**Frequent Alteration of Annexin A9 and A10 in HPV-Negative Head and Neck Squamous Cell Carcinomas: Correlation with the Histopathological Differentiation Grade**

**Cecilia Salom 1,†, Saúl Álvarez-Teijeiro 1,2,†, M. Pilar Fernández 3, Reginald O. Morgan 3, Eva Allonca 1,2, Aitana Vallina 4, Corina Lorz 2,5, Lucas de Villalaín 6, M. Soledad Fernández-García 4, Juan P. Rodrigo 1,2,\* and Juana M. García-Pedrero 1,2,\***


Received: 3 December 2018; Accepted: 4 February 2019; Published: 10 February 2019

**Abstract:** The annexin protein superfamily has been implicated in multiple physiological and pathological processes, including carcinogenesis. Altered expression of various annexins has frequently been observed and linked to the development and progression of various human malignancies. However, information is lacking on the expression and clinical significance of annexin A9 (ANXA9) and A10 (ANXA10) in head and neck squamous cell carcinomas (HNSCC). ANXA9 and ANXA10 expression was evaluated in a large cohort of 372 surgically treated HPV-negative HNSCC patients and correlated with the clinicopathologic parameters and disease outcomes. Down-regulation of ANXA9 expression was found in 42% of HNSCC tissue samples, compared to normal epithelia. ANXA9 expression in tumors was significantly associated with oropharyngeal location and histological differentiation grade (*p* < 0.001). In marked contrast, ANXA10 expression was absent in normal epithelium, but variably detected in the cytoplasm of cancer cells. Positive ANXA10 expression was found in 64% of tumors, and was significantly associated with differentiation grade (*p* < 0.001), being also more frequent in oropharyngeal tumors (*p* = 0.019). These results reveal that the expression of both ANXA9 and ANXA10 is frequently altered in HNSCC and associated to the tumor differentiation grade, suggesting that they could be implicated in the pathogenesis of these cancers.

**Keywords:** annexin A9; annexin A10; head and neck squamous cell carcinoma; differentiation grade; immunohistochemistry

### **1. Introduction**

Twelve annexins comprise a ubiquitous, multigene family in vertebrates with properties that enable binding interactions with calcium and cell membrane components, including anionic phospholipids, cytoskeletal proteins and extracellular matrix glycoproteins. Annexin-knockdown or annexin-knockout models have provided limited insight into the biological functions of different annexin proteins [1] and there are only indirect links based on statistical association with genetic diseases. They have been implicated in a variety of biological processes, including membrane organization, vesicle trafficking, calcium metabolism, cell adhesion, subcellular transport, growth and differentiation, and wound healing [2,3], many of which are relevant to cancer progression.

Annexins are characterized structurally by a conserved C-terminal core that consists of a tetrad of homologous annexin (ANX) domains, each 68–69 amino acids long, harboring ligands that can coordinate calcium ions in conjunction with membrane phospholipids, or bind to other proteins and carbohydrate-containing biomolecules. The binding properties and targets of each annexin are distinct, exemplified by the apparent calcium-independence of annexins A9 and A10 [4]. The N-terminal region of each annexin is unique, with a variable length and amino acid sequence that contributes to annexin conformation, protein interactions and non-overlapping functional specificity in the biological activity of different annexins [5,6].

More than 4000 annexins have been reported in different species, widely distributed among eukaryotes and prevalent in different forms of prokaryotes and unicellular eukaryotes [1,4]. The twelve annexins common to vertebrates are referred to as annexins A1–A13 (ANXA1–ANXA13) with ANXA12 remaining unassigned. There are 13 human annexin genes, including a unique duplication of ANXA8, ranging in size from 15 kb (ANXA9) to 96 kb (ANXA10) and spread throughout the genome on chromosomes 1, 2, 4, 5, 8, 9, 10 and 15 [1,7].

The expression pattern and tissue distribution of annexins vary widely. While annexins A1, A2, A4, A5, A6, A7 and A11 are ubiquitously expressed, others exhibit very restrictive expression such as ANXA3 in neutrophils, ANXA8 in placenta and skin, ANXA9 in the tongue, ANXA10 in the stomach and ANXA13 in the small intestine [7]. The promoter regulation of annexin A9 has been partially characterized [8], but distal DNA elements, regulatory RNAs and epigenetic changes are under current study in high-throughput experiments, so the molecular basis of its expression remains incomplete.

The term annexinopathy has been used to define those human diseases in which abnormal levels and pleiotropic effects of annexins contribute to the pathogenesis [9,10]. Although direct involvement of these proteins in the etiology of any genetic disease has not been demonstrated, they have been implicated in various pathologies such as diabetes, cardiovascular and autoimmune diseases, infection and cancer [10,11]. Mounting evidence shows that several annexins are frequently altered in cancers, suggesting a possible role in the process of tumorigenesis. Some annexins have been found overexpressed in specific types of tumors, while others consistently show loss of expression [9–11]. Emerging mechanistic studies are helping to relate annexin expression changes to tumor cell function, particularly tumor growth, invasion and metastasis, angiogenesis and drug resistance. The expression of individual annexins is associated with particular cancer types hence annexins could also be useful biomarkers in the clinic [10,11]. More precise localization of these proteins in different tissues could deepen our understanding of their pathophysiological functions, which continues to be a key area of investigation.

The overall goal of this study was to investigate the expression pattern and clinical significance of ANXA9 and ANXA10, specifically in head and neck squamous cell carcinomas (HNSCC). ANXA9 shows generally restricted tissue expression but is known to exhibit altered expression in breast cancer [12], colorectal cancer [13] and cutaneous melanoma [14]. It was also shown to be overexpressed in differentiating keratinocytes in pemphigus [15] and binds to other cytoskeletal proteins [16]. Several studies have been published to date on the expression ANXA10 in gastrointestinal cancers, and its overexpression in oral cancer is correlated with cell proliferation [17]. We focused our study on the expression and clinical significance of ANXA9 and ANXA10 specifically in HNSCC using immunohistochemistry techniques in a large homogeneous cohort of 372 surgically treated, HPV-negative, HNSCC patients.

### **2. Materials and Method**

### *2.1. Patients and Tissue Specimens*

Surgical tissue specimens from 372 patients with HPV-negative HNSCC who underwent resection of their tumors at the Hospital Universitario Central de Asturias between 1990 and 2009 were retrospectively collected, in accordance to approved institutional review board guidelines. All experimental protocols were approved by the Institutional Ethics Committee of the Hospital Universitario Central de Asturias and by the Regional CEIC (Comité Ético de Investigación Clínica) from Principado de Asturias (approval number: 81/2013 for the project PI13/00259). Informed consent was obtained from all patients. Representative tissue sections were obtained from archival, paraffin-embedded blocks and the histological diagnosis was confirmed by an experienced pathologist (M.S.F.-G).

All patients had a single primary tumor, microscopically clear surgical margins and received no treatment prior to surgery. Only fourteen patients were women, and the mean age was 58.6 years (range 30 to 86 years). All but twelve patients were habitual tobacco smokers, 198 moderate (1–50 pack-year) and 153 heavy (>50 pack-year), and 335 were alcohol drinkers. The stage of the tumors was determined according to the TNM system of the International Union Against Cancer (7th Edition). Two hundred and thirty (62%) of 372 patients received postoperative radiotherapy. Patients were followed-up for a minimum of 36 months. The mean follow-up for the whole series was 34.6 months (median, 21.5 months); for the patients without recurrence, 71 months (median, 67 months); and for the patients dead by the tumor, 18 months (median, 13.5 months). Recurrence was defined as relapse of the tumor in the five first years after treatment at any site: local recurrence, nodal metastasis, or distant metastasis. Information on HPV status was available for all the patients. HPV status was analyzed using p16-immunohistochemistry, high-risk HPV DNA detection by in situ hybridization and genotyping by GP5+/6+-PCR, as previously reported [18,19]. The characteristics of the studied cases are shown in Table 1.

### *2.2. Tissue Microarray (TMA) Construction*

Three morphologically representative areas were selected from each individual tumor paraffin block. Subsequently, three 1 mm cylinders were taken to construct TMA blocks, as described previously [20,21], containing a total of 372 HNSCC (134 tonsillar, 107 base of tongue, 64 hypopharyngeal and 67 laryngeal carcinomas). In addition, each TMA included three cores of normal epithelium as an internal negative control. The normal epithelium was obtained from adult male, non-smokers and non-drinkers, patients that were operated from tonsillectomy due to chronic tonsillitis, and patients operated from benign vocal cord lesions (e.g., polyps, cysts).


**Table 1.** Clinicopathologic characteristics of the tumors studied.

### *2.3. Immunohistochemical Study*

The formalin-fixed, paraffin-embedded tissue samples were cut into 3-μm sections and dried on Flex IHC microscope slides (Dako, Glostrup, Denmark). The sections were deparaffinized with standard xylene and hydrated through graded alcohols into water. Antigen retrieval was performed with proteinase K and the samples were placed for 15 min in hydrogen peroxide at 3%. Staining was done at room temperature on an automatic staining workstation (Dako Autostainer Plus) using the following primary antibodies (developed by Dr. MP Fernández, Department of Biochemistry, University of Oviedo [4]) and conditions: Anti-ANXA9 at a concentration of 1:100 for 30 min and anti-ANXA10 at a concentration of 1:100 for 45 min. Immunodetection was carried out with the Dako EnVision Flex + Visualization System (Dako Autostainer), using diaminobenzidine as a chromogen. Counterstaining with hematoxylin for 7 min was the final step.

After staining, the sections were dehydrated and set up in a slide in a standard medium. Negative controls were carried out without the primary antibody. The vascular endothelium, in which the expression of both annexins had previously been shown, was used as a positive control.

Since staining showed a homogeneous distribution, a semiquantitative scoring system based on staining intensity was applied. Immunostaining was scored blinded to clinical data by two independent observers as negative (0), weak to moderately (1+), and strongly positive (2+) based on staining intensity. Scores ≥ 1 were considered as positive expression.

### **3. Results**

### *3.1. Expression of ANXA9 and ANX10 in Normal Epithelia*

Non-keratinized stratified squamous epithelium showed different expression patterns for the two annexins studied. ANXA9 expression was absent in basal and parabasal cells, while expression increased towards the most differentiated layers of the epithelium (Figure 1A). Contrasting this, negative ANXA10 expression was detected in all cell layers of normal epithelium (Figure 1D).

**Figure 1.** Immunohistochemical analysis of annexins A9 (ANXA9) and A10 (ANXA10) expression in head and neck squamous cell carcinomas (HNSCC) tissue specimens. Representative examples of ANXA9 (**A**) and ANXA10 (**D**) expression in normal epithelium, positive ANXA9 (**B**) and ANXA10 (**E**) expression in carcinomas, and negative ANXA9 (**C**) and ANXA10 (**F**) expression in carcinomas. Original magnification ×40.

### *3.2. Expression of ANXA9 in HNSCC Tissue Specimens*

Immunohistochemical analysis of ANXA9 expression was successfully evaluated in 346 of 372 tumor samples. Two-hundred of them (58%) showed positive ANXA9 expression predominantly with a membranous pattern, although cytoplasmic expression was also observed in some cases (Figure 1B,C). The relationship between the expression of ANXA9 and clinicopathologic characteristics is shown in Table 2. Positive ANXA9 expression was strongly and significantly associated with the degree of differentiation of the tumors (*p* < 0.001). Thus, ANXA9 expression was mainly found in well-differentiated tumors whereas expression was reduced in moderately and poorly differentiated tumors (Figure 2A,C). We also observed differences in ANXA9 expression between the different HNSCC subsites, with ANXA9 expression being significantly higher in oropharyngeal tumors (*p* < 0.001).


**Table 2.** Relationship between ANXA9 and ANXA10 expression and clinicopathological parameters.
