A subset of equine gastric squamous cell carcinomas is associated with Equus caballus papillomavirus-2

Abstract

Squamous cell carcinoma (SCC) is the most common neoplasm of the equine stomach. The mechanisms underlying malignant transformation, however, are unknown. As Equus caballus papillomavirus-2 (EcPV-2) is a likely etiologic agent for some genital SCCs, we hypothesized that EcPV-2 is associated with a subset of equine gastric SCCs. To this aim, we performed PCR and in situ hybridization (ISH) for EcPV-2 E6/ E7 oncogenes on 11 gastric SCCs. As controls, we included 11 non-SCC gastric samples. PCR for EcPV-2 was positive in 7/11 (64%) gastric SCCs; non-SCC gastric samples were all negative. Intense hybridization signals for EcPV-2 E6/E7 nucleic acid were detected by ISH within tumor cells in 5/11 (45%) gastric SCCs, including distant metastases. No hybridization signals were detected within any of the non-SCC gastric cases. This study provides evidence for a potential association between EcPV-2 and a subset of equine gastric SCC.

Squamous cell carcinoma is the most common type of stomach cancer in the horse.⁹ These cancers are most often recognized late in the course of disease and are associated with a poor prognosis.⁹ Thus far, an underlying cause for this type of cancer has not been identified. Human papillomaviruses (HPV) are oncogenic viruses with a well-established causal association with several types of cancer, including squamous cell carcinoma.² Virtually all cases of human cervical cancers, as well as a subset of other genital and oropharyngeal cancers, are causally associated with certain types of HPV.² The HPV oncogenes E6 and E7 are upregulated within these cancers and drive carcinogenesis.² These oncogenes interfere with host cell cycle regulation and apoptosis resulting in unregulated cell growth and genetic instability.² Similarly, there is increasing evidence that equine papillomaviruses may play a role in oncogenesis of certain types of tumors.⁷ Equus caballus papillomavirus type 2 (EcPV-2) likely plays a causative role in development of a subset of genital squamous cell carcinomas.^3,4,6–8,10 Accumulating evidence for this association includes consistent detection of EcPV-2 DNA within genital SCC and precursor lesions, as well as identification of oncogenic E6 transcripts within genital SCC.^3,4,6–8,10 These lesions typically progress from precursor lesions, such as papillomatous plaques and papillomas, to carcinoma in situ and ultimately to SCC.⁶

Given the association of EcPV-2 with a subset of genital SCC, we hypothesized that EcPV-2 is also associated with a subset of gastric SCC. To this aim, we performed PCR and in situ hybridization for EcPV-2 E6/ E7 on 11 equine gastric SCC cases. Tissue samples were collected from archived biopsies and necropsies from multiple institutions (Table 1): 1) North Carolina State University’s College of Veterinary Medicine (archives searched 2008–2018) 2) Rollins Animal Disease Diagnostic Laboratory (archives searched 2012–2018) 3) University of Pennsylvania, School of Veterinary Medicine (archives searched 2007–2018) and 4) Animal Health Laboratory, University of Guelph (archives searched 1997–2018). A full necropsy was performed on all cases except for two (Cases 2 and 10). Median age at time of diagnosis was 19 years old, with an age range from 3 to 24 years old. All cases were male. Metastases were identified in seven out of the 9 (78%) cases with a full necropsy. Eleven non-SCC cases, including normal stomach, gastric hyperplasia, gastritis, gastric polyp, and leiomyoma, were included (cases 12–22, Suppl. Table S1). One equine genital SCC was included as a positive control.

Table 1.

PCR and in situ hybridization for EcPV2 E6/E7 viral nucleic acid within gastric squamous cell carcinomas.

Case #	Institution	Age (years)	Breed	Necropsy details	Metastasis	EcPV2 PCR	EcPV2 ISH
1	NC State	14	Paint horse	15cm gastric mass	Spleen, liver	+	+
2	NC State	21	Thoroughbred	Biopsy only	Unknown	+	+
3	U of Guelph	Unk	Unk	20cm gastric mass	Lung, mediastinum; carcinomatosis	+	+
4	U of Penn	19	Quarter horse	8cm gastric mass	Liver, kidney, spleen, lung; carcinomatosis	+	+
5	U of Guelph	20	Paso Fino	10cm gastric mass	None	+	+
6	U of Guelph	Unk	Unk	25cm gastric mass	Colon; carcinomatosis	+	−
7	U of Guelph	7	Unk	24cm gastric mass	Spleen, lymph nodes, liver, kidney; carcinomatosis	+	−
8	NC State	22	Quarter horse	30cm gastric mass	Liver, omentum, lung	−	−
9	U of Guelph	19	Paint horse	47cm gastric mass	Lungs, rib, kidney, liver, thoracic inlet	−	−
10	U of Guelph	3	Saddlebred	Biopsy only	Unknown	−	−
11	U of Guelph	22	Morgan horse	23cm gastric mass; additional stomach mass	None	−	−

EcPV-2, Equus caballus papillomavirus-2; ISH, in situ hybridization; Unk, unknown

Hematoxylin and eosin stained slides were evaluated to confirm the diagnosis and further characterize these cases. Gastric SCC were characterized by infiltrative, expansile, and non-encapsulated masses of neoplastic keratinocytes arranged in islands, trabeculae, and cords supported by scirrhous stroma. The neoplasms variably merged with the overlying, hyperplastic squamous mucosa and frequently invaded the submucosa and tunica muscularis. Foci of keratinization and abortive cornification (keratin pearls) were often present. Central degeneration, lytic necrosis, and neutrophilic infiltrates of neoplastic islands were common. Anisocytosis and anisokaryosis were mild to marked and mitotic index ranged from 1 to 11 (area: 2.37mm²). Robust lymphoplasmacytic inflammation infiltrated the supporting scirrhous stroma. Metastasis was observed in 7 cases, most commonly involving the liver (5/7), lung (4/7), peritoneal cavity (4/7), and spleen (3/7). Other sites of metastasis included the mediastinum (1/7), colon (1/7), kidney (2/7), thoracic inlet (1/7), omentum (1/7), lymph nodes (1/7), and rib (1/7). Vascular invasion was present in 2 of the primary SCC and 3 of the metastatic neoplasms.

PCR for the EcPV-2 E6 gene was performed on total DNA isolated from two 25μM scrolls of formalin-fixed, paraffin-embedded (FFPE) tissue samples using a Qiagen DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA) following manufacturer’s recommended protocols. PCR primers for EcPV-2 E6 gene were designed using Primer 3 design software based upon the published EcPV-2 sequence between nucleotides 5 and 622 (Genbank reference sequence EU503122.1). Primers amplified a 226 base pair (bp) product and are as follows: EcPV-2 E6-For (5’-CAT TTG GTG GAA CAC AGC AG-3’) and EcPV-2 E6-Rev (5’-TTT TTC GTC CCT GGT CAG TC-3’). Primers for equine GAPDH were designed using Primer 3 and are as follows: Eq-GAPDH-For (5’-GAT TGT CAG CAA TGC CTC CT-3’) and Eq-GAPDH-Rev (5’-AAG CAG GGA TGA TGT TCT GG-3’). Equine GAPDH primers amplified a 196 bp product. PCR reaction mixtures contained 1 X buffer, 200 uM of each dNTP, 1uM each primer, 1U Taq Polymerase (HotStarTaq DNA Polymerase, Qiagen) and 10ul (100ng) total DNA in a 50ul reaction mixture. A reaction without template DNA served as the negative control. PCR reaction conditions were as follows: 95°C for 15 minutes followed by 35 cycles of 95°C for 30 seconds, 58°C for 1minute, and 72°C for 1 minute, followed by 72°C for 10 minutes. PCR products were run on a 1% agarose gel and visualized using SYBR Safe DNA gel stain (Invitrogen, Thermo Fisher Scientific, Pittsburgh, PA). EcPV-2 DNA was amplified from 7/11 (64%) gastric SCC cases but none of the non-SCC control cases (Table 1). GAPDH was amplified from all gastric SCC samples and all non-SCC control samples.

In situ hybridization was performed using a previously validated probe set designed to detect EcPV-2 E6/E7 nucleic acid in genital SCCs (RNAscope; Advanced Cell Diagnostics, Hayward, CA).¹⁰ As stated in the original publication, this probe is designed to detect a double-stranded DNA virus, and may therefore detect both viral DNA as well as messenger RNA (mRNA).¹⁰ ISH was performed on 5-μm-thick FFPE sections using the RNAscope 2.5 RED assay kit according to manufacturer’s recommended protocols. Successful hybridization results in deposition of a red stain which is in direct correlation with the amount of EcPV-2 nucleic acid. A negative control was run concurrently using a predesigned probe to the bacterial gene DapB (ACD). The RNAscope 2.5 BROWN assay kit was used whenever the RED assay kit produced background on negative control sections. The BROWN assay kit produced superior results with reduced background on six cases.

Hybridization signals for EcPV-2 were detected within 5 gastric SCCs, including distant metastases (Table 1). Two of the 5 positive cases had overlying surface squamous epithelium to evaluate, which in both cases was characterized by mucosal hyperplasia (Figs 1A and 2A; Suppl. Fig 1A). These regions corresponded to intense nuclear staining for EcPV-2, predominantly within the upper layers of epithelium (Figs 1B and 2B; Suppl. Fig 1B). This distribution is congruent with the productive papillomavirus life cycle, where vegetative viral replication takes place within keratinocytes in the upper layers of the mucosal epithelium.² The presence of epithelial hyperplasia and intense nuclear ISH signals within the upper layers of epithelium suggests that SCC may develop from precursor hyperplastic lesions, as is the case for equine genital SCC and HPV associated cancers. Less intense dot-like signals were present within the basal epithelium, where only mRNA transcripts for E6/E7 would be expected (Suppl. Fig 1B). No hybridization signals were detected using the negative control probe (Fig 2C; Suppl. Fig 1C).

Figures 1–4.

Horse No. 1. Figures 1–3. Mucosal hyperplasia overlying squamous cell carcinoma (SCC); stomach. Figure 1. (a) Hyperplastic gastric squamous epithelium and islands of neoplastic squamous epithelium within the deeper submucosa. Hematoxylin and eosin (HE). (b) Intense hybridization signals (red) are present within the overlying mucosal and deeper islands of neoplastic cells. In situ hybridization (ISH), EcPV-2 E6/E7 probe. Figure 2. (a) Higher magnification of the hyperplastic gastric squamous epithelium. HE. (b) Intense hybridization signals are present within the nuclei of the hyperplastic epithelium. ISH, EcPV2 E6/E7 probe. (c) No hybridization signals are present using the negative control probe. ISH, DapB probe. Figure 3. (a) Higher magnification of invasive islands of neoplastic squamous epithelial cells within the submucosa. Dot-like hybridization signals are present within the nuclei and cytoplasm of neoplastic cells. Intense diffuse hybridization signals are present within the nuclei of some of the neoplastic cells. In situ hybridization (ISH), EcPV-2 E6/E7 probe. (b) No hybridization signals are present using the negative control probe. ISH, DapB probe. Figure 4. Metastatic SCC; spleen. (a) Island of neoplastic squamous epithelium within the spleen. HE. (b) Strong dot-like hybridization signals are present within the nuclei and cytoplasm of neoplastic cells. Intense diffuse hybridization signals are present within the nuclei of some of the neoplastic cells. In situ hybridization (ISH), EcPV-2 E6/E7 probe. (c) No hybridization signals are present using the negative control probe. ISH, DapB probe.

Invasive neoplastic cells contained variably intense dot-like hybridization signals within the nucleus and/or cytoplasm (Fig 3A; Suppl. Figs 2–5), occurring within 60–100% of the neoplastic cells. No signals were seen above background staining levels using the negative control probe (Fig 3B; Suppl. Figs 2–5C). Metastatic lesions were present in 3 of the 5 positive cases; ISH signals for EcPV-2 were also detected in metastatic lesions and characterized by variably intense dot-like hybridization signals within tumor cell nuclei and/or cytoplasm (Fig 4). Six gastric SCCs had no hybridization signals for EcPV-2 nucleic acid (Table 1). All non-SCC gastric samples were negative (Supp. Table 1).

All ISH positive gastric SCC samples were also PCR positive for EcPV-2. Two EcPV-2 PCR positive gastric SCCs, however, were ISH negative. While the exact cause for this discrepancy is unknown, it may reflect a difference in sensitivity between the two assays. Regardless, there were still a significant percentage of gastric SCC that were EcPV-2 negative. One possibility is that there is more than one mechanism for oncogenesis in gastric SCC, as has been suggested for genital SCC.¹⁰ Alternatively, it is possible that EcPV-2 is an essential initiating oncogenic event but not required in later steps of malignant progression (“hit-and-run” hypothesis). This hypothesis has been suggested for development of human cutaneous SCCs.² In this scenario, EcPV-2 could be present within some tumors but lost within others.

While between 30–60% of human oropharyngeal SCCs are causally associated with HPV, human gastric SCCs are rare with an unknown etiopathogenesis.^1,5 There is no reported association with HPV. Suggested hypotheses for oncogenesis include tumors arising from regions of squamous metaplasia, transformation of a stem cell, or squamous differentiation within an adenocarcinoma.¹ We identified EcPV-2 nucleic acid within 45% of equine gastric SCC, which is closer to the percentage identified in human oropharyngeal SCC than gastric SCC. This difference could possibly be attributed to the significant differences between the human and equine stomach. While the human stomach is lined entirely by glandular epithelium, the equine stomach is composed of both a glandular and a non-glandular squamous portion; it is this squamous portion that gives rise to the gastric SCCs.⁹ Given this distinction, it is possible the squamous portion of the equine stomach is more like an extension of the oropharyngeal macro- and micro- environments and thus susceptible to diseases more commonly seen within the oropharyngeal region. Provided that, however, EcPV-2 has only been detected within 3/20 (15%) and 4/15 (26.6%) orophargyngeal SCCs.^3,7 This lower detection rate within equine oropharyngeal SCC compared with gastric SCC could reflect the small sample size in these studies, differences in methodology to detect EcPV-2, differences in prevalence of EcPV-2, or even differences within oncogenic potential of EcPV-2 at these different sites.

In conclusion, this study provides evidence for a potential association between EcPV-2 and a subset of equine gastric SCC. Further studies are necessary to determine the role of EcPV-2 in development of gastric SCC.