In this exploratory analysis, we used the highly sensitive PEA technology to analyze plasma samples collected from glioma patients before therapy and compared the results with age- and sex-matched control patients with meningiomas. The details of the PEA technology are described in our recent review (7). Our laboratory has previously frequently used other proteomic technologies, such as mass spectrometry (MS) and ELISA for biomarker discovery (12–15) and we are in a good position to compare PEA, ELISA and MS technologies. Mass spectrometry multiplexing allows simultaneous analysis of 50–100 proteins by multiple reaction monitoring (MRM) (12), significantly less than PEA (> 3,000 proteins measured in parallel). Also, MS suffers from low precision (CVs ~ 30–60%) for direct serum analysis. PEA precision is much better at 10–15%. Individual ELISA assays usually require 100 ul sample volume per assay (or, by extrapolation, 100 mL for 1,000 assays, a volume which is unrealistic in clinical practice). PEA requires only 15uL of sample for > 3,000 assays. PEA is more specific than single ELISAs using the same antibodies due to dual recognition of both antibodies and nucleotides (7). To summarize, PEA is a new, ultra-sensitive, highly specific and precise technology that requires no sample pre-treatment and can be highly multiplexed with minimal sample volume requirements. These properties qualify PEA as superior to LC/MS/MS and single ELISAs for liquid biopsy-based biomarker discovery applications.
To our knowledge, our work is the first demonstration of the use of this technology for glioma biomarker discovery. Most of the identified candidates have been shown in the past to have some relationship to gliomas/brain tumors, suggesting that our findings are unlikely to represent false discovery and are quite relevant to gliomas.
In this study, we found associations between glioma and several candidate plasma biomarkers. GFAP had the highest discriminatory potential between gliomas and meningiomas (Fig. 1), followed by several other candidates (Table 2).
The combination of just two markers (GFAP and FABP4) results in a ROC curve with an AUC of 0.98 (Fig. 5). Due to the small number of patients, these data should be interpreted with caution until they are independently validated with a larger number of patients.
For a long time, the structural role of GFAP in astrocytes (the main type of glial cells in the central nervous system (CNS)) was acknowledged (16). More recently, GFAP has been demonstrated to be involved in numerous astrocyte functions. (17, 18). After brain surgery, in the early stages of recovery, it has been found that GFAP increases in response to astrocytic reaction to the brain injury. (19)
Recent emerging evidence supports the involvement of GFAP in glioblastoma multiforme (GBM). Serum GFAP was significantly increased in Grade 4 glioma and was detected in 63% of all Grade 4 patients compared to 13% of healthy controls, (20) indicating that glioma patients had higher GFAP levels, in accordance with our findings. (21, 22) Serum GFAP correlates with invasiveness and malignancy in astrocytomas and high-grade gliomas, compared to lower grade gliomas (16, 17). Thus, GFAP is a potential biomarker and possible therapeutic target for gliomas.
Another well-known glioma biomarker is NEFL (Neurofilament light polypeptide) also known as neurofilament light chain, a potential tumor suppressor (23). NEFL is involved in a variety of common human cancers such as breast, prostate, and head and neck cancers (24). Serum NFL concentration was higher in patients with CNS tumors with disease in progression versus CNS tumors with stable disease (25). In addition, serum NFL was higher in patients with metastatic solid tumors with known brain metastases than in those with metastatic tumors with no brain metastases (25).
Fatty acid-binding protein 4 (FABP4) is one of ten intracellular small molecular weight proteins that make up the FABP family (26, 27) and is found in adipose tissue, peripheral macrophages, and microglia (28). Furthermore, it is not found in normal brain blood vessels, although it has been found in certain endothelial cells or tumor cells in benign and malignant meningiomas (10, 29). FABP4 has a role in carcinogenesis in meningiomas by stimulating cell proliferation in a cell type-independent way. In this connection, rapamycin, a well-known inhibitor of the mTOR pathway, which is a master regulator of cell growth and metabolism, inhibits FABP4 production in endothelial cells (30). FABP4 is expressed in a significantly higher percentage of GBMs in comparison to both normal brain tissues and lower-grade glial tumors. Data suggest that FABP4 may play a role in angiogenesis associated with GBMs. Another study analyzed FABP4 expression in a cohort of paraffin-embedded meningioma specimens by immunohistochemistry and double immunofluorescence analyses (9). These results demonstrate that FABP4 is commonly expressed in meningioma vascular endothelial cells while tumor cell expression of FABP4 is primarily observed in anaplastic meningiomas. A combination of FABP4 immunostaining with histopathologic grading might provide a more accurate prediction of the biological behavior of meningiomas than histopathologic grading alone (29, 30).
MMP3 is involved in cell migration. PDGFR-alpha induces MMP3 gene expression and increased cell proliferation and cell migration upon stimulation by platelet-derived growth factor (PDGF) The induction of expression of MMP3 in glioblastoma cells triggers a cascade of gene expression, resulting in decreased cell adhesion and migration (31–33).
High plasma interleukin-8 (IL-8) was associated with short progression-free survival in newly diagnosed patients with Glioblastoma Multiforme (GBM). IL-8 was mostly secreted and expressed by mesenchymal GBM cell lines and expressed by vascular cells and immune cells. High plasma IL-8 at surgery was associated with short OS in newly diagnosed GBM (34).
Prokineticin 1 (PROK1) is a relatively conserved hypoxia-induced small protein mostly known for its ability to induce proliferation, migration and fenestration in the context of angiogenesis, pain perception and neurogenesis (35). In a recent glioma biomarker discovery study, PROK1 expression in glioma tissue was found to be significantly higher than that in normal tissue (P < 0.05) with higher expression in high grade gliomas compared to low grade ones. Interestingly, glioma patients with higher PROK1 expression had a significantly shorter progression-free survival time further suggesting putative prognostic value of PROK1 in human gliomas (36).
Aldehyde dehydrogenase 3A1 (ALDH3A1) is a brain expressed (midbrain, hypothalamus) drug metabolizing protein that has been implicated in temozolomide-mediated resistance of glioma tumors (37). Down-regulation of aldehyde dehydrogenase activity via Wnt/beta-catenin signaling blockade has been recently proposed as a possible means of reducing ALDH-mediated resistance to temozolomide (the chemotherapeutic drug currently used as standard treatment for glioblastoma). (38)
Interleukin 12 is a broad acting cytokine and a potent proinflammatory cytokine that is expressed by activated macrophages and acts as an essential inducer of Th1 responses. It is expressed in human brain (basal ganglia, hypothalamus, midbrain) and its expression in the context of glioma pathology it has been found to be mediated via insulin-like growth factor (IGF)-1 signaling networks (39). Of note, a recent comprehensive transcriptome analysis of GBM microarray and data from The Cancer Genome Atlas (TCGA) database revealed interleukin 12 as a direct target certain microRNAs involved in GBM development (39).
Indoleamine 2, 3-dioxygenases 1 (IDO-1) is a tryptophan catabolizing enzyme that permits the conversion of tryptophan into kynurenine, a pathway with clear immunosuppressive functions. IDO1 targeting has emerged as a novel therapeutic opportunity in modern cancer immunotherapy (40). Interestingly, while IDO1 is not generally expressed in the adult central nervous system, most of GBM patients do express significant levels of IDO1 (41). Accumulating evidence highlight an important putative role for IDO1 in regulating tumor immunological escape in brain tumors. (42, 43).
For the other identified candidate biomarkers (EDDM3B, LMOD1, GP2, SPINT3, CTRL, OXT) there is no specific literature linking them to gliomas or other brain tumors.
Similar to our own findings, the incidence of meningiomas is higher in females than males (Fig. 3) (29). One of the reasons could be that in females, endogenous sex hormone levels are significantly higher during childbearing years (44). Also, according to one of the previous studies, the risk of meningioma has increased because of reproductive variables or using exogenous sex hormones (45). Also, according to previous studies, males have a higher incidence for glioma than females although this is not important in the pathological diagnosis and clinical treatment (46).
There are important limitations to this study, mainly due to the small number of patients. Consequently, these results need to be further validated in a larger cohort of samples. Specifically, the nature of the interaction between GFAP and FABP4 and its role in glioma diagnosis and prognosis requires further study.