Papillomaviral skin diseases of humans, dogs, cats and horses – a comparative review. Part 2: Pre-neoplastic and neoplastic diseases

Abstract

Papillomaviruses (PVs) are well recognized to cause pre-neoplastic and neoplastic diseases in humans. Similarly, there is increasing evidence that PVs play a significant role in the development of pre-neoplastic and neoplastic diseases of the haired skin of dogs and cats and the mucosa of horses. As the mechanisms by which PVs cause neoplasia are well studied in humans, it is valuable to compare the PV-induced neoplasms of humans with similar PV-associated neoplasms in the companion animal species. In the second part of this comparative review, the pre-neoplastic and neoplastic diseases thought to be caused by PVs in humans, dogs, cats, and horses are described. This includes PV-induced cutaneous plaques, cutaneous squamous cell carcinomas (SCCs) and mucosal SCCs within the four species. The review concludes with a discussion about the potential use of vaccines to prevent PV-induced diseases of dogs, cats, and horses.

Introduction

In the second part of this comparative review, the pre-neoplastic and neoplastic diseases caused by papillomaviruses (PVs) in humans, dogs, cats, and horses are described. Unlike the hyperplastic diseases described in part 1, these diseases rarely resolve spontaneously. Diseases included in part 2 include cutaneous plaques, cutaneous squamous cell carcinomas (SCCs), and mucosal SCCs. Although cross-species infection of horses and cats by bovine PV types has been shown to cause equine and feline sarcoids (Chambers et al., 2003; Munday et al., 2015b), neither cross-species PV infections nor PV-induced mesenchymal neoplasia are recognized in people and these will not be discussed in this comparative review. The review ends with a brief discussion on the use of vaccines to prevent diseases caused by PVs in humans and the companion animal species.

Papillomaviral cutaneous plaques

Humans

PV-induced cutaneous plaques are rare in people. They are caused by numerous closely related PV types within the Betapapillomavirus genus (Kremsdorf et al., 1984). Infection by these PVs are ubiquitous, with people being infected in the first few days of life (Antonsson et al., 2000; Antonsson et al., 2003). In an immune competent person, replication of PVs is inhibited by the immune system and the PV causes minimal epithelial hyperplasia and remains asymptomatic. However, if the immune system is unable to mount a normal immune response, increased PV replication results in the development of visible lesions. While plaques are well-recognised to develop in patients that receive immunosuppressive therapy after organ transplantation, the pathogenesis of human PV-induced plaques is best illustrated by the rare genetic disorder epidermodysplasia verruciformis (EV) (Egawa and Doorbar, 2017; Orth, 2006). EV is due to inherited defects in the EVER1, EVER2 or CIB1 genes which code for proteins that contribute to keratinocyte intrinsic immunity (Beziat et al., 2021; Ramoz et al., 2002). As keratinocyte intrinsic immunity inhibits PV replication, loss of this immunity allows greater viral replication and the development of visible cutaneous plaques (Beziat et al., 2021). Interestingly, the impact of mutations within the EVER1, EVER2 and CIB1 genes appears to be specific to the immune response to PVs as people with EV do not show other signs of immunosuppression.

People with EV usually develop flat plaques in early childhood. They are most common on the trunk, neck and extremities and appear as hypo- or hyperpigmented plaques that can coalesce (Cubie, 2013). Larger plaques can appear verrucous, especially if they are exposed to UV light. Histologically, the plaques appear as thickened epidermis covered by increased keratin. PV-induced cellular changes are usually visible in a high proportion of keratinocytes within the plaques (Cubie, 2013).

If possible, restoration of normal immune function will result in lesion resolution. If restoration is not possible (for example in people with EV), plaques are expected to slowly become larger and more numerous throughout life (Sterling, 2016). Progression to SCC occurs in 20–30% of EV patients (Sterling, 2016). However, neoplastic transformation only occurs in plaques that are exposed to UV light (Sterling, 2016), suggesting the Betapapillomaviruses cannot cause neoplasia without a co-factor (McBride, 2022). Although there is some evidence from individual cases that treatment using oral retinol, oral interferon, or topical imiquimod may be beneficial, results have been inconsistent and no treatment is curative if normal immune function cannot be restored (Sterling, 2016).

Dogs

Viral cutaneous plaques (also referred to as pigmented plaques) are rare lesions in dogs. They are caused by a number of closely-related Chipapillomavirus types with CPV4 being the CPV type most often identified in these lesions (Lange et al., 2009a; Lange et al., 2009b; Lange et al., 2012; Luff et al., 2015; Luff et al., 2012a; Luff et al., 2012b; Tobler et al., 2006; Tobler et al., 2008; Yuan et al., 2012; Zhou et al., 2014). Cases are sporadic and contact with other affected dogs has not been reported. The epidemiology of infection by the Chipapillomaviruses has not been studied. However, it is possible that dogs could be frequently asymptomatically infected by Chipapillomaviruses and, as with human cutaneous plaques, the development of a canine cutaneous plaque could be primarily due to an inability to inhibit replication by a PV type that commonly infects the skin. This is supported by rare reports of plaques developing in dogs with possible immunosuppressive conditions (hyperadrenocorticism, putative hypothyroidism, and hypoglobulinemia) and receiving immunosuppressive medication (Stokking et al., 2004; Tobler et al., 2008). Furthermore, certain breeds, such as Vizslas and Pugs, are predisposed to developing plaques (Hansen et al., 2018; Nagata et al., 1995; Tobler et al., 2008) suggesting that it is possible that these dogs inherited an EV-like defect in keratinocyte intrinsic immunity. However, further research is required to determine if there truly is an underlying genetic component in these dogs, and if it parallels that of human EV patients.

Viral plaques have been reported to develop in dogs as young as 2-years-old but are more common in middle-aged dogs (Munday et al., 2022b; Nagata et al., 1995). They are typically dark, multiple, and 1–10mm in diameter. As in humans, they tend to be flat early in the clinical course then become more exophytic as the lesions progress (Munday et al., 2022b). They are most common on the ventrum and medial aspects of the limbs (Gross et al., 2005). In rare cases, lesions can coalesce and involve a large proportion of the body (Hansen et al., 2018; Munday et al., 2022b; Fig. 1). Most canine viral plaques do not significantly impact life and spontaneous regression is possible. However, extensive plaques can cause pruritis or pain (Knight et al., 2016; Munday et al., 2022b). Histology reveals moderately thickened epidermis covered by increased keratin. Lesions typically have a scalloped appearance with prominent epidermal and dermal pigment and keratohyalin granules (Fig. 2). While viral replication has been detected within plaques (Lange et al., 2013a), PV-induced cellular changes are rarely histologically visible (Munday et al., 2017c). This is in contrast to both human and feline plaques that often contain prominent evidence of PV replication.

Figure 1.

Papillomaviral cutaneous plaques, dog. The heavily pigmented plaques on this Chinese crested dog are predominantly confined to the skin of the limbs and ventrum. Larger plaques are mildly exophytic and plaques have coalesced in some areas (Image courtesy of Dr Anne Quain).

Figure 2.

Papillomaviral cutaneous plaques, dog. The plaque consists of heavily pigmented, hyperplastic epidermis covered by increased quantities of keratin. Large quantities of pigment and marked clumping of keratohyalin granules are visible within the thickened epidermis. Neither koilocytosis nor the presence of cells with expanded, blue-grey cytoplasm are visible within these lesions. PCR was used to amplify canine papillomavirus type 18 DNA from this plaque. Scale bar = 25μm. Haematoxylin and eosin.

Canine viral plaques have rarely been reported to progress to SCCs. Some evidence suggests neoplastic transformation is more likely in plaques caused by specific CPV types. Of these types, CPV16 is most frequently associated with neoplastic transformation, although other CPV types have also been identified in plaques that have undergone neoplastic transformation (Alves et al., 2020; Chang et al., 2020; Lange et al., 2009a; Luff et al., 2019; Luff et al., 2016; Munday et al., 2011d; Tobler et al., 2006). Similar to the high-risk human Alphapapillomaviruses, CPV16 has been shown to integrate into the host genome within the plaques, supporting the possibility that this may be a higher-risk CPV type (Luff et al., 2019). Whether or not sun exposure influences neoplastic progression of a canine plaque is currently unknown.

Treatment of dogs with pigmented plaques is often not required. However, if plaques become large and cause irritation to the dog, surgical excision of plaques is possible, although additional plaques usually develop. Other potential treatment options that act by destroying the infected epithelium include laser ablation and tigilanol tiglate gel (Hansen et al., 2018; Knight et al., 2016). Medical treatments that have been tried include oral retinol, topical imiquimod, azithromycin, and interferon alfa-2b although, as in humans with EV, there is no definitive evidence that these therapies are effective. Indeed, if canine plaques are caused by a defect in keratinocyte intrinsic immunity, medical treatments would appear unlikely to be able to restore immunity sufficiently to be curative.

Cats

PV-induced cutaneous plaques are rare in cats, typically developing in middle-aged or older animals. Depending on the presence of dysplasia within the epidermal cells these can be classified as feline viral plaques or Bowenoid in situ carcinomas (BISCs). Plaques are most often caused by infection with FcaPV2 (a Dyothetapapillomavirus) although other Taupapillomavirus FcaPV types can also cause these lesions (Demos et al., 2019; Kok et al., 2019; Lange et al., 2009c; Munday et al., 2017a; Munday and Thomson, 2021; Munday et al., 2018; Nespeca et al., 2006; Vascellari et al., 2019). While less is known about the other FcaPV types, infection by FcaPV2 is ubiquitous in cats, with kittens being infected during, or shortly after, birth (Geisseler et al., 2016; Thomson et al., 2015). As in humans, this suggests plaque development is primarily due to an immune dysfunction that allows increased replication by a normally asymptomatic PV type. Sphinx or Devon Rex cats develop plaques more frequently and at a younger age suggesting the likelihood of an inherited defect in the keratinocyte immune system (Munday et al., 2016a; Ravens et al., 2013). The immune defect that allows plaque development appears to be specific for cutaneous PV infection and other signs of immunosuppression have not been reported in affected cats.

Feline viral plaques present as multiple pigmented or non-pigmented non-painful, non-pruritic, slightly-raised lesions up to 2cm in diameter that are most common on the face, head, and neck (Wilhelm et al., 2006; Fig. 3). Histology reveals mild to moderate hyperplasia of cells within the deeper layers of the epidermis. Unlike human and canine plaques, the epidermal cells within feline plaques are often dysplastic. PV-induced cellular changes can be prominent in smaller plaques (Fig. 4), although these changes become less common in larger, more dysplastic, lesions (Gill et al., 2008; Wilhelm et al., 2006). The frequent dysplasia within feline viral plaques suggests the FcaPV types that cause plaques cause greater dysregulation of epidermal cells than either the human Betapapillomaviruses or the canine Chipapillomaviruses. Additionally, feline viral plaques consistently contain marked accumulation of p16^CDKN2A protein (p16) due to PV-induced degradation of the retinoblastoma protein (pRb) (Munday and Aberdein, 2012; Munday et al., 2011a). Such PV-induced changes in cell regulation are observed in human SCCs caused by the high-risk Alphapapillomaviruses, but are not seen in the human SCCs associated with Betapapillomavirus infection (Küsters-Vandevelde et al., 2009).

Figure 3.

Papillomaviral cutaneous plaques, cat. Multifocal pigmented and non-pigmented plaques are visible (Image courtesy of Dr Linda Vogelnest).

Figure 4.

Papillomaviral cutaneous plaque, cat. The epidermis is thickened and contains increased pigment. There is crowding of cells within the basilar layers. Papillomavirus-induced cellular changes are prominent including cells with expanded clear cytoplasm containing perinuclear bodies (arrows). PCR amplified Felis catus papillomavirus type 3 DNA from this plaque. Scale bar = 40μm. Haematoxylin and eosin.

Feline viral plaques can spontaneously regress (Munday JS, personal observation), presumably due to resolution of the immune dysfunction. However, regression appears to be rare and more often plaques persist and can become large and ulcerative, resulting in significant morbidity. Viral plaques can also progress to SCC, especially in Devon Rex and Sphinx cats in which the progression tends to be rapid and result in SCCs with high metastatic potential (Munday et al., 2016a; Ravens et al., 2013). Whether or not sunlight plays a significant role in neoplastic transformation is uncertain.

As in humans and dogs, treatment of feline plaques is local rather than systemic and surgical excision, imiquimod cream, and cryotherapy are all suggested treatment options (Gill et al., 2008). However, considering the likelihood of an underlying inability to control viral replication, development of additional lesions should be expected.

Horses

Horses are currently not recognized to develop widespread PV-induced plaques due to loss of immune control of a normally asymptomatic PV type. However, horses do frequently develop aural plaques on the inner surface of the pinnae (Torres et al., 2010). These plaques are thought to be caused by EcPV types in both the Dyoiotapapillomavirus and Dyorhopapillomavirus genera, different from the EcPV types that cause cutaneous and genital warts (Gorino et al., 2013; Lange et al., 2011a; Mira et al., 2018; Taniwaki et al., 2013; Torres and Koch, 2013). Equine aural plaques may be contagious, with the causative PV types hypothesised to be spread between horses by an insect vector (Fairley et al., 2014). Aural plaques do not spontaneously resolve and there is a single report of progression to SCC (Peters-Kennedy et al., 2020).

Histologically, aural plaques comprise moderately thickened epidermis covered by marked hyperkeratosis. Epidermal folding does not occur so that the thickened epidermis remains plaque-like rather than forming an exophytic wart (Scott and Miller, 2011). Aural plaques contain reduced melanin compared to surrounding skin. Clumped keratohyalin granules and koilocytes can be visible although PV-induced cellular changes are subtle within the plaques. As aural plaques are typically only a cosmetic concern, they are rarely treated. Surgical debulking can be successful although lesion recurrence is common (Peters-Kennedy et al., 2020; Scott and Miller, 2011). Treatment using repeated topical application of imiquimod has been reported to be curative although this treatment usually causes significant pain (Torres et al., 2010; Zakia et al., 2016).

Squamous cell carcinomas of the skin

Humans

As previously discussed in this review, HPVs can cause cutaneous SCCs due to progression of a viral cutaneous plaque or, less commonly, due to progression of a recalcitrant wart (Egawa and Doorbar, 2017; Sterling, 2016). However, both viral plaques and recalcitrant warts develop due to underlying immune defects and it is less clear whether or not PVs are significant causes of SCCs in people with intact immune systems and no pre-existing PV lesions (Smola, 2020). As Betapapillomaviruses asymptomatically infect the skin of all people (Antonsson et al., 2000), it is possible that they could influence regulatory pathways within the infected epidermal cells. For example, PV infection of a keratinocyte could reduce the ability of the cell to undergo apoptosis in response to UV light-induced DNA damage (Akgul et al., 2006). This reduced ability of a cell to undergo apoptosis could allow the accumulation of DNA mutations and subsequent neoplastic transformation. However, evaluating the role of Betapapillomaviuses in skin SCCs is difficult as PV DNA is detectible in both SCCs and normal skin samples (Forslund et al., 2003). Additionally, unlike mucosal SCCs caused by the high-risk Alphapapillomaviruses, cutaneous SCCs associated with Betapapillomaviruses do not consistently contain alterations in p53, pRb, p16 or any other marker that could indicate a PV etiology of the cancer (Küsters-Vandevelde et al., 2009; Smola, 2020). Intriguingly, it has recently been hypothesized that the presence of Betapapillomaviruses could actually protect against skin SCCs by promoting immune surveillance of the skin surface (Strickley et al., 2019). While the high-risk Alphapapillomaviruses are well established as causes of human cancers, these sexually-transmitted PVs have a strong mucosal trophism and are extremely rare causes of SCCs in haired skin (Amiraraghi et al., 2019).

Dogs

There are rare reports of canine cutaneous SCCs developing within viral cutaneous plaques or cutaneous warts (Luff et al., 2016; Thaiwong et al., 2018). Additionally, some studies of canine cutaneous SCCs have detected PV DNA in a proportion of neoplasms (Teifke et al., 1998; Waropastrakul et al., 2012; Zaugg et al., 2005). However, as in other species, asymptomatic infection of canine skin is common (Lange et al., 2011b) making the detection of PV DNA in a SCC difficult to interpret. Overall, there is currently little evidence that PVs are a significant cause of canine cutaneous SCCs.

Cats

Of the four species discussed in this review, there is the strongest evidence of a PV etiology of skin cancer in cats. While PV infections in cats have been associated with basal cell carcinomas and Merkel cell carcinomas (Ito et al., 2022; Munday et al., 2017b), the majority of studies have evaluated cutaneous SCCs. These studies suggest that feline cutaneous SCCs can be subdivided into those that are primarily caused by exposure to UV light and those that appear likely to be caused by PV infection (Munday et al., 2011b; Thomson et al., 2016). The proportion of SCCs associated with PV infection is variable depending on the area of the body. Within UV-protected areas of the body (haired pigmented skin) as many as 75% of SCCs may be caused by PV infection while only around 30% of SCCs from UV-exposed areas of the body (non-pigmented nasal planum, pinnae and eyelids) show evidence of a PV etiology (Munday et al., 2011b). Interestingly, the presence of these two distinct subtypes is similar to human oral SCCs which can be subdivided into those caused by tobacco and alcohol and those caused by infection by PVs. Feline cutaneous SCCs can develop as progression from a viral cutaneous plaque (Munday et al., 2016a; Ravens et al., 2013). However, it is currently uncertain if all PV-associated SCCs develop from plaques or whether some develop within normal skin without progressing through this precursor lesion.

Most studies report that FcaPV2 is the predominant type detectible in feline cutaneous SCCs (Altamura et al., 2018; Munday and Aberdein, 2012; Munday et al., 2009a; Munday et al., 2013; Munday et al., 2008; O’Neill et al., 2011; Thomson et al., 2016; Yamashita-Kawanishi et al., 2021a), although other FcaPV types have also been detected (Carrai et al., 2020; Munday et al., 2011b; Munday et al., 2018; Yamashita-Kawanishi et al., 2021b; Yamashita-Kawanishi et al., 2018). Histological evidence of PV infection is rarely detectible. However, SCCs contain FcaPV2 DNA more frequently, and have higher FcaPV2 loads, than normal skin (Munday et al., 2008; Thomson et al., 2016). In addition, a proportion of SCCs contain detectible FcaPV gene expression (Altamura et al., 2016; Thomson et al., 2016) that can be localized to the neoplastic cells within the SCCs (Hoggard et al., 2018). Additional evidence for a role of PVs in feline skin SCCs is derived from the observation that cats with PV-associated nasal planum SCCs survived longer than cats with nasal planum SCCs that did not contain evidence of PV infection (Munday et al., 2013). This supports the presence of two etiologically different subsets of SCCs and is consistent with human oral SCCs in which patients with PV-induced oral cancers survive longer than those with cancers caused by tobacco and alcohol (Lewis et al., 2010).

Evidence indicates that FcaPV2 causes neoplasia by the E7 protein degrading pRb and the FcaPV2 E6 protein interfering with normal p53 function (Altamura et al., 2016; Munday and Aberdein, 2012). The loss of pRb removes an important checkpoint preventing cell replication while impaired p53 function reduces the ability of the cell to recognize damaged DNA and undergo apoptosis. Loss of pRb also results in a marked increase in p16 within the cell that can be detected using immunohistochemistry (Munday and Aberdein, 2012) and PV-associated feline cutaneous SCCs consistently contain intense nuclear and cytoplasmic p16 immunostaining (Munday et al., 2013; Munday et al., 2011b; Fig. 5). Similarly, degradation of pRb and p53 are important mechanisms by which the high-risk human Alphapapillomaviruses cause cancer and resultant cancers consistently contain intense p16 immunostaining (Doorbar et al., 2012; Lewis et al., 2017).

Figure 5.

Cutaneous squamous cell carcinoma, cat. Neoplastic cells exhibit intense nuclear and cytoplasmic immunostaining to antibodies against the p16^CDKN2A protein. Note the presence of nests of neoplastic cells infiltrating from the overlying epidermis into the dermis. PCR amplified Felis catus papillomavirus type 2 DNA from the lesion. Scale bar = 75μm. Bond Refine Detection staining kit with haematoxylin counterstain.

While FcaPV2 shows some similarities to the high-risk Alphapapillomaviruses, an important difference is that cats are infected with FcaPV2 in the first days of life and infection is ubiquitous and lifelong (Thomson et al., 2015). As most cats do not develop cutaneous SCCs despite being infected, this suggests other factors determine the development of cancer. While these factors are poorly understood, there is variability in the viral loads of FcaPV2 between cats. Furthermore, the viral loads on an individual cat remain constant over an extended period of time (Thomson et al., 2019). This suggests individual variability in the ability of the immune system to control PV replication and it is possible that cats that are less able to prevent replication are predisposed to SCC development due to their chronically higher viral loads. This hypothesis is supported by the key role of the immune system in determining whether or not infection by a human high-risk Alphapapillomavirus will cause cancer (McBride, 2022). Alternatively, SCC development may be more dependent on the presence of co-factors such as UV exposure (Altamura et al., 2016; Munday and Kiupel, 2010). Overall, evidence suggests FcaPV2 (and possibly FcaPV3 and FcaPV4) promotes neoplasia to a much greater extent than the PVs that asymptomatically infect the skin of humans and dogs.

Horses

Non-genital skin SCCs have been rarely reported to develop due to progression from EcPV8-induced skin lesions and from aural plaques (Peters-Kennedy et al., 2020; Peters-Kennedy et al., 2019). However, there is currently little evidence PVs are a common cause of non-genital skin cancers of horses.

Mucosal neoplasia

Humans

The ability of the high-risk Alphapapillomaviruses to cause cancer is well established and these PVs are estimated to cause around 5% of all human cancers (Plummer et al., 2016). This includes almost all cervical and anal SCCs as well as around half of oral, penile, vaginal and vulval SCCs (Brianti et al., 2017). Infections are spread by sexual contact. Viral replication can result in inconspicuous flat plaques, but many infections remain asymptomatic (Doorbar et al., 2012; McBride, 2022). An immune response is generated and the infection is cleared within two years in most people (McBride, 2022). However, in a small proportion, the immune response is unable to resolve the PV infection predisposing to accidental integration of PV DNA into host cell DNA. Integration prevents viral replication but results in increased production of viral proteins and loss of normal pRb and p53 function. This creates a population of rapidly dividing, genetically unstable cells that are at increased risk of developing additional spontaneous DNA mutations that result in neoplastic transformation (Brianti et al., 2017; Doorbar et al., 2012).

While virtually all cervical cancers are considered to be caused by PV infection, oral SCCs can be subdivided into those caused by PV infection and those that are caused by tobacco and alcohol. The two subsets of SCCs have a different clinical presentation, with SCCs caused by PVs generally developing in the oropharynx of younger patients (Economopoulou et al., 2020). As the high-risk Alphapapillomaviruses so consistently cause neoplasia by degrading pRb and loss of pRb causes a marked increase in cell p16, intense nuclear and cytoplasmic p16 immunostaining confirms a PV etiology of a human oral SCC (Lewis et al., 2017).

Dogs

There are rare reports of canine oral SCCs developing within oral warts or plaques. Multiple flat plaques and invasive SCCs developed in the mouth of a dog in association with CPV17 (Munday et al., 2016b). Consistent with human PV-induced oral SCCs, the canine SCCs had marked p16 immunostaining, although unlike in human oral SCCs, PV-induced cellular changes were also prominent within the neoplastic cells (Munday et al., 2015c). There are also rare reports of progression from CPV1-induced oral warts to invasive oral SCCs in dogs (Regalado Ibarra et al., 2018; Thaiwong et al., 2018). However, PVs are not currently thought to be a significant cause of oral, or other mucosal, SCCs in dogs (Munday et al., 2015a; Porcellato et al., 2014).

Cats

Whether or not PVs play a significant role in the development of feline oral SCCs is currently uncertain. Many studies of feline oral SCCs have detected feline PV DNA only sporadically within these neoplasms (Chu et al., 2020; Munday and French, 2015; Munday et al., 2009b; Munday et al., 2011c; O’Neill et al., 2011). Immunohistochemistry revealed variable p16 immunostaining; however, this was not associated with the presence of PV DNA and no oral SCC contained the intense nuclear and cytoplasmic immunostaining that characterises PV-induced oral SCCs in humans (Munday and French, 2015; Munday et al., 2011c). In contrast, FcaPV2 DNA and gene expression were detected in 31% of feline oral SCCs in a recent study (Altamura et al., 2020). However, FcaPV2 DNA was also detected in 36% of non-neoplastic samples making it hard to determine the role of the PV in the SCCs. FcaPV2 DNA was also reported in 43% of a series of feline oral SCCs, although no non-neoplastic samples were included in this study (Yamashita-Kawanishi et al., 2021a). Recently in situ carcinomas of the mouth and third eyelid were reported in a cat (Munday et al., 2022a). Consistent with a PV etiology, these neoplasms contained FcaPV3 DNA and prominent PV-induced changes. Furthermore, like PV-induced oral SCCs in humans, all neoplastic cells contained intense nuclear and cytoplasmic p16 immunostaining (Munday et al., 2022a). Overall, there is evidence that FcaPVs infect the oral and other mucosa of cats (Altamura et al., 2020; Munday and French, 2015). As PVs cause a proportion of SCCs of feline skin, it appears possible they may also cause a proportion of oral SCCs in this species.

Horses

Of the companion animal species, the evidence supporting a role for PVs in the development of mucosal neoplasia is strongest in horses. As in humans, PVs are thought to contribute to penile, vulval and some oropharyngeal SCCs of horses (Sykora et al., 2017; Fig. 6). However, in contrast to humans in which cervical cancers are common and almost always caused by PV infection, SCCs of the internal genitalia are extremely rare in horses and have not been associated with PV infection. The most common location for PV-associated SCCs in horses is the penis and it is generally accepted that a significant proportion of these are caused by EcPV2 (Sykora and Brandt, 2017). Evidence supporting a role of this PV includes the detection of EcPV2 DNA using PCR and in situ hybridization (Bogaert et al., 2012; Greenwood et al., 2020a; Knight et al., 2011; Scase et al., 2010; van den Top et al., 2015; Zhu et al., 2015) and the detection of EcPV2 gene expression in the SCCs (Ramsauer et al., 2019a; Sykora et al., 2012). There are rare reports of detection of other EcPV types in penile neoplasms but the significance of this is unknown (Lange et al., 2013b). In addition to some oropharyngeal SCCs (Hibi et al., 2019; Knight et al., 2013; Sykora et al., 2017), EcPV2 is also present within a proportion of equine gastric SCCs (Alloway et al., 2020). Human gastric cancers have not been associated with PV infection, although this could be due to the absence of squamous epithelium lining the fundus of the human stomach. There is little evidence that EcPV2 significantly contributes to the development of equine periocular SCCs (Newkirk et al., 2014; Greenwood et al., 2020a).

Figure 6.

Vulvar squamous cell carcinoma, horse. Two ulcerated squamous cell carcinomas are visible within the vulva of this horse. Two small exophytic vegetative papillomas are also visible (arrows). PCR amplified Equulis caballus papillomavirus type 2 DNA from one of the SCCs (Image courtesy of Dr Hailey Flemming).

The mechanism of transmission of EcPV2 is unknown. However, asymptomatic infection by EcPV2 is detectable in 10 – 35% of penile samples and 0 – 15% of vulval samples and 36% of horses have serological evidence of previous EcPV2 infection (Bogaert et al., 2012; Fischer et al., 2014; Greenwood et al., 2020b; Knight et al., 2013). In one study, asymptomatic infection by EcPV2 was detected in late term fetuses and 2-day-old foals (Greenwood et al., 2020b) suggesting vertical transmission or infection from the mare during, or shortly after birth. It is possible that, as seen in children with anogenital warts (Padel et al., 1990), an EcPV2 infection acquired at birth could subsequently become localized to the genitals. Alternatively, sexual transmission of EcPV2 could be possible although lesions have been reported in geldings with no history of sexual contact (Bogaert et al., 2012). Additionally, unlike in humans in whom sexual transmission of the causative HPVs causes cancer more frequently of the internal than external genitalia, EcPV2-associated equine genital cancers are confined to the external genitalia. Other possible methods of transmission could include non-sexual direct contact between horses (muzzle to genital) or mechanical transmission by biting flies or human cleaning of the penis and prepuce (Sykora and Brandt, 2017). Such non-sexual transmission would explain the restriction of cancers to the external genitalia.

Like the HPVs that cause cervical cancer, integration of EcPV2 DNA within the cell DNA increases expression of viral E6 and E7 proteins (Ramsauer et al., 2019a). However, unlike the high-risk Alphapapillomaviruses, the EcPV2 proteins do not degrade either p53 (van den Top et al., 2015) or pRb (CG Knight, unpublished observation) and increased expression of p16 is not present within EcPV2-associated SCCs (Ramsauer et al., 2019a). Currently, although EcPV2 has been associated with altered host expression of genes associated with DNA replication, cell cycle replication, ECM-receptor integration and focal adhesion, the precise mechanisms by which EcPV2 promotes neoplasia are unclear (Ramsauer et al., 2019a). Whether co-factors are important in SCC development is also unknown.

In the authors’ experience, equine penile warts can progress to penile SCCs. However, it is currently unknown how often this happens and whether all penile SCCs develop this way or if some develop in clinically normal penile mucosa. The frequency with which penile warts regress or progress to SCC appears to be critical information as this will determine the aggressiveness of the treatment used for a penile wart. The ability of EcPV2 to cause both genital warts and genital SCCs is unexpected as human genital warts are caused by a different subset of PVs than those that cause cervical cancer. This observation adds evidence that EcPV2-induced genital warts less frequently regress, and more commonly progress, than human genital warts, and that equine genital warts may be best considered as pre-neoplastic lesions.

Possibly due to progression from a hyperplastic or pre-neoplastic lesion, equine SCCs tend to be more exophytic than SCCs that develop at other locations, although these can also appear as ulcerative masses on the penis. Histologically, the SCCs appear as typical SCCs with trabeculae and nests of invading epithelial cells with keratin pearls variably present (Fig. 7). While the neoplastic cells do not contain evidence of PV infection, some SCCs are surrounded by hyperplastic mucosa that may contain koilocytosis (Ramsauer et al., 2019b).

Figure 7.

Penile squamous cell carcinoma, horse. Large nests of neoplastic epithelial cells are visible infiltrating adjacent to smooth muscle bundles within the penis (asterisk). Many of the neoplastic cells have brightly eosinophilic cytoplasm consistent with keratinization. Scale bar = 60μm. Haematoxylin and eosin.

Prevention of PV diseases

Multivalent PV vaccines have been successfully used to prevent both genital warts and PV-induced mucosal SCCs in people (Pitisuttithum et al., 2015). However, for a vaccine to be effective it has to be given prior to first infection by the PV (Kreider, 1963). As the more important HPV types are sexually transmitted, this provides ample time to vaccinate people prior to first infection. In contrast, none of the PV types that infect dogs, cats, and horses are known to be sexually transmitted and the mechanism and age of infection is currently less clear for some of the more important PV types that infect the companion animal species.

Dogs are likely to be infected by the PV types that cause canine oral and cutaneous warts within the first 6 to 12 months of age and virus-like particle vaccines have been shown to effectively prevent warts after experiment inoculation of CPV1 (Suzich et al., 1995). However, as warts are self-resolving in the overwhelming majority of cases, it appears unlikely a vaccine would be commercially viable. It is unknown at what age dogs are infected by the Chipapillomavirus types that cause viral plaques. However, due to the similarities between these PVs and the human Betapapillomaviruses, it is possible that dogs could be infected in the first few days of life making it difficult to administer a vaccine prior to first infection.

In cats, an effective vaccine against FcaPV2 would prevent a proportion of cutaneous viral plaques and SCCs (Munday and Thomson, 2021). However, cats are infected by FcaPV2 within the first few days of life (Thomson et al., 2015) making vaccination prior to first infection impossible. It is possible that using a vaccine to stimulate high maternal antibodies could prevent infection of kittens. However, the observation that all kittens are infected despite around 20% of cats have detectible antibody titres against FcaPV2 (Geisseler et al., 2016) suggests maternal antibodies may not fully prevent infection. A study in which adult cats were vaccinated with a virus-like particle FcaPV2 vaccine showed the vaccine significantly increased antibody titres. However, the raised antibody titres did not lower viral loads suggesting that, if viral loads determine which cats develop SCCs, vaccinating adult cats is unlikely to prevent cancer (Thomson et al., 2019).

It is probable that vaccines against cutaneous warts and aural plaques could be administered prior to first infection in horses. However, as both warts and aural plaques generally cause only minor discomfort for the horse, it is uncertain whether such vaccines would be widely used. A vaccine against EcPV2 may be a valuable method to prevent genital and oropharyngeal SCCs in horses (Schellenbacher et al., 2015). However, as EcPV2 infections were detected in late term fetuses and in very young foals in one study (Greenwood et al., 2020b), additional research is required to determine when horses are infected by this PV and therefore the most appropriate way to use vaccines to prevent SCCs caused by EcPV2.

Conclusions

PV-induced cutaneous plaques in humans, dogs and cats are all likely to develop due to a defect in intrinsic keratinocyte immunity. This defect allows increased replication by PV types that normally remain as asymptomatic infections on the skin. The cause of the loss of normal immune function in dogs and cats is unknown, but the susceptibility of some breeds to plaque formation suggests a genetic basis. Feline plaques are unusual as they often contain cell dysplasia. This dysplasia can result in a classification of in situ carcinoma and feline viral plaques appear more likely to progress to neoplasia than plaques in humans and dogs. The best evidence for a significant role of PVs in the development of cutaneous SCCs is seen in cats. Cutaneous SCCs in humans, dogs, and horses could be influenced by the presence of PVs, but there is currently little evidence that PVs are a common cause of SCCs in these species. In contrast, PVs cause most cervical and anal SCCs in people as well as a proportion of oral and genital SCCs. There is similarly good evidence for a role of PVs in equine genital cancer. Oropharyngeal SCCs associated with PV infection have been reported in dogs, cats, and horses but additional evidence is required to determine what role the PVs play and how frequently PVs cause oropharyngeal SCCs in these species.