Introduction
Glioblastoma is a highly malignant type of glioma, and accounts for over 50% of emerging glioma cases [
1]. Treatment with postoperative chemoradiotherapy is currently the standard of care for glioblastoma. Even though some patients may benefit from the Stupp regimen [
2], the overall prognosis remains poor [
3], with less than 10% survivorship at 5 years [
4]. At present, prognostic predictions for patients with glioblastoma are mainly made based on patient age, Karnofsky Performance Status (KPS), prior treatment, resection range, methylation of O6-methylguanine-DNA methyltransferase (
MGMT) promoter methylation status, isocitrate dehydrogenase genes (IDH), telomerase reverse transcriptase (
TERT), and other molecular markers, such as alpha thalassemia/mental retardation syndrome X-linked (ATRX) gene [
5,
6]. It is of paramount significance to make overall assessment of patients, and to evaluate potential high-risk factors that affect patient prognosis, to improve the timely adjustment of treatment methods and the accuracy of prognostic assessments.
The World Health Organization Classification of Tumors of the Central Nervous System, fifth edition (WHO CNS5 classification; CNS5) states that an IDH-wildtype diffuse astrocytic glioma in an adult with microvascular proliferation or necrosis or
EGFR gene amplification or + 7/− 10 chromosome copy number changes or
TERT promoter mutation should be diagnosed as GBM (CNS5), even if the histological grade was considered low [
1]. There are different driver genes, molecular characteristics, and clinical prognosis associated with either IDH mutant or IDH-wildtype glioblastoma [
7]. GBM (CNS5) is considered as an independent genotyping for diagnosis based on the fifth edition of the 2021 World Health Organization classification of tumors of the central nervous system [
1], thus further advancing the role of molecular neuropathology in CNS tumor classification. Compared with the IDH mutant type, IDH wild type glioblastoma (IDHwt GBM, CNS4) exhibits higher invasiveness, has a poor prognosis, with a median patient survival time ranging from 6 to 15 months [
8]. As studies have shown, 30 to 50% of IDHwt GBM (CNS4) demonstrates methylation of the
MGMT promoter, which is associated with favorable clinical responses to TMZ, and is considered to be a poor prognostic factor [
9]. However, the utility of this biomarker may be limited by acquired drug resistance, and disease prognosis still varies greatly. The prognostic value of methylation of
MGMT promoter [
10], the
TERT promoter, and
EGFR [
11] in GBM (CNS5) remains controversial. It also remains unclear whether heritable factors can contribute to risk stratification for patients [
12], and there are likely other factors remaining to be identified that can stratify prognosis for patients with GBM (CNS5). Therefore, there still must be additional reliable biomarkers developed for patient stratification and disease prognosis of patients with GBM (CNS5).
During the progression of glioblastoma, factors such as the tumor microenvironment, and infiltration of non-tumor cells and immune cells influences the gene expression and transcription types of glioblastoma [
13], and can result in the interconversion of molecular subtypes. Neutrophils are important members of the tumor microenvironment. Neutrophils exhibit tumor-promoting activity by inducing angiogenesis [
14‐
17], inhibiting T cell activation (immunosuppression) [
18‐
22], inducing genetic instability [
23‐
25], and maintaining tumor cell proliferation [
26‐
29]. Tumor-associated neutrophils (TANs) are also prognostic markers for patients with tumors [
30‐
32], and are closely related to the prognosis of gastric carcinoma [
30], breast cancer [
33], cholangiocarcinoma [
34] and urothelial carcinoma [
32]. However, there are few studies on TANs in patients with glioma diagnosed by WHO CNS5 classification, especially for patients with GBM (CNS5), of which the prognostic value is currently unclear. What’s more, at present, studies on hematologic markers of glioma mostly center on preoperative peripheral blood samples [
35], and are often disturbed by many factors such as preoperative stress and postoperative infection, which can greatly limit the representation of the real postoperative condition of glioma. Correlative research on the influence of peripheral blood neutrophils on the overall survival of patients with glioma before postoperative radiotherapy has been reported less now, and its influence on the prognosis of glioma is of certain research value.
In the present study, RNA-sequencing (RNA-seq) expression profiles and clinical data from the TCGA database were used to measure the abundance of TANs in the tumor microenvironment by the CIBERSORTx algorithm, and to evaluate the relationship between TANs and clinical prognosis. Moreover, Gene Set Variation Analysis (GSVA) enrichment analysis was performed to explore differences in biological characteristics between the high and low TANs groups, and the CGGA database was used for external verification. In addition, a retrospective analysis was made on the levels of peripheral blood neutrophils before radiotherapy for patients with GBM (CNS5), and the prognostic significance of this metric was determined for GBM (CNS5).
Discussion
While the integrated WHO CNS5 classification has advantages for guiding clinical diagnosis compared with previous simple histological diagnosis, it also further increases the heterogeneity of GBM cohorts and sets higher requirements for evaluating prognosis. Despite some research efforts in IDHwt GBM (CNS4), it remains unknown as to whether TANs could serve as a prognostic biomarker in patients diagnosed as GBM (CNS5). Patients diagnosed with GBM (CNS5) were included in this study, and we found that high TANs level remains an independent prognostic factor for poor OS of GBM (CNS5) [TCGA cohort: HR (95%CI) = 1.621(1.004–2.619); CGGA cohort: HR (95%CI) = 1.526(1.029–2.323)]. Moreover, the level of TANs infiltration was significantly correlated with the expression of apoptotic genes, including
TNFRSF10C and
TNFRSF10D, and with expression of the neutrophil marker genes
CXCR1,
S100A9. In order to investigate the effect of peripheral blood neutrophils on the prognosis of GBM (CNS5), data from 143 patients was analyzed. Peripheral blood neutrophils before radiotherapy was an independent prognostic factor for OS [HR (95%CI) = 2.098 (1.055–4.172)]. Neutrophils are present in most solid tumors microenvironments [
44‐
49], and are important non-malignant cells found in the tumor microenvironment [
50]. Neutrophil infiltration influences the response to different anticancer therapies, and high neutrophil infiltration is associated with a poor response to radiotherapy [
51]. In this study, a subgroup analysis of 126 patients who received radiotherapy confirmed that high TANs infiltration was associated with shorter OS [HR (95%CI) = 1.753 (1.047–2.936)].
Most current studies on the prognostic significance of TANs do not agree on the relevant biomarkers of neutrophils, which may result in a bias in prognostic estimates. By analyzing three groups of operative specimens of patients with gastric cancer who received total or partial gastrectomy independently at two medical centers, Zhang et al. found that high infiltration of TANs in gastric tissue suggests a better prognosis [
30]. Zhao et al. [
52] demonstrated that high infiltration of TANs in gastric tissue suggests a poor prognosis. Causes for this difference may be that CD66b was used to mark neutrophils in the former study, while CD15 was used to mark neutrophils in the latter study. CD15 can be expressed not only in neutrophils, but also in monocytes, eosinophils, and tumor cells, among other cell types. As a consequence, the RNA-seq data of TCGA and CGGA datasets were analyzed in the present study by CIBERSORTx in an exploratory way, to infer the neutrophil infiltration levels and avoid potential biases introduced by evaluating only specific neutrophil markers.
TANs are involved in malignant transformation and angiogenesis in numerous preclinical and clinical studies [
53‐
57]. Arora et al. demonstrated that higher levels of
S100A8 (median survival: High vs. Low = 12.73 months vs. 15.1 months, respectively;
P = 0.0009) and
S100A9 (median survival: High vs. Low = 12.67 months vs. 15.03 months, respectively;
P = 0.0005) gene expression was associated with poor prognosis in GBM (CNS4) patients [
58]. By releasing angiogenic factors including S100A8 and S100A9, as well as activating vascular endothelial growth factors A (VEGFA) in the extracellular matrix and MMP9, tumor angiogenesis was maintained by neutrophils [
14‐
17]. This angiogenic effect was also found in hepatocellular carcinoma, gastric cancer, and nasal carcinoma [
59‐
61].
S100A8/S100A9 co-expression in hepatocellular carcinoma cells promotes malignant progression by induction of ROS, down-regulation of p38 MAPK signaling, cell survival, and resistance to tumor necrosis factor (TNF)-α-induced apoptosis [
62]. Li et al. report that high expression of
MMP9 is associated with the pathological grading of gliomas and predicts poor prognosis [OS: HR (95%CI) = 1.171(1.018–1.346), PFS: HR (95%CI) = 1.146(1.012–1.299)]. Patients with lower
MMP9 expression are more likely to benefit from TMZ treatment regardless of
MGMT-methylation status [
63]. Furthermore, neutrophil can stimulate dormant cancer cells through release of MMP9 which can produce epitopes that bind to tumor integrins and trigger the proliferation of cancer cells [
26,
27]. It has been reported that CXCR1 mRNA expression is significantly higher in patients with glioma than in normal individuals [
64]. The tumor-promoting activity of neutrophils was related to growth factors and chemotactic factors [
53,
65,
66], and CXCR was involved in promoting neutrophil maturation, survival, and recruitment [
18,
67‐
69]. TNFRSF10C is a protein that belongs to the TNFRSF family that binds to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) and inhibits intracellular apoptotic signaling pathways [
70].
TNFRSF10D expression is associated with prostate cancer and TNFRSF10D is a direct effector
p53 and ERK signaling pathways [
71]. Although the prognostic value of TNFSF10C and TNFRSF10D has not been previously investigated in glioma, these proteins have the potential to be used as novel biomarkers.
Neutrophils are classical congenital immune cells that are important members of the tumor immune micro-environment. Neutrophils in peripheral blood and tissues are of the same origin [
72,
73]. A clinical study of 1233 patients undergoing radical radiotherapy demonstrated a significant association between elevated blood neutrophil counts and reduced 3-year OS [
74]. In view of current glioma studies, the clinical studies to explore the prognostic value of neutrophils have mostly focused on preoperative peripheral blood samples, and most of evaluated the ratio of neutrophils to lymphocytes [
35], which may not truly reflect the prognostic value of peripheral blood neutrophils, given that this index is susceptible to lymphocyte interference. K. Takakura et al. [
75] demonstrated that NLR was significantly associated with high density CD20+ lymphocytes (
P = 0.031) and CD163+ macrophagocytes (
P = 0.023), but not with CD66b + neutrophils (
P = 0.397). Also, the correlation between neutrophils and prognosis may also be influenced by the location in tumors. Immunohistochemical studies on operative specimens of esophageal squamous carcinoma found that 5-year rates of DFS and OS were 20 and 26.7%, respectively, in patients with increased CD66+ intratumoral neutrophils, but 51.1 and 55.5%, respectively, in patients with decreased CD66+ neutrophils, suggesting that CD66+ neutrophils are an independent prognostic factor of DFS (HR = 2.174 (1.249–3.784),
P = 0.006) and OS (HR = 1.858 (1.038–3.325),
p = 0.037). No prognostic significance of peritumoral neutrophils was noted [
76]. The correlation between neutrophils and prognosis was also influenced by the time of specimen collection, especially peripheral blood specimens. Whereas most patients with glioma are treated with surgery, there may be differences in tumor burden status after operation compared with pre-operation. Meanwhile, neutrophil infiltration was shown to associate with radiotherapy sensitivity [
77]. Presently, there are few reports on the association with postoperative peripheral blood neutrophil before radiotherapy and OS of GBM (CNS5). Therefore, the time point before radiotherapy used in this study, with strict inclusion and exclusion criteria to avoid the influences brought by postoperative surgical stress or postoperative infection, may ensure better evaluation of the effects of the overall immune status of patients with glioma before radiotherapy. Our results showed that the level of peripheral blood neutrophils before radiotherapy was an independent risk factor that affects the prognosis of patients with GBM (CNS5) suggesting that immune status before radiotherapy affects the survival of patients with glioma.
There are several limitations of this study that need to be discussed. First, this study is an observational study, and it is unknown to what extent unmeasured confounders may have influenced the results. In order to reduce the interference of confounding factors, the study used multivariate analysis to adjust as many confounding factors as possible. Additionally, E-values were calculated to assess the impact of unmeasured confounders. However, confounding factors such as the precise types, dose, course and comedication of chemotherapy and radiotherapy were not fully documented in the database, and were therefore unable to be evaluated in this study. Secondly, levels of TANs evaluated by CIBERSORTx is calculated by mRNA-seq, which lacks data validation on a cell-by-cell level. Furthermore, the interactions between blood neutrophils and TANs and the tumor-promoting or tumor-inhibiting mechanisms of neutrophils were not been explored in depth. The results of this study need to be validated by prospective multi-center randomized trials with a larger patient population in the future.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.