iRhom2 in the pathogenesis of oral squamous cell carcinoma

iRhom2 is an inactive rhomboid protease involved in diverse signalling events. It has been implicated in the pathogenesis of a number of cancer types, including oesophageal and ovarian cancer, while its closely associated family member, iRhom1, is implicated in head and neck cancer. However, a role for iRhom2 in head and neck cancer has not been investigated. Immunoblotting for iRhom2 in 54 oral squamous cell carcinoma (OSCC) and 24 paired normal tissues demonstrated higher levels of iRhom2 protein in tumour compared with normal samples (P < 0.05). iRhom2 over-expression correlated with poor patient survival (P < 0.0005) but with no other clinicopathological variable. Increased cell migration was observed in stably over-expressing iRhom2 clones of OSCC cell lines in the absence of increased cell proliferation, but not in the normal oral keratinocyte cell line, NOK-hTERT, and this was abrogated by knock-down of iRhom2. iRhom2 protein expression is increased in a proportion of OSCC and this up-regulation is associated with faster cell migration and decreased patient survival. These data implicate iRhom2-controlled signalling events in the pathogenesis of this cancer.


Introduction
Rhomboids are intramembrane serine proteases, with their active and catalytic residues located within the cell membrane lipid bilayer, enabling them to easily cleave and activate / inactivate membrane proteins within their transmembrane helices [1]. Although lacking protease function, iRhom2 has been shown to be involved in diverse signalling events such as the trafficking/activation of important growth and signalling factors, e.g. ADAM17, EGFR, and TGFα [2,3]. iRhom2 has been implicated in the pathogenesis of a number of cancer types including oesophageal and ovarian cancer [4], while its closely associated family member, iRhom1, is implicated in EGFR signalling in head and neck cancer [5], but evidence for a specific role for iRhom2 in head and neck squamous cell carcinoma (HNSCC) has not been described.
HNSCC was reported to be the eighth most common cancer type in the UK in 2014. Worldwide, it is the sixth leading cancer type by incidence, with more than 550,000 cases and about 300,000 deaths in 2014 [6]. The 5 year overall survival in this cancer remains at approx. 50% [7] despite advances in surgical and chemoradiation treatments, implicating hitherto unexplored pathways.. This investigation focuses on the possible oncogenic role of iRhom2 in oral squamous cell carcinomas, the most common form of HNSCC, and its possible utility as a biomarker for survival.

Patients
Fresh, frozen tissues were obtained from 54 OSCC tumours and 24 paired normal adjacent tissues. Written consent was obtained from all patients (REC numbers: EC.47.01 & 10/ H1002/53). The tissues were snap frozen at point of surgery and stored at -80°C prior to use. Demographic and clinicopathological data was collected.
Frozen tissues were embedded in OCT and a 5 μm section assessed for tumour presence. Tumour samples with less than 60% tumour content were excluded. Normal samples with more than 5% tumour content were also excluded.

Cell lines and cell culture
The oral cancer cell lines, PE/CA-PJ15 [8] (source: ECACC) and Liv37K, a locally derived, low passage cell line, both showed detectable levels of iRhom2 expression and were selected to represent an established oral cancer cell model and a 'close to the tumour' oral cancer cell model, respectively. The normal oral keratinocyte cell line, NOK-hTERT [9] (Source: Prof Karl Münger, The Channing Laboratory, Brigham and Womens Hosptial, Boston, USA) was selected as a control. All cell lines were cultured at 37 °C in 5.0% carbon dioxide. Liv37K was used at or below passage 10 for all experiments reported here.
Cell monolayers were scraped into 10 ml PBS, followed by centrifugation to recover cell pellets which were stored at − 80 °C prior to protein extraction. The identity of the cell lines was confirmed by STR typing using GenePrint10® (Promega) and were mycoplasma negative by E-Myco PCR detection kit (CHEMBIO Ltd).

Protein extraction
5 mm 3 pieces of tissue were excised from areas with an increased number of tumour cells or lack of tumour as appropriate (determined from H&E sections) and re-suspended in protein extraction buffer (50 mM Tris-HCl, pH 6.8, 1% SDS, 1% EDTA, 0.25% glycerol, 0.25% β-mercaptoethanol with protease inhibitor (Roche Diagnostics), and sonicated for 30 s to extract total protein which was stored at − 80 °C until use. Recovered cell pellets were re-suspended in protein extraction buffer and similarly processed.
Densitometry was used to normalise and semi-quantitate the levels of protein expression in comparison with the actin control and the positive control sample run on each gel. Samples were ascribed a status of high (or overexpression) or moderate (or normal expression) based on a cut-off of 19.9 relative absorbance units, which was determined to be the boundary of the 75 th percentile from the cumulative frequency curve of the expressing samples. Samples demonstrating no measurable expression were classified as no expression.

Proliferation assay
A crystal violet assay was used to assess proliferation. Briefly, cells were seeded into 6 well plates and harvested daily for 5 days, then stained with 0.05% crystal violet solubilised in 10% acetic acid and absorbance read at 570 nm using a Spectramax-Plus284 absorbance microplate reader (Molecular Devices) against a 10% acetic acid blank.

Migration assay
Cells were seeded in triplicate into removable chambers (Ibidi, Thistle Scientific) placed in 6 well plates, then cultured until they reached approximately 85% confluency. The removable inserts were withdrawn and the plates transferred to culture chamber attached to a Nikon Eclipse TE300 microscope for image capture at ten minutes intervals for 30 h. Gap closure images were analysed on Image J (https :// image j.nih.gov/ij/) and T-Scratch [10].

Statistical analysis
Protein expression data from tissue was analysed using box plot analysis and paired sample t test (SPSS statistics, version 24). Chi-squared and the Kaplan-Meier survival tests were used to analyse associations between tissue expression and clinicopathological features.
No significant correlations were observed between iRhom2 protein expression in tumours and their clinicopathological features, except that high levels of iRhom2 expression were shown to correlate negatively with overall survival (P < 0.0005) (Fig. 2).

Impact of altered expression of iRhom2 in cell lines
In contrast to data in the literature, the rate of proliferation was not increased in any of the cell lines by overexpression of iRhom2 ( Supplementary Figs. 1 & 2), but was reduced in Liv37K and, to a lesser degree, in NOK-hTERT, while having no apparent effect on PE/CA-PJ15.
Despite the lack of increase in cell proliferation, PE/ CA-PJ15 and Liv37K clones with up-regulated iRhom2 migrated approximately twice as fast as their respective wild-type variants (Figs 3a-d), with the knockdown clone of PE/CA-PJ15 showing a slower rate of migration than the wild type variant (Fig. 3a, b). Neither iRhom2 overexpression nor shRNA knock-down appeared to alter the rate of migration of NOK-hTERT cells (Fig. 3e, f).

Discussion
A role for iRhom2 has been proposed in oesophageal and ovarian cancer [4], but evidence for a specific role in HNSCC has not been described. We have now shown that iRhom2 protein expression is often upregulated in OSCC in comparison to adjacent normal tissues (P < 0.05).   Furthermore, increased iRhom2 expression correlated with poor survival (P < 0.0005), but did not correlate with any other clinicopathological indicators of prognosis, such as extracapsular spread or tumour size, suggesting that it may be an independent prognostic indicator. However, the sample size was too small to assess this statistically. iRhom2 has been shown to be involved in diverse signalling events such as the trafficking/activation of important growth and signalling factors, e.g. ADAM17, EGFR, and TGFα (transforming growth factor α) [2,3]. ADAM17, also known as TACE (Tumour Necrosis Factor-α converting enzyme), is involved in the shedding of signalling molecules such as TNFα, TGFα, EGF, Notch1 and CD44, and adhesion molecules such as L-selectin, syndecans, CAMs (Cell Adhesion Molecules) and cadherins [11][12][13]. Following its synthesis, ADAM17 protein is thought to be largely stored in the ER in immature form, prior to maturation in the Golgi apparatus and onward trafficking to the cell membrane [14]. Therefore, only a fraction of the mature / activated form is demonstrated to be available at the cell membrane at any one time [14] and trafficking to the cell membrane is thought to be orchestrated by iRhom2 [2]. Increased ADAM17 expression has been shown to lead to poor survival in OSCC orthotopic animal models [15] and may thus be implicated as the downstream initiator of iRhom2.
To augment our observational study on clinical samples, we both over-expressed and knocked down iRhom2 expression in two oral cancer cell lines and one immortalised normal oral keratinocyte cell line. Interestingly, cell proliferation was not shown to be augmented by increased expression of iRhom2 and these data are in contrast to the literature that suggests a proliferative role for iRhom2 [16,17]. Upregulation of iRhom2 in the two cancer cell lines increased their rate of 2D migration in a wound healing assay but did not confer a similar advantage to a non-cancerous oral epithelial cell line. This data again implicates ADAM17 as the effector as this has been associated with poor prognosis and promotion of oral cancer cell line migration via inactivating cleavage of CXCL1, a cell adhesion regulator [18]. Furthermore, motility of cancer associated fibroblasts in culture has been shown to be dependent on iRhom2-dependent, ADAM17mediated cleavage of TGF beta receptor 1 [19].
These data implicate activation of the iRhom2-ADAM17 pathway in the pathogenesis of OSCC, and suggest that this system works by influencing or increasing rate of cell migration. Deregulation of ADAM17 has previously been implicated in head and neck squamous cell carcinoma (HNSCC) [20], with up-regulation of ADAM17, as well as increased sheddase activity, correlating with tumour aggressiveness, rate of growth and prognosis in this and other cancers [20][21][22][23]. Increased ADAM17 sheddase activity has also been demonstrated following increased expression of the protein in head and neck cancer cell lines [20]. Although the exact mechanism remains largely unclear, there is some evidence to suggest that upregulated ADAM17 in head and neck cancer plays key roles in tumour initiation and progression via proteolytic and/ or adhesive properties [20]. This may relate in turn to our data indicating poor survival in those patients whose cancers overexpressed iRhom2. However, additional research is required to completely unravel the signalling pathways involved and confirm whether modulation of iRhom2 expression may be a future target for therapy or useful as a prognostic marker.
Informed consent Informed consent was obtained from all individual participants included in the study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.