MALARIA AND SOME POLYOMAVIRUSES (SV40, BK, JC, AND MERKEL CELL VIRUSES) VOLUME 104 This publication represents the views and expert opinions of an IARC Working Group on the Evaluation of Carcinogenic Risks to Humans, which met in Lyon, 7-14 February 2012 LYON, FRANCE - 2014 IARC MONOGRAPHS ON THE EVALUATION OF CARCINOGENIC RISKS TO HUMANS INTRODUCTION TO POLYOMAVIRUSES 1. Discovery BK polyomavirus (BKV) and JC polyomavirus (JCV), were eventually found to chronically In 1953, Ludwik Gross reported that a filter- infect the great majority of humans worldwide able infectious agent could cause salivary cancer (reviewed in Abend et al., 2009; Maginnis & in laboratory mice (Gross, 1953; Stewart et al., Atwood, 2009). The apparent ubiquity of BKV 1957). The cancer-causing agent was found to and JCV makes it difficult to correlate seropos- be a non-enveloped DNA virus that was named itivity for BKV- or JCV-specific antibodies with murine polyomavirus (from the Greek roots specific disease states, such as cancer. poly-, which means “many,” and -oma, which Reports in the past four years have revealed the means “tumours”), for its ability to cause existence of seven more human polyomaviruses. tumours in multiple tissues in experimentally Perhaps the most intriguing of the new species, infected rodents (reviewed in Sweet & Hilleman, named Merkel cell polyomavirus (MCV), was 1960). The discovery spurred renewed interest in discovered through a directed genomic search the idea that viral infections might be a major of an unusual form of skin cancer, Merkel cell cause of cancer in humans. carcinoma (MCC) (Feng et al., 2008). Another By the late 1950s, various investigators had new polyomavirus, trichodysplasia spinulo- succeeded in developing cell culture systems for sa-associated polyomavirus (TSV), was isolated analysing the transforming activities of murine from a rare hyperplastic (but non-neoplastic) polyomavirus in vitro. This work set the stage trichodysplasia spinulosa skin tumour that can for the discovery of the primate polyomavirus occur in transplant patients (van der Meijden simian virus 40 (SV40), which was identified as et al., 2010). Little is currently known about a contaminant in primary cultures of monkey the cancer-causing potential of TSV. Five other kidney cells used to produce vaccines against recently discovered human polyomaviruses – poliovirus (Sweet & Hilleman, 1960; Eddy et al., named WU polyomavirus (WUV), KI polyoma- 1962). Millions of individuals were exposed to virus (KIV), human polyomavirus 6 (HPyV6), infectious SV40 virions present in contaminated HPyV7, and HPyV9 – have not so far been clearly polio vaccines administered between 1955 and associated with human disease states (Allander 1963 (reviewed in Shah & Nathanson, 1976; et al., 2007; Gaynor et al., 2007; Schowalter et al., Dang-Tan et al., 2004). 2010; Sauvage et al., 2011; Scuda et al., 2011). The In 1971, the first two naturally human-tropic remaining Monographs in this Volume will focus polyomaviruses were discovered in specimens on SV40, BKV, JCV, and MCV. from immunocompromised patients (Gardner et al., 1971; Padgett et al., 1971). The two viruses, 121 IARC MONOGRAPHS – 104 Fig. 1.1 Phylogenetic tree of known polyomavirus species Human-derived species are in bold type. Prepared by the Working Group. 122 Introduction to polyomaviruses 2. Taxonomy and phylogeny encompass all the currently known human- and primate-tropic species, while the third genus The exterior structure of the non-envel- includes species thought to be tropic only for oped capsids of members of the viral family birds. The four primate polyomavirus species Polyomaviridae is strikingly similar to the that have been proposed to be potentially impli- capsids of a different family of non-enveloped cated in human cancer (SV40, BKV, JCV, and viruses called the Papillomaviridae. Both families MCV) are all members of the proposed genus carry circular, double-stranded DNA (dsDNA) Orthopolyomavirus. genomes. Based on these considerations, the Although SV40, BKV, and JCV are separate two virus groups were originally classified viral species, they are very closely related to one within a single family, the Papovaviridae (a sigla another, sharing about 70–75% identity at the condensation of PApillomavirus POlyomavirus nucleotide level across the entire genome. MCV is simian VAcuolating). However, sequencing of only distantly related to the SV40 cluster, sharing the genomes of various polyomaviruses and less than 35% nucleotide identity across the papillomaviruses revealed essentially no detect- entire genome (Fig. 1.1). Phylogenetic analyses able sequence homology between the two virus of BKV and JCV isolates from different popula- groups and furthermore showed that the two tions worldwide have shown that the two virus groups had dramatically different genetic organ- species exhibit a geographical pattern of genetic ization. Since the sequencing results indicate that drift that closely resembles proposed patterns of polyomaviruses and papillomaviruses probably prehistoric human migration (reviewed in Yogo never shared a common viral ancestor, they were et al., 2004, 2009). MCV is somewhat more closely officially split into separate viral families. related to murine polyomavirus, with the two Interestingly, several recent reports suggest species sharing about 50% nucleotide identity that polyomaviruses and papillomaviruses may across the complete viral genome. Phylogenetic occasionally recombine with one another to trees based on the nucleotide or amino acid produce viable chimeric viruses with mixed sequences of individual viral gene products give genetic characteristics of both families (Woolford similar results, suggesting that the members et al., 2007). Another recent report documenting of the SV40 cluster diverged from one another a novel viral species that infects eels suggests relatively recently, while the split between MCV that polyomaviruses might also recombine with and members of the SV40 cluster occurred in the members of other families of DNA viruses to much more distant past. produce viable chimeric progeny (Mizutani et al., 2011). The taxonomy of these apparently chimeric viral species is currently undefined. 3. Structure of the virion A recent proposal that is currently being reviewed by the International Committee on The exterior surface of the polyomavirus Taxonomy of Viruses suggests that the various virion is a naked protein capsid composed entirely members of the family Polyomaviridae be of a single virally encoded protein called capsid grouped into separate genera (Johne et al., 2011). viral protein 1 (VP1) (Stehle et al., 1996) (see If approved, this would result in the division of cover photograph). The virion contains a total the current sole genus Polyomavirus into three of 72 pentameric VP1 capsomers arranged on a genera: Orthopolyomavirus, Wukipolyomavirus, T = 7d icosahedral lattice (Yan et al., 1996). VP1 and Avipolyomavirus. The first two genera subunits are folded into a classic eight-stranded β 123 IARC MONOGRAPHS – 104 jellyroll fold that is shared among a wide variety serological analyses of polyomavirus epidemi- of viral capsid proteins and the cellular protein ology (Hamilton et al., 2000; Carter et al., 2003; nucleoplasmin (Stehle et al., 1994; Carrillo-Tripp de Sanjose et al., 2003; Stolt et al., 2003). Purified et al., 2009). The jellyroll, which forms the core polyomavirus VLPs can be immunogenic when of the pentameric VP1 interface, is arranged administered to laboratory animals, raising perpendicular to the spherical virion, such that serum antibody responses capable of neutral- the assembled virion has a distinctive knobby izing native virions in vitro (Goldmann et al., appearance. An additional capsid protein, VP2, 1999; Velupillai et al., 2006; Randhawa et al., as well as its N-truncated isoform VP3, associate 2009). VLPs can even be humorally immuno- with the central lumen of each VP1 capsomer genic in mice in the absence of functional T-cell (Chen et al., 1998). immunity (Vlastos et al., 2003), suggesting that In all extensively studied examples, it appears VLP-based immunogens might be effective in that polyomavirus infectious entry requires inter- immunocompromised human subjects who actions between VP1 and one or more cellular may be at greater risk of polyomavirus-associ- glycans that carry at least one sialic acid residue. ated disease. This suggests that polyomaviruses In several distantly related polyomaviruses, the might be a suitable target for the development binding site for the sialylated glycan receptor is of VLP-based preventive vaccines, similar to the formed by a pocket along the outer rim of the highly successful VLP-based vaccines against apical portion of the VP1 capsomer knob (Stehle hepatitis B virus and human papillomaviruses and Harrison, 1996; Neu et al., 2008; Neu et al., (HPV). 2010). The loops that form the receptor binding pocket vary extensively between polyomavirus species, and even among closely related polyoma- 4. Genomic organization, gene virus subspecies (Luo et al., 2012). This variation products, and replication may reflect selective pressure to evade recog- nition by antibodies that occlude the receptor The circular ~5 kb dsDNA polyomavirus binding site. genome is roughly divided into two oppositely The floor of the canyons between the oriented transcriptional units separated by a capsomer knobs is formed primarily by N- and non-coding control region (NCCR) (reviewed C-terminal arms of VP1. The arms