Epithelial–mesenchymal transition (EMT) in vulvar cancer with and without inguinal lymph node involvement

Vulvar cancer is a rare gynaecologic malignancy accounting for 2–5% of all gynaecologic cancers (Rogers and Cuello 2018; Te Grootenhuis et al. 2018). Although it mostly affects elderly women, the incidence in younger patients is rising (Te Grootenhuis et al. 2018).

Vulvar squamous cell carcinoma (VSCC) is the most common histopathological subtype of vulvar cancer (Dasgupta et al. 2020; Rogers and Cuello 2018; Singh and Gilks 2020). There are two distinct molecular pathways that lead to VSCC, associated with either (1) VIN of the usual type (u-VIN) which is associated with high-risk HPV infections, mostly resulting in non-keratinizing VSCC and (2) VIN of the differentiated type (d-VIN), associated with p53 alterations often developing in a background of lichen sclerosus, mostly resulting in keratinizing VSCC (Dasgupta et al. 2020; Singh and Gilks 2020).

Regardless of the pathogenetic pathway, one of the most relevant prognostic factors for local recurrence of VSCC is the presence of inguinal lymph node metastases (Rogers and Cuello 2018; Te Grootenhuis et al. 2018). Risk factors for lymph node involvement of VSCC are the depth of invasion, tumor size and the ontogenetic tumor stage (Höckel et al. 2018; Julia and Hoang 2021) as well as morphologic features, such as infiltrative growth with dissociatively growing tumors and the presence of strong peritumoral desmoplastic change (Holthoff et al. 2016; Horn et al. 2012). Dissociative tumor growth and strong peritumoral stromal remodelling have been shown to be associated with epithelial-to-mesenchymal transition (EMT) in VSCC (Holthoff et al. 2016; Rodrigues et al. 2013). Furthermore, there is mounting evidence that EMT plays an essential role in lymphatic spread (Campo et al. 2015).

The present study was designed to evaluate the immuno-histochemical expression of EMT markers in correlation to the presence of inguinal lymph node involvement in VSCC.

Patients diagnosed with macro-invasive squamous carcinomas of the vulva were identified from our medical records. All patients had no previous vulvar surgery, had not received neoadjuvant therapy and were included in the prospective surgical trial for vulvar field resection based on the ontogenetic anatomy (German Clinical Trials Register, number DRKS00013358; (Höckel et al. 2018)). Patients were separated into those without and those with histopathologically proven inguinal lymph node involvement. Representative tumor tissue, using full slides, was evaluated immunohistochemically for p53, p16, MIB-1 and EMT markers (vimentin, cyclin D1, e-cadherin). Immuno-histochemical antibody information is provided in Table 1.

p16 was scored as positive if there was diffuse block-like cytoplasmic and nuclear staining and as negative for any lesser staining (such as patchy or absent staining) in accordance with the LAST (Lower Anogenital Squamous Terminology) project recommendations (Darragh et al. 2013) and recent studies (Tessier-Cloutier et al. 2020).

p53 was scored as wild type representing heterogeneous nuclear staining of variable intensity of scattered nuclear-positive tumor cells in the basal and parabasal layers and mid-epithelial with heterogeneously positively stained nuclei with notable basal negativity of the invasive tumor cell nests. Mutant staining patterns were defined as strong nuclear staining in at least 80% of the basal cells with or without significant parabasal extension, complete negative staining of the tumor cells with the presence of a positive staining in adjacent inflammatory and stromal cells (serving as positive internal control) as described recently for VSCC (Tessier-Cloutier et al. 2020).

Nuclear staining of Ki-67 was scored from 1–3 (Fig. 1) on the basis of a comparison with normal vulvar squamous epithelium (Podoll et al. 2017) and existing evidence on penile squamous cell carcinoma (Protzel et al. 2007; Stankiewicz et al. 2012): 1 = basal expression or low-grade pattern with nuclear Ki-67 staining of the basal/ outer cell layers of the invasive tumor cells nests with very limited suprabasal extension; 2 = increased expression or intermediate pattern with staining up to the middle of the infiltrating tumor cell nests; 3 = increased expression or high-grade pattern with diffuse staining of the tumor cell nests. All nuclear staining was recognised as positive, regardless of the staining intensity.

For EMT markers, an immuno-reactive score (IRS) was calculated by summarizing the percentage of positive cells and their staining intensity. Staining intensity (SI) was scored as weak, moderate or strong. The percentage of positive tumor cells at the front of invasion was scored as 0 = 0% positive tumor cells, 1 = 1% positive cells, 2 = 2– < 10% positive cells, 3 = 10– < 33% positive cells, 4 = 33– < 66% positive cells, 5 = 66–100% positive cells. The median values of IRS were evaluated in correlation to the presence of inguinal lymph node involvement. Additionally, the zonal staining pattern was evaluated, comparing the front of invasion with the tumor center (Stewart and McCluggage 2013).

The pattern of invasion and the grade of peritumoral desmoplastic reaction (DSR) and inflammatory response (PER) were evaluated in correlation to the lymph node status as previously described (Horn et al. 2012).

Median differences including 95% confidence intervals were calculated as effect measures following Hodges and Lehman. The significance level was determined at 5% for two-tailed tests. All analyses were performed by IBM SPSS Statistics, version 26.

An Institutional Review Board approval was obtained prior to the initiation of the study.

Lymph node involvement and recurrent disease play a critical role in the prognosis of VSCC (Rogers and Cuello 2018; Te Grootenhuis et al. 2018). Depending on the tumor stage, up to 50% of all VSCC patients will be present with lymph node involvement at the time of diagnosis (Höckel et al. 2018). Among those node positive patients, 30–50% will relapse (Melo Maia et al. 2013).

Epithelial–mesenchymal transition (EMT) has been studied in several malignancies (Bhangu et al. 2012; Toll et al. 2013), including gynecologic cancers (Stewart and McCluggage 2013). However, only few studies are available for vulvar cancer (Holthoff et al. 2016; Rodrigues et al. 2013; Zannoni et al. 2011). EMT represents a molecular shift within the tumor cells, leading to the adoption of a mesenchymal phenotype associated with loss of polarity, loss of cell adhesion and enhanced cellular motility (Toll et al. 2013), with a consecutively increased risk of metastatic spread and poor prognosis (Bhangu et al. 2012; Stewart and McCluggage 2013; Toll et al. 2013). While many immuno-histochemical markers (IHC) have been evaluated to identify cells undergoing EMT, some of the most validated markers include vimentin, e-cadherin and cyclin D1 (Holthoff et al. 2016; Stewart and McCluggage 2013; Zannoni et al. 2011).

It has been shown that EMT markers are highly upregulated in VSCC tumors representing a spray-like pattern of invasion (Holthoff et al. 2016), mostly associated with a strong peritumoral desmoplastic reaction and low peritumoral inflammatory response. As reported for squamous cell carcinomas of the uterine cervix (Horn et al. 2012), a spray-like pattern of invasion is also associated with a higher risk of lymphatic spread and poor prognostic outcome in VSCC (Holthoff et al. 2016; Jeffus et al. 2015). These results could not be confirmed examining our cases (Tab. 1), may be done by the limited number of cases examined. There was no correlation of the expression of the tested EMT markers (vimentin, cyclin D-1, e-cadherin) when the different patterns of invasion were compared. The missing correlation may be biased by the limited number of cases tested.

It has been shown that cells with an EMT phenotype are able to escape the immunological surveillance by host immune cells (Kudo-Saito et al. 2009). Holthoff et al. (2016) reported that VSCC with a strong peritumoral desmoplastic reaction displayed a 3.3-fold increase in the number of cases with positive staining for vimentin compared to patients with a strong peritumoral inflammatory reaction (p = 0.0031). In contrast, Rodrigues et al. (2013) described no correlation between the expression of e-cadherin, vimentin and other EMT-related markers and the inflammatory response in VSCC. In the present study, there was no correlation between the grade of peritumoral inflammatory response or the desmoplastic reaction and the expression of the tested EMT markers within the tumor cells (suppl. Table 2–4). In the study of Holthoff et al. (2016), there was no correlation between the loss of e-cadherin and peritumoral stromal response in VSCC. Regarding the contradictory results in a potential correlation between the peritumoral inflammatory response and the expression of EMT markers in VSCC, it may be hypothesized, that the presence of strong inflammatory response at the front of invasion may suppress the process of EMT in squamous cell carcinomas (Holthoff et al. 2016). Results supporting this idea are pending.

Examining HPV-negative and -positive cell lines from squamous cell carcinomas, it has been shown that HPV-/p16-positive cell lines represented a reduced EMT transition (Umbreit et al. 2014).

Rodrigues et al. (2013) described no correlation between e-cadherin and vimentin and a potential HPV infection in VSCC. However, a low β-catenin and high Slug expression were significantly associated with HPV-negative (p53-aberrant) tumors and a poor prognostic outcome. Although not reaching statistical significance, the median IRS value of vimentin was higher in VSCC with aberrant p53 staining when compared to wild type staining (4.5 vs. 0; p = 0.087; suppl. Table 1). For cyclin D1, a significantly increased staining was seen in p53-aberrant VSCC (p = 0.03). These results may indicate that different EMT markers may be expressed differently in VSCC depending on its pathogenetic pathways (HPV high-risk infection versus p53 mutation). In the present study, p16^non−block/p53^aberrant tumors, were associated with a spray-like pattern of invasion and low proliferative activity (Ki-67 labeling score 1; Fig. 5a), likely caused by the p53 mutation. In contrast, in p16^block/p53^wt VSCC, indicating a high-risk HPV infection, finger-like pattern of invasion and high proliferative activity (score 3 Ki-67 labeling index) were significantly more common (Fig. 5b).

In squamous cell cancer of the head and neck as well as in those of the female genital tract, it has been suggested that upregulated EMT may be associated with a higher tumor stage and the presence of metastatic spread (Stewart and McCluggage 2013; Wan et al. 2020). Holthoff et al. (2016) were unable to demonstrate any correlation between the immuno-histochemical expression of the EMT markers β-catenin, vimentin and e-cadherin and nodal involvement in VSCC. Low e-cadherin expression was associated with deeper invasion, higher tumor stage and ≥ 2 positive lymph nodes in VSCC (Rodrigues et al. 2013). In the same study, there was a significant association between positive vimentin staining and the depth of invasion, but as mentioned above, not with lymph node involvement. In the present study, there were no differences in the immuno-reactive scores of vimentin and cyclin D1 or the staining of e-cadherin between node negative and node positive cases. In the authors’ opinion, additional studies are required examining a larger number of cases to compare node negative and positive VSCC. Furthermore, the cases need to be separated based on the different pathogenetic pathways of VSCC (high-risk HPV- and p53-driven tumors).

In the present study, different microanatomical locations, i.e. topographic subpopulations of tumor cells were evaluated regarding the immuno-histochemical expression of EMT markers comparing the central tumor to its front of invasion. The vast majority of cases demonstrated pronounced cyclin D1 staining at the front of invasion while vimentin staining was seen exclusively in tumor cells at the invasion front (Fig. 7a, b). Contrary, more than two-thirds of the cases represented reduced or negative staining for e-cadherin at the front of invasion (Fig. 7c, d). Holthoff et al. (2016) reported that immuno-histochemical expression of vimentin was exclusively seen at the front of invasion in two-thirds of VSCC, and 77% displayed a loss of e-cadherin. In head and neck squamous cell carcinomas, Dal Vecchio et al. (2011) reported a significant correlation between the expression of vimentin and lymph node involvement. Rodrigues et al. (2013) found an increased vimentin expression at the front of invasion for vulvar carcinomas. In the same study, a significant loss of nuclear β-catenin staining (p = 0.013) was seen in tumor cells at the front of invasion, especially in HPV-negative VSCC. The infiltrative tumor cells in CIN 3-like squamous cell carcinomas of the uterine cervix showed a nuclear expression of cyclin D1 and loss of e-cadherin (Stewart and Crook 2017) whereas the central parts of the tumor demonstrated a reverse staining pattern. Both nuclear expression of cyclin D1 and loss of e-cadherin were also pictured at the front of invasion within the review by Stewart and MaCluggage (2013). Zannoni et al. (2011) reported a down-regulation of e-cadherin expression in VSCC compared to normal squamous epithelium at the front of invasion. Loss of e-cadherin leads to reduced cell adhesion which may result in an increased metastatic potential of tumor cells (Bakir et al. 2020; Jolly et al. 2017). The different zonal staining patterns suggest that the tumor cells at the front of invasion acquire an invasive and migratory behavior to support the infiltrative tumor growth and support the ability for metastatic spread. Different microanatomical staining patterns for EMT-associated markers have been reported for several gynecologic malignancies (Stewart and McCluggage 2013) including cervical squamous cell carcinomas (Koay et al. 2012; Lee and Shen 2012), suggesting that the EMT phenotype of cancer cells is a prerequisite for infiltrative tumor growth at the front of invasion as well as metastatic spread (Bakir et al. 2020; Jolly et al. 2017).

Meanwhile, EMT is believed to be a reversible process, affected by tumor microenvironment, and is followed by an inverse process, resulting in a mesenchymal–epithelial transition (MET; Bakir et al. 2020; Jolly et al. 2017; Stewart and McCluggage 2013). One hypothesis is that MET may be necessary for colonisation and tumor growth at the metastatic site (Guarino et al. 2007; Stewart and McCluggage 2013) as well as the realisation of metastatic deposits from circulating tumor cells (Bakir et al. 2020; Jolly et al. 2017). These ideas may be supported by the present study, showing an immuno-histochemical expression of e-cadherin in the tumor cells with on average 74.9% positively stained cells in inguinal lymph metastases of VSCC, accompanied by tumor cells that are negative for vimentin in 12/16 cases within the lymph node deposits. Additionally, one of the cases with a small tumor cell deposit within the lymph node was positive for vimentin (Fig. 6h), whereas the larger metastases were completely negative. However, the number of cases included in the present study was too small to draw definitive conclusions.

The present study showed no differences in the expression of EMT markers when node positive and negative VSCC were compared. However, different staining patterns were noted within different cellular subpopulations in their micro-anatomical context indicating an EMT-phenotype in the tumor cells at the front of invasion. Paired analyses of tumor cells within the vulvar tumor and its metastatic deposits within the inguinal nodes may indicate a mesenchymal–epithelial transition (MET). Although a limited number of cases was examined, our results and those obtained from the literature suggest that EMT may be more frequent in tumors associated with a p53 mutation compared to those caused by a high-risk HPV infection. The pronounced EMT features in VSCC representing p53 alterations may be responsible for the more aggressive clinical behaviour of that tumors, regardless of lower proliferative activity.

Additional studies are required examining a larger cohort of cases with separation VSCC according to its different pathogenetic pathways. A better understanding of the etiology for lymphatic spread and metastases in HPV- and non-HPV-associated vulvar squamous cell carcinoma may eventually help identify patients at high risk for lymphatic or metastatic spread.