System software, giving results as viral copies detected per nanogram of DNA in the reaction.

OtHV1 Genome Sequencing

Illumina MiSeq, Pacific Biosciences (PacBio) single molecule real time (SMRT), and Sanger sequencing platforms were used to obtain the OtHV1 genome from the cervical tumor DNA extract. This sample contained primarily host genomic DNA

(California sea lion, Zalophus californianus), however based on the high viral load seen on the novel OtHV1-specific qPCR, we estimated at least 1% of the total DNA would be of OtHV1 origin. We hypothesized this could provide appropriate coverage of the OtHV1 genome for de novo assembly. The entire MiSeq and PacBio runs were dedicated to total DNA extracted from this single tumor sample to minimize herpesvirus strain variability.

48

For the MiSeq run, DNA library for OtHV1 was constructed using Nextera XT sample preparation kit as directed by manufacture (Illumina). The library was quantified using a Qubit 2.0 spectrophotometer (Life Technologies, Carlsbad, CA) and analyzed for size distribution using a Bioanalyzer (Agilent). Sequencing was performed on an

Illumina MiSeq with 2 x 300-nt paired-end reads following standard Illumina protocols.

Preliminary bioinformatics analysis (see methods below) of these results produced:

~150Kbp for the OtHV1 composed of three large non-overlapping contiguous sequences (contigs). To address this, further attempts to bridge the gaps in the MiSeq generated OtHV1 contigs was performed using PacBio SMRT.

Total DNA extract from the California sea lion cervical tumor was stored at -80˚C prior to PacBio sequencing. The PacBio RS II platform (English et al. 2012) at the

University of Florida Interdisciplinary Center for Biotechnology Research (UF-ICBR) was used for genomic sequencing. The purpose was to generate the longest possible insert libraries for long-read sequencing, in order to facilitate OtHV1 genome assembly. DNA quality was evaluated using the Agilent TapeStation with a Genomic Tape. The average

DNA size had peak >60kb. Quantitation was performed by fluorescence (QUBIT,

ThermoFisher). Eight micrograms of high MW genomic DNA were applied to a G-tube

(Covaris, Inc.) using fragmentation conditions for 20 kb. AMPure magnetic beads (Cat#

A63881, Beckman Coulter) at 0.45:1:00 beads to sample ratio, were then used to clean the DNA before library construction reactions.

Large-insert (20 Kb) library construction was performed using 5 micrograms of G- tube fragmented DNA according to the PacBio protocol (P/N 100-286-000-07) with a few modifications. Briefly, SMRT bell adaptors were attached to the sample fragments

49

in four basic steps: DNA damage repair, DNA end repair, ligation of adaptors, and exonuclease III/VII digestion. The final library yield was approximately 1.2 micrograms

(~25% of the original mass).

The final library was further size-selected on the Electrophoretic Lateral

Fractionator (ELF, SageBioSciences), using the 0.75% Agarose (Native) Gel Cassettes v2 (Cat# ELD7510), specified for 0.8-18 kb fragments (3.61-hour run). This was done in order to maximize the average length for the “Reads of Insert” during sequencing. The final library was quantitated by fluorescence (QUBIT, ThermoFisher), and sized on the

Agilent TapeStation (genomic tape). A total of ~200 ng of approximately 24 kb library fragments were recovered by pooling fractions corresponding to wells 1-3 from the ELF.

This material was used to set up sequencing reactions in the PacBio RS II for two

SMRT cells (Single Molecule Real Time) according to the manufacturers protocol, using v3 SMRT cells and P6/C4 chemistry reagents, 6-hr movies. A 100pM on-plate loading concentration was used. Approximately 70,000 reads with an average polymerase read length of ~15kb were obtained per SMRT cell.

Additionally, OtHV1 gap closure was attempted using novel primers (design based on generated draft genome) for conventional PCR using platinum Taq DNA

Polymerase (Invitrogen, Carlsbad, CA, USA), GC RICH PCR System (Sigma-Aldrich,

Atlanta, GA, USA) and TaKaRa Ex Taq® Hot Start Version (Takara Bio, Mountain View,

CA, USA) and Sanger sequencing.

OtHV4 Genome Sequencing

Illumina MiSeq and Sanger sequencing platforms were used to sequence the

OtHV4 genome from total DNA extracted from a Northern fur seal vaginal swab. For the

MiSeq run, a DNA library was constructed using Nextera XT sample preparation kit as

50

directed by manufacture (Illumina). The library was quantified using a Qubit 2.0 spectrophotometer (Life Technologies, Carlsbad, CA) and analyzed for size distribution using a Bioanalyzer (Agilent). The library was then sequenced on an Illumina MiSeq with 2 x 300-nt paired-end reads following standard Illumina protocols. Preliminary bioinformatics analysis (see methods below) of these results produced: ~145Kbp for the

OtHV4 genome composed of 7 large contigs. Additional OtHV4 gap closure was attempted using novel primers (design based on generated draft OtHV4 genome) for conventional PCR using platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA,

USA), GC RICH PCR System (Sigma-Aldrich, Atlanta, GA, USA) and TaKaRa Ex Taq®

Hot Start Version (Takara Bio, Mountain View, CA, USA).

Bioinformatics and De Novo Assemblies

Sequence data for both OtHV1 and OtHV4 MiSeq runs were stored on

BaseSpace and transferred into CLC Genomics 9.5.3 for trimming, genome assembly and annotation. To generate the OtHV1 genome, a de novo assembly was performed on trimmed reads using CLC Genomics with default settings. Contigs larger than 2kbp were selected and a BLAST search for similarity with known herpesviruses in the nucleotide collection (nr/nt) database was performed to identify viral contigs from sea lion genomic DNA. The PacBio reads were then assembled using the MiSeq OtHV1 draft genome (3 large contigs) as scaffolding in an attempt to fill the gaps and confirm the MiSeq generated OtHV1 draft genome assembly.

For OtHV4, de novo assembly was performed using the same method described above for the OtHV1 MiSeq assembly. In addition, the Norther fur seal MiSeq paired- end reads were assembled using the OtHV1 draft genome as a scaffold. For both

OtHV1 and OtHV4 primer design and Sanger sequencing were used to confirm areas of

51

assembly as well as attempt to fill the gaps in the genomes. To predict putative protein coding genes, open reading frames (ORFs) with a start codon of AUG that were larger than 150 amino acids were identified using CLC Genomics. Additional ORFs larger than

100aa were also explored, but only included when associated with a E-value <1^-2

BLAST hit. These ORFs were then searched against the pfam database and Genome

Annotation Transfer Utility (GATU, http://athena.bioc.uvic.ca/virology-ca-tools/gatu/).

Additional ORFs were identified using Blast searches and a small number did not have a known function and/or predictable putative protein.

Gene Comparisons: Percent Identity and Differences of Select Genes

Percent identity and distances between selected genes comparing OtHV1,

OtHV4, DeGHV1, HHV8, HHV4, and Sameriine gammaherpesvirus 2 was performed using CLC genomics and default settings. Comparisons were made for the following core genes: DNA-dependent-DNA polymerase (dpol, ORF9), terminase (term, ORF7), glycoprotein B (GlyB, ORF8), and major capsid protein (MCP, ORF25). Additionally, the suspected host-derived gene vFLIP from OtHV1 and OtHV4 were also compared to the other viral vFLIP homologues, including representative viruses from Poxviridae,

Percaviruses, and Rhadinoviruses. In addition to vFLIP genes, host cellular FLIP

(cFLIP) genes were also included in the comparison from representative animal species including diverse vertebrates. The purpose of this comparison is to better understand the relationship between viral vFLIPs and host cFLIPs to elucidate the temporal acquisition of the host derived gene by the virus.

Phylogenic Analysis

Four herpesvirus core genes (polymerase, terminase, glycoprotein B, and major capsid protein) and one oncogene (vFLIP) were used for viral phylogenetic analysis.

52

Herpesviruses with complete or almost complete genomes were selected for this comparison. Twenty-seven herpesviruses representative of all known herpesvirus genera were used for phylogenetic analysis (Table 3-1 and Table 3-2). Herpesviruses with complete or almost complete genomes were selected for this comparison to allow for generation of a concatenated tree including four core herpesvirus proteins (aa sequence comparison). Predicted amino acid sequences were compared with sequences in the databases of GenBank (National Center for Biotechnology

Information, Bethesda, Maryland, USA) using BLASTP. (Altschul et al, 1990)

Homologous sequences were aligned using MAFFT

(https://www.ebi.ac.uk/Tools/msa/mafft). (Katoh and Toh, 2008)

Bayesian analyses of each alignment were conducted using Mr. Bayes 3.2.6 on the CIPRES server with mixed amino acid substitution models, gamma distributed rate variation, and a proportion of invariable sites. (Miller et al, 2015; Ronquist et al, 2012) A total of 4 chains were run, and statistical convergence was assessed via the average standard deviation of split frequencies and potential scale reduction factors of parameters. Human alphaherpesvirus 3 (GenBank Accession # NC_001348) was selected as the outgroup for the core gene analyses, and Latimeria_chalumnae FLIP

(GenBank Accession # XP_014345763) was used as the outgroup for the vFLIP analysis. The initial 25% of 2,000,000 iterations were discarded as burn in. When no significant differences in topology were seen in the core gene analyses, the core gene alignments were concatenated, and the concatenated alignment was used for Bayesian analysis using the same methodology.

53

Maximum Likelihood (ML) analyses of each alignment were performed using

RAxML-HPC2 version 8.0.24 on the CIPRES server, (Stamatakis et al, 2008) with a gamma distributed rate variation and proportion of invariable sites. Human alphaherpesvirus 3 (GenBank Accession # NC_001348) was selected as the outgroup for the core gene analyses, and Latimeria_chalumnae FLIP (GenBank Accession #

XP_014345763) was used as the outgroup for the vFLIP analysis. Bootstrap analysis was used to test the strength of the tree topology; 1000 resamplings were performed.

(Felsenstein, 1985) When no significant differences in topology were seen in the core gene analyses, the core gene alignments were concatenated and the concatenated alignment was used for maximum likelihood analysis using the same methodology.

Results

Histological Classification

Histologic examination of the California sea lion cervical tumor showed an infiltrative, unencapsulated and poorly defined mass extending from the cervical mucosa into the underlying muscularis. Similar neoplastic cells were present in surrounding lymphatic vessels. Intranuclear inclusions morphologically consistent with herpesvirus (OtHV1) were present in some neoplastic cells. This presumably represented the site of primary neoplasia. Metastatic lesions in the uterus, lung, liver and lymph nodes showed numerous variably sized islands and nests of similar neoplastic cells.

PCR and qPCR Results

The California sea lion cervical tumor sample was PCR and qPCR positive for

OtHV1 and negative for OtHV4. The OtHV1-specific qPCR indicated approximately 1.7 x 106 viral copies detected per nanogram of DNA in the cervical tumor. The Northern fur

54

seal vagina swab was PCR and qPCR positive for OtHV4 and negative for OtHV1. The

Northern fur seal vaginal swab sample had an OtHV4 viral load of approximately 3.44 x

105 viral copies detected per nanogram of DNA in the vaginal swab.

All 10 diagnostic samples from the known positive California sea lion cervical tumors were positive with OtHV1 PCR primers (confirmed to be OtHV1 with Sanger sequencing) as well as OtHV1-specific qPCR. All 10 OtHV1 negative California sea lion cervix samples were negative. All nonOtHV1 herpesvirus positive samples were also all negative on OtHV1 PCR and qPCR.

OtHV1 Genome Assembly

MiSeq, PacBio and Sanger sequencing were used to obtain the genomic sequence of OtHV1 from a cervical tumor collected from an adult California sea lion. De novo assembly of OtHV1 sequence data generated a draft OtHV1 genome of 3 non- overlapping contigs (from 5’ to 3’): contig1- 40,761 bp with an average coverage of

39.46, contig2- 92,701 bp with an average coverage of 59.05, contig3- 18,856 bp with an average coverage of 54.16 (totaling 152,318 bp). These 3 large contigs contained genes with high similarities to known herpesviruses (Figure 3-1). BLAST analysis revealed 100% sequence identity to the previously sequenced OtHV1 DNA-dependent-

DNA polymerase (dpol) gene (Genbank accession number AF236050.1), terminase gene (Genbank accession number AF236051.1), and glycoprotein B (GenBank accession number KP861870.1), confirming identification of this sequence as OtHV1.

The genome organization of OtHV1, typical of gammaherpesviruses, includes conserved gene blocks composed of mostly structural genes, lytic DNA replication genes, and packaging genes, as well as host-derived cellular homologous accessory genes and undefined hypothetical proteins. Initial assembly of the OtHV1 genome

55

revealed a gene order that is roughly co-linear to HHV8 and HHV4; therefore, contigs were arranged based on synteny with HHV8 and HHV4. Based on this arrangement, the two gaps lie between ORF11 and ORF17, and ORF69 and ORF75. Further results on

OtHV1 genes will be discussed in the “OtHV1 and OtHV4 annotation and gene comparison” section below.

Additional sequencing of the cervical tumor sample using PacBio SMRT sequencing provide long sequence reads (averaging 15Kbp) to bridge the gaps

(particularly areas of tandem repeats or high GC content) between the OtHV1 contigs and confirm the assembly of the MiSeq generated OtHV1 genome. A de novo assembly as well as an assembly using the OtHV1 contigs generated from the MiSeq run as a scaffold were performed and both produced similar results. Analysis of all PacBio reads showed OtHV1 sequences represent >6% of the total sequences generated on the 2

SMRT cells which generated 1.8 gigabasepairs (Gbp) with 172,930 reads having a N50 read length of 14,646 bp and had approximately 40x coverage to the OtHV1 reference genome. Upon further inspection, both the MiSeq assembly and the PacBio assembly had four areas of the OtHV1 genome had significantly higher coverage than others.

When factoring out these areas of high coverage, the average coverage across the

OtHV1 genome dropped to 10-15x. Further discussion of the source of these spikes in coverage is addressed in the “OtHV1 and OtHV4 annotation and gene comparison” section below.

OtHV4 Genome Assembly

The MiSeq paired-end Illumina run from the Northern fur seal vaginal swab DNA library generated a total of 1.84 gigabase pair (Gbp) of raw data sequence from

9,447,394 reads, with an average length of 195 bp per read. Assembly of the OtHV4

56

using the OtHV1 genome as a scaffold and a de novo assembly on trimmed reads both produced similar results: ~145Kbp for the OtHV4 genome composed of 7 large contigs.

These contigs had high similarity to known herpesviruses including OtHV1 scaffolding.

The following results will focus primarily on the assembly using OtHV1 as a scaffolding to help guide the order of the non-overlapping contigs.

Assembly of the OtHV4 reads to the OtHV1 draft genome as scaffolding resulted in a total of 21,186 mapped reads (0.22% of reads were mapped) with an average length of 158 bp per read. Thus, 0.18% of bases sequenced were of herpesviral origin in this sample. The low amount of viral DNA in the sample compared to the >99% host genomic DNA resulted in low numbers of OtHV4 reads in the run, causing low coverage in the OtHV4 assembly with several gaps. This resulted in an OtHV4 draft genome with

145,453 bp. In total there were 7 large contigs that compose the OtHV4 genome from the de novo assembly with average coverage ranging from 14.04 to 19.94 and several areas with <5x coverage. The OtHV4 draft genome had 4 areas with much higher coverage compared to the coverage throughout the majority of the genome (Figure 3-2), similar to the OtHV1 assembly. BLASTX analysis revealed 100% sequence identity with the previously sequenced OtHV4 polymerase and glycoprotein B, confirming identification of this sequence as OtHV4.

To predict putative protein coding genes, open reading frames (ORFs) larger than 150 amino acids were identified and searched against the pfam database and

Genome Annotation Transfer Utility (GATU, http://athena.bioc.uvic.ca/virology-ca- tools/gatu/). Additional ORFs were identified using BLASTX or BLASTP searches, and a small number did not have a predicted function and/or related putative protein. A

57

minimum of 74 open reading frames were identified in OtHV4 draft genome (Figure 3-

2). The coding capacity is likely much larger when considering smaller ORFs, dynamic viral mRNA editing, the use of alternative translation initiation codons, non-coding

RNAs, initiation/termination codons, and alternative splicing.

OtHV1 and OtHV4 Annotation and Genome Comparison

Typical of other gammaherpesviruses, the organization of OtHV1 and OtHV4 draft genomes includes a centrally located conserved gene block composed mostly of structural genes, lytic DNA replication and packaging genes (Figure 3-2). This area is flanked on either side by other core genes and accessory genes on the 5’ and 3’ ends, some of which are suspected host-derived homologous and undefined hypothetical proteins. Initial assembly of the OtHV1 genome revealed a gene order that is roughly co-linear to HHV8, therefore contigs were arranged based on synteny with HHV8.

Based on this arrangement, the two gaps in OtHV1 lie between ORF11 and ORF17, and between ORF69 and ORF75 (Figure 3-1). While the gaps in OtHV4 lie between

ORF11 and ORF17, Ot17 and Ot18, Ot19 and Ot2, and ORF 75 and Ot22 (Figure 3-2).

In OtHV1, the 5’ contig was 40,762 bp, which had 2 short areas of high coverage at the 3’ end (>350x in areas 39,304 to 39,355 and 39,530 to 39,567). These areas may represent tandem repeats that present a challenge for the assembly software, resulting in the gap between ORF11 and ORF17. Primers were designed to walk through this area, and conventional PCR using platinum Taq DNA Polymerase, GC RICH PCR

System, and TaKaRa Ex Taq® Hot Start Version were used in attempts to bridge the gap but failed. This may reflect errors in the assembly, or alternatively, contig1 and contig2 may not join at this site. On the 5’ end of contig1, there were several small

ORFs (<200bp) as well as several accessory genes, including ORFs with homology to

58

portions of immunoglobin superfamily and vBCL2 (Ot13). Towards the 3’ end of contig1, were herpesvirus core genes (ORF6 through ORF 11), including terminase (ORF7), glycoprotein B (ORF8), and DNA-dependant-DNA polymerase (dpol). The most 5’ contigs in OtHV4 had a similar gene order (Figure 3-2).

The middle contig(s) of both OtHV1 and OtHV4 represent core genes as seen in other gammaherpesviruses. This includes ORF17 through ORF69 (Figure 3-1).

Following ORF 69 there is an area of very high coverage (>1800x)- most of which have variant sequences. There were only 2 Pacbio reads bridging across the viral genome in this area, so we hypothesize the assembly may be correct; however, we suspect the huge spike in coverage does not come from viral template. Most likely these are cellular snippets present from host genomic DNA. Our initial BLAST search of the ORF in that area (Ot17) showed strong similarity to two sea lion genes: Zalophus californianus

Ovarian cancer immunoreactive antigen domain-containing protein 1 (OCIAD1,

GenBank accession number FJ692792.1, query coverage 58%, E-value 8^-149, and identity of 88%) and Zalophus californianus procollagen lysine 2-oxoglutarate 5- dioxygenase 2 (Plod2) gene (GenBank accession number FJ692836.1, query coverage

58%, E-value 7^-156, and identity of 90%). In humans, OCIAD1 is a gene that plays a role in tumor metastasis through effects on cell adhesions and is a marker for malignancy and recrudescence in ovarian cancer, thyroid cancer, and several others

(Sengupta et al. 2008, Wang et al. 2010, Yang et al. 2012, Nagata et al. 2012). PLOD2 is a gene associated with collagen synthesis and plays a role in metastatic progression by laying a “highway” for the cancer cells to travel along during cell invasion and migration (Gilkes et al. 2014, Du et al. 2017). PLOD2 is used as a prognostic factor

59

used in hepatocellular carcinoma, breast cancer, sarcomas, bladder cancer and renal cell carcinoma (Eisinger-Mathason et al. 2013, Miyamoto et al.2016, Kurozumi et al

2016, Noda et al.2012, Gilkes et al 2013). On further inspection, however, the publication associated with these genes was focused on clarifying the conflicting phylogenetic hypotheses of caniformian families using noncoding nuclear markers (Yu et al. 2011). Thus, this area contains both introns and exons of the OCAID1 and PLOD2 genes. Additional BLAST searches of this area also had a few significant hits with endogenous viral elements (EVE). Some EVEs have been documented to have oncogenic potential (Katzourakis and Gifford 2010). EVE are sequences that are derived from ancient retroviruses after being inserted a portion of retroviral genome in both the viral genome and in several places in the host genome. This has also been described in double stranded DNA viruses that become integrated into the host genome as a stage in their life cycle (Katzourakis and Gifford 2010). If these viral insertions occur in the genome of germ line cells the insertions become inheritable. From this point forward, we will refer to this ORF (Ot17) as viral EVE (vEVE) and will explore its expression in cervical tumors in the following chapter. Past ORF69 in KSHV is a stretch of miRNAs followed by LANA- the primary latently expressed protein with significant impact on the viral genome, as well as having cellular transforming properties. We suspect this area in OtHV1 contains similar miRNAs; however, it was not evaluated for miRNA. Also in this area was a smaller ORF (Ot19) that had distant similarity to EBNA1 and LANA when a BLAST search was performed, so we suspect there is a

LANA/EBNA1-like gene in this area. LANA has a highly repetitive sequence which may have resulted in poor assembly of this area in OtHV1, and this is where contig2 ends.

60

Between contig2 and contig3 there were only 2 Pacbio reads bridging across the viral genome in this area with a high GC content (70.5%). We hypothesize that due to this high GC content, sequencing was impeded and have included this portion in the draft genome. Two previously described herpesvirus oncogene analogs were identified in contig3: vFLIP and CDK4. Therefore, several known viral oncogene homologs were identified in OtHV1, including vBCL2, vFLIP, and EBNA1/LANA-like region, and two other genes, vCDK4 and vEVE, may also be important transforming factors.

Expression patterns of these OtHV1 genes in cervical tumors will be further explored in

Chapter 4.

OtHV1 and OtHV4 Percent Identity and Distance Comparisons

Using percent identity and distance comparisons, four core genes and one accessory gene (vFLIP) were more closely analyzed to determine relationships between

OtHV1, OtHV4, DeHV1, HHV4, HHV8 and Samriine gammaherpesvirus 2. Alignments and comparison tables were created to assess the percent identity of 4 core genes

(Figures 3-3, 4,5,6) and suspected host-derived gene vFLIP (Figure 3-7).

Briefly, OtHV1 and OtHV4 had the highest percent identities and lowest distance in all core gene comparisons (terminase 90.48% and 0.06, glycoprotein B 91.66% and

0.09, dpol at 94.48%, and 0.04, major capsid protein 97.11% and 0.01). Next OtHV1 was most closely related to DeGHV1 (terminase 45.34% and 0.73, glycoprotein B

46.28% and 0.69, dpol 56.12% and 0.52, major capsid protein 63.30% and 0.44).

OtHV1 and HHV8 percent identity and distance were similar (terminase 41.80% and

0.80, glycoprotein B 37.34% and 0.84, dpol 52.73% and 0.59, major capsid protein

59.36% and 0.50).

61

OtHV1 vFLIP had the closest percent identity and lowest distance with walrus cFLIP (67.04% and 0.36), similar to OtHV4 and walrus cFLIP (62.71 and 0.51). In a related pattern, both DeGHV1 and HHV8 were more closely related to their host cFLIP.

OtHV1 and OtHV4 Represent New Oceanic Gammaherpesvirus Genus

Both Bayesian and Maximum Likelihood values indicate that OtHV1 and OtHV4 cluster together with a posterior probability (pp) of 100% and ML bootstrap value of

100%. These viruses cluster within the Gammaherpesvirinae with a pp of 100% and an

ML bootstrap value of 100%. After the most basal gammaherpesvirus divergence, the genus Lymphocryptovirus, they are the next most basal divergence of the remaining gammaherpesviruses (Figure 3-8). Additionally, Delphinid gammaherpesvirus 1, a herpesvirus sequenced from a genital lesion in a common bottlenose dolphin (Tursiops truncatus), groups with the two pinniped herpesviruses with a pp of 100% and an ML bootstrap value of 100%.

The phylogenetic analysis of the vFLIP gene finds that OtHV1 and OtHV4 cluster together with a pp of 99.3% and ML bootstrap value of 74.8%. They cluster most closely with host FLIP from the Caniformia (Figure 3-9), the suborder containing the family

Otariidae, with a pp of 90.7% and ML bootstrap value <50%. They do not cluster with other herpesviral vFLIP genes, which, together with a vFLIP from the Molluscipox, form a clade with a pp of 99.6% and ML bootstrap value of 69%. Bottlenose dolphin gammaherpesvirus 1 clusters with host cetacean vFLIP genes, with a pp of 99% and an

ML bootstrap bvalue <50%.

Discussion

Using next generation sequencing, we were able to obtain draft genomes for

OtHV1 and OtHV4 from a California sea lion cervical tumor and a Northern fur seal

62

vaginal swab, respectively. The majority of the sequenced reads were of host origin

(94% in the sea lion sample and >99% in the Northern fur seal sample). However, sequencing data provided enough viral reads to compose almost completed drafts of both OtHV1 and OtHV4 genomes. Both genomes still have gaps, most likely due to the low coverage, or sequencing bias and assembly errors from high GC content or tandem repeats. The draft genomes provide valuable information about the phylogenetic relationship between these herpesviruses and other more distantly related herpesviruses, as well as potential oncogenes within OtHV1. This supports use of this technique for viral discovery in swabs and tissue samples from non-model species.

One of the significant findings was the novel genus of marine mammal herpesviruses identified. Evidence that this early branching group represents a new herpesvirus genus including both pinniped herpesviruses (OtHV1 and OtHV4) as well as a herpesvirus obtained from a common bottlenose dolphin (DeGHV1). It appears that these herpesviruses in marine mammals have branched from their terrestrial herpesvirus counterparts early on in the subfamily’s evolution, perhaps stemming from a common terrestrial ancestor that moved into the aquatic environment. We proposed this new genus of marine mammal gammaherpesviruses be called Marmamvirus.

A second phylogenetic analysis of the suspected host derived gene, vFLIP

(Figure 3-8), was performed to determine when this gene was acquired by OtHV1 and if it was closer to host cFLIP or other viral vFLIPs. Additional viral vFLIP homologues were included in this comparison. Poxviridae, Percavirus and Rhadinovirus vFLIPs branched together very early in the tree. OtHV1 and OtHV4, however, grouped together with Caniformia cFLIPs. Interestingly, DeGHV1 grouped with Cetacea cFLIP. The

63

combination of the core genes grouping away from the Rhadinovirus genus early on and the vFLIP from OtHV1, OtHV4 and DeGHV1 grouping with their host cFLIP indicate that OtHV1 and OtHV4 acquired vFLIP in an independent acquisition compared to

DeGHV1 after splitting from their more terrestrial-associated herpesvirus family members. vFLIP is one of several host-derived viral genes commonly found in gammaherpesviruses that is recognized for its ability to induce cancer in humans and other animals. In addition to vFLIP, other potential oncogenes identified in OtHV1 included: LANA/EBNA1-like gene, vBCL2, vCDK4 and vEVE. Acting on various cellular pathways, these genes are thought to contribute to viral propagation, persistent infections, and impairment of apoptosis, a vital step in tumorigenesis. Although further work must be done to better understand their functional roles, the finding that OtHV1 contains several well-known herpesvirus oncogenes provides support that it may play a significant role in the pathogenesis of urogenital carcinoma in California sea lions.

Several viral oncogenes involved in altering signal transduction pathways, blocking apoptosis and inactivating tumor suppressor pathways have been identified in

HHV8 and HHV4 (Damania et al. 2004). Current in vitro studies have identified a number of potent regulatory and immune modulatory viral genes, but their role in tumorigenesis in vivo is limited due to the lack of appropriate lab animal models. One way to fill the lack of traditional laboratory models is to study naturally occurring gammaherpesvirus-associated cancers in other species. These newly identified potential OtHV1 oncogenes in combination with the high prevalence of urogenital carcinomas in California sea lions offers this opportunity. Studying naturally occurring, virally induced cancers such as urogenital carcinoma in California sea lions can provide

64

models for investigating viral/host interplay and give insight into basic cellular processes perturbed in these cancers in a naturally developing disease. Additionally, some findings may lead to new anti-viral therapeutic approaches, offering new perspectives on cancer treatments in humans and animals alike.

65

Table 3-1. Herpesvirus species list for concatenated phylogenetic tree. Virus name Abbreviation Accession # Subfamily Genus Human alphaherpesvirus 3 Human HV3 NC_001348.1 Alpha Simplex Suid alphaherpesvirus 1 Suid AHV1 NC_006151.1 Alpha Elephantid betaherpesvirus 1 Elephantid BHV1 NC_020474.2 Beta Human betaherpesvirus 7 Human HV7 NC_001716.2 Beta Roseolovirus Callitrichine herpesvirus 3 Callitrichine HV3 NC_004367.1 Gamma Lymphocryptovirus Human gammaherpesvirus 4 Human HV4 NC_007605.1 Gamma Lymphocryptovirus Macacine gammaherpesvirus 4 Macacine GHV4 NC_006146.1 Gamma Lymphocryptovirus Alcelaphine gammaherpesvirus 1 Alcelaphine GHV1 NC_002531.1 Gamma Macavirus Bovine gammaherpesvirus 6 Bovine GHV6 NC_024303 Gamma Macavirus Ovine GHV2 NC_007646.1 Gamma Macavirus Porcine lymphotropic herpesvirus 2 Porcine lymphotrophic HV2 AA0012350 Gamma Macavirus Equid GHV2 NC_001650.2 Gamma Percavirus Equid gammaherpesvirus 5 Equid GHV5 NC_026421.1 Gamma Percavirus Ateline gammaherpesvirus 3 Ateline GHV3 NC_001987.1 Gamma Rhadinovirus Bovine gammaherpesvirus 4 Bovine GHV4 NC_002665.1 Gamma Rhadinovirus Ceropithecine herpesvirus 17 Ceropithecine HV17 NC_003401 Gamma Rhadinovirus Cricetid gammaherpesvirus 2 Cricetid GHV2 NC_015049.1 Gamma Rhadinovirus Human gammaherpesvirus 8 Human HV8 NC_009333.1 Gamma Rhadinovirus Murid HV4 NC_001826 Gamma Rhadinovirus Retroperitoneal fibromatosis-associated herpesvirus Macaque RFHVMn KF703446 Gamma Rhadinovirus Saimiriine gammaherpesvirus 2 Saimiriine GHV2 NC_001350.1 Gamma Rhadinovirus Delphinid gammaherpesvirus 1 Delphinid GHV1 NC_035117.1 Gamma Unclassified Felis catus gammaherpesvirus 1 Felis catus GHV1 NC_028099.1 Gamma Unclassified Harp seal herpesvirus Harp Seal HV KP136799.1 Gamma Unclassified Myotis gammaherpesvirus 8 Myotis GHV8 NC_029255 Gamma Unclassified Otarine herpesvirus 1 OtHV1 submitted Gamma Unclassified Otarine herpesvirus 4 OtHV4 submitted Gamma Unclassified

66

Table 3-2. Herpesvirus gene list for concatenated phylogenetic tree. Polymerase Terminase Glycop B Major capsid Virus name (ORF 9) (ORF 7) (ORF 8) (ORF 25) Human alphaherpesvirus 3 NP_040151 NP_040153 NP_040154 NP_040163.1 Suid alphaherpesvirus 1 YP_068333 YP_068331 YP_068330 YP_068356.1 Elephantid betaherpesvirus 1 YP_007969814 YP_007969816 YP_007969815 YP_007969788.1 Human betaherpesvirus 7 YP_073778 YP_073780 YP_073779 YP_073799.1 Callitrichine herpesvirus 3 NP_733857 NP_733855 NP_733856 NP_733870 Human gammaherpesvirus 4 YP_401712 YP_401715 YP_401713 YP_401697 Macacine gammaherpesvirus 4 YP_068007 YP_068010 YP_068009 YP_067994 Alcelaphine gammaherpesvirus 1 NP_065512 NP_065510 NP_065511 NP_065524 Bovine gammaherpesvirus 6 YP_009041990 YP_009041988 YP_009041989 YP_009042004 Ovine gammaherpesvirus 2 YP_438136 YP_438134 YP_438135 YP_438149 Porcine lymphotropic herpesvirus 2 AAO12282 AAO12280 AAO12281 AAO12367 Equid gammaherpesvirus 2 AIU39456 AIU39454 AIU39455 NP_042621 Equid gammaherpesvirus 5 YP_009118399.1 YP_009118397.1 YP_009118398.1 YP_009118415.1 Ateline gammaherpesvirus 3 NP_047983 NP_047981 NP_047982 NP_047996 Bovine gammaherpesvirus 4 NP_076501 NP_076499 NP_076500 NP_076517 Ceropithecine herpesvirus 17 NP_570750.1 NP_570748.1 NP_570749.1 NP_570765.1 Cricetid gammaherpesvirus 2 YP_004207849.1 YP_004207847 YP_004207848.1 YP_004207861.1 Human gammaherpesvirus 8 YP_001129355 YP_001129353.1 YP_001129354 YP_001129378 Murid gammaherpesvirus 4 NP_044849 NP_044847 NP_044848 NP_044863 Retroperitoneal fibromatosis-associated herpesvirus AGY30688 AGY30686 AGY30687 AGY30708 Saimiriine gammaherpesvirus 2 NP_040211 NP_040209 NP_040210 NP_040227 Dephinid gammaherpesvirus 1 YP_009388514.1 YP_009388512.1 YP_009388513.1 YP_009388526.1 Felis catus gammaherpesvirus 1 YP_009173887.1 YP_009173885.1 YP_009173886.1 YP_009173900.1 Harp seal herpesvirus AJG42938.1 AJG42936.1 AJG42937.1 AJG42951.1 Myotis gammaherpesvirus 8 YP_009229846 YP_009229844 YP_009229845 YP_009229859 Otarine herpesvirus 1 submitted submitted submitted submitted Otarine herpesvirus 4 submitted submitted submitted submitted

67

A

B

C

Figure 3-1. Annotated draft genome of OtHV1. ORFs in red represent core herpesvirus genes, ORFs in yellow represent OtHV1 accessory genes, and the gap where the PacBio read bridged contig2 and contig3 is in blue. A: contig1 was 40,761bp with an average coverage of 40x; B: contig2 was 92,701bp with an average coverage of 59x; and C: contig3 was 18,856 bp with an average coverage of 54x. This produced a 152,318bp draft genome with a minimum of 86 ORFs.

68

Figure 3-2. OtHV4 (lower annotated sequence) MiSeq contig assembly to OtHV1 draft genome (upper annotated sequence) with OtHV4 contig map at the bottom of the figure. Note the four areas of increased coverage, all near areas where there is a gap in the assembly of OtHV4.

69

A

B

Figure 3-3. Amino acid alignment (A) and comparison (B) for herpesvirus core gene Terminase (ORF7). A) Amino acid alignment with herpesvirus listed on left and length of amino acid region on right (horizontal scale from aa 0- 748aa) with black line indicating sequence and gaps representing non-conserved area. Bar graph below alignment shows 0-100% conserved region. B) Comparison chart between OtHV1 and other related gammaherpesviruses amino acid coding regions. Upper right area of the chart displays percent identity with color gradient showing maximum percent identity (red) to minimum percent identity (blue). The lower left area of the chart displays minimum (blue) and maximum (red) distances between respective viruses.

70

A

B

Figure 3-4. Amino acid (aa) alignment (A) and comparison (B) for herpesvirus core gene Glycoprotein B (ORF8). A) Amino acid alignment with herpesvirus listed on left and length of amino acid region on right (horizontal scale from aa 0-748aa) with black line indicating aa sequence and gaps showing non-conserved area. Bar graph below alignment shows 0-100% conserved region. B) Comparison chart between OtHV1 and other related gammaherpesviruses amino acid coding regions. Upper right area of the chart displays percent identity with color gradient showing maximum percent identity (red) to minimum percent identity (blue). The lower left area of the chart displays minimum (blue) and maximum (red) distances between respective viruses.

71

A

B

Figure 3-5. Amino acid (aa) alignment (A) and comparison (B) for herpesvirus core gene DNA Polymerase (ORF9). A) Amino acid alignment with herpesvirus listed on left and length of amino acid region on right (horizontal scale from aa 0-748aa) with black line indicating aa sequence and gaps showing non-conserved area. Bar graph below alignment shows 0-100% conserved region. B) Comparison chart between OtHV1 and other related gammaherpesviruses amino acid coding regions. Upper right area of the chart displays percent identity with color gradient showing maximum percent identity (red) to minimum percent identity (blue). The lower left area of the chart displays minimum (blue) and maximum (red) distances between respective viruses.

72

A

B

Figure 3-6. Amino acid (aa) alignment (A) and comparison (B) for herpesvirus core gene Major Capsid Protein (ORF25). A) Amino acid alignment with herpesvirus listed on left and length of amino acid region on right (horizontal scale from aa 0-748aa) with black line indicating aa sequence and gaps showing non-conserved area. Bar graph below alignment shows 0-100% conserved region. B) Comparison chart between OtHV1 and other related gammaherpesviruses amino acid coding regions. Upper right area of the chart displays percent identity with color gradient showing maximum percent identity (red) to minimum percent identity (blue). The lower left area of the chart displays minimum (blue) and maximum (red) distances between respective viruses.

73

A

B

Figure 3-7. Amino acid (aa) alignment (A) and comparison (B) for herpesvirus gene vFLIP as well as various host cFLIP. A) Amino acid alignment with herpesvirus vFLIPs and host cFLIPs listed on left and length of amino acid region on right (horizontal scale from aa 0-188aa) with black line indicating aa sequence and gaps showing non- conserved area. Bar graph below alignment shows 0-100% conserved region. B) Comparison chart between OtHV1 and other related herpesviruses and host FLIP amino acid coding regions. Upper right area of the chart displays percent identity with color gradient showing maximum percent identity (red) to minimum percent identity (blue). The lower left area of the chart displays minimum (blue) and maximum (red) distances between respective viruses.

74

Figure 3-8. Concatenated Bayesian/Maximum Likelihood analysis phylogram based on 4 herpesvirus core genes (polymerase, terminase, glycoprotein B, and major capsid protein) using predicted amino acid sequences. Bayesian posterior probabilities of each branch is expressed as % (left) and maximum likelihood (ML) bootstrap values as % (right). Branches are colored based on grouping subfamilies (olive green-Alphaherpesvirinae, blue-Betaherpesvirinae) and genus of respective herpesviruses (fern green-cryptoviruses, maroon- macaviruses, pink-percaviruses, orange-rhadinoviruses, and red- novel genus (proposed name: marmamvirus). Most notable, Otarine HV1, Otarine HV4 and Delphinid GHV1 branches very early from other known gammaherpesviruses.

75

Figure 3-9. Concatenated Bayesian/Maximum Likelihood analysis phylogram based on host FLIP/FADD-like amino acid sequences and virus vFLIP host derived amino acid sequences (see Table 3-2 for list of genes and GenBank Accession numbers). Red lines indicate viral vFLIP; green and black lines represent host FLIP/FADD-like sequence. Red arrows indicate OtHV1 and OtHV4 which group with Caniforma host FLIP and Bottlenose Dolphin Gammaherpesvirus (DeGHV1) vFLIP groups with Cetacea host FLIP. All terrestrial viral vFLIP anologues group together indicating an early acquisition of this gene from their hosts. These findings support that OtHV1, OtHV4 and DeGHV1 were independently acquired from their host after branching from their ancestrally terrestrial herpesvirus.

76

CHAPTER 4 OTHV1 GENE EXPRESSION IN UROGENITAL CARCINOMA

Background

Establishing a causal relationship between a microbe (virus, , parasite, etc.) and a specific disease process can prove challenging due to the presence of other factors also influencing disease occurrence. Correlation between the presence of a microbe and a disease does not necessarily indicate a causal role. Historically, Koch’s postulates have been used to definitively prove causality versus correlation regarding a pathogen’s presence in a given disease process. Briefly, Koch’s postulates state: 1) microorganism presence must be confirmed in every case of the disease but not in healthy individuals, 2) the microorganism must be cultured from the diseased individual

3) the cultured microorganism must be inoculated into a healthy individual and recapitulate the disease, and 4) the microorganism must be isolated from this newly infected individual and be the same as the original microbe. Although these criteria initially proved critically important for scientifically establishing causal relationships between pathogens and hosts, strict adherence to this approach limits the ability to identify a variety of pathogens in their role of inducing a given disease. Disease often has a multifactorial pathophysiology, including immune competence, environmental influence, genetic predispositions, contaminant exposure and coinfections. Viewing microbes in binary categories of pathogenic or nonpathogenic, with pathogenic organisms as the sole cause of infectious disease is misleading. Additionally, the inability to culture and infect a laboratory animal with a particular pathogen to induce disease limits researchers’ ability to prove Koch’s postulates.

77

The limitations of Koch’s postulates are particularly apparent when applied to fastidious organisms, such as most viruses. Host cellular machinery is required for viral replication, meaning viruses cannot be cultured in a cell-free media, and host cell types are typically few (Rivers 1937). To circumvent these inherent limitations, revisions of

Koch’s postulates incorporating the complex nature of disease pathophysiology and more recent advances in microbiology and molecular biology have been described

(Falkow 1988; Fredricks and Relman, 1996). This approach is particularly relevant in the indication of genetic involvement in the pathophysiology of viral infections. This has allowed for a better understanding of the complex components to the multistep process of infectious disease and cancer.

Herpesviruses have diverse animal hosts, including mollusks, fish, reptiles, birds and mammals (McGeoch et al. 2008). In humans, close to 100% of adults carry one or more of the 8 herpesviruses endemic in people (Looker et al 2015a, Looker et al 2015b,

Jansen et al. 2016). This high prevalence of herpesvirus infection is also typical in non- human species. (McGeoch et al, 2006). Typically, latent herpesviruses in their endemic hosts rarely cause clinically significant disease after initial infection. However, under some conditions, viral recrudescence or latent herpesvirus proteins are responsible for inducing significant disease in their hosts. The high penetrance of these viruses combined with latent and lytic stages of infection makes it challenging to establish a causal relationship with lesions, particularly in species not typically studied in a laboratory environment.

Herpesviruses are very host specific. This makes it challenging to study herpesviruses in traditional laboratory animal models that are not natural hosts for the

78

virus of interest. Much has been learned from studying herpesvirus pathogenicity in traditional animal models and in vitro studies, but much still remains to be understood about the multifactorial components of the pathogenicity of herpesvirus-associated cancers. Cofactors, such as immune suppression, can lead to increased risk of developing disease, as a functional immune system is critical for maintenance of a latent state. For example, diagnosis of herpesvirus-induced Kaposi’s sarcoma is 500 times more likely in human AIDS patients compared to the general population

(Hernández-Ramírez et al. 2017). The high prevalence of herpesvirus infections and the lower prevalence of disease complicates establishing causality in this multifactorial disease process.

Oncogenic herpesviruses in humans, mice, and monkeys have been well studied, and a variety of mechanisms behind their tumorigenicity have been established. Expression patterns of key viral genes, as well as ncRNAs, have been implicated in determining whether a herpesvirus remains a latent, benign, and subclinical, or results in cellular transformation- sending the host cell down a path of malignancy. Examples of these genes include latently expressed viral proteins such as

LANA in Kaposi’s sarcoma-associate herpesvirus (KSHV) and its homologue EBNA1 in

Epstein-Barr Herpesvirus (EBV). Other herpesvirus genes that will be discussed in this chapter are vFLIP, vBCL2, vCDK4 and vEVE, (see Chapter 1 for more details). Briefly, in KSHV infected cells both LANA and vFLIP are expressed during latency. LANA and vFLIP act as viral promoters and host transcriptional factors influencing viral gene expression and host cellular pathways involved in cell proliferation, apoptosis, immune modulation, epigenetic modification, and signal transduction pathways (Renne et al.

79

2001). vBCL2 and viral cyclins (e.g. in OtHV1 vCDK4) are expressed in the lytic stage and are well known host derived herpesvirus genes. vBCL2 interferes with apoptotic pathways and blocks autophagy- which in turn helps virally infected cells evade immune surveillance. Viral cyclins is a transcriptional regulator that help initiate host cell cycling

(Mittnacht and Boshoff 2000). Lastly, endogenous viral elements (EVE) have been identified in both host and viral genomes and have been associated in some instances with inducing malignancies (Moore and Chang 2010). Homologues to these host or viral oncogenes have been identified in OtHV1, as discussed in the previous chapter, and here we will look at expression levels of these genes in California sea lion urogenital carcinoma using RISH.

Molecular techniques such as PCR, qPCR and next generation sequencing can quantify and assess the phenotypic characteristics of oncogenes; this is typically done on tumor biopsy homogenates. The inherent heterogeneity of tumors, and the potential for surrounding unaffected healthy tissue to be included in these biopsies, can influence results and the spatial/contextual information about the viral association with tumors may be diluted or lost. Advanced in techniques such as laser capture and single cell

RNAseq have improved accuracy, however these techniques can be expensive and challenging when working with limited or degraded RNA.

Recently, a modification of RNA in situ hybridization (RISH), called RNAscope, has become available. In contrast to immunohistochemistry, this technology is particularly powerful when studying non-traditional or novel species because instead of relying on availability of commercial antibody or expensive and time-consuming generation of novel antibodies to visualize a protein of interest, RNA specific probes can

80

be rapidly and economically generated by using the known sequence of the gene being studied. More recently, a technique with higher specificity has been developed called

BaseScope. This technology has been refined to allow specific identification of the level of detecting single nucleotide polymorphisms (SNPs), differentiating between various isoforms, and distinguishing between very closely related homologous genes.

One of the greatest advantages to this technology is the ability to retain the spatial and structural integrity of the tumor while localizing the virus in affected tumor cells. In addition, levels of expression of viral RNAs can be visualized in both healthy and tumor tissues.

This research on California sea lion cancer used RISH to localize OtHV1 RNA in primary cervical lesions from animals with carcinoma, and in healthy cervix from animals without urogenital carcinoma. Five OtHV1 genes, previously implicated in cellular transformation in other herpesviruses, were assessed for expression patterns to determine the level of expression of these viral oncogenes in cancer cells.

Materials and Methods

Cases and Tissue Preparation

From 2011 to 2017, proximal vagina and/or cervix were collected between 5 minutes and 24 hours post-mortem from 25 adult California sea lions at The Marine

Mammal Center in Sausalito, California (Table 4-1). Tissues were immediately fixed in

10% neutral buffered formalin for 24 hours (n=11) or archived at -80’C then fixed in 10% neutral buffered formalin for 24 hours (n=14). Following formalin fixation, tissues were embedded in paraffin, sectioned at 5m thickness, stained with hematoxylin and eosin

(H&E), and evaluated at the University of Illinois College of Veterinary Medicine

81

Zoological Pathology Program by a pathologist with expertise on sea lion urogenital carcinoma.

Animals selected as controls had no gross or histologic reproductive tract lesions. Animals died or were euthanized due to non-cancer related clinical disease

(e.g. trauma, domoic acid intoxication, pneumonia, other). Healthy cervix was defined as described in Colegrove et al (2009b). Briefly, normal cervical epithelium was composed of simple columnar to stratified squamous epithelium (Figure 4-1). Cervical and vaginal epithelial thickness has been shown to vary depending on seasonality related to synchronized estrus in summer months. Thicker cervical and vaginal epithelium corresponded to the breeding season in the summer months, when the seasonal influences of estrogen are at their highest (Colegrove et al. 2009b). All cases of normal cervix were grouped together regardless of seasonality and considered to be representative of healthy cervical epithelium.

Urogenital carcinoma lesions were classified as either cervical intraepithelial lesions (CIN) or invasive lesions, based on definitions established by Colegrove et al.

(2009a), which approximate those in an established grading scale used in human cervical and vaginal neoplasms (Kurman et al. 1992). Briefly, urogenital carcinoma lesions were classified as CIN when mild dysplasia was confined to the lower third of the epithelial layer (equivalent to low grade epithelial lesions in women, also known as

CIN 1), or when the lesion was moderate to markedly dysplastic with atypical parabasal cell proliferation extending from one third to the entire thickness of the epithelium

(considered high-grade intraepithelial lesions in women, similar to CIN 2 and 3). All lesions classified as CIN were grouped together, regardless of the degree of dysplasia

82

or CIN level. When a lesion invaded beyond the basal epithelial layer through the basement membrane into the underlying stroma it was classified as invasive. In many cases, both CIN and invasive lesions were present, and representative areas of CIN and invasive lesions were collected and grouped with the respective lesion type.

Additionally, gross and histologic evidence of metastases was noted, as well as any intravascular neoplastic rafts of cells observed in the area of the primary lesions.

Basescope in situ Hybridization

To localize gene expression of OtHV1 in urogenital carcinoma, five RISH probes were designed to specifically bind to OtHV1 genes that were thought to have oncogenic potential. Additionally, two positive controls designed to bind to sea lion housekeeping genes (peptidyl prolyl isomerase B [PPIB, also known as cyclophilin B] and DNA- dependent RNA polymerase II) and a negative control dihydrodipicolinate reductase

(dapB) of Bacillus subtilis were used for quality control. The OtHV1 genes selected for

RISH were vBCL2, vFLIP, vCDK4, vEVE, and LANA-like transcript. To optimize specificity, we utilized the Basescope platform to ensure host-derived OtHV1 gene would not cross react with homologous host cellular genes.

Visualization of five OtHV1 gene transcripts was performed using custom

Basescope probes and BasescopeTM Red Reagent Kit (Advanced Cell Diagnostics

(ACD), Hayward, California, USA, cat #322910) following the manufacturer’s protocol with minor modification. Briefly, eight serial FFPE sections from each case were cut onto four slides (two sections per slide). Slides were baked for 1 hour at 60˚C then dewaxed in xylene twice for 5 minutes, followed by dehydration in 100% ethyl alcohol twice for 2 minutes, and air-dried. RNAscope hydrogen peroxide (ACD, Hayward,

83

California, cat #322381) was applied to each section and incubated for 5 minutes at room temperature, then slides were rinsed five times in double distilled water (ddH2O).

Slides were placed in RNAscope 1X Target Retrieval Reagent (ACD, cat #322000) and incubated at a slow boil for 30 minutes. Slides were rinsed at room temperature in ddH2O, followed by 100% ethanol rinse and air-dried. A hydrophobic barrier was placed around each tissue section (ACD, ImmEdge TM hydrophobic Barrier Pen, cat #310018) and each section was treated with Protease III digestion buffer (ACD, Cat. #322381) for

30 minutes at 40˚C. Slides were washed in ddH20 and target or control probes applied and incubated for 2 hours at 40˚C. Basescope custom probes targeting five viral genes

(Table 4-1) included: OtHV1-LANA-like (4ZZ probe targeting 198-434), OtHV1-vFLIP

(4ZZ probe targeting 223-504), OtHV1-vBCL2 (4ZZ probe targeting 3-240), OtHV1- vCDK4 (4ZZ probe targeting 3-855), OtHV1-vEVE (4ZZ probe targeting 39-261). To assess RNA quality, two additional custom positive control probes (aimed at cellular housekeeping genes) and one proprietary negative probe (aimed at a non-specific bacterial transcript) were used: PPIB (3ZZ probe to sea lion genome provided by the

Broad Institute targeting 68-221), polR2A (DNA-dependent RNA polymerase II, 3ZZ probe targeting sea lion, 555-705), and dapB (dihydrodipicolinate reductase, 3ZZ probe,

ACD, cat #701011) of Bacillus subtilis. Following incubation, slides were washed in wash buffer (ACD, cat #310091) for 2 min at room temperature. Signal amplification reagents (AMP) 0-5 were applied as follows: AMP 0 for 30 minutes at 40˚C, AMP 1 for

15 minutes at 40˚C, AMP 2 for 30 minutes at 40˚C, AMP 3 for 30 minutes at 40˚C,AMP

4 for 15 minutes at 40˚C, and AMP 5 for 30 minutes at room temperature. Slides were rinsed for 2 minutes with wash buffer between amplification reagents. Positive signal

84

was visualized using Fast RedTM dye incubated for 10 minutes at room temperature, rinsed with tap water, counterstained with Gill’s hematoxylin, dried at room temperature, then coverslipped.

Image Analysis

Each tissue section was imaged using cellSens Entry 1.11 software (Olympus) slide scanner on an Olympus BX46 with a 40X objective (area in field of view=0.34 mm2).

Images were imported into ImageJ. Regions of interest were defined as normal epithelium in controls or neoplastic epithelium in cancer cases. Representative areas of normal cervix, CIN and invasive urogenital carcinoma lesions were selected using the

ImageJ polygon tool, avoiding regions of inflammation, necrosis, underlying stroma or poor-quality staining- for instance, at the cut edge of the sections (Figure 4-3). Using the

IHC tool, the software was trained to recognize individual positive-stained pink areas by selecting ten representatives punctate, well-defined pink dots from all probes in every case (healthy and cancer) and saved as a Reader User Model to analysis all images.

The trained model was selected, and color space segmentation was used to define the positively stained areas. The image type was then converted to 8-bit, and identical thresholds of 200 was set for all images. The total positively labelled area (representing any positive stained pink dot) was divided by the total area of interest and defined as the percent positive labeling. Control cervix and vagina had a thinner simple columnar epithelial layer compared to neoplastic lesions. To ensure similar sized areas were being evaluated in normal epithelium, the percent positive labeling was measures over

3 40x fields and averaged. Therefore, for controls three representative areas between

160,870 to 1,300,000 per 40x field of normal epithelium were selected and added, totaling a region of interest for healthy controls ranging from 482,610 to 3,900,000. In

85

CIN and invasive lesions an area between 523,000 to 3,200,000 was selected within on

40x field and percent positive labeling was calculated as described above. In cases with both CIN and infiltrative lesions, both categories were selected and analyzed separately for comparison of gene expression between these lesion types.

Data Analysis

Descriptive tables (quartiles) were generated from the percent positive labeling to determine extent of probe binding for healthy, CIN, and invasive samples. To compare proportion of positive labeling in non-cancer animals to CIN and invasive lesions in cancer animals, a generalized linear mixed model with a binomial distribution and a log- odds (logit) link function was used, as data were not normally distributed. A Tukey test was used to investigate the differences between groups (Healthy, CIN, and Infiltrative) in order to determine significance with a p-value of <0.01.

Results

Gross Findings

A total of 25 stranded California sea lions were necropsied following humane euthanasia for terminal disease or natural death while undergoing treatment at The

Marine Mammal Center in Sausalito, California (Tables 4-1 and 4-2). Causes of death/euthanasia included domoic acid intoxication (n=8), sarcocystasis (n=1), metastatic urogenital carcinoma (n=13) or undetermined (n=1). In the 13 cases with advanced stage urogenital carcinoma, nine had visible cervical or vaginal lesions noted on gross necropsy. Lesions were pink to beige, roughened, plaque-like masses ranging in size from <0.5cm to 8 cm diameter, some of which completely effaced the cervical architecture. In four cases with severe metastatic lesions, no primary lesions of the

86

vaginal or cervical mucosa were visible on gross examination, and 11 animals had no gross metastatic lesions or evidence of primary lesions on the cervix or vagina.

Histological classification

A histologic diagnosis of normal cervical or vaginal epithelium was confirmed in nine of the 25 cases examined, and these cases were regarded as controls. Two cases were euthanized for domoic acid toxicosis, had grossly normal appearing vaginal and cervical mucosa and no metastatic lesions, but on histologic examination, there was early stage carcinoma in the cervix, with intraepithelial neoplasia (CIN) and rare nests of infiltrative neoplastic cell in the underlying stroma. The remaining 14 animals had vaginal or cervical intraepithelial lesions and metastases with morphological features characteristic of urogenital carcinoma (Gulland et al. 1996; Lipscomb et al. 2000;

Colegrove et al. 2009a). Figure 4-2 shows a representative H&E image of a CIN (Figure

4-2a) and an invasive lesion (Figure 4-2b). All CIN lesions had a thickened epithelial layers that lacked normal differentiation and squamous maturation, with most cases

(n=15) having all normal epithelium completely replaced by neoplastic epithelial cells comprised of moderate-sized polygonal neoplastic epithelial cells (CINs). Nuclei were large and oval with coarsely clumped chromatin and typically had 1-2 large, central nucleoli. Anisocytosis and anisokaryosis were mild to moderate in these CIN lesions and mitotic figures ranged from 1 to 10 per high power field. There are low to moderate scattered necrotic neoplastic cells, lymphocytes, and plasma cells in and underlying the abnormal epithelium of most CINs.

All invasive lesions in both the vagina and cervix had multifocal cords or nests of neoplastic cells that extended into the submucosa with no apparent delimiting basement membrane. Neoplastic cells within invasive lesions often were large and had abundant

87

pale eosinophilic cytoplasm. Nests of neoplastic cells were abundant within the submucosa. The larger cells in the invasive lesions resulted in less cells per area coverage compared to the more tightly packed neoplastic cells within the CIN lesions.

Surrounding and underlying CIN and invasive lesions in the submucosa were moderate numbers of lymphocytes and plasma cells.

Of the 16 total animals with urogenital carcinoma lesions, seven had only CIN with no evidence of invasion, seven had both CIN and invasive lesions, and two had only invasive lesions. Metastases were confirmed with histology in 14 of the cases and two cases had no evidence of distant metastasis.

OtHV1 Gene Expression in Sea Lion Urogenital Carcinoma

All 16 cases with CIN and invasive lesions had strong positive labeling for all

OtHV1 genes assessed. The nine control vaginal/cervical samples were negative for all

RISH OtHV1 probes, and had similar polR2 positive control (host cellular housekeeping gene) labeling intensity as the cancer cases. In the 16 cancer cases, all OtHV1 genes had positive labeling in the CIN and infiltrative lesions (Figure 4-3). Positive labeling by all probes appeared as punctate to coalescing bright pink staining primarily localized to the nucleus and to a lesser extent within the cytoplasm of neoplastic epithelial cells.

Most animals with urogenital carcinoma had diffuse lesions throughout the cervical/vaginal epithelium, with no normal appearing epithelium in the sections examined. These cases had strong, evenly distributed, positive labeling throughout the intraepithelial and invasive lesions. However, one case with both normal and neoplastic cervical epithelium had positive labeling for all OtHV1 genes only within the neoplastic lesion but not in surrounding normal epithelium (Figure 4-2b) and the positive control pol2A was evenly distributed throughout the normal and neoplastic epithelium.

88

The percent of positive labeling by each of the five probes were not normally distributed. Descriptive statistics (minimum, first quartile, median, second quartile, and maximum) for percent positive labeling are presented for all OtHV1 probes, negative and positive probe controls were calculated for healthy (Table 4-3), CIN (Table 4-4) and invasive (Table 4-5) lesions. The generalized linear model and a Tukey test were used to determine significance. Figure 4-3 shows a graphical representation of all samples as well as the significance between healthy, CIN, and invasive lesions for each RISH probe.

All control and cancer cases had no binding with the negative control RISH probe

(dpol). The positive control (polR2A) ranged in healthy tissue from 0.075% to 1.43% with a median of 0.53%, CIN ranged from 0.32% to 2.32% with a median of 0.60%, and invasive lesions ranged from 0.061% to 1.86% with a median of 0.66%. There was no significant difference in polR2A labeling between the healthy, CIN, or invasive lesions; thus, this cellular housekeeping positive control is an appropriate positive control for

California sea lions with and without urogenital carcinoma, including CIN and invasive lesions.

In contrast, the second positive control probe used, PPIB, did have significant differences in staining between healthy and CIN (P=0.00121), healthy and invasive

(P=1.95e-06), and between CIN and invasive (P<1e-09) samples. In healthy cervix/vagina, the PPIB positive percent labeling ranged from 0.07% to 2.75% (median

1.29%), in CIN, PPIB ranged from 0.43% to 8.2% (median 3.9%), and in invasive it ranged from 1.11% to 61.0% (median 13.8%). Because of the significant differences in expression of PPIB in healthy vs. urogenital carcinoma, this cellular housekeeping gene

89

is not an appropriate positive control for California sea lions when comparing healthy cervix/vagina to urogenital carcinoma lesions.

For all OtHV1 RISH probes (EBNA1-like, vFLIP, vBCL2, vCDK4, and vEVE) there was a significant difference (P<2e-16) among all groups (healthy vs. CIN, healthy vs. invasive, and CIN vs. invasive). CIN lesions had significantly higher expression levels of all OtHV1 genes assessed compared to invasive and healthy tissue, and invasive tissue had significantly higher levels of expression as compared to healthy but lower levels of expression than in CIN for all OtHV1 RISH probes. In healthy cervix/vagina, the EBNA1-like positive labeling ranged from 0.00% to 0.023% (median

0.001%, all <0.06% negative threshold), CIN ranged from 8.2% to 69.6% (median

49.6%), and invasive ranged from 22.0% to 73.6% (median 37.3%). In healthy cervix/vagina, vFLIP ranged from 0.0% to 0.14% positive labeling (median 0.002%),

CIN ranged from 21.4% to 71.5% (median 54.3%), and invasive ranged from 24.5% to

61.4% (median 43.6%). In healthy cervix/vagina, vBCL2 ranged from 0.00% to 0.52%

(median 0.001%), CIN ranged from 40.1% to 80.0% (median 54.2%). In healthy cervix/vagina, vCDK4 ranged from 0.00% to 0.005% (median 0.001%), CIN ranged from

23.2% to 75.5% (median 54.7%), and invasive ranged from 17.3% to 64.8% (median

39.8%). Lastly, vEVE ranged from 0.00% to 0.26% (median 0.052%) in healthy tissue,

CIN ranged from 24.2% to 75.7% (median 54.0%), and invasive ranged from 32.3% to

70.4% (median 41.4%).

Discussion

These findings strongly support that OtHV1 plays an important role in the etiology of urogenital carcinoma in California sea lions. Well-characterized viral oncogene homologues were localized directly within neoplastic sea lion genital epithelium in all

90

urogenital carcinoma cases. Not only were these OtHV1 genes localized to neoplastic epithelium, but the use of OtHV1 RISH provides support that OtHV1 genes are actively expressed at very high levels in neoplastic lesions but absent in control cases, and in normal cervical epithelium in cancer animals. This evidence strongly supports our hypothesis that urogenital carcinoma in California sea lions is a virally induced cancer and is an appropriate model for studying virally induced cancer in a host with naturally occurring disease.

The multifactorial pathophysiology of virally-induced cancers is well known and presents a persistent challenge in studying these diseases in traditional laboratory animals and in vitro models. The import roles of host microenvironment and immune system influences in driving herpesvirus gene expression are essential parts of the real- life factors that induce disease. Strong support already exists that contaminant exposure and genetic factors influence the potential of sea lions to develop urogenital carcinoma. Combined with the viral component of urogenital carcinoma, this very closely parallels the multifactorial nature of herpesvirus-induced cancers in people.

Studying this virally-induced cancer in sea lions can provide a natural disease model to help elucidate factors that cause herpesviruses to induce cellular transformation, malignancy and metastasis in their hosts. Developing a better understanding of factors that cause these ubiquitous viruses to cause disease in a select group of the infected population may provide novel approaches to treatment and prevention of these diseases.

In all cases of urogenital carcinoma, there were high expression levels of all

OtHV1 genes examined within the neoplastic epithelium of the vagina and cervix. No

91

healthy cervical or vaginal epithelium had positive RISH labeling for any OtHV1 genes examined. OtHV1 gene expression levels were significantly higher in CIN compared to invasive lesions, which concurs with findings of other oncogenic viruses. Early viral activity may initiate cellular transformation but have a lesser role as the malignancy progresses. However, significantly higher expression levels of OtHV1 genes in the CIN lesions compared to invasive lesions may also have been the result of the invasive lesions having larger, less tightly packed cells with lower nucleus: cytoplasm ratios within the same sized area when compared to the more tightly-packed and slightly smaller neoplastic cells in the CIN lesions. Regardless of the cause of higher OtHV1 gene expression in CIN versus invasive lesions, the high levels of expression in both lesion types, in combination with the lack of OtHV1 gene expression in normal cervical epithelium, supports the important role of OtHV1 in urogenital carcinoma.

It can be challenging to find an appropriate host cellular housekeeping gene as a positive control in cancer tissues; there are alterations in various cellular pathways of cancer cells as well as virally infected cells. We used two positive controls, polR2A and

PPIB, in an attempt to identify a good candidate. The expression level of polR2A was not significantly different in healthy, CIN, or invasive lesions, and thus provided a good positive control. The PPIB gene, however, had a stepwise increase, with the lowest expression in the control cervix/vagina epithelium (median=1.29%, range 0.07%-2.75%) compared to CIN (median=3.9%, range=0.43%-8.23%) and the highest expression levels in invasive lesions (median= 13.8%, range 1.1%-61.0%). PPIB acts as an intracellular chaperone and is a member of the immunophilin family of proteins, often localized to the endoplasmic reticulum but also contains a nuclear localization signal

92

and can be found on the plasma membrane of cells (Prince et al. 1994). PPIB has been found to participate in a number of fundamental cellular processes that involve protein folding and plays an important role in intracellular transport. High expression levels of

PPIB have also been linked to viral infections. PPIB has been found to enhance HIV infection (DeBoer et al. 2016), with HIV infected cells having significant PPIB upregulation (Haverland et al. 2014), and it plays a role in HIV viral replication (Brass et al. 2008). PPIB has been found to be essential for infection of additional viruses, including human papillomavirus 16 (Bienkowska-Haba et al. 2009), virus

(Watashi et al. 2005), and Japanese encephalitis virus (Kambara et al. 2011). Thus, this increase in sea lion PPIB expression in California sea lion urogenital carcinoma may be the result of increased viral replication. Because of the known involvement of PPIB in viral replication, and the concurrent increase in PPIB with OtHV1 gene expression, the data suggest that this host housekeeping gene may be useful for assessing viral replication in California sea lion urogenital carcinoma, and perhaps other virally induced cancers. Further work would be needed to validate PPIB as a marker for viral replication.

Urogenital carcinoma has been described as having multicentric development within the urogenital tract (Colegrove et al. 2009a). This was supported in all cases examined for this study, as every case of urogenital carcinoma had evidence of multiple or diffuse intraepithelial lesions in the cervical and/or vaginal epithelium. Interestingly, all but one case with only CIN (no evidence of deep or localized invasion) also had distant metastases. Conversely, invasion did not guarantee metastatic disease, as in one case lacking metastatic lesions, both CIN and invasive lesions were identified in the cervical

93

tissue sections examined. This demonstrates the challenge of trying to grade these lesions.

Collecting biopsies antemortem or using small samples of the reproductive tract during postmortem examination may not give a true understanding of the extent or severity of the cancer. Thus, multiple biopsies and histologic examination from various areas of the proximal vaginal and cervix are recommended to increase the likelihood of identifying invasive lesions. These portions of the reproductive tract consist of thick, convoluted vaginal and cervical folds, making lesion identification grossly challenging in some cases- even during necropsy. Fixation of the whole reproductive tract in 10% neutral buffered formalin can make gross lesions more apparent. Additional diagnostics such as ultrasound, radiographs or complete necropsy are highly recommended to stage the cancer. In cases where metastatic lesions were present without evidence of invasive lesions, we propose there may have been invasive lesions in areas of cervix and/or vagina not sampled.

In conclusion, our findings show that OtHV1 is an important contributing factor in urogenital carcinoma of California sea lions and these animals can be a useful model for studying virally-induced caners. Future research including viral gene expression in primary and metastatic lesions, as well as studying the impacts of immune competence, contaminants, genetic factors, and hormone influences may help elucidate the various intrinsic and extrinsic factors that drive virally-induced cancer development and progression. Moreover, the availability of post-mortem samples from various stages of cancer can provide a valuable model for studying naturally occurring metastatic disease, which currently lacks good laboratory models that recapitulate this process.

94

A B

Figure 4-1. H&E. Histology sections of normal cervix from two adult California sea lions at different times of year. A) Normal epithelium showing pseudostratified columnar epithelial layer with under laying stoma collected in the spring (case 1004-1), and B) normal stratified squamous epithelium showing areas of and metaplasia secondary to estrogen effects during pupping season in the summer (case 1119-1).

A B

Figure 4-2. H&E. Histology sections of urogenital carcinoma. A) cervical intraepithelial neoplasia (CIN, case 1041-11) and B) invasive urogenital carcinoma lesions (case 1094-1) from the cervix of an adult California sea lion.

95

Table 4-1. Normal cervical or vaginal cases from California sea lions that died or were euthanized for non-cancer related causes. (control cases). Collection Collection Pathology Cause of Death Patient ID Histology description site date ID CSL 11220 Cervix 7/9/14 171041-3 Normal epithelium Domoic acid (euthanized) CSL 12902 Cervix 3/4/16 171003-1 Normal epithelium Domoic acid and severe malnutrition CSL 13002 Cervix 4/6/16 171004-1 Normal epithelium Undetermined (died in treatment) CSL 13383 Cervix 6/19/17 171091-1 Normal epithelium Suspect sarcocystis (euthanized) CSL 13399 Vagina 7/5/17 171095-1 Normal epithelium Undetermined (euthanized) CSL 13453 Vagina 7/26/17 171100-1 Normal epithelium Domoic acid (euthanized) CSL 13483 Cervix 8/9/17 171114-1 Normal epithelium Domoic acid (euthanized) CSL 13491 Cervix 8/9/17 171116-1 Normal epithelium Domoic acid (euthanized) CSL 13467 Cervix 8/14/17 171119-1 Normal dysplasia Domoic acid (euthanized)

96

Table 4-2. Urogenital carcinoma cases from California sea lions that died or were euthanized with urogenital carcinoma as primary cause of death, or had lesions identified incidentally on histologic examination. Collection Collection Pathology Metastases Cause of Death Patient ID Histology description site date ID 0=no, 1=yes CSL 10337 Vagina 6/19/12 171041-12 1 Carcinoma in situ Carcinoma (euthanized) CSL 10462 Cervix 11/8/12 171041-16 1 Carcinoma in situ Carcinoma (died in treatment) CSL 10611 Vagina 4/9/13 171041-18 1 Carcinoma in situ Carcinoma (euthanized) CSL 10675 Cervix 7/24/13 171041-20 1 Carcinoma in situ Carcinoma (euthanized) CSL 10707 Cervix 8/27/13 171041-22 1 Carcinoma in situ Carcinoma (euthanized) CSL 10778 Cervix 12/17/13 171041-26 1 Carcinoma in situ Carcinoma (euthanized) CSL 13473 Cervix 8/1/17 171101-2 0 Carcinoma in situ Domoic acid (died in treatment) CSL 13479 Cervix 8/14/17 171120-1 0 CIN and invasive UGC Domoic acid (euthanized) CSL 10240 Cervix 11/13/11 171041-10 1 CIN and invasive UGC Carcinoma (died in treatment) CSL 10273 Cervix 1/29/12 171041-11 1 CIN and invasive UGC Carcinoma (euthanized) CSL 10482 Vagina 12/20/12 171041-17 1 CIN and invasive UGC Carcinoma (euthanized) CSL 10689 Cervix 8/2/13 171041-21 1 CIN and invasive UGC Carcinoma (euthanized) CSL 12597 Cervix 7/2/15 171041-30 1 CIN and invasive UGC Carcinoma (euthanized) CSL 13337 Cervix 5/19/17 171073-1 1 CIN and invasive UGC Carcinoma (euthanized) CSL 13325 Cervix 5/08/17 171071-1 1 All invasive UGC Carcinoma (euthanized) CSL 13385 Cervix 6/22/17 171094-1 1 All invasive UGC Carcinoma (euthanized)

97

Table 4-3. Basescope custom probes targeting 5 viral genes, 2 positive controls and a negative control. Proposed mechanism of action (MOA) for genes and number (#) zz pairs provided.

Probe name Homologous gene Proposed MOA for genes # ZZ pairs

Anti-apoptotic, induces cellular proliferation: causes OtHV1-EBNA1 Epstein-Barr Nuclear Antigen 1-like 4ZZ chromosomal instability Viral Fas-associated death domain-like Anti-apoptotic; binds Caspase 8 and blocks death OtHV1-vFLIP interleukin-1β-converting enzyme- 4ZZ receptor-induce apoptosis inhibitory protein

OtHV1-vBCL2 Viral B-cell lymphoma 2 Anti-apoptotic and immune system evasion 4ZZ

OtHV1-vCDK4 Viral cyclin-dependent kinase 4 Induces cellular proliferation 4ZZ

Suspected retroviral element inserted in viral and host OtHV1-vEVE Viral endogonous viral element 4ZZ genome Zalophus californianus peptidyl prolyl Regulate protein folding of type I collagen: viral Zc-PPIB 3ZZ isomerase B replication Zalophus californianus DNA-dependent Catalyzes the transcription of DNA to mRNA and most Zc-polR2A 3ZZ RNA polymerase II snRNA and microRNA Bacillus subtilis dihydrodipicolinate Component of the lysine biosynthetic pathway in bacteria dapB 3ZZ reductase and higher plants

98

A

B

C Figure 4-3. Histology sections of cervix from California sea lions labeled with RISH positive control probe polR2A. Pink punctate areas indicate positive labeling and black lines indicate margins of epithelium/lesion of interest selected for image analysis of RISH positive percent labeling. A) control case with normal cervical epithelium (case #); and B) CIN (case 1120-1) and C) invasive (case 1094-1) lesions from animals with urogenital carcinoma.

99

Table 4-4. Quartile results for percent positive labeling in control cases for all OtHV1 RISH probes and positive controls from 9 California sea lions with no evidence of urogenital carcinoma grossly or on histopathologic examination. Healthy OtHV1 BaseScope Probes Positive controls (n=9) EBNA1 vFLIP vBCL2 vCDK4 vEVE PPIB polR2A Minimum 0.000 0.000 0.000 0.000 0.000 0.069 0.075 First quartile 0.000 0.001 0.000 0.000 0.002 0.680 0.357 Median 0.001 0.002 0.001 0.001 0.052 1.288 0.531 Third quartile 0.003 0.010 0.006 0.001 0.079 1.441 0.648 Maximum 0.023 0.139 0.052 0.005 0.259 2.750 1.433

Table 4-5. Quartile results for percent positive labeling in urogenital carcinoma intraepithelial neoplasia (CIN) lesions for all OtHV1 RISH probes and positive controls from 14 urogenital carcinoma cases with histologic finding of CIN (CIN only=7, CIN with invasion lesions=7). OtHV1 BaseScope Probes Positive controls CIN (n=14) EBNA1 vFLIP vBCL2 vCDK4 vEVE PPIB polR2A Minimum 8.195 21.430 40.147 23.487 24.202 0.433 0.321 First quartile 39.122 47.746 49.316 45.246 46.039 2.483 0.449 Median 49.551 54.320 54.191 54.744 53.979 3.909 0.596 Third quartile 56.197 64.469 58.212 60.065 63.894 6.051 1.609 Maximum 69.639 71.489 79.967 75.510 75.738 8.226 2.232

Table 4-6. Quartile results for percent positive labeling in invasive urogenital carcinoma lesions for all OtHV1 RISH probes and positive controls from 9 California sea lions (CIN and invasive lesions=7, 2=invasive only). Invasive OtHV1 BaseScope Probes Positive controls (n=9) EBNA1 vFLIP vBCL2 vCDK4 vEVE PPIB polR2A Minimum 22.034 24.484 23.740 17.319 32.263 1.110 0.061 First quartile 31.238 31.476 31.527 29.839 35.189 8.736 0.304 Median 37.268 43.612 34.984 39.831 41.394 13.839 0.661 Third quartile 47.499 50.575 49.084 46.083 53.586 18.968 1.034 Maximum 73.571 61.410 66.312 64.768 70.415 61.027 1.855

100

Figure 4-4. Representative series of H&E (top row) and RISH labeled tissue from control cervix (left column), CIN lesion (middle column), and invasive lesion (right column) with OtHV1 RISH probes (vBCL2, vCDK4, EBNA1, vFLIP, and vEVE), positive controls (PPIB and polR2) and negative control (dapB).

101

Figure 4-4. Continued.

102

A B

C D

Figure 4-5. Series of early, non-invasive urogenital carcinoma of the cervix with normal epithelium transitioning into cervical intraepithelial carcinoma (CIN). Notice the OtHV1 vEVE RNA in situ hybridization probe (B) only binds to neoplastic epithelial cells with in the CIN lesion and not to the surrounding normal epithelial cells. This pattern was consistent for all OtHV1 RISH probes. A) H&E of normal epithelium (left) with dysplastic epithelium (right) extending into the stroma with significant inflammatory reaction surrounding CIN, B) OtHV1 vEVE RISH probe is positively staining in CIN but not in neighboring normal epithelium or stroma, C) host housekeeping gene DNA-dependent RNA polymerase II RISH positive control is staining homogenously across both CIN and healthy epithelium and D) negative control dihydrodipicolinate reductase of Bacillus subtilis (dapB) has no binding to normal epithelium, CIN or surrounding stroma.

103

vEVE

Figure 4-6. Dot plot showing mean percent positive staining of BaseScope probes for all healthy cervix, CIN and invasive tumors. A) negative control (dapB), B) positive control (polR2A), C) positive control PPIB, D) vBCL2, E) vCDK4, F) vEVE, G) EBNA1, H) vFLIP. *=significant difference >0.01

104

CHAPTER 5 CONCLUSION

This work shows that urogenital carcinoma continues to be a common cancer in

California sea lions along the western coast of the United States. It was diagnosed in 1 of every 4 adult sea lions that presented to The Marine Mammal Center from 2005 to

2015. We sequenced the genome of OtHV1 from a cervical tumor collected from a sea lion with metastatic urogenital carcinoma and identified several potential viral oncogenes within this OtHV1 draft genome. Novel OtHV1 RNA in situ hybridization probes targeting the potential oncogenes showed significantly higher expression within cervical tumors than in normal cervical epithelium. These findings strongly support the hypothesis that OtHV1 plays a role in the etiology of urogenital carcinoma.

Additionally, California sea lion OtHV1 and Northern fur seal OtHV4 were found to be most closely related to common bottlenose dolphin DeGHV1- a herpesvirus identified in bottlenose dolphin genital lesions. Phylogenetic analysis showed that these three marine mammal herpesviruses group together and branch off early in the herpesvirus lineage, consistent with a new genus in Herpesviridae. Interestingly, two of these viruses (OtHV1 and DeGHV1) are associated with neoplastic lesions. Our findings support the hypothesis that urogenital carcinoma in California sea lions is a virally induced cancer and can be studied as a comparative model for naturally occurring, virally-induced cancers.

Like other virally induced cancers, urogenital carcinoma is likely the result of several contributing factors. This multifactorial pathophysiology is challenging to recapitulate in traditional laboratory models. In HHV8 and HHV4, it is known that the host’s immune system plays a critical role in the progression of these infections to

105

induce malignancy. We propose that similar cofactors are likely at play in the development of urogenital carcinoma in sea lions. Although studying diseases in wildlife and non-traditional species has limitations in the ability to control specific parameters, these opportunistic samples can still provide valuable models for studying disease pathophysiology in a real-world setting. This is particularly true with current advances in molecular biology, such as next generation sequencing and RNA in situ hybridization, as used in this study. Future work using RNAseq to explore the impacts OtHV1 has on host gene expression in primary and metastatic lesions may also provide valuable insights into the important cellular pathway protuberances that result in malignancy and metastasis. The oncogenic mechanisms employed by OtHV1 in urogenital carcinoma more realistically represent the pathophysiology of the virally-induced cancer progression than traditional laboratory models. Future research is needed on the effects of contaminants on sea lion immune function, as well as the role of host genetics, as factors contributing to the progression of this disease.

106

LIST OF REFERENCES

Acevedo-Whitehouse, K., F. Gulland, D. Greig, and W. Amos. 2003. Inbreeding: Disease susceptibility in California sea lions. Nature 422:35.

Alcami, A. 2003. Viral mimicry of cytokines, chemokines and their receptors. Nat Rev Immunol 3:36-50.

Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J Mol Biol 215:403-410.

Aswad, A., and A. Katzourakis. 2015. Convergent capture of retroviral superantigens by mammalian herpesviruses. Nat Commun 6:8299.

Ballestas, M. E., P. A. Chatis, and K. M. Kaye. 1999. Efficient persistence of extrachromosomal KSHV DNA mediated by latency-associated nuclear antigen. Science 284:641-644.

Barragán-Vargas, C., J. Montano-Frías, G. Ávila Rosales, C. R. Godínez-Reyes, and K. Acevedo-Whitehouse. 2016. Transformation of the genital epithelial tract occurs early in California sea lion development. R Soc Open Sci 3:150419.

Bienkowska-Haba, M., H. D. Patel, and M. Sapp. 2009. Target Cell Cyclophilins Facilitate Human Papillomavirus Type 16 Infection. PLoS Pathogens 7:e1000524.

Boshoff, C., T. F. Schulz, M. M. Kennedy, A. K. Graham, C. Fisher, A. Thomas, J. O. McGee, R. A. Weiss, and J. J. O'Leary. 1995. Kaposi's sarcoma-associated herpesvirus infects endothelial and spindle cells. Nat Med 1:1274-1278.

Bowen, L., B. M. Aldridge, R. Delong, S. Melin, E. L. Buckles, F. Gulland, L. J. Lowenstine, J. L. Stott, and M. L. Johnson. 2005. An immunogenetic basis for the high prevalence of urogenital cancer in a free-ranging population of California sea lions (Zalophus californianus). Immunogenetics 56:846-848.

Brass, A. L., D. M. Dykxhoorn, Y. Benita, N. Yan, A. Engelman, R. J. Xavier, J. Lieberman, and S. J. Elledge. 2008. Identification of host proteins required for HIV infection through a functional genomic screen. Science 319:921-926.

Browning, H. M., K. Acevedo-Whitehouse, F. M. D. Gulland, A. J. Hall, J. Finlayson, M. P. Dagleish, K. J. Billington, K. Colegrove, and J. A. Hammond. 2014. Evidence for a genetic basis of urogenital carcinoma in the wild California sea lion. Proceedings of the Royal Society Biological Sciences Series B 281.

Browning, H. M., F. M. D. Gulland, J. A. Hammond, K. M. Colegrove, and A. J. Hall. 2015. Common cancer in a wild animal: the California sea lion (Zalophus californianus) as an emerging model for carcinogenesis. Philosophical Transactions of the Royal Society of London B Biological Sciences 370.

107

Buckles, E. L., L. J. Lowenstine, R. L. DeLong, S. R. Melin, R. K. Vittore, H. N. Wong, G. L. Ross, J. A. St Leger, D. J. Greig, R. S. Duerr, F. M. Gulland, and J. L. Stott. 2007. Age- prevalence of Otarine Herpesvirus-1, a tumor-associated virus, and possibility of its sexual transmission in California sea lions. Vet Microbiol 120:1-8.

Carretta, J. V., E. M. Oleson, J. Baker, D. W. Weller, A. R. Lang, K. A. Forney, M. M. Muto, B. Hanson, A. Orr, J. Huber, H. Lowry, M. S. Barlow, J. Moore, J. E.Lynch, D. Carswell, L. Brownell, R. De Long. 2016. U.S. Pacific marine mammalmstock assessments: 2015. 1- 7.

Cheng, E. H., J. Nicholas, D. S. Bellows, G. S. Hayward, H. G. Guo, M. S. Reitz, and J. M. Hardwick. 1997. A Bcl-2 homolog encoded by Kaposi sarcoma-associated virus, human herpesvirus 8, inhibits apoptosis but does not heterodimerize with Bax or Bak. Proc Natl Acad Sci U S A 94:690-694.

Colegrove, K. M., F. M. Gulland, D. K. Naydan, and L. J. Lowenstine. 2009a. The normal genital tract of the female California sea lion (Zalophus californianus): cyclic changes in histomorphology and hormone receptor distribution. Anat Rec (Hoboken) 292:1801- 1817.

Colegrove, K. M., F. M. Gulland, D. K. Naydan, L. J. Lowenstine, 2009b. Tumor morphology and immunohistochemical expression of estrogen receptor, progesterone receptor, p53, and Ki67 in urogenital carcinomas of California sea lions (Zalophus californianus). Vet Pathol 46:642-655.

Cortés-Hinojosa, G., F. M. Gulland, R. DeLong, T. Gelatt, L. Archer, and J. F. Wellehan. 2016. A novel gammaherpesvirus in northern fur seals (callorhinus ursinus) is closely related to the california sea lion (zalophus californianus) carcinoma-associated otarine herpesvirus-1. J Wildl Dis 52:88-95.

Cotter, M. A., and E. S. Robertson. 1999. The latency-associated nuclear antigen tethers the Kaposi's sarcoma-associated herpesvirus genome to host chromosomes in body cavity- based lymphoma cells. Virology 264:254-264.

Damania, B. 2004. Oncogenic gamma-herpesviruses: comparison of viral proteins involved in tumorigenesis. Nat Rev Microbiol 2:656-668.

De Guise, S., A. Lagacé, P. Béland. 1994. Tumors in St. Lawrence Beluga Whales (Delphinapterus leucas). Vet Pathol 31:444-449.

De Guise, S., D. Martineau, P. Béland, and M. Fournier. 1995. Possible mechanisms of action of environmental contaminants on St. Lawrence beluga whales (Delphinapterus leucas). Environ Health Perspect 103 Suppl 4:73-77.

108

DeLong, R. L., W. G, Gilmartin, J. G. Simpson. 1973. Premature births in California sea lions: association with high organochlorine pollutant residue levels. Science 181:1168-1170.

Del Toro, L., G. Heckel, V. F. Camacho-Ibar, and Y. Schramm. 2006. California sea lions (Zalophus californianus californianus) have lower chlorinated hydrocarbon contents in northern Baja California, México, than in California, USA. Environ Pollut 142:83-92.

Deming, A. C., K. M. Colegrove, P. J. Duignan, A. J. Hall, J. F. X. Wellehan, and F. M. D. Gulland. 2018. Prevalence of Urogenital Carcinoma in Stranded California Sea Lions (Zalophus californianus) from 2005-15. J Wildl Dis 54:581-586.

Du, H., M. Pang, X. Hou, S. Yuan, and L. Sun. 2017. PLOD2 in cancer research. Biomed Pharmacother 90:670-676.

Eisinger-Mathason, T. S., M. Zhang, Q. Qiu, N. Skuli, M. S. Nakazawa, T. Karakasheva, V. Mucaj, J. E. Shay, L. Stangenberg, N. Sadri, E. Puré, S. S. Yoon, D. G. Kirsch, and M. C. Simon. 2013. Hypoxia-dependent modification of collagen networks promotes sarcoma metastasis. Cancer Discov 3:1190-1205.

Elde, N. C., and H. S. Malik. 2009. The evolutionary conundrum of pathogen mimicry. Nat Rev Microbiol 7:787-797.

Falkow, S. 1988. Molecular Koch's postulates applied to microbial pathogenicity. Rev Infect Dis 10 Suppl 2:S274-276.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791.

Feng, Y., M. N. E. Forsell, B. Flynn, W. Adams, K. Lore, R. Seder, R. T. Wyatt, and G. B. K. Hedestam. 2013. Chemical cross-linking of HIV-1 Env for direct TLR7/8 ligand conjugation compromises recognition of conserved antigenic determinants. Virology 446:56-65.

Fredricks, D. N., and D. A. Relman. 1996. Sequence-based identification of microbial pathogens: a reconsideration of Koch's postulates. Clin Microbiol Rev 9:18-33.

Ganem, D. 1997. KSHV and Kaposi's sarcoma: the end of the beginning? Cell 91:157-160.

Gilkes, D. M., S. Bajpai, C. C. Wong, P. Chaturvedi, M. E. Hubbi, D. Wirtz, and G. L. Semenza. 2013. Procollagen lysyl hydroxylase 2 is essential for hypoxia-induced breast cancer metastasis. Mol Cancer Res 11:456-466.

Gilkes, D. M., G. L. Semenza, and D. Wirtz. 2014. Hypoxia and the extracellular matrix: drivers of tumour metastasis. Nat Rev Cancer 14:430-439.

109

Greig, D. J., F. M. D. Gulland, and Kreuder, Christine. 2005. A Decade of Live California Sea Lion (Zalophus californianus) Strandings Along the Central California Coast: Causes and Trends, 1991-2000. Aquatic Mammals 31:11-22.

Greig, D. J., G. M. Ylitalo, A. J. Hall, D. A. Fauquier, and F. Gulland. 2007. Transplacental transfer of organochlorines in California sea lions (Zalophus californianus). Environ Toxicol Chem 26:37-44.

Grinde, B. 2013. Herpesviruses: latency and reactivation - viral strategies and host response. J Oral Microbiol 5.

Gross, A., J. M. McDonnell, and S. J. Korsmeyer. 1999. BCL-2 family members and the mitochondria in apoptosis. Genes & Development 13:1899-1911.

Grulich, A. E., M. T. van Leeuwen, M. O. Falster, and C. M. Vajdic. 2007. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis. Lancet 370:59-67.

Gulland, F. M., J. G. Trupkiewicz, T. R. Spraker, and L. J. Lowenstine. 1996. Metastatic carcinoma of probable transitional cell origin in 66 free-living California sea lions (Zalophus californianus), 1979 to 1994. J Wildl Dis 32:250-258.

Haverland, N. A., H. S. Fox, and P. Ciborowski. 2014. Quantitative proteomics by SWATH-MS reveals altered expression of nucleic acid binding and regulatory proteins in HIV-1- infected macrophages. J Proteome Res 13:2109-2119.

Henderson, S., D. Huen, M. Rowe, C. Dawson, G. Johnson, and A. Rickinson. 1993a. Epstein- Barr virus-coded BHRF1 protein, a viral homologue of Bcl-2, protects human B cells from programmed cell death. Proc Natl Acad Sci U S A 90:8479-8483.

Hernández-Ramírez, R. U., M. S. Shiels, R. Dubrow, and E. A. Engels. 2017. Cancer risk in HIV-infected people in the USA from 1996 to 2012: a population-based, registry-linkage study. Lancet HIV 4:e495-e504.

Howlader, N., A. B. Mariotto, C. Besson, G. Suneja, K. Robien, N. Younes, and E. A. Engels. 2017. Cancer-specific mortality, cure fraction, and noncancer causes of death among diffuse large B-cell lymphoma patients in the immunochemotherapy era. Cancer 123:3326-3334.

Jansen, M. A., D. van den Heuvel, S. H. Bouthoorn, V. W. Jaddoe, H. Hooijkaas, H. Raat, P. L. Fraaij, M. C. van Zelm, and H. A. Moll. 2016. Determinants of Ethnic Differences in Cytomegalovirus, Epstein-Barr Virus, and Type 1 Seroprevalence in Childhood. J Pediatr 170:126-134.e121-126.

110

Kambara, H., H. Tani, Y. Mori, T. Abe, H. Katoh, T. Fukuhara, S. Taguwa, K. Moriishi, and Y. Matsuura. 2011. Involvement of cyclophilin B in the replication of Japanese encephalitis virus. Virology 412:211-219.

Kannan, K., N. Kajiwara, B. J. LeBoeuf, S. Tanabe. 2004. Organochlorine pesticides and polychlorinated biphenyls in California sea lions. Enviro Pollu 131:425-434.

Katoh, K., and H. Toh. 2008. Improved accuracy of multiple ncRNA alignment by incorporating structural information into a MAFFT-based framework. BMC Bioinformatics 9:212.

Katzourakis, A., and R. J. Gifford. 2010. Endogenous viral elements in animal genomes. PLoS Genet 6:e1001191.

King, D. P., M. C. Hure, T. Goldstein, B. M. Aldridge, F. M. Gulland, J. T. Saliki, E. L. Buckles, L. J. Lowenstine, and J. L. Stott. 2002. Otarine herpesvirus-1: a novel gammaherpesvirus associated with urogenital carcinoma in California sea lions (Zalophus californianus). Vet Microbiol 86:131-137.

Klein, G. 1972. Herpesviruses and oncogenesis. Proc Natl Acad Sci USA 69:1056-1064.

Kurman, R. J., H. J. Norris, E. J. Wilkinson, Armed Forces Institute of Pathology (U.S.), and Universities Associated for Research and Education in Pathology. 1992. Tumors of the cervix, vagina, and . Armed Forces Institute of Pathology, Washington, D.C.

Kurozumi, A., M. Kato, Y. Goto, R. Matsushita, R. Nishikawa, A. Okato, I. Fukumoto, T. Ichikawa, and N. Seki. 2016. Regulation of the collagen cross-linking enzymes LOXL2 and PLOD2 by tumor-suppressive microRNA-26a/b in renal cell carcinoma. Int J Oncol 48:1837-1846.

Kvansakul, M., and M. G. Hinds. 2013a. Structural biology of the Bcl-2 family and its mimicry by viral proteins. Cell Death Dis 4:e909.

Laake, J.L., M.S. Lowry, R.L. DeLong, S.R. Melin, J.V. Carretta. 2018. Population growth and status of California sea lions. J Wildl Manage 82:583-595.

Lair, S., L. N. Measures, and D. Martineau. 2016. Pathologic Findings and Trends in Mortality in the Beluga (Delphinapterus leucas) Population of the St Lawrence Estuary, Quebec, Canada, From 1983 to 2012. Vet Pathol 53:22-36.

Lancaster, W. D., and C. Olson. 1982. Animal papillomaviruses. Microbiol Rev 46:191-207.

Leathlobhair, M. N., F. M. D. Gulland, and E. P. Murchison. 2017. No evidence for clonal transmission of urogenital carcinoma in California sea lions (Zalophus californianus). Wellcome Open Res 2:46.

LeBouef, B.J, and M. L. Bonnell. 1971. DDT in California sea lions. Nature 234:108-109.

111

Lipscomb, T. P., D. P. Scott, R. L. Garber, A. E. Krafft, M. M. Tsai, J. H. Lichy, J. K. Taubenberger, F. Y. Schulman, and F. M. Gulland. 2000. Common metastatic carcinoma of California sea lions (Zalophus californianus): evidence of genital origin and association with novel gammaherpesvirus. Vet Pathol 37:609-617.

Looker, K. J., A. S. Magaret, M. T. May, K. M. Turner, P. Vickerman, S. L. Gottlieb, and L. M. Newman. 2015a. Global and Regional Estimates of Prevalent and Incident Herpes Simplex Virus Type 1 Infections in 2012. PLoS One 10:e0140765.

Looker, K. J., A. S. Magaret, K. M. Turner, P. Vickerman, S. L. Gottlieb, and L. M. Newman. 2015b. Global estimates of prevalent and incident herpes simplex virus type 2 infections in 2012. PLoS One 10:e114989.

Marshall, W. L., C. Yim, E. Gustafson, T. Graf, D. R. Sage, K. Hanify, L. Williams, J. Fingeroth, and R. W. Finberg. 1999. Epstein-Barr virus encodes a novel homolog of the bcl-2 oncogene that inhibits apoptosis and associates with Bax and Bak. J Virol 73:5181- 5185.

Mayer, J., M. Sauter, A. Rácz, D. Scherer, N. Mueller-Lantzsch, and E. Meese. 1999. An almost-intact human endogenous retrovirus K on human chromosome 7. Nat Genet 21:257-258.

McGeoch, D. J., F. J. Rixon, and A. J. Davison. 2006. Topics in herpesvirus genomics and evolution. Virus Res 117:90-104.

Mettenleiter, T. C., B. G. Klupp, and H. Granzow. 2009. Herpesvirus assembly: an update. Virus Res 143:222-234.

Mi, S., X. Lee, X. Li, G. M. Veldman, H. Finnerty, L. Racie, E. LaVallie, X. Y. Tang, P. Edouard, S. Howes, J. C. Keith, and J. M. McCoy. 2000. Syncytin is a captive retroviral envelope protein involved in human placental morphogenesis. Nature 403:785-789.

Michaelis, M., J. Geiler, D. Klassert, H. W. Doerr, and J. Cinatl. 2009. Infection of human retinal pigment epithelial cells with A viruses. Invest Ophthalmol Vis Sci 50:5419-5425.

Miller, M. A., T. Schwartz, B. E. Pickett, S. He, E. B. Klem, R. H. Scheuermann, M. Passarotti, S. Kaufman, and M. A. O'Leary. 2015. A RESTful API for Access to Phylogenetic Tools via the CIPRES Science Gateway. Evol Bioinform Online 11:43-48.

Mittnacht, S., and C. Boshoff. 2000. Viral cyclins. Rev Med Virol 10:175-184.

Miyamoto, K., N. Seki, R. Matsushita, M. Yonemori, H. Yoshino, M. Nakagawa, and H. Enokida. 2016. Tumour-suppressive miRNA-26a-5p and miR-26b-5p inhibit cell aggressiveness by regulating PLOD2 in bladder cancer. Br J Cancer 115:354-363.

112

Moore, P. S., and Y. Chang. 2010. Why do viruses cause cancer? Highlights of the first century of human tumour virology. Nat Rev Cancer 10:878-889.

Nagata, C., H. Kobayashi, A. Sakata, K. Satomi, Y. Minami, Y. Morishita, R. Ohara, H. Yoshikawa, Y. Arai, M. Nishida, and M. Noguchi. 2012. Increased expression of OCIA domain containing 2 during stepwise progression of ovarian mucinous tumor. Pathol Int 62:471-476.

Newmeyer, D. D., and S. Ferguson-Miller. 2003. Mitochondria: Releasing power for life and unleashing the machineries of death (vol 112, pg 481, 2003). Cell 112:873-873.

Ní, M., Leathlobhair, M., F. M. D. Gulland, and E. P. Murchison. 2017. No evidence for clonal transmission of urogenital carcinoma in California sea lions ( Zalophus californianus). Wellcome Open Res 2:46.

Noda, T., H. Yamamoto, I. Takemasa, D. Yamada, M. Uemura, H. Wada, S. Kobayashi, S. Marubashi, H. Eguchi, M. Tanemura, K. Umeshita, Y. Doki, M. Mori, and H. Nagano. 2012. PLOD2 induced under hypoxia is a novel prognostic factor for hepatocellular carcinoma after curative resection. Liver Int 32:110-118.

Odell, D. K. 1975. Breeding biology of the California sea lion, Zalophus californianus. Rapports et Proces-verbaux de Rdunions Conseil International pour L'exploration de la Met 169:374-378.

Oltvai, Z. N., C. L. Milliman, and S. J. Korsmeyer. 1993. BCL-2 heterodimerizes in-vivo with a conserved homolog, Bax, that accelerates programmed cell-death. Cell 74:609-619.

Pagano, J. S., M. Blaser, M. A. Buendia, B. Damania, K. Khalili, N. Raab-Traub, and B. Roizman. 2004. Infectious agents and cancer: criteria for a causal relation. Seminars in Cancer Biology 14:453-471.

Parkin, D. M. 2006. The global health burden of infection-associated cancers in the year 2002. Int J Cancer 118:3030-3044.

Parrish, N. F., and K. Tomonaga. 2016. Endogenized viral sequences in mammals. Curr Opin Microbiol 31:176-183.

Ploegh, H. L. 1998. Viral strategies of immune evasion. Science 280:248-253.

Polster, B. M., J. Pevsner, and J. M. Hardwick. 2004. Viral Bcl-2 homologs and their role in virus replication and associated diseases. Biochimica Et Biophysica Acta-Molecular Cell Research 1644:211-227.

113

Prince, L. S., R. B. Workman, and R. B. Marchase. 1994. Rapid endocytosis of the cystic fibrosis transmembrane conductance regulator chloride channel. Proc Natl Acad Sci U S A 91:5192-5196.

Puglielli, M. T., N. Desai, and S. H. Speck. 1997. Regulation of EBNA gene transcription in lymphoblastoid cell lines: characterization of sequences downstream of BCR2 (Cp). J Virol 71:120-128.

Radford, A. D., D. Chapman, L. Dixon, J. Chantrey, A. C. Darby, and N. Hall. 2012a. Application of next-generation sequencing technologies in virology. J Gen Virol 93:1853-1868.

Randhawa, N., F. Gulland, G. M. Ylitalo, R. DeLong, and J. A. K. Mazet. 2015. Sentinel California sea lions provide insight into legacy organochlorine exposure trends and their association with cancer and infectious disease. One Health 1:37-43.

Rappoport, N., and M. Linial. 2012. Viral proteins acquired from a host converge to simplified domain architectures. PLoS Comput Biol 8:e1002364.

Reeves, R. R., B. S. Stewart, and S. Leatherwood. 1992. The Sierra Club handbook of seals and sirenians. Sierra Club Books, San Francisco, CA.

Renne, R., C. Barry, D. Dittmer, N. Compitello, P. O. Brown, and D. Ganem. 2001. Modulation of cellular and viral gene expression by the latency-associated nuclear antigen of Kaposi's sarcoma-associated herpesvirus. J Virol 75:458-468.

Rivers, T. M. 1937. Viruses and Koch's Postulates. J Bacteriol 33:1-12.

Ronquist, F., M. Teslenko, P. van der Mark, D. L. Ayres, A. Darling, S. Höhna, B. Larget, L. Liu, M. A. Suchard, and J. P. Huelsenbeck. 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol 61:539- 542.

Rous, P. 1911. A sarcoma of the fowl transmissible by an agent separable from the tumor cells. J Exp Med 13:397-411.

Sarid, R., T. Sato, R. A. Bohenzky, J. J. Russo, and Y. Chang. 1997. Kaposi's sarcoma- associated herpesvirus encodes a functional bcl-2 homologue. Nat Med 3:293-298.

Sengupta, S., C. M. Michener, P. Escobar, J. Belinson, and R. Ganapathi. 2008. Ovarian cancer immuno-reactive antigen domain containing 1 (OCIAD1), a key player in ovarian cancer cell adhesion. Gynecol Oncol 109:226-233.

Stamatakis, A., P. Hoover, and J. Rougemont. 2008. A rapid bootstrap algorithm for the RAxML Web servers. Syst Biol 57:758-771.

114

VanDevanter, D. R., P. Warrener, L. Bennett, E. R. Schultz, S. Coulter, R. L. Garber, and T. M. Rose. 1996. Detection and analysis of diverse herpesviral species by consensus primer PCR. J Clin Microbiol 34:1666-1671.

Vogt, P. K. 1965. Avian tumor viruses. Adv Virus Res 11:293-385.

Wang, C., C. M. Michener, J. L. Belinson, S. Vaziri, R. Ganapathi, and S. Sengupta. 2010. Role of the 18:1 lysophosphatidic acid-ovarian cancer immunoreactive antigen domain containing 1 (OCIAD1)-integrin axis in generating late-stage ovarian cancer. Mol Cancer Ther 9:1709-1718.

Watashi, K., N. Ishii, M. Hijikata, D. Inoue, T. Murata, Y. Miyanari, and K. Shimotohno. 2005. Cyclophilin B is a functional regulator of RNA polymerase. Mol Cell 19:111-122.

Yang, A. H., J. Y. Chen, and C. H. Lee. 2012. Expression of NCAM and OCIAD1 in well- differentiated thyroid carcinoma: correlation with the risk of distant metastasis. J Clin Pathol 65:206-212.

Ylitalo, G. M., J. E. Stein, T. Hom, L. L. Johnson, K. L. Tilbury, A. J. Hall, T. Rowles, D. Greig, L. J. Lowenstine, and F. M. Gulland. 2005. The role of organochlorines in cancer- associated mortality in California sea lions (Zalophus californianus). Mar Pollut Bull 50:30-39.

Yu, L., P. T. Luan, W. Jin, O. A. Ryder, L. G. Chemnick, H. A. Davis, and Y. P. Zhang. 2011. Phylogenetic utility of nuclear introns in interfamilial relationships of Caniformia (order Carnivora). Syst Biol 60:175-187. zur Hausen, H. 2009. Childhood leukemias and other hematopoietic malignancies: Interdependence between an infectious event and chromosomal modifications. International Journal of Cancer 125:1764-1770.

115

BIOGRAPHICAL SKETCH

Dr. Alissa Deming is an aquatic animal veterinarian and molecular biologist, with research interests in comparative infectious disease diagnostics, novel pathogen identification, and host/microbe evolution and ecology. Dr. Deming completed her

Bachelor of Science in microbiology and biotechnology in 2003. After working as a sea turtle biologist for West Palm Beach Environmental Resource Management, she pursued her Master of Science in molecular biology at Florida Atlantic University. Her thesis work was focused on stress protein expression and biomarkers in sea turtle fibropapilloma, a common cancer in juvenile green turtles associated with a herpesvirus and closely linked to pollution.

Her aquatic animal focus continued throughout veterinary school at the University of Florida, where she obtained her Doctor of Veterinary Medicine and a certificate in aquatic animal health in the Spring of 2012. After veterinary school she completed a one-year small animal emergency and surgery internship at VCA Emergency Animal

Hospital and Referral Center in San Diego, California, followed by specialty training in aquatic animal medicine at the Navy Marine Mammal Program and SeaWorld San

Diego. Dr. Deming was selected as the Geoffrey Hughes Research Fellow at The

Marine Mammal Center in Sausalito, California, supporting this PhD work understanding the role of OtHV1 California sea lion urogenital carcinoma.

In addition to her research, Dr. Deming is a clinical veterinarian for the University of Florida Marine Animal Rescue Network as well as The Marine Mammal Center rehabilitation hospital. She provides veterinary support for stranding response, rehabilitation, health assessments and various field projects in the US and Mexico.

116