Synthesis and Significance of Arachidonic Acid, a Substrate for Cyclooxygenases, Lipoxygenases, and Cytochrome P450 Pathways in the Tumorigenesis of Glioblastoma Multiforme, Including a Pan-Cancer Comparative Analysis

Simple Summary Glioblastoma multiforme is a brain tumor with a very unfavorable prognosis, where the vast majority of patients do not survive a year after diagnosis. One line of research that may help in designing more successful therapeutic approaches is the synthesis and metabolism of arachidonic acid, which is then converted into a large number of different lipid mediators, including prostaglandins and leukotrienes (by cyclooxygenases and lipoxygenases, respectively). In this paper, we discuss the synthesis of arachidonic acid in glioblastoma multiforme tumors as well as the significance of lipid mediators synthesized from arachidonic acid, which can increase the proliferation of glioblastoma multiforme cancer cells, cause angiogenesis, inhibit the anti-tumor response of the immune system, and be responsible for resistance to treatment. Abstract Glioblastoma multiforme (GBM) is one of the most aggressive gliomas. New and more effective therapeutic approaches are being sought based on studies of the various mechanisms of GBM tumorigenesis, including the synthesis and metabolism of arachidonic acid (ARA), an omega-6 polyunsaturated fatty acid (PUFA). PubMed, GEPIA, and the transcriptomics analysis carried out by Seifert et al. were used in writing this paper. In this paper, we discuss in detail the biosynthesis of this acid in GBM tumors, with a special focus on certain enzymes: fatty acid desaturase (FADS)1, FADS2, and elongation of long-chain fatty acids family member 5 (ELOVL5). We also discuss ARA metabolism, particularly its release from cell membrane phospholipids by phospholipase A2 (cPLA2, iPLA2, and sPLA2) and its processing by cyclooxygenases (COX-1 and COX-2), lipoxygenases (5-LOX, 12-LOX, 15-LOX-1, and 15-LOX-2), and cytochrome P450. Next, we discuss the significance of lipid mediators synthesized from ARA in GBM cancer processes, including prostaglandins (PGE2, PGD2, and 15-deoxy-Δ12,14-PGJ2 (15d-PGJ2)), thromboxane A2 (TxA2), oxo-eicosatetraenoic acids, leukotrienes (LTB4, LTC4, LTD4, and LTE4), lipoxins, and many others. These lipid mediators can increase the proliferation of GBM cancer cells, cause angiogenesis, inhibit the anti-tumor response of the immune system, and be responsible for resistance to treatment.

In humans, there are six representatives of iPLA 2 : iPLA 2 β to iPLA 2 η [16]. All of these enzymes belong to the GVI PLA 2 . They are activated by ATP [27], and their activity is independent of Ca 2+ levels and reduced by calmodulin [28]. Enzymes in this group show different specificities for cleaving fatty acids from phospholipids at the sn-2 position. Depending on the enzymes, they show a higher ability to release a given fatty acid, e.g., oleic acid C16:1n-9 [27] or ARA C20:4n-6 [29].
Seventeen different groups of PLA 2 have been classified to date, which includes sPLA 2 [16]. Some sPLA 2 groups consist of only the sPLA 2 found in the venom of snakes, insects such as bees, and scorpions [16,[30][31][32]. In humans, there are nine representatives of sPLA 2 [16]. These enzymes cleave fatty acids from phospholipids at the sn-2 position without showing specificity to a particular fatty acid [16,33]. Once secreted into the intercellular space, sPLA 2 not only cause the release of ARA C20:4n-6 but can also activate their receptor PLA 2 R1 [34].
In humans, there are six representatives of iPLA2: iPLA2β to iPLA2η [16]. All of these enzymes belong to the GVI PLA2. They are activated by ATP [27], and their activity is independent of Ca 2+ levels and reduced by calmodulin [28]. Enzymes in this group show different specificities for cleaving fatty acids from phospholipids at the sn-2 position. Depending on the enzymes, they show a higher ability to release a given fatty acid, e.g., oleic acid C16:1n-9 [27] or ARA C20:4n-6 [29].
Seventeen different groups of PLA2 have been classified to date, which includes sPLA2 [16]. Some sPLA2 groups consist of only the sPLA2 found in the venom of snakes, insects such as bees, and scorpions [16,[30][31][32]. In humans, there are nine representatives of sPLA2 [16]. These enzymes cleave fatty acids from phospholipids at the sn-2 position without showing specificity to a particular fatty acid [16,33]. Once secreted into the intercellular space, sPLA2 not only cause the release of ARA C20:4n-6 but can also activate their receptor PLA2R1 [34].
After fatty acids are cleaved from phospholipids by PLA2, free fatty acids are formed, most commonly ARA C20:4n-6 and lysophosphatidylcholine (LPC) if PC was the reaction substrate ( Figure 2). LPC can then be converted to lysophosphatidic acid (LPA) by the action of enzymes with lysophospholipase D (lysoPLD) activity [35,36]. An extracellular enzyme with lysoPLD activity is autotaxin (ATX)/ENPP2 [35,36]. Importantly, if the substrate for PLA2 is phosphatidic acid (PA), then LPA is formed directly [37]. LPA is a lipid mediator that acts through its six receptors (from lysophosphatidic acid receptor 1 (LPAR1) to LPAR6) [38]. Figure 2. Importance of PLA2 in metabolism of ARA and production of lipids mediators from ARA. ARA C20:4n-6 is cleaved from PC by PLA2. This reaction also produces LPC, which can be converted in the intercellular space to LPA by ATX. LPA can be considered a lipid mediator because its biological activity is related to the activation of its specific receptors: LPAR1-LPAR6. Free Figure 2. Importance of PLA 2 in metabolism of ARA and production of lipids mediators from ARA. ARA C20:4n-6 is cleaved from PC by PLA 2 . This reaction also produces LPC, which can be converted in the intercellular space to LPA by ATX. LPA can be considered a lipid mediator because its biological activity is related to the activation of its specific receptors: LPAR 1 -LPAR 6 . Free ARA C20:4n-6, on the other hand, can be used for eicosanoid production in either the COX pathway or the LOX pathway. ↑-higher expression of given enzymes in GBM tumor relative to healthy tissue; ↓-lower expression of given enzymes in GBM tumor relative to healthy tissue.

Cytosolic Phospholipase A 2 and Calcium-Independent Phospholipase A 2 in Glioblastoma Multiforme
Expression of cPLA 2 α/PLA2G4A is upregulated in GBM tumors compared to healthy brain tissue [39]. This is also confirmed by bioinformatics analysis on the GEPIA portal [9] and the transcriptomics analysis by Seifert et al. [8]. At the same time, the expression of cPLA 2 β/PLA2G4B is lower, and the expressions of cPLA 2 γ/PLA2G4C, cPLA 2 δ/PLA2G4D, cPLA 2 ε/PLA2G4E, and cPLA 2 ζ/PLA2G4F are unchanged, according to GEPIA [9]. The expression of cPLA 2 γ/PLA2G4C is lower, and cPLA 2 ζ/PLA2G4F is not different in GBM tumors relative to healthy brain tissue, according to the transcriptomics analysis by Seifert et al. [8]. For six of the iPLA 2 , expression in GBM tumor does not differ compared to healthy brain tissue, according to GEPIA [9]. The expression of iPLA 2 β/PLA2G6 and iPLA 2 δ/PNPLA6 is lower in GBM tumor than in the healthy brain, according to the transcriptomics analysis by Seifert et al. [8]. Expressions of the remaining iPLA 2 do not differ between GBM tumors and healthy brain tissue.
In the case of iPLA 2 η/PNPLA4, higher expression in GBM tumors is associated with a worse prognosis for the patient, according to GEPIA (Table 1) [9]. For iPLA 2 ζ/PNPLA2, there is a trend (p = 0.087) of worse prognosis and higher expression of this gene in the GBM tumor. Red background-higher expression in the tumor; blue background-lower expression in the tumor; red background-worse prognosis with higher expression of a given PLA 2 .
cPLA 2 are activated in GBM cells, in particular, by sPLA 2 enzymes [40,41]. This is associated with the induction of cPLA 2 phosphorylation via MAPK kinase cascades as well as with an increase in cytoplasmic Ca 2+ levels via phospholipase C-γ (PLC-γ) activation. cPLA 2 α increases the proliferation of GBM cells, although the effect is not large. The most significant property of cPLA 2 α in GBM cells is causing chemoresistance to temozolomide (TMZ) and other chemotherapeutics, such as doxorubicin and 5-fluorouracil [39]. At the same time, the increased activity of cPLA 2 may also decrease the viability of GBM cells, where TMZ induces the phosphorylation of cPLA 2 . This increases the activation of this enzyme [42] and thus leads to an increase in the level of free ARA 20:4n-6, whose excess reduces the viability of GBM cells. The reason for this may be in the activation of PPAR by this fatty acid [7,43,44] and the generation of reactive oxygen species (ROS) [45].
PLA 2 may also be important in the interaction of GBM cells with endothelial cells. GBM cells cause an increase in the expression and activity of cPLA 2 and iPLA 2 in endothelial cells [46,47]. An increase in cPLA 2 activity in endothelial cells can also be caused by radiation therapy [48]. A rise in the activity of cPLA 2 and iPLA 2 leads to the production of LPA [49]. GBM cancer cells may also increase COX-2 expression in endothelial cells, which increases the production of prostanoids including prostaglandin E 2 (PGE 2 ) [47]. LPA and Cancers 2023, 15, 946 7 of 56 PGE 2 increase the proliferation and migration of endothelial cells [46,47,49]. This is also a mechanism of angiogenesis as a side effect of GBM radiotherapy [47,48]. At the same time, angiogenesis can be inhibited by pericytes [47].
Dying endothelial cells in a GBM tumor can secrete PGE 2 that increases the proliferation of GBM cells [50]. This is associated with the processing of iPLA 2 β by caspase 3 [16,51], which increases the activity of this iPLA 2 and, thus, leads to an increase in PGE 2 production [50].

Secretory Phospholipase A 2 in Glioblastoma Multiforme
Analyses on the GEPIA portal indicate that PLA2G5 expression is higher in GBM tumors [9]. There is also elevated expression of PLA2G2A, PLA2G12A, and PLA2G15 but no other sPLA 2 in GBM tumors [9]. The transcriptomics analysis by Seifert et al. showed that the expressions of PLA2G2A and PLA2G5 are higher in GBM tumors than in healthy brain tissue [8]. This is the same as the data from the GEPIA web server. However, Seifert et al. showed that the expression of PLA2G12A and of the other sPLA 2 enzymes is not different in GBM tumors relative to healthy brain tissue [8]. Wu et al. also showed that PLA2G5 expression is higher in gliomas than in healthy tissue and increases with tumor grade [52].
Higher expression of certain sPLA 2 in GBM tumors is associated with a worse prognosis. According to GEPIA, these include PLA2G1B and PLA2G15 [9]. Wu et al. showed a higher number of sPLA 2 affecting prognosis. In particular, worse prognoses in patients with GBM are associated with higher expression of PLA2G1B, PLA2G2E, PLA2G3, and PLA2G5 [52].
PLA2G5 is significant for tumorigenesis in low-grade gliomas and GBM. This suggests that a high expression of this sPLA 2 is associated with a worse prognosis in patients with GBM and low-grade gliomas ( Table 2) [52]. Analyses on the GEPIA portal show no significant association between the expression of the aforementioned sPLA 2 and the GBM patient prognosis [9]. sPLA 2 are secreted outside the cells where they perform their function. They have their own receptor, PLA 2 R1, from the C-type lectin superfamily and mannose receptor family [34], located in the cell membrane, through which it passes once. According to both GEPIA [9] and Seifert et al. [8], PLA 2 R1 expression does not differ between GBM tumors and healthy brain tissue. An above-average expression of this receptor in a GBM tumor is associated with a worse prognosis for the patient [9], indicating that sPLA 2 may act on PLA 2 R1 and be pro-tumorigenic. sPLA 2 may act by participating in the production of LPA, a lipid mediator that has six different receptors [38]. According to GEPIA, LPAR 3 expression is downregulated in GBM tumors relative to healthy brain tissue [9], whereas LPAR 5 and LPAR 6 expression is upregulated in GBM tumors. The expression of other LPA receptors is not altered in GBM tumors. The transcriptomics analysis by Seifert et al. shows that LPAR 1 expression is lower, and LPAR 6 expression is higher in GBM tumors relative to healthy brain tissue [8]. The expression of other LPA receptors does not differ between GBM tumors and healthy brain tissue. sPLA 2 also have the same catalytic properties as other PLA 2 . They cause the release of ARA 20:4n-6 from cell membrane phospholipids; this reaction produces free ARA 20:4n-6 and LPC. The latter is converted into LPA in the intercellular space by ATX [53], which is secreted by GBM cancer cells [54,55] and whose expression in GBM tumors is higher than in healthy brain tissue [53] and is elevated by interaction with microglial cells [55]. At the same time, GEPIA reports that ATX expression is not altered in GBM tumors [9], and Seifert et al. showed that it is lower [8] than in healthy brain tissue. The level of ATX expression in the tumor is not associated with prognosis severity for patients with GBM [9].
Another important source of ATX in the GBM tumor microenvironment is microglial cells [55], where ATX expression is upregulated by GBM cells, especially under hypoxia. Microglial cells also express the LPAR 1 receptor and can respond to LPA [55].
Increased expression of various sPLA 2 [52] and ATX [53] in GBM tumors also results in increased LPA production. GBM cancer cells show a loss of primary cilia, which leads to an increase in the distribution of LPAR 1 in the plasma membrane of these cells and to an enhancement of signal transduction by this receptor as a result of a greater association of G proteins with this receptor [56].
LPA causes GBM cells to migrate [53][54][55]57,58] due to the activation of LPAR 1 , which results in the activation of protein kinase C (PKC)α. This is responsible for the phosphorylation of the progesterone receptor at the Ser 400 residue [59,60]. GBM cancer cell migration is also facilitated by the LPA-induced decrease in oligodendrocyte adhesion [54]. It is also worth mentioning that in addition to LPAR 1 , the receptor for advanced glycation end products (RAGE) may be another important receptor causing GBM cancer cell migration [61].
LPA increases the proliferation of GBM cancer cells [55]. The effect of LPA on proliferation depends on LPAR 1 receptors [55] and RAGE [61], and it occurs via the activation of two signaling pathways. The first is the Rho → sodium-hydrogen antiporter 1 (NHE-1) pathway, which leads to an increase in intracellular pH and, thus, the proliferation of GBM cancer cells [62]. The second pathway is the activation of extracellular signal-regulated kinase (ERK) mitogen-activated protein kinase (MAPK) by the phosphatidylinositol-4,5bisphosphate 3-kinase (PI3K) → PKC pathway [62], which can also be initiated by epidermal growth factor receptor (EGFR) transactivation. Studies on PLA 2 G2A have shown that this sPLA 2 increases GBM cancer cell proliferation via EGFR transactivation [63][64][65]. This is associated with the activation of PKC, which activates EGFR [64]. EGFR activation results in the activation of the Src → ERK MAPK → Akt/PKB → mammalian target of rapamycin (mTOR) → ribosomal protein 70 S6 kinase (p70S6K) pathway [63,65]. Its consequence is an increase in the proliferation of GBM cancer cells. sPLA 2 can also increase GBM cancer cell proliferation indirectly through the activation of cPLA 2 inside a GBM cell [40]. This process is independent of LPA.
Phosphorylation of the progesterone receptor by LPA increases vascular endothelial growth factor (VEGF) expression in GBM cancer cells [60], the most important growth factor in angiogenesis. LPA is also important in radiotherapy-induced angiogenesis in GBM tumors [58]. An increase in tumor vascularization during exposure to ionizing radiation can be inhibited by ATX inhibitors, which could have some clinical application in future therapies against GBM [58].
The aforementioned actions of LPA were carried out on various models of specific GBM cell lines. Significantly, the action of LPA may be more pronounced in GBM cancer stem cells than non-cancer stem cells, as the former show much higher expression of LPAR 1 and LPAR 3 [67]. LPAR 1 is important in the development of GBM. Higher expression of this receptor in GBM tumors is associated with a worse prognosis [55]. At the same time, an analysis on the GEPIA portal did not link LPAR 1 and LPAR 3 expression to prognosis severity for GBM patients [9]. In addition, it did not show that the expression of the other LPA receptors had an effect on the prognosis for GBM patients.

Pan-Cancer Analysis of Phospholipase A 2 Genes and Comparison of GBM Expression against Other Cancers
We also performed a pan-cancer analysis of the expression of the PLA 2 genes with the GEPIA portal [9].
In GBM, but not in lower grade gliomas, there is higher expression of cPLA 2 α/PLA2G4A compared to healthy brain tissue [8,9]. Among the analyzed 31 tumor types, only four more had higher expression of this PLA 2 , and eight other types showed a decrease. For this reason, higher expression of this enzyme in GBM tumors can be considered characteristic for this cancer.
In GBM, the expression of cPLA 2 β/PLA2G4B is decreased relative to healthy brain tissue [9], similar to lower grade glioma and 19 other types of cancer. This indicates that the decreased expression of this PLA 2 in tumor is a hallmark of cancer.
Seifert et al. also indicates that cPLA 2 γ/PLA2G4C expression may be downregulated in GBM tumors relative to healthy brain tissue [8]. According to a pan-cancer analysis based on the GEPIA, cPLA 2 γ/PLA2G4C expression is downregulated in nine types of tumors but not in GBM or lower grade gliomas, whereas it is upregulated in seven types of tumors [9]. Changes in cPLA 2 γ/PLA2G4C expression in GBM tumors could be a hallmark of cancer.
Seifert et al. also showed a decrease in the expression of iPLA 2 β/PLA2G6 and iPLA 2 δ/PNPLA6 in GBM tumors relative to healthy brain tissue [8]. According to GEPIA, iPLA 2 β/PLA2G6 expression is downregulated in 15 tumor types (Table 3) [9], whereas iPLA 2 δ/PNPLA6 expression is only downregulated in three types. For this reason, it can be thought that decreased iPLA 2 β/PLA2G6 expression may be a hallmark of cancer. In contrast, reduced expression of iPLA 2 δ/PNPLA6 is characteristic of GBM.

Name of Cancer
Cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) Red background, ↑-expression higher in tumors than in healthy tissue; blue background, ↓-expression lower in tumors than in healthy tissue; gray background, =-expression does not differ between tumors and healthy tissue.
PLA2G2A expression is downregulated in 18 out of 31 types of cancer, indicating that it is generally downregulated in cancer (Table 4). In contrast, increased expression of PLA2G2A may occur in GBM [8,9], which may be characteristic of GBM. On the other hand, in 17 out of 31 cancers, there is a higher expression of PLA2G7 in the tumor than in healthy tissue. Its expression in a GBM tumor is not different from its expression in healthy brain tissue [8,9]. Table 4. Pan-cancer analysis of gene expression of sPLA 2 and sPLA 2 receptors. PLA2G1B  PLA2G2A  PLA2G2D  PLA2G2E  PLA2G2F  PLA2G3  PLA2G5  PLA2G7  PLA2G10  PLA2G12A  PLA2G12B  PLA2G15  PLA2G16

Name of Cancer
Red background, ↑-expression higher in tumor than in healthy tissue; blue background, ↓-expression lower in tumor than in healthy tissue; gray background, =-expression does not differ between tumor and healthy tissue.

Lysophospholipid Acyltransferases in Glioblastoma Multiforme
When discussing the importance of PLA 2 in tumorigenesis in GBM, it is also important to mention enzymes that catalyze the opposite reaction to the enzymes in question. An example of this is lysophosphatidylcholine acyltransferases (LPCAT), which catalyze the opposite reaction towards PC [68]. LPCAT causes the formation of PC from LPC and fatty acyl-CoA. For this reason, LPCAT decreases the level of LPA, a lipid mediator important in cancer processes in GBM. According to the GEPIA portal, GBM tumors have higher expressions of LPCAT1, LPCAT2, and LPCAT3, but lower expression of LPCAT4/LPEAT2 relative to healthy brain tissue [9]. In addition, according to Seifert et al., the expression of LPCAT1 and LPCAT3 is higher in GBM tumors than in healthy brain tissue [8]. In contrast, LPCAT4 expression is lower in GBM tumors. This confirms the results obtained from the GEPIA database. An increase in the expression of the aforementioned enzymes may contribute to a decrease in LPA level but also contribute to the intense remodeling of phospholipids in the cell membranes of GBM cells. At the same time, according to the GEPIA database, the expression of the mentioned enzymes does not affect the prognosis severity of GBM patients [9]. 4.6. Acyl-CoA Thioesterases and Arachidonic Acid C20:4n-6 in Glioblastoma Multiforme The most important pathway for the formation of free ARA C20:4n-6 is through PLA 2 activity. However, free ARA C20:4n-6 can be formed from hydrolysis of arachidonyl-CoA by acyl-CoA thioesterases (ACOT) [69], a group of nine enzymes that cause hydrolysis of fatty acyl-CoA to free fatty acid and CoA [69,70]. An example of an enzyme from this group is ACOT7, which shows activity towards arachidonyl-CoA and saturated fatty acyl-CoA [69][70][71]. According to GEPIA and Seifert et al., there is a reduction in ACOT7 expression in GBM tumors relative to healthy brain tissue [8,9], where higher expression of this enzyme is associated with a worse prognosis for a GBM patient [9], suggesting the involvement of ACOT7 in tumorigenesis in GBM.
According to GEPIA and Seifert et al., there is also elevated expression of ACOT9 in GBM tumors [8,9], an enzyme showing the highest activity to myristoyl-CoA [69,70,72] and low activity to longer acyl-CoA. Importantly, the expression level of ACOT9 is not associated with the prognosis for a patient with GBM [9]. According to GEPIA, the expression of other ACOT does not differ between GBM tumors and healthy brain tissue [9]. In addition, Seifert et al. indicate that the expression of ACOT4 and ACOT8 in GBM tumors is lower than in healthy brain tissue [8].
COX-1 (another name is prostaglandin-endoperoxide synthase 1 (PTGS1)) is a constitutive enzyme with a constant level of expression [84]. A second enzyme with the same activity is COX-2 (another name is prostaglandin-endoperoxide synthase 2 (PTGS2)) [85], an inducible enzyme that is regulated at the transcriptional level and is characterized by rapid degradation of the COX-2 protein [86]. The half-life of the COX-2 protein is only 5 h.
Sometimes, cyclooxygenase-3 (COX-3), a variant of COX-1 that retains intron 1 in its mRNA, is also mentioned in the context of conversion to prostanoids [87]. Although there is expression of the COX-3 protein, which is longer than COX-1, this enzyme has the same activity as the other cyclooxygenases. In mice and dogs, COX-3 is more sensitive to the inhibitors acetaminophen and phenacetin. Humans also have a variant of COX-1, but it is as sensitive to these inhibitors as standard COX-1 [88]. PGH 2 is unstable and undergoes spontaneous nonenzymatic conversion, mainly with PGE 2 and, in smaller amounts, with prostaglandin D 2 (PGD 2 ) [78]. In the synthesis of PGE 2 , we can distinguish three synthases: membrane-bound prostaglandin E synthase-1 (mPGES-1)/PTGES [89][90][91], membrane-bound prostaglandin E synthase-2 (mPGES-2)/PTGES2 [92], and cytosolic prostaglandin E synthase (cPGES)/PTGES3 [93]. These synthases are dependent on glutathione, which serves to reduce the endoperoxide bridge in PGH 2 with the formation of a single hydroxyl group. In addition, cPGES forms a complex with heat shock protein 90 (Hsp90), which is important in the activity of this PGE 2 synthase [94]. mPGES-1 and mPGES-2 bind with either COX-1 or COX-2 [92,95,96], whereas cPGES binds only with COX-1 [93,97]. mPGES-1 is an inducible enzyme whose expression under the influence of inflammatory reactions increases following the expression of COX-2 [96]. mPGES-2 [96] and cPGES [93] are constitutive enzymes, meaning that their expression is not altered by inflammatory reactions.

Cyclooxygenase Pathway and Glioblastoma Multiforme
After ARA C20:4n-6 is released from cell membrane phospholipids, it is processed with COX and LOX. In the healthy brain, ARA C20:4n-6 is processed mainly with LOX, whereas in GBM tumors, it is processed mainly with COX, as shown by experiments on C6 cells [140].
COX-1 expression [141] and COX-2 expression [141,142] are elevated in GBM tumors compared to healthy brain tissue, whereas according to GEPIA and Seifert et al., just COX-1 expression is elevated [8,9]. The expression of all three PGE 2 synthases, i.e., mPGES-1, mPGES-2, and cPGES, is also elevated in GBM [143], although according to GEPIA, only cPGES expression is higher compared to its expression in the healthy brain [9]. In contrast, Seifert et al. showed no change in PGE 2 synthase expression in GBM tumors [8]. cPGES is enzymatically bound with just COX-1 [93,97]. Therefore, it is possible that COX-1-cPGES may play an important role in the production of PGE 2 in GBM tumors. According to the GEPIA portal, there are also changes in the expressions of other prostaglandin synthases. In a GBM tumor, there is increased expression of H-PGDS but decreased expression of L-PGDS [9], both synthases involved in PGD 2 synthesis. In contrast, Seifert et al. showed that the expression of H-PGDS and L-PGDS in a GBM tumor is lower than their expressions in a healthy brain [8]. According to GEPIA in a GBM tumor, there is also increased expression of AKR1B1, decreased expression of AKR1C1 and AKR1C2, and no change in AKR1C3 expression [9]. Similarly, Seifert et al. showed that in a GBM tumor, there is higher expression of AKR1B1 and decreased expression of AKR1C1, but there is no difference in the expressions of AKR1C2 or AKR1C3 between the GBM tumor and healthy brain tissue [8]. AKR1B1 is involved in the synthesis of PGF 2α [115], whereas AKR1C1 and AKR1C2 are involved in the conversion of PGE 2 into PGF 2α [114]. Expression of the TxA 2 synthesizing synthase TBXAS1 [8,9,144] is also elevated in GBM tumors, which may explain the increased expression and production of TxA 2 and the higher TxA 2 /PGI 2 ratio in GBM tumors than in healthy brain tissue [145,146].
As for receptors for prostaglandins, according to the GEPIA portal, there is an elevated expression of PTGER 4 /EP 4 and TBXA 2 R/TP in the tumor relative to healthy brain tissue [9], these two being receptors for PGE 2 and TxA 2 , respectively. In contrast, Seifert et al. showed that the expression of prostanoid receptors in GBM tumors did not differ relative to the healthy brain [8].
According to the GEPIA portal, the expression of MRP4/ABCC4 [9], a transporter responsible for the secretion of prostaglandins from the cell, is also increased in GBM tumors. The transcriptomics analysis by Seifert et al. did not confirm this [8]. GEPIA and Seifert et al. show no change in the expressions of PGT/SLCO2A1, 15-PGDH, 12-HEDH/PTGR1, and PTGR2 [8,9]-the first is a transporter that takes prostaglandins into the cell, and the second, third, and fourth are prostaglandin-degrading enzymes.
COX-2 is important in GBM tumor function. Its expression in GBM tumors is upregulated by hypoxia [23] and EGFR activation [147,148] as well as the action of epidermal growth factor receptor variant III (EGFRvIII) [147] and hepatocyte growth factor (HGF) [149]. COX-2 expression and biosynthesis of the most important product of this enzyme, PGE 2 , is present in GBM cancer cells. However, PGE 2 in GBM tumors may not come mainly from GBM cancer cells but rather from tumor-associated macrophages (TAM) [150].
COX-2 is also important for GBM cancer stem cells. COX-2 expression, and with it, the production of PGE 2 , is higher in GBM cancer stem cells than in differentiated GBM cells [158,159]. This lipid mediator activates the Wnt pathway in GBM cancer stem cells, leading to the self-renewal and proliferation of these cells. PGE 2 induces angiogenesis in GBM tumors. Therefore, COX-2 expression is positively correlated with microvessel density in GBM tumors [160]. Notably, PGE 2 causes vasculo-genic mimicry of GBM cells, which promotes angiogenesis [161]. In GBM cells, PGE 2 also increases the expression of CXCL8/IL-8 [153], which has pro-angiogenic properties [162]. PGE 2 causes cancer immune evasion. Through EP 4 , PGE 2 increases the expression of tryptophan-2,3-dioxygenase (TDO) [163], an enzyme that converts tryptophan into a signaling molecule that reduces immune cell activity. PGE 2 also affects tumor-associated cells which are important in cancer immune evasion. PGE 2 increases the recruitment of myeloid-derived suppressor cells (MDSC) to the tumor niche in GBM [164] and interferes with the cytotoxic function of various immune cells, as shown by experiments in other cancer models. When acting chronically, PGE 2 impairs the cytotoxic function of natural killer (NK) cells [165,166], dendritic cells [167], and T cells [168]. PGE 2 also causes M2 polarization of macrophages [169], immunosuppressive cells that promote tumor growth. PGE 2 also causes radiation resistance [170,171] and TMZ resistance in GBM [172]. COX-2 expression and PGE 2 production in GBM cancer cells are upregulated by TMZ [173] and ionizing radiation [170], which is related to caspase 3 activation in damaged cells and subsequent NF-κB activation [174]. Then, NF-κB increases COX-2 expression and, thus, the production of PGE 2 that trans-activates EGFR and activates the β-catenin pathway, which has a pro-survival effect and leads to resistance to further therapy [170]. Through EP 1 and EP 3 , PGE 2 increases the intensity of β-oxidation and tricarboxylic acid cycle activity in mitochondria [172], leading to TMZ resistance. In response to ionizing radiation, healthy brain tissue also induces increased production of PGE 2 and pro-inflammatory cytokines [175], which increases GBM cell migration as well as causes tumor recurrence [176]. PGD 2 is also produced in GBM tumors [177]. At physiological concentrations, this prostaglandin increases the proliferation and migration of GBM cells, but, at concentrations of several micromoles, it decreases the viability and inhibits the proliferation of the GBM cells studied [177][178][179]. This effect may be due to 15d-PGJ 2 , which has anti-cancer properties [180]. PGD 2 is non-enzymatically converted to 15d-PGJ 2 [100]. High concentrations of PGD 2 result in an accumulation of 15d-PGJ 2 to a level that causes a measurable reduction in the viability of GBM cancer cells. Cyclopentenone prostaglandins, particularly PGJ 2 , ∆ 12 -PGJ 2 , and 15d-PGJ 2 , have anti-tumor properties, as demonstrated in in vitro studies on GBM cells. These prostaglandins inhibit tumor cell proliferation through PPARγ activation [181,182]. TxA 2 may also play an important role in tumorigenic processes in GBM. In GBM cells, TxA 2 increases the expression of IL-6, which participates in tumorigenesis [183]. TBXAS1 inhibitors induce apoptosis and inhibit the migration and proliferation of GBM cancer cells [144,184,185], indicating an autocrine effect of TxA 2 . In addition, in an in vivo model, TBXAS1 inhibitors inhibited angiogenesis and GBM tumor growth [185]. The described inhibitors increased the sensitivity of GBM cells to alkylation chemotherapy [185] and radiotherapy [186].
Given the role of COX-2 in tumorigenesis in GBM, high COX-2 expression in GBM tumors is associated with poorer patient prognoses [160,187,188], although the GEPIA data showed no correlation between COX-1 and COX-2 expression and patient prognosis severity [9]. In addition, the expression of other prostanoid metabolism enzymes worsens the prognosis for GBM patients, in particular, high expression of mPGES-1, the synthase responsible for the production of PGE 2 [121]. This is confirmed with the GEPIA data [9], although the expression levels of other PGE 2 synthases are not associated with prognosis severity [9,121]. Of the other prostaglandin synthases, high expression of AKR1B1, a PGF 2αproducing synthase [115], in GBM tumors is associated with poorer patient prognoses [9].
According to the GEPIA portal, expression of MRP4/ABCC4, a transporter that secretes prostaglandins from the cell, does not affect the prognosis for GBM patients [9]. Higher expression of certain prostaglandin receptors worsens the prognosis for patients with GBM. In particular, a worse prognosis is associated with higher expression of PTGER 1 /EP 1 and PTGIR/IP [9], which are receptors for PGE 2 and PGI 2 , respectively.
Higher expression of G2A/GPR132 is also associated with a worse prognosis (p = 0.052) in GBM patients [9]. G2A/GPR132 is a receptor for 9-HODE [83], a product of the activity of COX that process linoleic acid 18:2n-6 [80]. The role of this receptor in GBM has not been thoroughly investigated, although studies on fibroblasts have shown that G2A/GPR132 is an oncogene [189].
Prognosis severity is also affected by the expression level of enzymes involved in prostaglandin inactivation. High expression of 15-PGDH in GBM tumors is associated with a better prognosis [121]. The opposite is true for the expression of 12-HEDH/PTGR1, the enzyme that catalyzes the second prostaglandin inactivation reaction [121]. PGT/SLCO2A1 expression levels are not associated with prognosis severity. On the other hand, according to the GEPIA portal, the expressions of PGT/SLCO2A1, 15-PGDH, PTGR1, and PTGR2 do not affect the prognosis for patients with GBM [9].
Prostaglandin levels in GBM tumors may also be associated with a worse prognosis, particularly higher levels of PGE 2 and PGF 2α [121]. PGD 2 levels in GBM tumors do not affect prognosis severity [121]. At the same time, these lipid mediators are often unstable, transforming into other lipid mediators with lesser or different properties within a short time after synthesis. For this reason, they may act locally in the immediate vicinity of the site of their synthesis.
Relating enzyme expression and levels of the discussed prostaglandins to prognosis makes it possible to estimate the significant impact of a particular pathway on cancer processes. In GBM tumors, higher expressions of production enzymes and levels of PGE 2 (COX-2, mPGES-1) and PGF 2α (COX-2, AKR1B1) are responsible for worse prognoses [9,121]. On the other hand, higher expression of the prostaglandin-inactivating enzyme, 15-PGDH, is associated with better prognoses (Table 5) [121]. For this reason, NSAIDs are being investigated as either potential drugs [190,191] or agents with chemopreventive properties against GBM. Various meta-analyses inconclusively discuss the chemopreventive properties of NSAIDs, such as aspirin. Depending on the meta-analyses cited, regular use of NSAIDs, including aspirin, may either reduce the risk [192,193] or have no effect [194] on the risk of developing glioma or GBM. Nevertheless, the COX pathway produces prostaglandins that exhibit pro-cancer and anti-cancer properties. A better option may be to develop drugs that specifically target only particular prostaglandins relevant to tumorigenic processes in GBM, namely PGE 2 and PGF 2α [9,121]. It may be possible to develop drugs that are specific inhibitors of mPGES-1.

Pan-Cancer Analysis of Genes Related to the COX Pathway and GBM
Changes in the expression of various genes in GBM tumors relative to healthy tissue may be the result of tumor-specific neoplastic processes or specific mechanisms found only in GBM. For this reason, we performed a pan-cancer analysis of the expression of the genes involved in the COX pathway based on the data available in the GEPIA web server [9]. It showed that increased or decreased expression of a given gene relative to healthy tissue does not occur in all types of cancer. At the same time, in some cases, a certain trend of changes in the expressions of the genes studied can be observed. An example of this is TBXAS1, whose expression is increased in nine types of cancer but decreased in another four types of cancer. Similarly, the expression of mPGES-1/PTGES is increased in eight types of cancers but decreased in three types of cancers. Some genes tend to undergo decreased expression in tumors. An example of this is 15-PGDH/HPGD, whose expression is reduced in 18 types of cancer but increased in two types of cancer relative to healthy tissue. Another example is the expression of PGIS/PTGIS, decreased in 17 types of cancer but elevated only in pancreatic adenocarcinoma.
According to GEPIA, there is an increase in COX-1/PTGS1 expression in GBM tumors, which is the same as in lower grade glioma. In seven types of tumors, this gene is overexpressed, but in seven more types, its expression is reduced. This indicates that the increased expression of COX-1/PTGS1 in gliomas (GBM and lower grade gliomas) is specific to these diseases. Some studies also show increased expression of PGE 2 synthases (mPGES-1/PTGES, mPGES-2/PTGES2 and cPGES/PTGES3) [9,143], although GEPIA confirms it is only for cPGES/PTGES3 [9]. According to GEPIA, in lower grade glioma, there are no changes in the expression of PGE 2 synthases relative to healthy brain tissue. According to GEPIA, expression of cPGES/PTGES3 is increased in 11 types of tumors but is decreased in one type. For this reason, the increase in cPGES/PTGES3 expression in GBM can be considered cancer-specific, just like mPGES-1/PTGES, which has increased expression in eight types of cancer and decreased in three. According to GEPIA, only four types of cancers have increased expression of mPGES-2/PTGES2, which shows that this enzyme may not be cancer-specific.
According to GEPIA in GBM, there is also increased expression of H-PGDS/HPGDS but decreased expression of L-PGDS/PTGDS [9]. At the same time, Seifert et al. showed that the expression of both PGD 2 synthases is decreased in GBM tumors [8]. H-PGDS/HPGDS expression is also upregulated in lower grade glioma. H-PGDS/HPGDS expression is downregulated in five tumor types and upregulated in an equal number of tumor types. Changes in H-PGDS/HPGDS expression can be specific to gliomas. L-PGDS/PTGDS expression is lower in GBM compared to healthy brain tissue [8,9]. L-PGDS/PTGDS expression is decreased in almost all types of tumors and, thus, can be deemed specific to cancer.
In GBM, as in lower grade glioma, there is increased expression of TBXAS1 [8,9,144]. The expression of this enzyme is elevated in nine types of tumors, which means it may be cancer-specific.
In GBM tumors, there is also upregulation of AKR1B1 expression but downregulation of AKR1C1 and AKR1C2 expressions relative to healthy brain tissue [8,9]. Lower grade gliomas show no changes in the expressions of these enzymes. The expression of AKR1B1 increases in nine types of tumors. AKR1C1 and AKR1C2, on the other hand, have decreased expressions in 14 types of tumors, which indicates that these changes may be cancer-specific.
PGIS/PTGIS expression is downregulated in 17 types of tumors. At the same time, in GBM tumors, PGIS/PTGIS expression does not differ from healthy brain tissue [8,9].
In GBM and lower grade glioma, there is an increase in MRP4/ABCC4 expression [9]. This transporter also has increased expression in another four types of tumors but decreased expression in two types of tumors. Changes in MRP4/ABCC4 expression may be specific to gliomas.
Finally, 15-PGDH/HPGD expression is often downregulated in tumors (Table 6). This was shown by a pan-cancer analysis in which 18 out of 31 cancers had decreased expression of this enzyme. At the same time, in gliomas (GBM and lower grade glioma), there were no changes in 15-PGDH/HPGD expression relative to healthy brain tissue.

Lipoxygenases Pathway
In addition to the COX pathway, PUFA can be transformed with LOX. These enzymes exhibit dioxygenase activity, catalyzing the insertion of a hydroperoxyl group into a PUFA, most commonly ARA 20:4n-6. Hydroperoxyeicosatetraenoic acids (HpETE) are then formed from ARA 20:4n-6, which are further processed in the lipoxygenase pathway. The names of LOX enzymes are related to their sites of formation and the configuration of the hydroperoxyl group in ARA 20:4n-6. In humans, there are six LOX: The ALOX5 gene is found on chromosome 10. The other LOX form a gene cluster on 17p13.1 [195,196]. There is also a mouse 8-LOX [197], whose sequence is 78% identical to that of human 15-LOX-2/ALOX15B [197,198]. It is likely that mouse 8-LOX and human 15-LOX-2/ALOX15B are derived from a common ancestor, which was indirectly confirmed by mutagenesis experiments on these two enzymes. Changing only two amino acids in either mouse 8-LOX or human 15-LOX-2/ALOX15B alters the catalytic properties of these two enzymes in 15-LOX and 8-LOX, respectively [197].
Glutathione can be attached to LTA 4 by LTC 4 synthase (LTC 4 S) ( Figure 4) [229,230]. LTC 4 is then formed. LTC 4 S combines with 5-LOX and FLAP to increase the efficiency of LTC 4 production with ARA 20:4n-6 [231]. Subsequently, amino acids from the conjugated glutathione in LTC 4 can be removed. As a consequence of this, LTC 4 is converted into other leukotrienes, namely LTD 4 , LTE 4 , and LTF 4 . All of these leukotrienes, together with LTC 4 , form a group called cysteinyl leukotrienes. LTD 4 is then formed from LTC 4 with the involvement of γ-glutamyltransferase 1 (GGT1) and γ-glutamyltransferase 5 (GGT5) [232]. Subsequently, LTD 4 can be converted to LTE 4 with the participation of dipeptidase 1 (DPEP1) and dipeptidase 2 (DPEP2) [233,234]. LTC 4 can also be converted to LTF 4 with the participation of carboxypeptidase A [235]. Amino acids can be attached back to cysteine in cysteinyl leukotriene, as exemplified by the conversion of LTE 4 to LTF 4 with the participation of an enzyme with γ-glutamyltranspeptidase activity [236]. LTF 4 , however, has a much weaker effect than LTE 4 , and the latter reaction can be considered an inactivation of LTE 4 .
The receptors for cysteinyl-leukotrienes are CysLTR 1 [249] and CysLTR 2 [250,251]. Both receptors show a 38% similarity in amino acid sequence [250]. CysLTR 1 shows a high affinity for LTD 4 and low affinity for LTC 4 and LTE 4 , and it shows no affinity at all for LTB 4 [249]. CysLTR 2 has the best affinity for LTC 4 and LTD 4 and a very low affinity for LTE 4 , and it shows no affinity at all for LTB 4 [250,251]. A receptor specific for LTE 4 is 2oxoglutarate receptor 1 (OXGR1)/GPR99 [252], which is also the receptor for 2-oxoglutarate. This receptor has a lower affinity for LTC 4 and LTD 4 . Another identified receptor for cysteinyl-leukotrienes specifically for LTC 4 and LTD 4 is G protein-coupled receptor 17 (GPR17) [253], which is also activated by uridine diphosphate (UDP), UDP-glucose, and UDP-galactose [253]. Further studies have not confirmed that GPR17 is a receptor for UDP, LTC 4 , and LTD 4 [254,255]. This receptor can, independently of its ligand, downregulate CysLTR 1 [256], which means it can reduce the action of cysteinyl leukotrienes. is a receptor for UDP, LTC4, and LTD4 [254,255]. This receptor can, independently of its ligand, downregulate CysLTR1 [256], which means it can reduce the action of cysteinyl leukotrienes. Figure 4. 5-LOX pathway. ARA C20:4n-6 is converted to 5-HpETE with 5-LOX. This enzyme also catalyzes the next step in leukotriene biosynthesis. It converts 5-HpETE into LTA4, which can then be converted into LTB4 with LTA4H, into LTC4 with LTC4S, or into 5-oxo-ETE. 5-HpETE can also be converted to 5-oxo-ETE. LTC4 can be converted to other cysteinyl leukotrienes. LTC4 can be converted to LTF4 with the involvement of carboxypeptidase A or to LTD4 with the involvement of GGT1 and GGT5. Subsequently, LTD4 can be converted into LTE4 with the participation of DPEP1 and DPEP2, and then converted into LTF4 with γ-glutamyltranspeptidase. ↑-higher expression of given enzymes in GBM tumor relative to healthy tissue.
The ALOX12B gene is only 38% similar to the ALOX12 gene. The highest expression of this enzyme is found in the skin, and it is much lower in the prostate and adrenal gland [196,199,284]. 12R-LOX is important in skin function; mutations in the ALOX12B gene lead to ichthyosis [208,210,285], as do mutations in the ALOXE3 gene. 12R-LOX/ALOX12B and eLOX3/ALOXE3 participate in a common pathway in lipid mediator production. 12R-LOX produces 12R-HpETE, which is converted to 11,12-bis-epi-HxA 3 with eLOX3 ( Figure 5) [200]. Under the influence of eLOX3/ALOXE3, 12-oxo-ETE is also formed from 12R-HpETE in small amounts [200]. 11,12-bis-epi-HxA3 with eLOX3 ( Figure 5) [200]. Under the influence of eLOX3/ALOXE3, 12-oxo-ETE is also formed from 12R-HpETE in small amounts [200]. Figure 5. 12-LOX pathway. ARA C20:4n-6 is converted to 12S-HpETE and 12R-HpETE with 12S-LOX and 12R-LOX, respectively. Either 12-oxo-ETE or the corresponding 12-HETE can be formed from these compounds. 12S-HpETE can also be converted to HxA3 or HxB3 with hemin and lipoxygenases: eLOX3, 12S-LOX, or 15-LOX-1. 12R-HpETE can undergo a similar conversion to 11,12-bis-epi-HxA3. HxA3 may undergo further transformations. HxA3 can be conjugated to glutathione. HxA3-C is then formed, from which amino acids can be detached-HxA3-D is then formed in a reaction similar to the transformation of cysteinyl-leukotrienes. HxA3 can also be converted to TrXA3. Arrows next to enzymes: higher or lower expression of given enzymes in GBM tumor relative to healthy tissue. ↓-lower expression of given enzymes in GBM tumor relative to healthy tissue.
There are also cysteinyl lipoxins, which, just like cysteinyl leukotrienes, are lipoxins with conjugated glutathione at carbon 6 [294]. They are synthesized from 15-HETE, from which, with the participation of 5-LOX/ALOX5, 15-hydroxy-5,6-epoxy-eicosatetraenoic acid is formed, a compound similar in structure to LTA 4 . The epoxy group from these two compounds is converted to a hydroxyl group and conjugated glutathione [294]. However, it is not known whether cysteinyl lipoxins are essential lipid mediators or merely arise as a result of the nonspecificity of enzymes conjugating glutathione to various compounds.

Lipoxygenases in Glioblastoma Multiforme
In GBM tumors, ARA C20:4n-6 is mainly processed by COX, as shown by experiments on the C6 cell line [140]. In contrast, in the healthy brain, this PUFA is mainly processed by the LOX pathway. This shows that in GBM tumors, the LOX pathway may not be as important as the COX pathway, although it is still important in tumor mechanisms in GBM tumors.

5-Lipoxygenase Pathway in Glioblastoma Multiforme
The expression of 5-LOX/ALOX5 in a GBM tumor is higher than in non-tumor brain tissue [300][301][302]. This is also confirmed by data obtained from the GEPIA portal [9] and from Seifert et al. transcriptomics analysis [8].
Expression of 5-LOX/ALOX5 in the GBM tumor is found in macrophage and microglial cells as well as in other cells, such as cancer cells [301,302]. It is higher in GBM cancer stem cells than in other GBM cancer cells [303]. According to GEPIA, higher expressions of FLAP/ALOX5AP, LTC 4 S, LTA 4 H, GGT5, and DPEP1 but not DPEP2 [9], the enzymes that synthesize LTB 4 and LTE 4 from the product of 5-LOX/ALOX5 activity, were also found in GBM tumors [224,225,229,230,232,234]. Seifert et al. showed that there are higher expressions of FLAP/ALOX5AP, LTA 4 H, and GGT5 in GBM tumors than in healthy brain tissue [8]. In contrast, LTC 4 S, DPEP1, and DPEP2 are not affected. The higher expression of enzymes responsible for leukotriene biosynthesis increases the production [304] and levels [305] of these lipid mediators further in GBM tumors than in healthy brain tissue, particularly cysteinyl-leukotrienes.
The expression level of 5-LOX/ALOX5 in GBM tumors does not affect prognosis [9,188], although simultaneous high expression of COX-2 and 5-LOX/ALOX5, two major ARA C20:4n-6 processing enzymes, is associated with a worse prognosis [188]. This shows that the two pathways in cooperation can impinge on prognosis severity.
The expression levels of most enzymes involved in leukotriene production and metabolism do not affect prognosis [9]. Only for GGT1, higher expression in GBM tumors is associated with a worse prognosis [9]. GGT5 expression showed a positive trend (p = 0.055) toward a worse prognosis. GGT1 and GGT5 are enzymes that catalyze the transformation of LTC 4 into LTD 4 [232], demonstrating that the transformation of cysteinyl leukotrienes may be important in tumorigenesis in GBM.
In addition, higher expression of 12-HEDH/PTGR1, an enzyme that degrades LTB 4 , as well as prostaglandins, may be associated with worse prognoses for GBM patients [121], although GEPIA did not confirm such a link [9]. In addition, GEPIA and Seifert et al. did not show that 12-HEDH/PTGR1 expression differs between GBM tumors and healthy brain tissue [8,9]. According to GEPIA [9] and Seifert et al. [8], expression levels of receptors for leukotrienes LTB 4 R1, LTB 4 R2, CysLTR 1 , CysLTR 2 , GPR17, and OXGR1/GPR99 do not differ between GBM tumors and healthy brain tissue. In addition, the expression levels of these receptors in GBM tumors do not affect prognosis [9].
Leukotrienes as well as the entire 5-LOX pathway are important in tumorigenesis in GBM. They may also be important in the onset of GBM and in the first stages of tumorigenesis. The GA genotype of rs2291427 in the ALOX5 gene is associated with a higher risk of GBM in men [306].
Expression of 5-LOX/ALOX5 is higher in GBM cancer stem cells than in other GBM cancer cells [303]. The products of 5-LOX/ALOX5 activity induce proliferation and selfrenewal of GBM cancer stem cells. The effects of 5-LOX/ALOX5 on GBM cancer stem cells are autocrine in nature. LTB 4 also increases the proliferation of GBM cells [307]. This is associated with an increase in Ca 2+ levels in the cytoplasm of GBM cells [307]. Studies of various cell lines show that 5-LOX/ALOX5 expression is present in only a portion of them [308,309]. Expression of 5-LOX/ALOX5 causes an autocrine increase in the proliferation of such a line and, thus, makes culture growth dependent on 5-LOX/ALOX5 activity. All GBM lines express LTA 4 H, LTB 4 R1/BLT1, LTB 4 R2/BLT2, and CysLTR 2 , but only some lines express LTC 4 S [309], indicating heterogeneity in the production of cysteinyl-leukotrienes and 5-HETE by GBM cancer cells.
The dependence of the proliferation of some GBM cancer cell lines on the 5-LOX pathway may be a potential therapeutic target for GBM treatment in personalized therapy. For this reason, the pan-LOX inhibitor Nordy [303,310], 5-LOX inhibitors such as caffeic acid [307], A861 [311], AA-863, and U-60,257 (pyriprost) [312], LTA 4 H inhibitors such as bestatin [311], and CysLTR 1 and CysLTR 2 receptor inhibitors such as montelukast and zafirlukast [313] have anti-tumor properties against GBM and inhibit proliferation. This is associated with decreased ERK MAPK activation and induction of apoptosis as a result of decreased expression of anti-apoptotic Bcl-2 and increased expression of pro-apoptotic Bax [308].
Cysteinyl leukotrienes may have anticancer properties by increasing the bioavailability of various chemotherapeutics. In the brain, as well as in GBM tumors, there is a blood-brain barrier (BBB) that is poorly permeable to many substances, including anticancer drugs [314]. However, cysteinyl leukotrienes have BBB permeability, as shown by experiments on rat RG-2 glioma tumors [315]. BBB permeability is highest for LTE 4 [315], with cysteinyl leukotrienes not causing BBB permeability in healthy brain tissue [315,316]. For this reason, the administration of LTC 4 prior to the administration of chemotherapeutics that pass poorly through the BBB increases the bioavailability of drugs such as cisplatin [317]. However, this method does not increase the bioavailability of all chemotherapeutics, as exemplified by paclitaxel [318].
The receptor for cysteinyl leukotrienes is GPR17 [253]. According to GEPIA [9] and Seifert et al. [8], the expression level of this receptor does not differ between GBM tumors and healthy brain tissue. Higher GPR17 expression is associated with better prognosis in patients with low-grade gliomas, according to the Chinese Glioma Genome Atlas (CGGA) [319] and GEPIA [9], but the expression of this receptor is not associated with prognosis in a GBM patient [9]. GPR17 expression is also higher in low-grade gliomas than in healthy brain tissue [319]. Activation of this receptor by the ligand inhibits proliferation in the G 1 phase and induces apoptosis of GBM cell lines LN-229 and SNB-19 [319]. In addition, GPR17 ligands inhibit tumor growth, as shown by experiments using patient-derived xenograft mouse models. The action of GPR17 is associated with a decrease in the levels of cyclic adenosine monophosphate (cAMP) and Ca 2+ in the cytoplasm, which reduces the activation of the PI3K → Akt/PKB pathway [319,320]. An increase in GPR17 expression can cause the proliferation and migration of GBM cells [321], particularly with an increase in the expression of this receptor by long non-coding RNA (lncRNA) colorectal neoplasia differentially expressed (CRNDE) in low-grade glioma cells [321].
The receptor for 5-HETE, and also other lipid mediators, is G2A/GPR132 [83]. Higher expression of this receptor, according to GEPIA, is associated with a worse prognosis for a GBM patient (p = 0.052) [9], yet there is no significant upregulation of this receptor expression in GBM tumors [8,9].
5-oxo-ETE may also play an important role in tumorigenic mechanisms in GBM. The receptor for this lipid mediator is OXER1/GPR99 [220][221][222]. The expression of this receptor does not differ between GBM tumor and healthy brain tissue [8,9]. According to GEPIA, higher expression of OXER1/GPR99, the receptor for 5-oxo-ETE, is associated with a worse prognosis for a GBM patient [9]. OXER1/GPR99 is also a receptor for 2-oxoglutarate, LTC 4 , and LTD 4 [252]. There is a lack of thorough research on the importance of 5-oxo-ETE in tumorigenesis in GBM tumors.

12-Lipoxygenase Pathway in Glioblastoma Multiforme
In GBM tumors, expression of 12S-LOX/ALOX12 and 12R-LOX/ALOX12B is not different from healthy brain tissue [8,9], nor is it associated with prognosis severity [9], nor is the expression of the receptor for 12S-HETE, i.e., GPR31, elevated and affecting prognosis [8,9]. In contrast, the expression of eLOX3/ALOXE3 in GBM tumors is lower than in other brain tissue [9,205]. On the other hand, the transcriptomics analysis by Seifert et al. showed no differences between eLOX3/ALOXE3 expression levels in GBM tumor and healthy brain tissue [8]. Downregulation of eLOX3/ALOXE3 expression in GBM tumor is associated with increased expression of miR-18a, which downregulates eLOX3/ALOXE3 expression [205]. At the same time, eLOX3/ALOXE3 expression is also not related to the prognoses of GBM patients [9].
12-LOX is involved in tumorigenesis in GBM. Studies on various cell lines have shown that 12-LOX expression is common in GBM cancer cells [309]. For this reason, 12-LOX inhibitors inhibit proliferation and reduce the viability of GBM cells [309,322]. 12-LOX inhibitors also inhibit the migration of GBM cells because they reduce the expression of matrix metalloproteinase 2 (MMP2) in these cells [309]. However, the exact mechanism of 12-LOX action on tumorigenic processes in GBM is poorly studied. The fact that eLOX3/ALOXE3 is anticancer in nature [205] suggests that a lipid mediator not formed by eLOX3/ALOXE3 is responsible for the pro-cancer properties of 12-LOX. Perhaps it is 12-HETE, a lipid mediator with proven pro-cancer properties in other cancers [323,324]. In addition, higher expression of G2A/GPR132, a receptor for 5-HETE, 12-HETE, 15-HETE, and 9-HODE, is associated with a worse prognosis for a GBM patient (p = 0.052) [9]. The oncogenic properties of G2A/GPR132 were also demonstrated in a study on fibroblasts [189], although there is no higher expression of G2A/GPR132 in GBM tumors than in healthy brain tissue [8,9].
12-LOX may also have anti-cancer properties. It converts ARA 20:4n-6 into 12-HpETE, a lipid from the hydroperoxyl group, and for this reason, it can cause lipid peroxidation, which, when free ARA 20:4n-6 is in excess and this PUFA is over-processed, has a destructive effect on the cell [325].
The products of eLOX3/ALOXE3 activity are hepoxilins and trioxilins [200,203], lipid mediators of physiological importance. However, there is a lack of studies on the importance of these lipid mediators in tumorigenesis in GBM.
Analysis on the GEPIA portal [9] and the transcriptomics analysis by Seifert et al. [8] showed no differences in the expression of EPHX2, the enzyme responsible for converting hepoxilins into trioxilins, between GBM tumors and healthy brain tissue [276,280]. At the same time, according to GEPIA, higher EPHX2 expression in GBM tumors is associated with a tendency toward a worse prognosis (p = 0.072), which may indicate that hepoxilins and trioxilins may have some role in neoplastic processes in GBM.

15-Lipoxygenase Pathway in Glioblastoma Multiforme
GEPIA [9] and Seifert et al. [8] showed no differences in the expression of 15-LOX-1/ALOX15 and 15-LOX-2/ALOX15B between GBM tumors and healthy brain tissue. According to GEPIA, the expression level of these enzymes does not affect the prognosis for patients [9]. Studies on various GBM lines have shown differences in the expression of 15-LOX-1/ALOX15 and 15-LOX-2/ALOX15B in GBM cancer cells [309]. 15-LOX is important in the function of GBM cancer cells, and 15-LOX inhibitors reduce the viability and migration of GBM cancer cells [309]. On the other hand, increasing the expression and activity of 15-LOX-1/ALOX15 throughout the body may have an anti-tumor effect against GBM, as shown by gene therapy using an adenovirus transducing the ALOX15 gene [329]. This effect may depend on 13-HODE and 15-HETE.
15-HETE can activate G2A/GPR132 [83]. Higher expression of this receptor. according to GEPIA. is associated with a worse prognosis for a GBM patient (p = 0.052) [9]. At the same time, the importance of this receptor in GBM has not been thoroughly investigated. Studies in other models have shown that G2A/GPR132 is an oncogene [189]; that is, 15-HETE through activation of G2A/GPR132 has a pro-cancer effect. At the same time, there is no significant upregulation of this receptor expression in GBM tumors [8,9].
The significance of lipoxins in GBM tumors has not been thoroughly investigated. The expression level of the LXA 4 receptor ALX/FPR2 does not differ between GBM tumors and healthy brain tissue (Table 7) [8,9]. The expression level of this receptor in GBM tumors does not affect prognosis. However, it may be important in tumorigenesis in GBM tumors. Studies on U-87 MG cells have shown that silencing ALX/FPR2 reduces the proliferation and migration of the cells tested [331]. In addition, cells with silenced ALX/FPR2 showed lower expressions of VEGF, a major pro-angiogenic factor. However, this receptor is activated not only by LXA 4 but also by other factors [332]-for this reason, the importance of LXA 4 in tumorigenic processes in GBM cannot be determined. Red background-higher expression in the tumor; blue background-lower expression in the tumor; red background-worse prognosis with higher expression.
The expression levels of various LOX are not associated with prognoses for GBM patients [9]. This indicates that the LOX pathway is not as relevant to cancer processes as other pathways. For this reason, drugs targeting LOX may show poor efficacy in GBM therapy. At the same time, the analyses performed in this study show that higher expression of OXER1 (the receptor for 5-oxo-ETE) and higher expression of G2A/GPR132 (the receptor for various HETE) are associated with poor prognosis [9]. This indicates a therapeutic target for future drugs developed for the treatment of GBM. In addition, higher expression of GGT1 in GBM tumors is associated with worse prognosis, and higher expression of GGT5 and EPHX2 is associated with a trend of worse prognosis for GBM patients. This indicates a future direction for research into tumor mechanisms in GBM.

Pan-Cancer Analysis of Genes Related to LOX Pathway and GBM
Similar to the COX pathway, we performed a pan-cancer analysis of the expression of the genes involved in the LOX pathway using the data from the GEPIA web server [9].
The expression of eLOX3/LOXE3 is reduced in GBM tumors. At the same time, there is no change in the expression of this enzyme relative to healthy brain tissue in lower grade gliomas. It is also reduced in two more types of tumors. For this reason, a decrease in eLOX3/LOXE3 expression may be considered specific to GBM.
In GBM tumors, there is elevated expression of 5-LOX/ALOX5 and FLAP/ALOX5AP relative to healthy brain tissue, which is similar to lower grade gliomas [9]. Expression of these proteins is elevated in 9 and 11 tumor types, respectively. In a similar number of tumor types, there is a reduction in the expressions of 5-LOX/ALOX5 and FLAP/ALOX5AP. This indicates that the elevated expressions of 5-LOX/ALOX5 and FLAP/ALOX5AP may be glioma-specific.
The expression of other LOX is not altered in GBM and lower grade gliomas, which is similar to most other types of cancer. In GBM tumors, there are elevated expressions of LTA 4 H/LTA4H and LTC 4 S/LTC4S relative to healthy tissue [9]. In lower grade gliomas, there is higher expression of only LTC 4 S/LTC4S [9]. According to Seifert et al., in II and III grade gliomas, there are higher expressions of LTA 4 H/LTA4H but not LTC 4 S/LTC4S relative to healthy brain tissue [8]. LTA 4 H/LTA4H expression is elevated in 4 out of 31 analyzed tumor types. LTC 4 S/LTC4S is upregulated in six tumor types but downregulated in eleven types [9]. Therefore, the elevated expression of LTA 4

H/LTA4H and LTC 4 S/LTC4S
can be considered as specific to GBM and glioma, respectively.
GGT5 expression is upregulated in GBM and lower grade gliomas [8,9]. It is downregulated in eleven tumor types and upregulated in seven. Therefore, the elevation of GGT5 expression can be considered characteristic for gliomas.
DPEP1 expression is elevated in GBM tumors but not in lower grade gliomas (Table 8) [9]. It is decreased in six types of tumors but increased in four types, including GBM. For this reason, it can be thought that changes in DPEP1 expression are characteristic of GBM. EPHX2 expression is often decreased in tumors. In a pan-cancer analysis, 17 types of tumors had a reduced expression of this enzyme relative to healthy tissue. At the same time, in GBM tumors, EPHX2 expression does not differ relative to healthy brain tissue [8,9].

Cytochrome P450 Pathway
In addition to the processing of ARA C20:4n-6 by COX and LOX, this fatty acid can also be converted into lipid mediators with cytochrome P450. It results in the formation of epoxyeicosatrienoic acids (EET) and HETE [333].
Another important property of EET is that it enters the cell membrane and intracellular membranes. This is as a result of the incorporation of EET into glycerophospholipids at the sn-2 position [357][358][359]. EET can also be metabolized by EPHX1 and EPHX2 [357,360]. This is the same enzyme that catalyzes the conversion of hepoxilins (a hydroxy-epoxy derivative of ARA) to trioxilins [276,280]. EET are then converted to dihydroxyeicosatrienoic acid (DHET). In this form, particularly 14,15-DHET, they can activate PPARα [348,361].
HETE and EET are the direct products of cytochrome P450 activity. However, cytochromes p450 are not only involved in the production of these ARA-derived lipid mediators. In addition, CYP4F and CYP4A cause ω-hydroxylation of the already discussed eicosanoids formed in COX and LOX pathways. CYP4A and CYP4F8 are responsible for the ω-hydroxylation and (ω-1)-hydroxylation of prostaglandins, respectively [126,333], and CYP4F is responsible for the transformation of LTB 4 and lipoxins [333]. The aforementioned reactions often result in the inactivation of these lipid mediators.
It should be mentioned that the aforementioned cytochromes P450 are not only involved in the metabolism of ARA C20:4n-6. They can also metabolize other fatty acids [336], such as linoleic acid [367], and many drugs, including anticancer drugs [349,368].  Figure 7. Cytochrome P450 pathway. ARA 20:4n-6 can be converted in the cytochrome P450 pathway, resulting in the formation of various ETT and HETE. ETT can undergo further transformations where they are incorporated into glycerophospholipids in the sn-2 position; in this form, they build the cell membrane and intracellular membranes. In addition, the epoxide bond in ETT can be transformed by EPHX1 and EPHX2 into two hydroxyl groups, resulting in the formation of various DHET. ETT can also undergo ω-hydroxylation, which results in the formation of various HEET. ETT can be converted with COX. 5,6-EET then produces 5,6-epoxy-PGH2, whereas 8,9-EET produces either 8,9,11-EHET or 8,9,15-EHET. ↑-higher expression of given enzymes in GBM tumor relative to healthy tissue; ↓-lower expression of given enzymes in GBM tumor relative to healthy tissue.

Cytochrome P450 Pathway in Glioblastoma Multiforme Tumors
ARA C20:4n-6 is converted to 20-HETE [369], which increases the proliferation of GBM cells [370]. 20-HETE may also be an important pro-angiogenic factor in GBM tumors by acting on endothelial cells [369] and enhancing vascular mimicry of GBM cells [371]. Importantly, 20-HETE may not be produced by GBM cells [372] but by TAM and endothelial progenitor cells (EPCs) [373]. CYP2U1 [336,374], whose expression in GBM tumors is elevated relative to healthy brain tissue [8,9], may be responsible for 20-HETE production in GBM tumors. Nevertheless, there is very little research focused on 20-HETE production in GBM tumors. . Cytochrome P450 pathway. ARA 20:4n-6 can be converted in the cytochrome P450 pathway, resulting in the formation of various ETT and HETE. ETT can undergo further transformations where they are incorporated into glycerophospholipids in the sn-2 position; in this form, they build the cell membrane and intracellular membranes. In addition, the epoxide bond in ETT can be transformed by EPHX1 and EPHX2 into two hydroxyl groups, resulting in the formation of various DHET. ETT can also undergo ω-hydroxylation, which results in the formation of various HEET. ETT can be converted with COX. 5,6-EET then produces 5,6-epoxy-PGH 2 , whereas 8,9-EET produces either 8,9,11-EHET or 8,9,15-EHET. ↑-higher expression of given enzymes in GBM tumor relative to healthy tissue; ↓-lower expression of given enzymes in GBM tumor relative to healthy tissue.

Cytochrome P450 Pathway in Glioblastoma Multiforme Tumors
ARA C20:4n-6 is converted to 20-HETE [369], which increases the proliferation of GBM cells [370]. 20-HETE may also be an important pro-angiogenic factor in GBM tumors by acting on endothelial cells [369] and enhancing vascular mimicry of GBM cells [371]. Importantly, 20-HETE may not be produced by GBM cells [372] but by TAM and endothelial progenitor cells (EPCs) [373]. CYP2U1 [336,374], whose expression in GBM tumors is elevated relative to healthy brain tissue [8,9], may be responsible for 20-HETE production in GBM tumors. Nevertheless, there is very little research focused on 20-HETE production in GBM tumors.
According to the GEPIA [9] and to Seifert et al. [8], the expression of most of the discussed cytochromes P450 do not differ between GBM tumors and healthy brain tissue. Both sources only show higher expression of CYP2U1 and lower expression of CYP4X1 in GBM tumors compared to healthy brain tissue. CYP2U1 is the cytochrome P450 producing 20-HETE and 19-HETE [336], which shows a possible source of these two lipid mediators in GBM tumors.
GEPIA, in contrast to Seifert et al. shows reduced expression of CYP2C8 in GBM tumors (Table 9). According to the GEPIA [9], the expression of this cytochrome P450 was not linked to the prognosis of GBM patients. Expression of the receptor for 20-HETE, i.e., GPR75, does not differ in GBM tumors compared to healthy brain tissue. The expression level of GPR75 is not associated with prognosis. Table 9. Significance of cytochromes P450 and GPR75 receptors in ARA metabolism and tumorigenic processes in GBM. The expression of EPHX1 and EPHX2, enzymes involved in the conversion of EET to DHET, does not differ between GBM tumor and healthy brain tissue [8,9]. In addition, the expression levels of these enzymes are not associated with the prognosis of a GBM patient.

Pan-Cancer Analysis of Cytochrome P450 Genes and Comparison of GBM Expression against Other Cancers
Changes in the expression of various genes in GBM tumors relative to healthy tissue could be the result of tumor-specific neoplastic processes or specific mechanisms found only in GBM. For this reason, a pan-cancer analysis of the expression of the cytochromes P450 genes described above was performed using the GEPIA portal [9].
CYP2C8 expression was lower in GBM tumors relative to healthy brain tissue [9], similar to lower grade gliomas (Table 10). Downregulation of CYP2C8 expression occurs in a variety of tumors. Out of 31 analyzed cancers, seven show decreased expression of this enzyme, which shows that reduced expression of CYP2C8 is common in cancers. In 11 types of cancers out of 31, there is an increase in CYP2J2 expression. However, in GBM and lower grade gliomas, there is no change in the expression of this cytochrome P450. GEPIA also shows that in 8 out of 31 cancers, including GBM tumors, there is higher expression of CYP2U1 compared to healthy tissue. This indicates that elevated CYP2U1 expression may be associated with cancerous processes. In GBM and lower grade gliomas, there is lower expression of CYP4X1 compared to healthy brain tissue [8,9]. In the other seven types of tumors, there is also a decrease in the expression of this cytochrome p450, which suggests that decreased CYP4X1 expression in tumor may be a common feature of cancer. Table 10. Pan-cancer analysis of the expression of the cytochromes P450 and GPR75 receptor genes in question. Red background, ↑-expression higher in tumor than in healthy tissue; blue background, ↓-expression lower in tumor than in healthy tissue; gray background, =-expression does not differ between tumor and healthy tissue.

Conclusions
The importance of the most important ARA C20:4n-6-derived lipid mediators in cancer mechanisms in GBM is very well understood. These compounds, particularly PGE 2 and leukotrienes, cause the proliferation and migration of GBM cancer cells, are important in the function of GBM cancer stem cells, cause angiogenesis, and by acting on cells of the immune system, inhibit the body's anti-tumor response. However, the importance in GBM cancer processes of lesser-known ARA C20:4n-6-derived lipid mediators has not yet been investigated. We are talking, for example, about EET, lipoxins, hepoxilins, and some prostanoids, including PGF 2α and TxA 2 . Investigating the function of these compounds will provide a better understanding of GBM tumor function. It may also contribute to the development of new therapeutic approaches.