Clonally transmissible cancers

Mark N. Ziats1,2,3,†, Owen M. Rennert1

1National Institute of Child Health and Human Development, NIH, Bethesda, MD, USA

2Robinson College, University of Cambridge, Cambridge, UK

3Baylor College of Medicine MSTP, Houston, TX, USA

† Corresponding author, email: ziatsm@mail.nih.gov

Clonally transmissible cancers are tumors or cancerous cells that originally developed in one individual organism (clonality) but then passed on to different individuals through physical transfer of the cells (transmissibility). There are two well-established instances of clonally transmissible cancers -- devil facial tumor disease (DFTD), which occurs in the marsupial Tasmanian devil via biting, and canine transmissible venereal tumor (CTVT), which occurs in dogs through transmission during sexual contact. While CTVT has been purported to be a clonally transmissible cancer for decades, definitive scientific evidence that CTVT and DFTD are clonally transmissible cancers has only recently been established using computational techniques of comparative genomic sequencing. No known instances of clonally transmissible cancers currently exist in humans, but researchers are interested in understanding the properties of DFTD and CTVT to further research on human cancer and immunology.

Overview: Features of Clonally Transmissible Cancers
While the two known clonally transmissible cancers are distinct from each other in a number of ways (Table 1) having arisen in different species thousands of years apart from each other--they also share in common many genetic, cellular, and pathological traits that presumably underlie their unique ability to be propagated between individual animals of their respective populations.

Foremost, a clonally transmissible cancer must be genetically identical in each individual organism harboring the tumor (Figure 1). By definition, this implies that the tumor arose in one individual with that specific genotype and was subsequently passed onto other individuals with different genotypes, as opposed to each individual giving rise to their own tumors (which would then have different genotypes). Recognition of this requirement, and the computational ability to analyze many tumor genotypes and compare them to their host’s [[wp:genome|genomes], provided the first scientifically rigorous evidence that DFTD and CTVT are clonal in origin (A-M Pearse & Swift, 2006; Hannah V Siddle et al., 2007).

Additionally, DFTD and CTVT have been demonstrated to share a number of cellular and pathological features with each other that are somewhat unique to clonally transmissible cancers. These properties are likely in aggregate to represent the distinguishing features that allow a clonally transmissible cancer to propagate within a population. First, the genomes of both DFTD and CTVT are incredibly disorganized, likely the result of one catastrophic genomic event that placed genes in particular places in the genome relative to each other, enabling the emergent properties of clonal transmissibility. These properties include: a unique ability to evade detection and elimination by the hosts’ immune system, a malignancy virile enough to propagate through generations of animals without being so malignant as to kill those it affects before spread is possible, and the capacity to establish, in the case of CTVT, the oldest known malignant cell line in continuous existence in nature.

Furthermore, there are unique aspects of canine and devil population structure and behavior that have been proposed as properties necessary for a clonally transmissible cancer to develop and thrive. Both devils and canines have experienced population bottlenecks in their evolutionary histories (Pearse et al., 2012), thereby significantly diminishing the variability of genotypes throughout their populations. This has been postulated to be one mechanism allowing CTVT and DFTD to evade immune destruction in their recipient hosts, because presumably their host genomes are so similar that they do not react vigorously to the transplantation of another individual’s genome (Hannah V Siddle & Kaufman, 2012). Additionally, the behavior of canines and devils likely contributes to the spread of these clonal tumors. DFTD is mainly a cancer of the mouth and face and is spread through biting. Tasmanian devils, in the course of mating, are known to bite one another in the face area severely enough to cause open wounds, and therefore provide a means of transfer for the tumor cells (Hawkins et al., 2006). Similarly, CTVT is transferred when the tumor is passed through genital contact during mating. As canines have multiple copulations with many partners as part of their normal mating habits, this allows many opportunities for the clonal tumor to survive in the canine population.

Despite their similarities, devil facial tumor disease and canine transmissible venereal tumor have distinct histories, genomic properties, clinical and pathological features, and implications to their respective species, as is detailed further below. Additionally, the uniqueness of clonally transmissible cancers has generated intense research into these tumors in order to understand properties that are applicable to human cancer genomics and tumor immunology.

Devil Facial Tumor Disease (DFTD)
Devil Facial Tumor Disease (DFTD) is a cancer affecting the face and mouth of the marsupial Tasmanian devil (Sarcophilus harrisii), which inhabits the island of Tasmania off the mainland of Australia. DFTD is a newly evolved clonally transmissible cancer, and its rapid spread throughout the island of Tasmania has put the devil population in serious danger of extinction (McCallum et al., 2009).

Historical Aspects
Devil facial tumor disease was first recognized in a photograph taken by a wildlife photographer in the northeast region of Tasmania in 1996. Although unrecognized by biologists at the time, this represented the first observation of DFTD. Subsequently, a number of other wildlife biologists encountered devils with similar facial lesions, and DFTD was recognized as a notifiable disease in September 2006 under the Animal Health Act. Based on these initial observations and subsequent genomic profiling experiments, DFTD is believed to have arisen on the island of Tasmania within the past 25 years (Murchison et al., 2011). Only recently, parallel efforts in genomics, veterinary medicine, and conservation biology has yielded a thorough understanding of the disease’s properties and its potential detrimental impact on the devil population has now been fully realized.

Pathological and Clinical Features
Tasmanian devils are the largest surviving carnivorous marsupials and are endemic to Tasmania. DFTD is characterized by primary tumors that often exceed 3 cm in size, and are usually found on the face, neck and inside the mouth of the animal. The tumors are ulcerating soft tissue neoplasms with pleomorphic round or spindle-shaped cells on microscopy (Hawkins et al., 2006; Loh et al., 2006). The cells themselves are large with a central nucleus and completely unremarkable ultrastructure (Loh et al., 2006). The primary frequently metastasizes to regional lymph nodes and viscera. Biting is the most likely route of transmission for DFTD (HAMEDE, MCCALLUM, & JONES, 2008; Pyecroft et al., 2007). Tasmanian devils, especially adults, are known to bite each other on the head often, usually during feeding interactions (HAMEDE et al., 2008). DFTD appears to have an incubation period of 6 months but once gross manifestations appear, most devils do not survive beyond a further 6 months (Lachish, Jones, & McCallum, 2007; Lachish, McCallum, & Jones, 2009).

Tasmanian Devil Genome Sequencing Projects
The advent and application of next-generation sequencing technologies has significantly advanced our understanding of the Tasmanian devil and its clonally transmissible tumor. There have been two projects till date that have established a Tasmanian devil reference genome and sequenced the DFTD genome. In the most recent project, genomic DNA from a normal fibroblast cell line derived from a 5-year old female Tasmanian devil was sequenced using Illumina paired end sequencing with a yield of 2.87×109 pairs of 100 base pair short-insert reads. This process was complemented by the generation of mate pair libraries of 3-10 kilo base pair long-insert reads that were circularized. The reads were then clustered independently and assembled into contigs and supercontigs or scaffolds for the final step of de novo whole genome assembly. Flow cytometry was used to sort each of the seven devil chromosomes from the fibroblast cell line. Copies of each sorted chromosome were amplified, sequenced, and aligned with the supercontigs to assign the supercontigs to chromosomes. The order of supercontigs on each chromosome was determined using spectral karyotyping that compared conservation and homology with the opossum genome. The Ensembl pipeline was used to identify and annotate devil genes (Curwen et al., 2004). Small RNA reads that had been generated previously were aligned to the assembled genome to annotate microRNAs (miRNAs) (Murchison et al., 2010).

The earlier devil genome project sequenced two individuals – a male from the northwest and a female from southeast of Tasmania (Miller et al., 2011). The project leveraged four data types: Roche/454 Life Sciences GS FLX Titanium chemistry paired reads, Titanium unpaired reads, 454 XL+ chemistry unpaired reads and Illumina Genome Analyzer IIx reads (which was also the platform used by (Murchison et al., 2011). Interestingly, this project found that the two sequenced individuals had about a million single nucleotide polymorphisms – that is, genomic positions where the two individuals had distinct nucleotides. This variation is only about a quarter of that observed between genetically distant human genomes, for instance, although the two genomes (human and Tasmanian devil) are of comparable size. This suggests that overall genetic variation in the devil population is low. By combining next-generation sequencing of whole nuclear genomes, characterizing 14 mitochondrial genomes from current and historic specimens, and SNP genotyping 175 additional devils across Tasmania, this project helped establish a resource for future conservation work based on captive breeding.

Origin and Evolution of DFTD
Chromosome painting and copy number analysis have also indicated that the founder Tasmanian devil was female. The DFTD cell lines appear to carry two copies of the X chromosome that have the same number of single-base substitution variants as the normal female devil X chromosome, and twice the number of variants as the normal male devil X chromosome suggesting that they did not arise from a X duplication event but are in fact descendants of the original homologous X chromosomes of a female founder. Tasmanian devils, like humans, have X and Y chromosomes, and the males are heterogametic.





The same study also shed light on the phylogeography of DFTD, that is, evolution of the tumor clones across Tasmania, by genotyping over a hundred tumors sampled from different parts of the island between 2004 and 2010 and also by characterizing tumor mitochondrial genomes (Figures 2 and 3). There was clear evidence of a founder effect for DFTD on the Forestier Peninsula, a part of Tasmania isolated from the rest of the island where a single clone entered the population and gave rise to subclones that are unique to the area. Among these Forestier Peninsula subclones, the authors also detected one subclone that has, as of 2010, replaced other subclones by a selective sweep. Finally, even within some individual hosts, DFTD was found to exhibit two distinct subclones that were the result of inoculations or cell transfers from bite injuries that were temporally separated indicating lack of protection from prior DFTD inoculation, which has important implications for immunotherapy for DFTD.

Establishment of Clonality
Pearse & Swift, in 2006, were the first to propose the clonal origins of DFTD. By studying 11 DFTD tumors collected over a year across eastern Tasmania, the authors found that the DFTD karyotype had 13 chromosomes as against the 14 diploid chromosomes present in normal devil cells. Strikingly, an identical pattern was observed in all the tumors – five chromosomes were missing and four unidentifiable chromosomes were gained by each DFTD sample. They also shared other complex rearrangements. Early studies of the MHC locus and four microsatellite loci yielded further evidence that the tumors were identical to each other and genetically distinct from their hosts (Hannah V Siddle et al., 2007). Most recently, genotyping DFTD tumors, their host animals, and unaffected Tasmanian devils across several microsatellite loci and sequencing at the mitochondrial locus control region has provided definitive confirmation of the genetic consistency between the tumors and their differences compared to the host and unaffected animals, establishing DFTD as a monophyletic, clonally transmissible cancer (Murchison et al., 2011).

Murchison et al., 2011, sequenced cell lines derived from DFTD tumors of two Tasmanian devils from the north and the south-east of Tasmania. They aligned the cell sequences to the reference genome to identify single-base substitutions and indels. They then compared the two DFTD genomes to each other, and with the reference genome and the genome of another normal male devil that was also sequenced to assess variation. By looking at variation that was common to each of these comparisons, they identified germline substitution variants and indels that were probably present in the animal that first gave rise to the DFTD clonal lineage (the “founder” Tasmanian devil) as well as somatic variants and indels acquired by the tumor since then (Figure 2). Next, the authors identified approximately 17,000 single-base substitutions that were present in either one of the two DFTD lineages, but not in the other. These substitutions were almost entirely absent in 110 other normal Tasmanian devils that were subsequently genotyped. Furthermore, they had high nonsynonymous to synonymous ratios suggesting strongly that they were somatic mutations picked up by the tumor sub-clones (the two sequenced DFTD cell lines) since divergence from a common ancestral clone. The comparisons made by Murchison et al., 2011 demonstrated that it was unlikely that the true number of somatic mutations in DFTD exceeded 17,000. This number is more than the average of 5,000 somatic mutations carried by most human cancers, but less than that of human cancers caused by known carcinogenic exposures and DNA mismatch repair defects (lung cancer and melanoma). The transversion profile of these somatic mutations also supports the presence of an underlying DNA mismatch repair defect.

Genomic Stability
Cytogenetic analysis of one of the DFTD cell lines in the same study revealed heterozygous deletions on chromosomes 1, 2 and 3 and trisomy 5p, few detectable hemizygous deletions, and no major detectable amplifications. These cytogenetic analyses were also undertaken using the same spectral karyotyping or chromosome painting method that had previously been used to align supercontigs on normal chromosomes to establish the reference sequence of the Tasmanian devil by comparison with the opossum genome. Chromosome painting is a method that identifies gross homologies between two sets of chromosomes, which in the present case were the normal and DFTD chromosomes. Another chromosome painting study across several DFTD strains had additionally detected rearrangements on chromosomes 4 and X of the DFTD genome besides confirming the rearrangements on chromosome 1 (Deakin et al., 2012). These results collectively suggest that DFTD has surprising genomic stability compared to most human solid tumors. It is believed that the rearrangement on chromosome 1 was essential to the establishment of clonality and now helps continue tumor propagation, while further rearrangements are minimized by DFTD to maintain its current tumorigenicity.

Telomere Length and its Control
The order Dasyuromorphia, that includes Tasmanian devils, is the only known order in the animal kingdom to display extreme differences in telomere length between homologous chromosomes (Bender et al., 2012). Chromothripsis or the occurrence of tens to hundreds of chromosomal rearrangements in a single catastrophic cellular crisis is a phenomenon now believed to initiate 2-3% of all cancers (Stephens et al., 2011). Breakage-fusion-bridge (BFB) cycles triggered by telomere loss that is repaired by microhomology-mediated or nonhomologous end-joining mechanisms may underlie chromothripsis. It has been speculated that the chromosomes with the shortest telomeres in the Tasmanian devil are predisposed to BFB events that lead to the original DFTD tumor (Deakin et al., 2012). However, so far there is no conclusive evidence to support such speculation. In contrast to the telomeres of healthy devil cells, DFTD cell telomeres are homogeneously short and demonstrate the high telomerase activity that is characteristic of several highly proliferative and immortal cell lines including many human cancers. High telomerase activity prevents DFTD cells from entering replicative senescence by lengthening the telomeres when necessary. It is driven by a striking up-regulation of the TERT gene that encodes the catalytic subunit of telomerase (Ujvari et al., 2012). DFTD cells also overexpress TINF2, a gene coding for the TRF1-interacting nuclear factor 2, a protein that bridges the components of the shelterin complex. Shelterin is the key negative regulator of telomere length (Palm & De Lange, 2008). This suggests that the short, uniform DFTD telomeres are the result of a finely maintained balance between telomerase and shelterin activity and help maintain DFTD genomic stability.

The DFTD Transcriptome
Deep sequencing of the transcriptome of DFTD and diverse normal tissues from these animals has provided remarkable insights into the origins of this cancer from Schwann cells of the peripheral nervous system (Murchison et al., 2010). Of the 20 most significantly overexpressed genes in DFTD, 9 belonged to the myelination pathway. Moreover, 8 of these 20 genes are known to play a role in Schwann cell differentiation. Myelin is the insulating membrane that covers the axons of neurons and is produced in the peripheral nervous system by Schwann cells. Another gene found to have highly variable expression in DFTD was POMC that codes for proopiomelanocortin, the protein cleaved to form adrenocorticotropic hormone (ACTH). ACTH stimulates cortisol release, which was also found to be elevated in diseased Tasmanian devils raising the intriguing prospect of genetic perturbations in the tumor actually controlling aspects of the animal’s behavior such as aggression which indirectly may influence biting habits. Profiling of microRNAs in the same study demonstrated up-regulation of miRNAs previously implicated in control of tumor invasiveness, and miR-222 that may influence DFTD immune evasion (Ueda et al., 2009), and down-regulation of miRNAs with tumor suppressor activity.

DFTD is evolving on the epigenetic front as well, and samples collected over the last few years are characterized by progressive loss of methylation. This marked hypomethylation is being driven by active demethylation by methyl-CpG binding domain proteins rather than by a lack of DNA methyltransferase I activity that maintains normal methylation (Ujvari et al., 2013).

One of the key translational outcomes of the transcriptomic profiling of DFTD has been the identification of a highly sensitive and specific diagnostic marker for this cancer in the form of the myelination pathway gene periaxin (PRX) that was shown to stain all primary and metastatic tumors in the original study and subsequently validated in a confirmatory analysis (Tovar et al., 2011). The availability of this immunohistochemical marker will allow additional investigation of uncommon non-facial DFTD primaries as well as metastatic tumors of uncertain origin.

DFTD Immunogenetics
A remarkably unique feature of DFTD is the ability of the tumor, which is an allograft in the recipient animal, to avoid immunologic rejection as other foreign allografts would normally be treated. This had led to much research on the strategies employed by DFTD to avoid immune detection in the host, resulting in a number of current hypothesis about how this may happen.

Major histocompatibility complex (MHC) genes are vital to differentiate between self and non-self, and therefore crucial to immunosurveillance and in the response to infectious agents. Low major histocompatibility complex (MHC) diversity is an inherent characteristic of the Tasmanian devil population. Recent analyses of historic and prehistoric devil tissue samples has shown no significant changes in the currently observed limited diversity of MHC class I alleles since at least the Mid-Holocene period (Morris, Austin, & Belov, 2013). It is convenient to hypothesize that this reduced MHC diversity leads to the recognition of DFTD cells as self by the genetically similar host devil cells. However, transplantation experiments have demonstrated the devils’ ability to mount a robust T-cell response and reject MHC-matched skin allografts (Kreiss et al., 2011). Further, the recent emergence of DFTD as against the ancient origins of the lack of MHC diversity suggests that the tumor evades immunosurveillance by mechanisms other than MHC polymorphism. Indeed, primary and cultured DFTD cells express very low levels of MHC class I molecules on their surface. This has been linked to down-regulation of the TAP1 and TAP2 genes responsible for transporting MHC class I molecules into the endoplasmic reticulum (ER) and decreased expression of β2-microglobulin (β2m) that stabilizes the class I heavy chain within the ER (Hannah V Siddle & Kaufman, 2012; H. V. Siddle et al., 2013). This down-regulation appears to be under epigenetic control since it can be reversed in vitro by treatment of DFTD cells with a histone deacetylase inhibitor and by interferon-gamma (IFN-γ). The IFN-γ mediated restoration of MHC class I activity is driven by the MHC class II transactivator that is also known to promote expression of MHC class I genes. Interestingly in rare cases, even in vitro, natural killer (NK) cells (CD3+ lymphocytes) were shown to approach the DFTD tumor and trigger β2m expression via IFN-γ (H. V. Siddle et al., 2013).

Apart from MHC class I, the MHC class II DA gene family also shows restricted allelic diversity hinting at the occurrence of a selective sweep in the devil population in the past possibly due to an infectious disease that fixed certain favorable alleles (Cheng, Sanderson, Jones, & Belov, 2012). Genes belonging to both MHC classes also exhibit significant copy number variation with one deletion that turns a class Ia gene into a pseudogene being particularly important since its geographic distribution mirrors the route of spread of DFTD (Cheng, Stuart, et al., 2012). The recent identification of Tasmanian devil orthologs of the Natural Killer Complex and Leukocyte Receptor Complex is helping characterize devil genes that encode receptors of the C-type lectin and immunoglobulin superfamilies and providing insights into DFTD immunobiology that will be critical to the possible development of a vaccine against this cancer in the future (Van der Kraan, Wong, Lo, Ujvari, & Belov, 2013).

Canine Transmissible Venereal Tumor (CTVT)
Canine Transmissible Venereal Tumor (CTVT, a.k.a. Sticker’s sarcoma) is a sexually transmitted clonally transmissible cancer that affects the genitalia of domestic dogs, wild canids, and wolves (Cockrill, 1979). The disease has a wide geographical distribution and is endemic in many populations. CTVT prognosis is good - the tumor is typically self-limited in healthy dogs and it is very sensitive to chemotherapy.

Historical Aspects
CTVT has been recognized since the early modern history of western veterinary medicine. The earliest description of a widespread canine genital is found in Delabere Blaine's 1810 veterinary medical text, On the Diseases of Horses and Dogs (Blaine, 1810; Figure 4). In that era, the disease was common in London and was noted as one of two cancers that affect dogs. In 1876, Russian veterinarian A.M. Novinsky described a transmissible canine cancer with clinical characteristics very similar to modern CTVT (Novinsky, 1876).

Due to its robust horizontal infectivity, CTVT was used as an animal model for cancer research (Murgia, 2006). However this use has largely been abandoned in favor of inbred rodent strains because their identical genetic makeup facilitates reproducible outcomes and tumor transfer from one individual to another.







Pathological and Clinical Features
CTVT is endemic in many parts of the world and has been documented on 6 continents (Murchison, 2008) (Figure 5). Modern cases are most common in feral canine populations in tropic and subtropic locales, but cases do occur in pet dogs (vonHoldt, 2006). The vast majority of cases are sexually transferred. CTVT initially manifests as small, firm nodules at the base of the penis in male dogs, or at the vaginal vestibulum in females (Murchison, 2008). Left untreated, tumors often grow rapidly and form large, ulcerated, exudative masses (Figure 6) (Murchison, 2008). Less common modes of transmission include sniffing, licking, and biting of tumor-affected tissues (Cohen, 1985; Das and Das, 2000). This can lead to primary tumor engraftment in unconventional anatomical locations, including the eyelid (Abbott, 1966), oral cavity (Rensburg, 1980; Bright, 1983), or nasal cavity (Parent, 1983). The disease usually remains localized to the primary site, but it does occasionally metastasize to lymph nodes, visceral organs, or the central nervous system (Murchison, 2008).

The tumor is typically self-limited in healthy dogs with intact immune defenses. CTVT initially grows quickly, but is then controlled and eventually eliminated by the host’s immune system of the course of several months. Immunosuppressed, very young, and very old dogs are less likely to clear the disease without treatment (Yang, 1989). Dogs that have cleared the tumor are not susceptible to reinoculation (Mizuno, 1989). This lengthy contagious interval and minimal overall health impact is thought to have significantly contributed to the tumor clone’s prolific survival and spread (vonHoldt, 2006).

In the past, dogs with severe CTVT were frequently euthanized for humanitarian reasons (Blaine, 1810). More recently, progress in cancer treatment has mitigated the morbidity and mortality associated with the disease. the tumor is extremely responsive to several classes of chemotherapeutic agents and the rate of treatment success is high (Brown, 1980). In metastatic or drug-resistant cases, adjuvent surgery or radiation therapy can decrease disease burden.

On histopathologic examination, CTVT cells exhibit a round to polyhedral shape with central nuclei, prominent nucleoli, and cytoplasmic vacuoles (Murchison, 2008). Immunohistological marker expression patterns suggest that the tumor may have arisen from a macrophage cell lineage (Murchison, 2008). This conjecture is supported by the fact that both tumors and macrophages can be infected by the parasite Leishmania infantum (Albanese, 2002). More definitive studies of the CTVT cell type of origin are complicated by the fact tumor cells have undergone thousands of years of evolution since establishment of the clone.

Establishment of Clonality
The earliest documented CTVT research showed that the tumor could be transmitted from one dog to another (Novinski, 1876; Sticker, 1906). According to this model, tumor cells from an affected dog are physically dislodged and transferred by contact, followed by implantation, replication, and establishment of cancer in the newly-infected canine. However, clonal cancer spread via this mechanism had not been previously documented. Furthermore, microscopy of CTVT tumor cells occasionally detected virus-like particles, which yielded an alternative hypothesis that the disease is transmitted by viruses (Ajello and Gimbo, 1965). Therefore clinical tumor engraftment studies and genetic experiments were designed to test the idea of clonal transmissibility.

Clinical engraftment studies sought to identify the agent of tumor spread. Transplantation of live cells was shown to be necessary and sufficient to transmit CTVT, while killed cells and cell-free filtrates do not transfer the disease (Cohen, 1985). This finding strongly suggested that live tumor cells are the transmissible oncogenic agent and contradicted the viral disease transfer hypothesis.

The CTVT Cancer Genome
Genetic studies of clonal transmissibility are based on the premise that cancer arose in a single, long-deceased individual and subsequently spread through the worldwide canine population. If CTVT originated and spread in this manner, all tumors should be nearly genetically identical to each other, but distinct from the inherited genetics of the host canine. Accordingly, comparisons of tumor versus host genetics potently tested for clonal transmissibility.

Two key genetic studies corroborated the idea of clonal transmissibility. First, analysis of the number and morphology of chromosomes showed that the karyotype differs markedly between CTVT tumor cells and normal cells from affected individuals. Strikingly, CTVT cases around the world all share an extraordinarily similar abnormal aneuploid karyotype - while normal dogs have 78 chromosomes, all CTVT tumors have 57-59 chromosomes (Murchison, 2008). Furthermore, minor geographic region-specific karyotypic differences were observed (Weber, 1965). While aneuploidy is common in cancer, it is vanishingly unlikely that the same complex aneuploid karyotype with consistent geographical differences would arise independently in tumors across the globe. It is more probable that all the tumors were derived from a single common ancestor and subsequently underwent divergent evolution to produce region-specific subclonal differences. Second, the genome of all tested tumor cells contains an insertion of a long interspersed nuclear element (LINE-1) near the oncogene c-myc (Katzir, 1985). It is conceivable that inheritance of the LINE-1 insertion is a genetic risk factor that predisposes canines to contracting CTVT. This possibility is excluded by further studies that have thus far detected the LINE-1 insertion in tumor cells from all cases, while non-tumor DNA from the same dogs does not contain the insertion (Katzir, 1987; Liao, 2003, Murgia, 2006). Accordingly, polymerase chain reaction detection of the LINE-1 insertion is used as a diagnostic biomarker for CTVT (Liao, 2003). While genetic insertions and deletions are common in cancer, it is unlikely that spontaneously arising tumors would all acquire the exact same insertion at precisely the same genomic location. Instead, it is hypothesized that the LINE-1 element was part of the genetic makeup of the founding CTVT case or that it was inserted in the tumor genome prior to the divergence of modern tumor subclones (Murgia, 2006).

More recently, higher resolution genetic techniques have allowed more sophisticated investigations of the clonal origins of CTVT. In a 2006 study, Murgia, et al. applied three tests of genes that differ between individuals to tumor and normal tissue from 40 affected canines. The authors assayed the highly polymorphic dog leukocyte antigen (DLA) genes (the canine equivalent of the Major Histiocyte Compatibility genes), microsatellite markers, and mitochondrial genes. All of these experiments supported the clonal transmissibility model by conclusively demonstrating that the tumors were genetically similar to each other, but clearly distinct from the affected individuals. Interestingly, the mitochondrial DNA sequencing revealed two separate subgroups with different copy numbers of the gene DQA1. This phenomenon is likely explained by the genesis of multiple subclonal CTVT lines sometime after the first tumor arose (Murgia, 2006).

Theories of Origin
The information richness in the tests of Murgia, et al. allowed investigation of CTVT’s origins by superimposing tumor data onto the results of other canine genetics studies. Microsatellite genotypes had previously been leveraged to examine the population structure and relatedness of canine groups including dog breeds and wolves (Parker, 2004). CTVT microsatellite genotypes were added to this analysis using a model-based clustering algorithm, which showed that the index CTVT tumor may have arisen in the a wolf (Murgia, 2006; Figure 7). More recent analyses favor origins in an ancient dog breed.



The age of the CTVT clone has been estimated several times using the microsatellite and mitochondrial genotypes from tumors and modern canines. This approach uses projected mutation rates to infer the most recent common ancestor of the tumors and canines. This method assumes that both groups have evolved independently since divergence. However, recent studies have shown that CTVT occasionally uptakes mitochondria from its host (Rebbeck, 2010). Accordingly, estimates that rely on mitochondrial evolution are confounded by these horizontal mitochondrial transfers. The most recent estimate postulates that the CTVT clone arose 2,000-20,000 years ago (Rebbeck, 2009). This makes the tumor the oldest known continually-propagated mammalian cell line.

In summary, CTVT tumors are believed to have derived from a single clonal ancestral tumor that subsequently divided into two closely-related sub-types. Their estimation of the age of this cell line would make it the oldest known mammalian somatic cell in continuous propagation. They suggest that CTVT can be thought of as a somatic mosaic parasite, with implications future studies of transmissible cancer. In the future, genetic sequencing technologies may yield improved models of the age and origin of CTVT. These strategies produce base-wise genotypes across the accessible genome, which would provide many more loci for analysis and avoid reliance on cross-contaminated mitochondrial sequences.

Transmitted Cancers in Humans
There are no reported human cases of clonally transmissible cancers that can independently survive and be transferred in the manner of CTVT and DFTD. However, instances of one human passing cancer cells to another through physical contact have been reported. The difference of these incidences from a true clonally transmissible cancer, however, is that the cases of human transfer were one-off events-- the actual tumor did not have properties allowing it to be continually transferred throughout the population, and therefore the tumor clone went extinct with the death of the host individuals.

Surgical Transplants
A number of case reports have described events where one person inadvertently transmitted cancer cells to another, usually in the context of a surgery or other medical procedure. For instance, it was reported that a cancer surgeon accidently was inoculated with his patient’s malignant sarcoma cells during the surgery when the surgeon inadvertently cut his hand (Gärtner, et al, 1996). The surgeon had the transplanted tumor cells removed, and remained cancer free two years later. While other cases have also been reported (Scanlon et al, 1965; Gugel and Sanders, 1986), overall the incidence of such events is exceedingly rare. Furthermore, these cases represent distinct events from those of CTVT and DFTD, as the tumor cells did not have endogenous properties making transmission part of their lifecycle, rather the transmission was the result of human intervention, and the clones died out along with their recipient hosts.

Vertical Transmission Between Mother and Fetus
The other known instances of human cancer cell transfer has occurred between pregnant mothers who had or developed cancer during pregnancy, and her unborn fetus (termed vertical transmission). Between 1866 and 2003, only fourteen cases of vertically transmitted cancers had been reported in the medical literature (Tolar and Neglia, 2003), although it is possible the actual rate is slightly higher due to unrecognized cases.

Instances of clonal fetal cancer cells being transferred back to the pregnant mother have not been reported, although three instances of rare choriocarcinoma transfer to the mother have been described (Buckell and Owen, 1954; Daamen 1968). Choriocarcinoma is a cancer derived from shared maternal and fetal tissue, and thus does not represent a true transfer of clonal cancer cells from the fetus to the mother.

Distinction from Virus-Associated Cancers
There are a number of virus-associated cancers in humans that are often confused with clonally transmissible cancers. The distinction relies in what is transmitted: in clonally transmissible cancers, the cancer cells themselves are transferred between individuals; whereas in virus-associated cancers, a virus is transferred between individuals and the virus independently causes two different cancers clones, one in each person affected. The most common forms of virus-associated cancer in humans are Human Papilloma Virus (HPV), which is known to cause cervical and penile squamous cell carcinoma, and Hepatitis C Virus, which can cause hepatocellular carcinoma. However, in all instances, no actual cancerous cells are transmitted between individuals, making these cancers distinct from clonally transmissible cancers.

Conclusions
The recent emergence of low-cost, high-throughput, genome sequencing technologies, combined with sophisticated comparative genomics pipelines have definitively established CTVT and DFTD as unique instances of clonally transmissible cancers. CTVT and DFTD share many features in common that in aggregate likely represent features allowing for clonal transmissibility, such as severe genomic rearrangement, bottleneck population structure, behavioral characteristics of the population, and immune evasion strategies of the tumor. Yet, they also have unique evolutionary histories, genomic and clinical features, and their implications to their respective species are distinct. The recent and rapid emergence of DFTD has placed the Tasmanian devil population in serious danger of extinction, while CTVT has been recognized for over one hundred years and has likely existed in the canine population for millennia.

Current research on these two diseases is focused on understanding both the underlying cancer genomics and the conservation-biology implications of these diseases to their respective species. The rapidly expanding body of research on these clonally transmissible cancers has generated some intriguing questions:

Unanswered Questions in the field of Clonally Transmissible Cancers
 * Do other unidentified clonally transmissible cancers exist?
 * Could clonally transmissible cancers develop in humans?
 * What genomic properties of canines and Tasmanian devils have allowed transmissible cancers to develop only in these species?
 * How do transmitted cancers evade immunosurveillance in their new hosts?
 * Will it be possible to develop a vaccine against clonally transmissible cancers?

Research on the underlying genomic properties and resultant clinical and immunologic properties of the tumor are of interest to researchers to gain a better understanding of human tumor immunology, and to attempt to develop vaccines against these tumors to protect devils and dogs. Furthermore, CTVT and DFTD may serve as complementary experimental models for human cancer researchers, in a parallel way to cancer cell lines or single-gene deletion rodent models.