Human leukocyte antigen-G overexpression predicts poor clinical outcomes in low-grade gliomas

Highlights

• High HLA-G expression was found to serve as a potential biomarker for predicting aggressive tumor grades.

• We found HLA-G expression could also serve as a potential predictor of the histological subtype in low-grade gliomas.

• Elevated HLA-G expression was first indicated as an independent predictor of poor clinical outcomes in patients with low-grade gliomas.

Abstract

Overexpression of human leukocyte antigen-G (HLA-G), a non-classical major histocompatibility complex class-I molecule associated with immunosuppression, has been reported in various human malignancies. In the present study, we examined the role of HLA-G in gliomas. Clinical characteristics, mRNA expression microarrays and follow-up data pertaining to 293 patients with histologically confirmed gliomas were analyzed. The expression levels of HLA-G were compared between different grades of gliomas and correlated with progression-free survival (PFS) and overall survival (OS) to evaluate its prognostic value. We found that HLA-G was overexpressed in gliomas as compared to that in normal brain tissue samples (− 1.288 ± 0.265). The highest expression levels were in glioblastomas (GBMs), anaplastic gliomas (AGs) and low-grade gliomas (LGGs), in that order (0.328 ± 0.778, 0.176 ± 0.881, − 0.388 ± 0.686, respectively). Significant inter-group differences were observed between low-grade and high-grade glioma tissues (p < 0.001 and p < 0.001, t-test, AGs and GBMs, respectively). More astrocytoma patients exhibited increased HLA-G expression as compared to other LGG patients (p = 0.004, Chi-square test). Significant differences were observed with respect to PFS and OS (p = 0.009 and 0.032, log-rank test, for PFS and OS, respectively) between the high- and low-expression subgroups in patients with LGGs. On Cox regression analysis, overexpression of HLA-G appeared to be an independent predictor of clinical outcomes (p = 0.007 and 0.026, for PFS and OS, respectively). Our results suggest that HLA-G expression may serve as a potential biomarker for predicting aggressive tumor grades of gliomas and for histological subtype of LGGs. Elevated HLA-G expression could serve as an independent predictor of poor clinical outcomes in patients with low-grade gliomas.

Introduction

Glioma, the most common and most lethal primary tumor of the central nervous system, accounts for more than 70% of all brain tumors (DeAngelis, 2001, Ohgaki and Kleihues, 2005). Gliomas are typically very difficult to resect completely, and often recur after surgery. Thus, they represent a challenge in neuro-oncology (DeAngelis, 2001, Ohgaki and Kleihues, 2005). For the most frequent and aggressive type of glioma, glioblastoma multiforme (GBM), the median duration of survival after standard treatment with maximal surgical resection and subsequent radio-chemotherapy is only 14.6 months, and the two-year survival rate is below 30% (Stupp et al., 2009, Stupp et al., 2005). Although recent studies have contributed to the knowledge of gliomas, yet the molecular basis of this cancer needs to be fully elucidated.

Evasion of immune surveillance and suppression of the immune system are hallmarks of cancer. Schreiber first proposed and improved the concept of cancer immunoediting, which results from the interaction between innate and adaptive immune cells in the tumor microenvironment, local inflammation and the immune response of the host (Dunn et al., 2002, Vesely et al., 2011, Vesely and Schreiber, 2013). The concept of cancer immunoediting refers to the complex interaction between tumor cells and the immune system, and to the dual role of immune system in the development of cancer, i.e., both tumor suppression and tumor progression (Manjili, 2011, Mittal et al., 2014, Schreiber et al., 2011). Up to 20% of all clinical tumors are related to an underlying chronic inflammation, wherein the immune cells stimulate tumor growth and progression (Grivennikov et al., 2010). The immunosuppressive property of human glioma is well-characterized (Wiendl et al., 2002). However, the underlying mechanisms which allow it to evade the host immune system are not completely understood.

The major histocompatibility complex (MHC) refers to a group of coding genes for the cell surface proteins which modulate the immune response. The human MHC is also called human leukocyte antigen (HLA) complex. HLA gene is located at the 6p21.31 region on the short arm of human chromosome 6. It is a 4000 kb gene complex which consists of hundreds of gene groups. It is known for its maximum allelic polymorphisms and is closely related to the regulation of human immune system (Reipert, 2014, Rouas-Freiss et al., 2003). Based on the gene sequence, it is generally divided into three groups: HLA-I (e.g., HLA-A, HLA-B, HLA-C, HLA-E, and HLA-G), HLA-II (e.g., HLA-DR, HLA-DP, and HLA-DQ) and HLA-III (Nakajima et al., 2010). Altered expression of HLA has been documented in a variety of inflammatory or malignant conditions. The association of HLA expression with certain human diseases has also been demonstrated. For instance, in the early 1970's a strong correlation between HLA-B27 and ankylosing spondylitis was detected in more than 90% patients (Schlosstein et al., 1973). Similar association has also been demonstrated between HLA and human cancers (e.g. cervical cancer and breast cancer) (Jia et al., 2015, Kochan et al., 2013).

The association between glioma and HLA gene expression has been reported. A positive correlation between HLA-E expression and survival status was found in patients with glioblastoma (Kren et al., 2011). In a recent study, overexpression of HLA-DR was shown to be associated with poor prognosis in glioma patients (Diao et al., 2015). In our previous study, we identified an association between HLA-G expression and radiologic morphology in low-grade gliomas (LGGs) (Wang et al., 2015). To the best of our knowledge, no study has yet investigated the role HLA-G plays in different grades of gliomas. We assessed the mRNA expression level of HLA-G in glioma patients and explored its potential correlation with different clinicopathological parameters.