Expression of the Non-classical HLA-E, -F, -G Molecules in the Tumour Microenvironment
Article Category: Research Article
Publicado en línea: 03 ago 2023
Páginas: 53 - 62
Recibido: 20 sept 2022
Aceptado: 09 ene 2023
DOI: https://doi.org/10.2478/fco-2022-0008
Palabras clave
© 2022 Anastasia Ormandjieva, published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Cancer is among the main causes of death worldwide, with lung cancer being the leading cause of cancer death, followed by breast cancer, prostate cancer, and colorectal cancer in 2022[1]. There is an urgent need to develop new strategies for minimising the cancer burden, like combining early detection with appropriate treatment[2].
Tumour cells escape immune responses by different mechanisms and human leukocyte antigen (HLA) molecules play a central role. HLA class I molecules (HLA-Ia) have the ability to present antigens to cytotoxic T lymphocytes (CTLs), which recognise tumour antigens and eliminate them. Tumour cells can escape CTLs by downregulating HLA-Ia, which allows the tumour to grow and eventually form metastasis[3]. HLA class II molecules present peptides to T helper lymphocytes, which have a supporting function, thus HLA class II antigen expression of tumour cells results in tumour antigen-specific immune answers[4].
Non-classical HLA-Ib (HLA-E, -F, -G) molecules are encoded by HLA class I genes and their expression on the tumour cells’ surface allows the tumour cells to escape T and natural killer (NK) cells’ immune response [5]. Recent investigations suggest that the expression of those molecules can be used as tumour markers because their expression is usually increased in the tumour microenvironment[6]. The mechanisms underlying tumour-elicited immunosuppression are still only partially understood and have been the topic of many investigations. The presence of non-classical HLA class I molecules is suggested to be a predictor of poor prognosis[6] and a growing amount of data highlights the important role of HLA-Ib molecules in immune regulation and cancer development. Clarification of the precise mechanisms of the interactions between the tumour microenvironment and the HLA-Ib molecules will create an opportunity for developing immunotherapies that are highly effective with fewer side effects.
The genes encoding non-classical HLA-Ib molecules are localised on the short arm of chromosome 6 at the telomere side[7]. In contrast to classical HLA molecules, the polymorphic diversity of non-classical Ib molecules—HLA-E, - F, and -G—is limited. They display a limited polymorphism with less encoded proteins by fewer alleles (Table 1).
Alleles and proteins of HLA class I molecules[8].
| 7,562 | 9,000 | 7,513 | 311 | 56 | 104 | |
| 4,399 | 5,414 | 4,157 | 121 | 11 | 35 | |
| 392 | 312 | 324 | 8 | 0 | 5 | |
HLA-E has eight exons. The first encodes the leader peptide sequence. Exons 2, 3, and 4 encode the HLA immunoglobulin-like α domains 1, 2, and 3. Exon 5 encodes the transmembrane domain and 6 and 7 – encode the cytoplasmic tail[7] (Figure 1). Peptides are anchored by five hydrophobic pockets in the grooves of HLA-E via 2, 3, 6, 7, and 9 residues, hydrogen bonding between the heavy chain and the main chain[7].
Figure 1:
Structure of the HLA-E, -F, and -G genes.
HLA-E presents peptides derived from leader sequences of classical class I HLA-molecules and HLA-G and thus provides an important “self signal” to the immune system[9]. HLA-E is expressed in skin, kidney, liver, thyroid, bladder, stomach, endometrium, and spleen tissues; lymph nodes; endothelial cells; and in B cells, T cells, monocytes, macrophages, and megakaryocytes[6]. HLA-E presents peptides and binds to the activation CD94/NKG2 receptors on T and NK cells[6]. HLA-E interacts with NK cells through the inhibitory receptor CD94/NKG2A and with activating receptor CD94/NKG2C. Its affinity to bind the inhibitory receptor is six-fold higher than that to the activating receptor[10]. The CD94/NKG2A receptor binds to the peptide-binding groove (CD94 subunit binds to the α1-helix and NKG2A to the α2-helix)[9].
Cells with a low level of expression of HLA class I molecules generate HLA class I derived peptides and display a low level of HLA-E, which allows NK cell lysis. HLA-E is also co-expressed with HLA-G in physiological conditions and they interact to modulate the immune response[9].
HLA-F has eight exons and its organisation is similar to HLA-E (Figure 1). The first exon encodes the leader peptide; exons 2, 3, and 4 encode the HLA immunoglobulin-like α domains; exon 5 encodes the transmembrane domain; and exons 6 and 7 encode the cytoplasmic tail. HLA-F also forms a complex with β2-microglobulin[7].
HLA-F has recently been studied more and there are a few reports about its protein expression. HLA-F is physiologically expressed on the surface of B cells and monocytes, and constitutive expression is found in the cytoplasm of B, T, and NK cells and monocytes. It binds to the inhibitory receptors—Ig-like transcript-2 (ILT-2) and Ig-like transcript-4 (ILT-4)—which suggests a role in the regulation of immune cell function[6].
In overall structure, HLA-G is also similar to HLA-E (Figure 1) but has seven isoforms, generated by alternative splicing: four membrane-bound (HLA-G1, -G2, -G3, -G4) and three soluble forms (HLA-G5, -G6, -G7)[6]. Novel isoforms predicted by Tronik-Le Roux et al., HLA-GL1 and sHLA-G1 are generated by alternatively spliced transcripts of HLA-G1[10]. HLA-G1 represents the full-length version of the HLA-G molecule[7]. HLA-G2 lacks exon 3, HLA-G3 lacks exons 3 and 4, and HLA-G4 does not have exon 4. Soluble isoforms have exon 4 with a stop codon, which results in a truncated protein, encoding the transmembrane domain. HLA-G forms heterodimers with β2-microglobulin, but can also form tetramers or be in a homodimeric state. Hetero dimers and tetramers bind to their respective receptors with the highest affinity[7].
HLA-G expression is tissue-restricted and can be found mainly in immune-privileged organs, organs during development, and in cells of the hematopoietic lineage[11]. Like HLA-E, HLA-G is a ligand for NK inhibitory receptors, but its role in immunity is more specialised. This is because of the predominant expression of HLA-G on foetal extravillous trophoblasts, which serves as a signal to NK cells, macrophages, and monocytes. HLA-G binds to leukocyte immunoglobulin-like receptors (LILRB1, LIR-1, ILT-2), LILRB2 (LIR-2, ILT-4), and killer immunoglobulin-like receptor 2DL4 (KIR2DL4). These receptors can transfer activation or inhibitory signals and are differentially expressed in B and T lymphocytes, peripheral NK cells, and monocytes. The binding of HLA-G to these receptors maintains maternal tolerance and promotes foetal development. HLA-G inhibits KIR2DL4 binding, which is similar to killer cell immunoglobulin–like receptor (KIR) recognition by classical HLA, which means that HLA-G has an important role in immune modulation[9].
HLA-G binds proteins coming from intracellular proteolysis, equivalent to classical HLA class I antigenic peptides[7]. The peptide-binding groove of HLA-G is different from that of HLA-E, regardless of similarities in the main chain conformations of their platform domains[9].
The process of transformation of normal cells to tumour cells is accompanied by changes in the level of expression of HLA classical and non-classical molecules[12]. These changes could be significant through the clinical course of the disease because HLA molecules are the bridge between tumour cells and the immune system. Expression patterns of HLA molecules are associated with the prognosis of particular cancer diseases, the disease-free determination, and the survival period[12].
Changes in the HLA expression appear as a result of structural gene abnormalities and defective regulation of HLA gene transcription and translation. The molecular mechanisms also include epigenetic processes arising with tumour development and progression, which are the basis of changes in the antigen processing machinery and tumour antigen expression in malignant cells[12].
In the process of evasion of immune surveillance, tumour cells manage to escape recognition by CTLs, which can destroy the tumour with the recognition of tumour-associated antigens (TAAs) on the tumour surface presented by classical HLA class I molecules. When tumour cells downregulate HLA class I molecules, or there is a complete loss of these molecules, CTLs are not recognised and their function is impaired because of low TAA expression[12].
Non-classical HLA-E and -G molecules are also an important part of immune surveillance. They provide an inhibitory signal to NK cells, after which they become unrecognisable, resulting in tumour escape[13].
The third mechanism by which tumour cells can escape the immune answer is attracting immunosuppressive regulatory T cells (Tregs) in the tumour microenvironment[13].
HLA-E is upregulated in cancers compared to healthy tissues, which suggests that this involves a gene transcription mechanism. The surface expression of HLA-E is regulated by post translational mechanisms, featuring conserved leader peptides, a transporter associated with antigen processing (TAP) and proteolytic enzymes. It is important to note that HLA-E is also able to present alternative peptides that are derived from tumour proteins, not only HLA class I leader peptides[14].
The peptide-loaded HLA-E molecule suppresses NK cell activity through ligation of the NK inhibitory receptor NKG2A[15] (Figure 2). This receptor transduces an inhibitory signal throughout two inhibitory immune-receptor tyrosine based inhibition motifs (ITIMs) after ligation with the peptide-loaded HLA-E molecule. This leads to suppression of the NK cytokine expression and cytotoxicity[15].
Figure 2:
HLA-E expressed on a tumour cell, interacting with immune cells.
Soluble HLA-E is also able to help tumour cells escape T and NK attacks by inducing apoptosis in these cells with the Fas/Fas ligand-mediated pathway[16].
Like HLA-E, HLA-F expression in tumour cells is also upregulated, which indicates a possible role in tumour development[17]. Elevated HLA-F levels in cancer lesions and peripheral blood are associated with poor survival in patients[18]. HLA-F/β2m-tetramers bind to the immune inhibitory receptors—killer cell Ig-like receptor, three Ig domains, and long cytoplasmic tail 2 (KIR3DL2), ILT2, ILT4 on NK cells, indicating the role in immune regulation[17] (Figure 3). HLA-F can also communicate with T cells by presenting peptides, but transportation of this non-classical molecule from the endoplasmic reticulum (ER) is slightly different; it is independent of tapasin and TAP. HLA-F export from the ER is possible because of a C-terminal valine residue and RxR motifs in the cytoplasmic tail that are Golgi localisation-related[17].
Figure 3:
HLA-F expressed on a tumour cell, interacting with NK and T cell receptors.
HLA-G expression has been studied the most compared to the other non-classical HLA molecules[11] and its relevance to cancer prognosis has been intensively investigated. HLA-G can decrease host antitumour immune responses in direct or indirect pathways[19]. The direct immune suppression caused by HLA-G includes inhibition of CTL, NK, and B cells, neutrophils, and dendritic cells, through interaction with ILT2 and ILT4 inhibitory receptors (Figure 4). HLA-G also alters dendritic cell (DC) maturation, migration, and antigen presentation[20]. Indirect immunoregulatory functions of HLA-G involve inducing tolerogenic Treg cells, suppressor T cells, myeloid-derived suppressor cells (MDSC), DC-10 (tolerogenic DC), and impairment of chemotaxis[21].
Figure 4:
HLA-G expressed on a tumour cell, interacting with immune cell receptors.
HLA-G expression can interfere with each of the three E phases (elimination, equilibrium, and escape) of tumour development. In the elimination phase, HLA-G expression can allow some tumour cells to evade the immune response from the host. The equilibrium phase is affected by control of the expression of HLA class II molecules by DC. In the escape phase, tumour cells express only HLA-G on their surface because they have lost molecules important for immune recognition[22]. Tumour progression is provided and affected by HLA-G, which enhances the expression of tumour metastasis-related factors (e.g., matrix metalloproteinases [MMPs])[21].
Tumour cells that have lost expression of classical HLA class I molecules are susceptible to NK-mediated lysis (the “missing self hypothesis”) and thus HLA-G plays an important role in the process of escaping NK-mediated death. Besides the described direct inhibitory action of NK cell inhibition, HLA-G impairs NK/DC crosstalk, which also inhibits NK cell cytotoxicity[22].
Regulation of the expression of HLA-G is an epigenetic, transcription, and post-transcription controlled process. The polymorphic 5′ variations in the upstream regulatory 5′-3′ region (URR) and untranslated regions (UTRs) in the HLA-G gene are reported to affect the HLA-G mRNA translation. They also affect the HLA-G expression by microRNA binding and modifying mRNA stability[22].
There are many reports that the expression of non-classical HLA-Ib molecules supports the progression of solid and malignant tumours by favouring escape of the tumour cells from immune surveillance, leading to a poor prognosis for the patient.
Expression of HLA-G in cancer was first studied in melanoma cell lines by Paul et al. in 1998. HLA-G expression in melanoma was observed on the cells’ surface, in a secretory form and in tumour-derived exosomes[23]. Derre et al. showed in their study that metastatic melanomas express HLA-E much less than primary tumours (10–20 percent vs. 30–70 percent)[24]. Also, serum levels of HLA-E were increased significantly, compared to healthy controls, in a study by Allard et al.[25].
In basal cell carcinoma (BCC), HLA-G expression is higher in aggressive tumours, compared to non-aggressive and nodular BCC[26]. Urosevic et al. observed that HLA-G expression decreases after radiotherapy. It is suggested that it induces switching off immune-tolerogenic molecules and tumours develop less efficient strategies for growth[27].
Esophageal squamous cell carcinoma (ESCC) is the most frequent esophageal cancer. In a study by Zheng et al., 70 percent of ESCC lesions were expressing HLA-G. Those authors also reported that sHLA-G was increased in the plasma of patients compared to that of healthy controls[28]. In another study by Yie et al., HLA-G expression was 90.9 percent in ESCC and associated with poor prognosis[29].
HLA-F expression in tumours was investigated by Lin et al. in non-small-cell lung cancer (NSCLC) lesions[18]. HLA-F expression was found in 24.1 percent of the NSCLC primary lesions and these patients had a poor prognosis, because of which HLA-F could be an independent prognostic factor for patients with this tumour type[18]. Lung adenocarcinomas are a subtype of NSCLC, increasing in incidence globally in both males and females, smokers and non-smokers. They are the cause for almost 50 percent of deaths attributable to lung cancer[30]. Yazdi et al. found that HLA-E is expressed in more than 70 percent of pulmonary adenocarcinoma cases. Low HLA-E expression is associated with a reduced risk of death and overall survival (OS)[31].
Hepatocellular carcinoma (HCC), the most common form of liver cancer, accounts for around 90 percent of cases[32]. In a study by Xu et al., HLA-F expression was significantly correlated with the degree of lymphatic or venous invasion in HCC[33].
Adenocarcinoma, the most common type of gastric cancer, is often diagnosed at its advanced stages, corresponding with a poor prognosis for patients[34]. Survival rates of HLA-G negative patients (44 percent for five years) are much higher than those of HLA-G positive patients (11 percent). This is observed in all tumour stages, but it is most significant in Stage II. In a study by Murdaca et al., HLA-G positive tumour cells were present in 25.5 percent of the cases[35]. HLA-G expression is associated with the histological grade and clinical stage of the disease and OS in a study by Zeestraten et al., and was found in 71 percent of primary gastric carcinomas[36].
In the case of loss or downregulation of classical class I HLA molecules, HLA-E and -G determine the prognosis of colon cancer patients in the Zeestraten et al. study[36]. HLA-G expression was found in 70.6 percent of tumour specimens in a study by Guo et al.[37]. Ye et al. observed HLA-G expression in 64.6 percent of primary colorectal carcinomas[38].
Breast cancer is currently the most common tumour in females[39]. In a study by Meiner et al., HLA-G was observed in 66 percent of primary lesions and correlated with tumour size, depth of invasion, lymph nodal metastasis, clinical stages, and OS[40]. In another study by Kruijf et al., HLA-E and -G are similarly expressed in 50 percent and 60 percent of patients, respectively[41]. Harada et al. provided the first evidence that in Stage II breast cancer, HLA-F positive patients have poorer postoperative outcome compared to those who are HLA-F negative[42].
In cervical cancer, HLA-E expression increases with lesion progression. High expression results in a lower rate of disease-free survival of the patients. OS of patients is influenced by HLA-E expression and in a Gooden et al. study it was expressed in 83.7 percent of cervical cancers[43]. Increased HLA-G expression is also observed in human papillomavirus (HPV)-induced cervical cancers and causes T cell dysfunction[44].
Ovarian cancer is the most lethal female reproductive cancer[45]. OS is strongly influenced by HLA-E. CTLs infiltration is associated with better survival only when HLA-E expression is low. In 89.4 percent of ovarian cancer cases, Gooden et al. observed expression of HLA-E[43]. HLA-G was detected in 50 percent of tumour lesions in a Menier et al. study[40]. In another study by Babay et al., HLA-E and -G are also highly expressed in ovarian carcinoma tissues (96.8 percent and 72.4 percent)[46]. There is also an association between the HLA-EG (*01:03) allele and the disease. Women carrying this allele are more susceptible to serous ovarian cancer[47].
Almost 80 percent of endometrial cancers are endometrioid endometrial adenocarcinoma (EEA), diagnosed early with good prognosis. However, there is still a subset of patients who fail to reach five-year overall survival, even for those with low-grade or early-stage EEA[48]. In Bijen et al.’s study, upregulation of HLA-G was detected in 39.8 percent of patients, and in Barrier et al.’s study this value was 55 percent. The degree of expression of HLA-G correlates with the stage of the disease in the Barrier et al. study[49].
Glioma is characterised by its lethality and profound genomic alterations (isocitrate dehydrogenase [IDH] mutation, 1p19q co-deletion, mutation in the telomerase reverse transcriptase [TERT] promoter)[50]. Wu et al. observed differences in HLA-E expression between low-grade and high-grade glioma[51]. They found that HLA-E expression was higher in diffuse astrocytomas, compared to oligodendrogliomas. Wu et al. also found a correlation between lower HLA-E mRNA expression and better progression-free survival (PFS) and OS in low-grade glioma patients[51]. In a recent study, Bukova et al. found that a 14 nucleotide insertion in the homozygous and heterozygous state of the HLA-G 14 bp ins/del polymorphism is more frequent in patients with glioma, and higher serum levels of sHLA-G are negatively associated with OS[52].
Immune effector cell functions are generally suppressed in chronic lymphocytic leukemia (CLL)[53]. Nuckel et al. observed that HLA-G expression of leukemic cells in B cell CLL is between 1 and 54 percent[54]. PFS is lower in patients in whom tumour cells express below 23 percent of HLA-G. These patients have a significantly higher survival rate compared with HLA-G positive patients. The decrease in survival rate is associated with an immune response deficiency, a CD4/CD8 T cells ratio, IgG reduction, and increase of sHLA-G molecules[53, 54].
In a study by Wagner et al., HLA-EG (*01:03) allele in patients with CLL is found to be associated with a high level of sHLA-E and increased mortality rates[53]. HLA-EG (*01:03) homozygous genotype is found to be associated with a higher risk of developing CLL and high levels of sHLA-E[53]. HLA-G serum levels are also increased, especially in patients with B and T acute lymphoid leukemia (ALL). In acute myeloid leukemia (AML), patients’ expression of HLA-G is 18.5 percent[55]. In a study by Lin, CLL patients with the HLA-G del/del genotype have shorter survival, compared to patients with the ins/del or ins/ins genotypes[20].
Expression of classical HLA molecules and HLA-G are altered in non-Hodgkin’s lymphoma (NHL) and corresponds to transformation, tumour relapse, or both. In a study by Sebti et al., increased sHLA-G levels were found in 65 percent of B-NHL and 58 percent of T-NHL patients[56].
Diffuse large B-cell lymphoma (DLBCL) is known as the most frequent and severe type of lymphoma, but HLA-G expression is found to be weak. However, Kupnicka et al. established that the survival rate correlates with HLA-G expression[57]. DLBCL patients with the del/del genotype have a shorter OS than those with the ins/del or ins/ins genotypes. The HLA-G genotype 725C/C in the 5′ URR has higher OS than those with HLA-G 725C/G/T[57].
In classical Hodgkin’s lymphoma, the expression of HLA-G was studied by Diepstra et al.[58] and Caocci et al.[59]. HLA-ER (*01:01) genotype is associated with a lower risk of Epstein-Barr virus-related classical Hodgkin’s lymphoma[60].
The significant role of non-classical HLA class Ib molecules in cancer has been pointed out in many studies. They function as indicators of cellular transformation, directly or through the presentation of endogenous antigens expressed in tumours. Immune evasion strategies of these molecules also include the generation of soluble versions that block the corresponding receptors on lymphocytes. Because of that, these molecules are considered a future target for anticancer immunotherapy. In recent studies, HLA-G is considered to be a new immune checkpoint in cancer[61, 62]. Although the relationship between non-classical HLA-Ib molecules and tumourogenesis still needs to be clarified, there are new drugs that can downregulate the expression of HLA-G in tumours. This strategy could perhaps restore the immune cells’ capability to destroy tumour cells. Palma et al. used CRISPR/Cas9 gene editing to block HLA-G expression in renal cell carcinoma and choriocarcinoma cell lines and were able to completely silence the gene expression[63].
There is a need for more studies in order to clarify the complex mechanisms with which HLA-Ib molecules control immune evasion in cancer, but the first steps in development of novel immunotherapies have already been taken.