The role of non-classical and chain-related human leukocyte antigen polymorphisms in laryngeal squamous cell carcinoma

Laryngeal squamous cell carcinoma (LSCC) is the major pathological subtype of laryngeal cancer. It has been shown that alterations of the expression of non-classical human leukocyte antigens (HLA) and the chain-related MIC molecules by malignant cells can lead to escape from the immune system control and certain allele variants may participate in immune editing and therefore be associated with modulation of cancer risk. The aim of the present study was to investigate the role of non-classical HLA class Ib and chain-related MIC polymorphisms, determined at the allelic level by next-generation sequencing (NGS), in patients from the Bulgarian population, diagnosed with LSCC. In the present study DNA samples from 48 patients with LSCC were used. Data was compared to 63 healthy controls analysed in previous studies. HLA genotyping was performed by using the AlloSeq Tx17 early pooling protocol and the library preparation AlloSeq Tx17 kit (CareDx). Sequencing was performed on MiniSeq sequencing platform (Illumina) and HLA genotypes were assigned with the AlloSeq Assign analysis software v1.0.3 (CareDx) and the IPD-IMGT/HLA database 3.45.1.2. The HLA disease association tests revealed a statistically significant predisposing association of HLA-F*01:01:02 (Pc = 0.0103, OR = 24.0194) with LSCC, while HLA-F*01:01:01 (Pc = 8.21e-04, OR = 0.0485) has a possible protective association. Additionally we observed several haplotypes with statistically significant protective and predisposing associations. The strongest association was observed for F*01:01:01-H*01:01:01 (P = 0.0054, haplotype score=-2.7801). Our preliminary study suggests the involvement of HLA class Ib in cancer development and the possible role of the shown alleles as biomarkers of LSCC.


Introduction
Larynx cancer is the twentieth most common cancer worldwide in 2020 [1]. Laryngeal squamous cell carcinoma (LSCC) is one of the most common head and neck malignancies and is the major subtype (95%) of laryngeal cancer [2]. LSCC has the highest rate of mortality and morbidity in head and neck cancer [3]. The 5-year survival rate has decreased in past years, despite the progress in diagnostics and treatment of patients [4]. Patients with early stage LSCC have a 5-year survival rate of 70-90%, but those with advanced stage have only 30% [5]. The age of onset of LSCC is between 50 and 70 years, and male/female ratio is 4:1.
Smoking and alcohol drinking are the main risk factors for the occurrence and development of LSCC. The survival rate of these patients is lower than those that do not smoke and/ Anastasia Ormandjieva anastasia.ormandjieva@gmail.com 1 or drink [5]. According to the location of the tumor, LSCC is divided into supraglottis, glottis ad subglottis. The prognosis can be affected by the site of the primary tumor [5], i.e. women are more likely to have LSCC of the supraglottis, rather than of the glottis [6]. Supraglottic primary tumors reoccur more often compared to glottis ones. Supraglottic tumors also have higher rates of regional nodal metastasis [6]. Comorbidities in patients with LSCC can also have an impact on their overall condition and survival [6]. Precision medicine is the key to improve the life of LSCC patients [7]. Immunotherapy strives to activate the host's immune defences in order to identify and destroy cancer cells [8]. For understanding the molecular characteristics of LSCC, assessing prognostic factors and biomarkers will contribute to evaluating the response to treatment and better outcome [3,6].
Human leukocyte antigen (HLA) genes reside in a highly polymorphic gene region on chromosome 6 (6p21.3) and encode proteins that present foreign antigens to the immune system [9]. Non-classical HLA molecules are characterized by lower genetic diversity, in contrast with classical ones, they also have different expression pattern, structure and functions [9]. Mechanisms of immune evasion involve the impaired ability of HLA to maintain immune surveillance [10]. Tumor cells can down-regulate HLA classical class I molecules and up-regulate non-classical molecules. HLA Ib molecules, recognized by inhibitory receptors show higher expression in tumour microenvironment and therefore are involved in tumor immune escape [11].
MHC class I polypeptide-related sequences A and B (MICA, MICB) are ligands of NKG2D, known as stress-induced ligands [12]. MIC molecules are expressed on different tumors and MICA interacts with NKG2D ligands and promotes cytolysis and IFN-secretion [13]. MICA expression is also induced by different cytokines [13] and its upregulation is correlated with the grade of the tumor and indicated poor prognosis. MICB is downregulated by the RNA binding protein IMP3, which facilitates immune evasion [12]. MIC molecules in soluble form also have a role in immune evasion [14]. Serum levels of sMIC correlate with cancer progression [15].
Clarification of the precise mechanisms of interaction between the tumor microenvironment and the HLA-Ib and MIC molecules will create an opportunity for developing immunotherapies that are highly effective, with fewer side effects. The aim of the present study was to investigate the role of non-classical HLA class Ib and chain-related MIC polymorphisms, determined at the allelic level by next-generation sequencing (NGS), in patients from the Bulgarian population, diagnosed with LSCC.

Materials & methods
The study included 48 patients from the Bulgarian population, diagnosed with LSCC after biopsy and pathophysiological examination. Patients were diagnosed in the Department of Dental, Oral and Maxillofacial Surgery, FDM, MU-Sofia; Specialized Hospital in Maxillofacial Surgery, Sofia; Department of Maxillofacial Surgery, University Hospital "Aleksandrovska"; Department of Otorhinolaryngology, University Hospital "Tsaritsa Yoanna", Sofia. The control group included 63 healthy individuals from the Bulgarian population, typed from a previous study of the research team [16]. Demographic data of LSCC patients and healthy controls is presented in Table 1. For all patients, the stage and grade of the tumors (Table 1) were assessed according to the TNM staging system, where T indicates the size and extent of the main tumor, N -the number of nearby lymph nodes that have cancer and M -whether the cancer has metastasized [17]. All patients and healthy controls provided informed consent for genetic testing as part of the institutional review board's approved standard-operating procedures at our institutions. The principles of the Declaration of Helsinki were strictly followed during the study.

DNA extraction
Genomic DNA was extracted from sodium citrate whole blood, by using the iPrep (Thermo Fisher Scientific, USA) automated system. Genomic DNA samples were stored at − 20 °C before being used for this study.

Statistical analysis
The association analysis was performed with generalized linear modelling with age and gender as covariables. The analysis was performed by using the PyHLA package for Python (version 1.1.1) [18] for association analysis, zygosity and interaction tests, and the haplo.stats package for R (v1.8.9) [19] and Arlequin (v3.5.2.2) [20], for estimating haplotype frequencies with estimation-maximization (EM) algorithms. Plots were generated by using GraphPad Prism (v8.0.1) for Windows, GraphPad Software, San Diego, California USA, www.graphpad.com. P-values for allele associations are adjusted for multiple testing (with Bonferroni correction) and values below 0.05 were considered significant.
The interaction tests suggested that E*01:01:01 and F*01:01:01 alleles have combined action (P = 7.18e-05, OR = 0.06). They are not in linkage disequilibrium (LD) in cases and control groups. The analysis also revealed that there is a combined action between F*01:01:01 and H*01:01:01 (Pc = 1.51-e06, OR = 0.05), but F*01:01:01 is associated with the disease in H*01:01:01 positive and negative samples. They are not in LD in cases and control groups.

Discussion
Various mechanisms of immune evasion have been characterized and are known as the hallmarks of cancer. As a consequence, there are numerous hypothesized and found biomarkers for the purposes of anti-cancer immunotherapy [21]. It is known that tumor progression is the result of imbalance of the host's immunity, [21] hence the extreme necessity to estimate prognostic factors and biomarkers that will predict responses and outcome to immunotherapies [6].
The immune system of patients diagnosed with larynx carcinoma is usually altered, but this may depend on various factors like alcohol use, viral infections, malnutrition and aging [6]. Numerous studies have reported the main role of immune regulation in carcinogenesis and treatment of LSCC [7]. However, how to reduce the mortality and increase the 5 year survival rate in LSCC patients remains an unsolved problem [3]. Elucidating the molecular characteristics and mechanisms in LSCC progression through genomic, transcriptomic and protein analysis, will enable the establishment of possible targets for prognostic and therapeutic purposes [6].
HLA-E presents peptides derived from leader sequences of classical class I HLA molecules and HLA-G. It is also able to present alternative peptides derived from tumor proteins [22]. The peptide loaded HLA-E molecule binds to the activation CD94/NKG2 receptors on T and NK cells. The expression of HLA-E is upregulated in cancer cells, which suggests the participation of gene transcription mechanisms. Soluble HLA-E also has a role in the tumor escape inducing apoptosis in T and NK cells, with the Fas/Fas ligand-mediated pathway [23]. Expression of HLA-F is also upregulated in tumor cells, associated with poor survival prognosis in cancer patients [24,25]. HLA-F/ß2m-tetramers inhibit NK cells through various receptors (KIR3DL2, ILT2, ILT4) and presents peptides to T cells [25]. HLA-G molecules inhibit antitumor responses via interaction of ILT2 and ILT4 recep-    HLA genotype associations with cancer are related to the peptide binding affinities of specific genotypes. Thus there is a distinctive presentation of peptides by the HLA molecules, depending on the matching sequences to the amino acids in the structure of the peptide binding site [31]. Better patient survival may be a result of presentation of tumor-associated antigens and neoantigens, after which T cell recognition and cytotoxic activities are heightened [31].
Our work is the first study to explore the role of nonclassical HLA Ib (E, F, G, H) and MIC (MICA and MICB) polymorphisms in this disease. In addition, haplotype analysis provides insightful information, considering the linkage disequilibrium of HLA.
Our results show a strong association of LSCC with two HLA-F alleles. Unfortunately, the overall function of HLA-F is still not fully investigated. To date, 59 HLA-F alleles, encoding 11 proteins, have been described. HLA-F*01:01 is known as the predominant allelic variant. In our patient cohort, HLA-F*01:01:01 is the prevailing allele (allele frequency of cases = 0.414; allele frequency of controls = 0.742). HLA-F is detected in different types of cancer, including oesophageal squamous cell carcinoma (ESSC) [32]. Zhang X. et al. found that elevated levels of HLA-F are correlated with poor survival in patients with ESSC [32]. The HLA-F bound with peptides has classical HLA class I (Ia) features, but selection of the peptides is more strict because of the more limited polymorphism of HLA-Ib molecules. HLA-F*01:01 presents peptides mainly from the cytoplasm and nucleoplasm. The most frequently bound are 16-mer (16 AAs/16 residues length). In a study by Ho G. et al. they report that the N-terminal site of the peptides is not anchored into the peptide binding region (PBR) [33]. They also found that interaction with the KIR3DS1 receptor on NK cells leads to delayed disease progression (in the context of HIV infection) [33], but this may also explain the protective role of HLA-F*01:01:01 in cancer. On the contrary, HLA-F can bind to inhibitory receptors ILT-2 (expressed on NK cells) and ILT-4 (expressed on T and B cells) and other immune cells (monocytes, macrophages and dendritic cells) [32]. Zhang et al. hypothesize that these bindings have direct inhibitory effect on the immune cells and have an impact on their immune functions in different stages of the immune response [32].
The two HLA-F alleles (F*01:01:01 and F*01:01:02) that have different associations with LSCC risk in our study, differ in the third field that shows a synonymous DNA substitution within the coding region. Synonymous substitutions are viewed as silent since they only affect the DNA and mRNA level and are not associated with amino acid modifications. However, they have significant impact on splicing, RNA stability, RNA folding, translation or co-translational protein folding. According to Sharma Y. et al. [34], factors i.e. matrix metalloproteinases [26]. Finally, HLA-G inhibits NK cells directly and indirectly through impairing the NK/DC crosstalk. HLA-H is known as a pseudogene, because it is deleted from chromosomes which carry HLA-A*23 and A*24 alleles [27], and it's amino acid sequence predicts a truncated non-functional protein. However, different studies report that this gene has transcriptional activity. Like the other HLA Ib molecules, HLA-H is involved in immune response and inhibition [28]. HLA-H may have a role, via its signal peptide, by mobilizing HLA-E to the cell surface, or as a transmembrane molecule. From this follows the hypothesis that HLA-H is tolerogenic, and its absence when the immune system is impaired might lessen tolerogenicity [28].
Head and neck squamous cell carcinoma (HNSCC) in general is the main focus of many studies and not specifically larynx cancer. In addition, most of the research emphasizes on the level of expression of non-classical HLA molecules (mainly HLA-E and -G) in LSCC tumor tissues. Also, association studies of HLA and HNSCC report associations with classical class I and class II alleles, but not nonclassical ones. Reports of MIC molecules are based on the levels of soluble MIC and on exon 5 short tandem repeats (STRs) polymorphism.
Overexpression of HLA-E is associated not only with laryngeal cancer, but with invasive lesions, so it is suggested as a biomarker for the disease [11]. HLA-G high expression is observed in premalignant larynx lesions and it is decreased in malignant ones, so its absence is suggested as a biomarker for malignant lesions [11]. Information on the clinicopathological significance of HLA-F is more limited, but it is proposed that this molecule has a similar immunosuppressive role in cancer. Nevertheless, the presence of HLA Ib molecules is a predictor for poor prognosis [9].
HLA-H is a pseudogene and there are rarely studies, in comparison to the functional HLA genes. However, HLA-H has transcriptional activity, according to different studies. It is hypothesized that HLA-H has a tolerogenic function and participates in the immune homeostasis, like HLA-G, -E and -F [9].
In regards to MIC associations with cancer, MICA is the more studied molecule. MICA is known as the most polymorphic non-classical HLA class I gene [29]. MICA expression levels have been correlated with different type of tumors, including HNSCC [29]. Upregulated MICA expression has been correlated with histological grade, involvement of the lymph nodes and vascular invasion [29]. It is also known that MICA expression has impact on the tumorigenesis and progression of oral squamous cell carcinoma (OSCC). Surface and soluble (sMICA and sMICB) levels of expression of MIC molecules are present in sera of patients, with different types of malignancies [30]. escape mechanisms in these type of malignancies. Our study showed the complex role of non-classical HLA and MIC genes in cancer predisposition and protection. This could be relevant for establishing novel biomarkers for diagnosis and therapy of LSCC.

Data Availability
The data that support the findings of this study are available on request from the corresponding author.

Conflict of interest None.
Ethics approval statement The principles of the Declaration of Helsinki were strictly followed during the study. Ethical approval was granted for this study by Research Ethics Committee at Medical University -Sofia with the approval number №15/19.05.2021.
Patient consent statement All patients and healthy controls provided informed consent for genetic testing as part of the institutional review board's approved standard-operating procedures at our institutions. synonymous substitutions can affect the level of expression of proteins, even their folding. Since synonymous substitutions have a functional impact, we speculate that the peptide binding affinity is different in HLA-F*01:01:01 and HLA-F*01:01:02 which in turn probably leads to their different role in LSCC risk. However such mechanism should be further clarified.
Barakat G. et al. investigated the association of LSCC with HLA-G polymorphisms [36]. HLA-G has 8 exons and a 14-bp insertion or deletion polymorphism can be observed in the 3' untranslated region (UTR), responsible for the stability and splicing of the mRNA isoforms [36]. They found a significant association between the 14-bp deletion HLA-G allele and LSCC [36]. In our study HLA-G alleles were not found to be significantly associated with the disease, but HLA-G is present in a few of the significantly associated haplotypes (Table 2).
MIC polymorphisms, associated with HNSCC, are mainly studied in the context of STRs [37,38]. Tani R. et al. found that MICA A5.1 homozygous plays a role in susceptibility to OSCC and increases the risk of early onset oral cancer, but have better prognosis than patients with other MICA genotypes [38]. We did not find significantly associated alleles, but MICA is also involved in haplotypes, involving the protective HLA-F allele.
To the best of our knowledge, this current study is the first to link HLA non-classical Ib and MIC polymorphisms, on allelic and haplotype levels, with LSCC. Of note, our study is preliminary and it is of need to enlarge the group of patients, and the control group, in a subsequent analysis that may also involve tumor stage and grade, in order to understand the mechanisms underlying the pathogenesis of LSCC.

Conclusion
Although significant advancements have been made in the fields of diagnostics and LSCC patients care, the long terms survival rate remains relatively low. That emphasizes the necessity for further studies on the molecular and immune manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted