Main

Squamous cell scrotal cancer (SCSC) is a rare malignancy and was the first cancer associated with occupational carcinogens including soot, lubricating and cutting oils (Melicow, 1975; Sorahan et al, 1989). Since the 1970s, occupational-related SCSC incidence has decreased in the United Kingdom as working conditions have improved (Sorahan et al, 1989). More recent reports, however, indicate a steady incidence in the Netherlands and increasing incidence in the USA (Wright et al, 2008; Verhoeven et al, 2010). Other factors may be involved.

Human papillomaviruses (HPV) play an important aetiological role in many anogenital and oropharyngeal carcinomas (http://www.hpvcentre.net) and have been occasionally linked to skin tumours outside the genital region (Riddel et al, 2011). Human papillomavirus-DNA has been found in tissue sections containing SCSC, precancers and normal scrotum, but biological evidence is missing that directly demonstrates type-specific HPV-DNA in tumour cells, producing oncogenic RNA transcripts (Guran and Pak, 1999; Nielson et al, 2009; Matoso et al, 2014, 2016). We tested 6 cases for 72 mucosal and cutaneous HPV-genotypes. If HPV-DNA was detected, laser-capture microdissection-PCR (LCM-PCR) was performed to assign or not an individual HPV type specifically to cancer cells as we have done for cervical (pre)-cancer and other HPV-related cancers (Quint et al, 2012; Guimera et al, 2013; María José Fernández-Nestosa et al, 2017). Viral transcriptional activity was confirmed by detection of viral E6*I oncogene mRNA. Immunohistochemical (IHC) expression of p16INK4a was also studied as in previous studies, and is a sensitive but not specific marker of oncogenic HPV-E7 gene activity. IHC expression of p53 was analysed as an indicator that p53 mutations may play a causal role in the development of HPV-negative scrotal cancers.

Materials and methods

Case selection

Formalin-fixed paraffin-embedded (FFPE) tissue blocks of six cases of invasive SCSC were collected from Spain (n=2), Australia (n=2) and Nigeria (n=2) (Table 1). Local and Catalan Institute of Oncology ethics committees approved all protocols and an international steering committee oversaw study progress.

Table 1 invasive scrotal cancers by country, year of/age at diagnosis, histological category, HPV DNA detection and genotyping by 4 assays, HPV histological localization by LCM, HPV16 E6*I mRNA detection, p16INK4a and p53 overexpression
Full size table

Histology and immunohistochemistry

Cases were sectioned using a sandwich technique (de Sanjose et al, 2010; Guimera et al, 2013). At least four pathologists evaluated the haematoxylin-eosin (H&E) slides, reaching a consensus diagnosis of the histological type of SCSC: Warty, basaloid or usual. This classification was based on the criteria defined for penile carcinomas by the WHO (Moch et al, 2016). P16INK4a staining was performed on one FFPE-section (4 μm) (Guimera et al, 2013) and was considered positive if more than 50% of tumour cells showed expression. P53 staining was performed (Halec et al, 2014) and a tumour was considered driven by p53 mutation if >75% of the tumour cells showed p53 expression.

Laser capture microdissection (LCM)

Human papillomavirus detected in whole tissue section was localized to tumour cells by LCM-PCR (Quint et al, 2012). A polyethylene-napthalate (PEN)-membrane slide, stained only with Hematoxylin was used to optimise sensitivity.

DNA isolation and HPV-DNA genotyping

The DNA was isolated using proteinase K solution and HPV-DNA was tested with the SPF10-DEIA-LiPA25 system (SPF10 -HPV-LiPA, version1 (Labo-Bio-Medical Products, Rijswijk, the Netherlands) (Kleter et al, 1998, 1999; Quint et al, 2012). In all samples, the endogenous human gene RNaseP served as a positive amplification control using real time PCR (Luo et al, 2005). On whole tissue sections, the HPV genotype detected was confirmed by sequencing of the SPF10 inter-primer region (Geraets et al, 2012). As SPF10 is based on viral L1 open reading frame (ORF) cases were also analysed with MPTS 123 PCR-Luminex assay (van Alewijk et al, 2013) to detect the HPV E6 ORF from 16 different genotypes. The presence of cutaneous HPV from five genera was evaluated by Beta HPV assay (25 beta HPV genotypes) and Cutaneous Wart-Associated HPV (23 alpha, gamma, mu and nu genotypes) assays (de Koning et al, 2006, 2010). The presence of 72 HPVs was analysed for L1, E1 or E6 HPV ORFs.

RNA isolation and E6*I mRNA detection

RNA extraction, HPV16 E6*I mRNA Reverse Transcription (RT)-PCR and luminex genotyping system were used as previously described (Halec et al, 2013). An endogenous human transcript, ubiquitin C (ubC) served as a positive amplification control (Halec et al, 2014).

Results

All six cases were primary SCSC (Table 1). One primary basaloid SCSC showed metastases at presentation. No cases were immuno-compromised.

One warty and two basaloid SCSC (case 3, 5 and 6) contained HPV16 DNA as a single infection by both L1 and E6 ORFs assays. E6*I mRNA assay confirmed the presence of HPV16 E6*I transcripts and immunohistochemistry showed diffuse staining for p16INK4a, while p53 was negative. Moreover, LCM-PCR localised HPV16 DNA only in the invasive tumour (Table 1 and Figure 1). From one case with HPV-positive basaloid carcinoma (case 6), HPV16 was confirmed in metastases in three inguinal lymph nodes.

Figure 1
figure 1

H&E histological images, p16INK4a and p53 expression pattern of the three HPV16 associated scrotal cancers.The quality of Case C was very poor and the expression of p53 was difficult to analyse.

Full size image

In the three usual-type SCSC specimens (case 1, 2 and 4), no HPV DNA or RNA was detected although quality controls confirmed the presence of human DNA and RNA. In HPV-DNA-negative cases 1 and 4, p53 was expressed in >75% of cancer cells and p16INK4a was completely negative (Table 1 and Figure 2). In case 2, p53 was expressed in around 25% of cancer cells and p16INK4a was expressed only patchily (Table 1 and Figure 2). The median age at diagnosis was 59 years (range: 42–66 years) for HPV-positive cases and 62 years for HPV-negative cases (range: 46–71 years).

Figure 2
figure 2

H&E histological images, p16INK4a and p53 expression pattern of the three HPV negative scrotal cancers.

Full size image

Discussion

Treatments for skin diseases (for example, arsenic and light therapies-UVA/UVB radiation) and HPV have been identified as possible carcinogens to explain the continuing occurrence of SCSC (Stern, 1990; Andrews et al, 1991; Guran and Pak, 1999; Matoso et al, 2016). Previous studies have found HPV-DNA in tissue specimens of SCSC. The current study demonstrates that HPV16 is present specifically in tumour cells of warty and basaloid type SCSC, where it is expressing the HPV-E6 oncogene. In addition, these HPV-positive SCSC show strong expression of p16INK4a. The specific localisation of HPV16 to SCSC tumour cells required the use of LCM-PCR, which excluded potential contamination from adjacent tissue as has been described in studies of cervical cancer and precancer and of penile cancer (Quint et al, 2012; Guimera et al, 2013; María José Fernández-Nestosa et al, 2017). The very sensitive and specific methodology for HPV-DNA detection (Kleter et al, 1998, 1999; Quint et al, 2012), employed in this study together with more specific data on viral transcriptional activity than p16INK4a staining provides new definitive evidence confirming the oncogenic role of HPV16 in some histological types of SCSC (Orihuela et al, 1995; Guran and Pak, 1999; Matoso et al, 2014, 2016).

In two out of three HPV-negative cases of usual type SCSC, p53 was expressed in >75% of the cancer cells indicating the presence of a mutation in this tumour suppressor gene, but p16INK4a was not expressed strongly.

This study clearly demonstrates the two separate aetiological pathways for SCSC suggested previously (Matoso et al, 2014): Human Papillomavirus and non-HPV related, associated with separate histological tumour types. The same pattern has been described by our group using LCM-PCR in penile neoplasia (María José Fernández-Nestosa et al, 2017) and is also seen in vulvar carcinomas, where the HPV prevalence is 33% and 29%, respectively (de Sanjose et al, 2013; Alemany et al, 2016). The association between histology and HPV in SCSC appears the same as that described in penile carcinoma and precancer where neoplasia related to HPV shows a warty/basaloid histology and differentiated PeIN and usual squamous cancer is associated with a HPV-unrelated pathway. In SCSC, HPV16 is the predominant genotype as in all non-cervical sites such as penile, vulval, vaginal, anal or oropharyngeal tumours (de Sanjose et al, 2013; Alemany et al, 2014, 2016; Castellsague et al, 2016). Human papillomavirus16 is also predominant in other cutaneous in situ or invasive SCC outside the genital area such as in periungual and/or subungual skin from immunocompetent patients (Riddel et al, 2011). Scrotal cancer like external vulvar cancer arises in an epithelium that is morphologically genital skin. The embryologic differentiation of scrotal cells resembles that of the vulvar labia majora in women, and may have the same stem-like cells vulnerable to HPV16 infection (Moore et al, 2013).

Although limited by the small number of cases, this study provides biological evidence of an aetiological relationship between HPV16 and SCSC. Larger epidemiological studies are needed to determine if other HPV types apart from HPV16 are involved, the relative role of HPV compared to other environmental and genetic factors and the possible changing role of HPV in scrotal carcinogenesis. Such studies could also determine the potential impact on scrotal cancer of HPV vaccines. Studies of the mechanism and transformation of the scrotal cells by HPV would also contribute to understanding why these events are rare compare to other genital sites, particularly the cervix.