Radiotherapy is the most often applied treatment after surgical resection of glioblastoma. Application PUFAs as adjuncts enhance eradication of glioma cells. Numerous in vitro and in vivo experiments have shown that PUFAs may increase the tumoricidal effect of radiotherapy [19, 22, 26]. PUFAs have little or no cytotoxic effect on normal cells, and at the same time, they diminish the deteriorative effect of irradiation. In our study, we treated U87 MG cells with UFAs (AA, DHA, GLA, OA, EPA) at different concentrations and cell viability, LDH activity, cell growth, cell morphology and gene expression changes were analyzed. Besides classical end-point assays (LDH measurements and MTS assay) we used the RT-CES system for real-time cellular analysis. This label-free and non-invasive method measures impedance and determines cell index, an indicator of cell number, proliferation, viability, adherence and cell growth [23, 25]. We demonstrated that AA, DHA, GLA and OA treatment decreased the proliferation rate of U87 MG glioma cells and in correlation with the cytotoxic effects, decreased the total LDH activity that could be recorded (Figure 1, 2, 3 and 5). EPA is an exception: it did not decrease the proliferation rate and LDH activity at the tested concentration range (Figure 4).
AA treatment dramatically decreased cell viability and LDH activity after 72 hours (Figure 1). Based on similar effects against glioma cells AA was considered a possible therapeutic PUFA agent . When cells are irradiated and treated with AA at the same time, LDH activity, mitochondrial dehydrogenase activity, were significantly decreased (Figure 1). We also detected a decrease in normalized cell index, which is an indicator of cell proliferation. This was more pronounced when AA was applied in combination with irradiation. From these results we assume that AA treatment would hold promise in glioblastoma radiotherapy as an adjunct.
Previously it was published that 20-50 μM DHA was cytotoxic to Neuro2a cells, and the concentration range below 10 μM inhibited apoptosis, without any detectable toxic effects . We made similar observations: 25-75 μM DHA diminished the proliferation rate and altered the metabolism of U87 MG cells (Figure 2). DHA treatment had a distinct effect on medulloblastoma (DAOY and D283) cells compared to glioma cells (U87 MG and U138) regarding cell proliferation: it did not affect glioma cells, while it inhibited proliferation of medulloblastoma cells .
In concert with our results related to GLA (Figure 5), similar proliferation inhibition was reported with C6 glioma cells . They found that in tumors treated with GLA and EPA the mitochondrial membrane potential, an indicator of apoptosis, decreased significantly [30, 31]. In our study GLA diminished cell viability and LDH activity of U87 MG cells, and increased the radio sensitivity of this cell line. Similarly, GLA was found to be cytotoxic to rat 36B10 astrocytoma cells in other studies . 10-50 μM GLA significantly increased cell proliferation at the outer layer of glioma spheroids, enhancing invasion . In contrast, we found that 50-75 μM GLA did not alter the proliferation rate of U87 MG cells (Figure 3). It was reported that GLA selectively induced apoptosis in spheroids and concentrations that exceed 100 μM inhibited proliferation, thus reduced invasion . Similarly, we found that 100 μM GLA diminished the proliferation rate of U87 MG cells. Interestingly, when 50-75 μM GLA was applied as adjunct to radiotherapy, proliferation and LDH activity of U87 MG cells were reduced (Figure 3). Previously, it was found that GLA acted selectively on tumor cells, it had low neurotoxicity and it may even protect normal tissue from the cytotoxic effect of irradiation or chemotherapy [10, 11]. Therefore, the additive effects of GLA with irradiation, its possible selectivity against tumor cells and even the protection of normal tissues against irradiation would make GLA an ideal candidate for combined therapy as previously indicated earlier [10, 11].
Interestingly, we found that at lower OA concentrations (100-200 μM) normalized cell index increased (Figure 5). This suggested elevated proliferation, although this was not confirmed with the end-point assays. At higher concentrations (400 μM) OA diminished cell proliferation soon after treatment as recorded by using the real-time cellular analysis (Figure 5). A similar concentration of OA (500 μM) influenced cell proliferation in a different manner depending on cell type: it inhibited cell growth on LNCaP prostate cells, it enhanced cell proliferation in case of breast cancer cell lines (MCF-7 and MDA-MB-231) and it had no effect on a non-tumorogenic epithelial cell line (MCF10A) [34, 35]. Our results showed no benefit using OA along with irradiation.
Previously, it was reported that EPA, similarly to GLA, protected rat hippocampus from the harmful effect of LPS-induced inflammation , therefore in case of additive effects of EPA and irradiation one could predict enhanced therapeutic effects. Under our conditions EPA treatment (50-100 μM) did not affect LDH activity and cell viability. Moreover, when it was used as an adjunct with 5 Gy or 10 Gy a significant, but very moderate change in cell metabolism could be detected (Figure 4). As assessed by using real-time cell analysis technology, EPA had no effect on normalized cell index of U87 MG cells even when it was applied in combination with 5 or 10 Gy irradiation (Figure 5). From these results we assume that EPA treatment would not be good candidate as an adjunct in glioblastoma radiotherapy.
Based on our observations on cell proliferation measurements and previously published data we could conclude that among the UFA we studied, DHA, GLA and AA may provide benefit as therapeutic adjuncts in the treatment of malignant brain tumor with radiation (results are summarized in Additional file 2: Table S2).
Morphological analysis of glioma cell line treated with AA, DHA, GLA and irradiation
Holographic microscopy permits the label-free and non-invasive visualization of living cells. Furthermore, it allows the determination of cell number and confluence. An integrated image analysis algorithm makes it possible to measure more than forty parameters of each cell in a holographic image (cell volume, cell thickness, cell shape convexity, cell perimeter length, cell optical length, etc.) which reflects cytotoxicity . During apoptosis, cell membrane permeability increases and the optical density of cells decreases, this changes their texture and the contrast becomes lower (http://www.phiab.se/products/holomonitor).U87 MG glioma cells were exposed to 25 μM AA, 25 μM DHA or 50 μM GLA alone or in combination with irradiation (10 Gy) and holographic and phase contrast images were recorded to detect morphological alterations following treatment (Figure 6). Our results showed that PUFAs as adjuncts to a dose of 10 Gy significantly diminished cell number, confluence, and average cell irregularity, while average cell thickness increased significantly (Figure 7). The latter described parameters indicate cell rounding and loss of adherence, which indicates that the treatment had a cytotoxic effect on U87 MG cells.
Our results concerning cell number, confluence, average cell thickness and average cell irregularity imply that combined treatment of glioma cells with AA, DHA or GLA and radiotherapy would have inhibitory effects on invasion and metastasis.
Gene expression analysis of PUFA treated and irradiated U87 MG cells
Several molecular targets for glioma treatment are subjects of clinical trials and under development [5, 29, 38]. Due to the complexity of glioma pathogenesis the application of more than one molecular target could be a solution for proper therapy. The foundations of an effective therapy would be the better knowledge of the affected genes and miRNAs in glioma pathogenesis. Because PUFAs are supposed to be radio sensitizing agents in glioblastoma treatment, the mRNA and miRNA expression analysis presented here emphasize several potential molecular targets (our results are summarized in Additional file 2: Table S3).
We found that AA, significantly increased c-MYC expression, just like 10 Gy, and combined exposure of U87 MG cells had an increased effect (Figure 8). Determination of c-MYC expression may serve as a prognostic value in glioblastoma, its expression was increased in approximately 70% of the cases . Alteration of c-MYC expression influenced apoptosis, cell cycle progression and carcinogenesis . In Jurkat and Raji cells oleic acid and linolenic acid induced over-expression of c-MYC after 24 hours [41, 42]. On U87 MG cells we detected significant over-expression only in case of treatment with 25 μM AA (Figure 8). Although c-MYC is an oncogene, its overexpression is correlated with a higher survival probability (P < 0.0001) . This result suggests that combined therapy of AA and irradiation may be beneficial for glioblastoma treatment (Figure 8).
According to previous findings DHA did not change the total levels of TP53, impaired DNA binding of TP53 was observed in endothelial cells . Under our conditions 10 Gy significantly increased the expression of TP53 on U87 MG cells, while GLA and DHA did not influence its expression (Figure 8). If AA was added as adjunct to radiotherapy the expression of TP53 was significantly decreased (Figure 8). In our previous paper we investigated the effect of a three-four times higher concentration of AA, DHA and GLA applied for a shorter incubation period on glioma cell lines . We found that they altered the expression of TP53 in GBM5 and U373 glioma cell lines, but not in GBM2 cell line , similarly to U87 MG cells observed in the present study. Differences in TP53 expression changes could be due to the different TP53 status, the variability of overall TP53 expression and relative levels of isoforms as these differences in glioblastoma are well documented .
One explanation of the beneficial effect of PUFAs would be that they may increase the activity of antioxidant enzymes . The excess of reactive oxygen species induce lipid peroxidation and hydroperoxide generation in glioma cells, which decrease their viability and their sensitivity to irradiation [6, 32, 45]. Therefore, we evaluated the expression of HMOX1, AKR1C1 and NQO1 genes which have a role in the defense mechanism against oxidative stress.
HMOX1 is a heat-shock protein; it degrades heme to biliverdin, CO and iron . HMOX1 inhibits apoptosis and inflammation, diminishes oxidative stress, enhances the rate of proliferation and playes a role in resistance to irradiation or chemotherapy [46–49]. HMOX1 is a potential therapeutic target, it is over-expressed and facilitates angiogenesis in glioma and may influence the outcome of the disease [47, 50]. Irradiation induced HMOX1 expression on pancreatic cancer cells . We observed the same effect when we irradiated U87 MG cells with 10 Gy (Figure 8). Exposure to AA or DHA or 10 Gy combined with AA or with DHA also increased its expression in a significant manner (Figure 8).
AKR1C1 encodes a drug-metabolizing enzyme; the level of expression of this gene may influence the prognosis of different cancers . Temozolomide treatment significantly increased the expression of AKR1C1 in U373 and T98G glioblastoma cells . We noticed the same effect when U87 MG cells were exposed to irradiation and AA or GLA treatment (Figure 8).
When AA, DHA or GLA was added as adjunct, NQO1 expression increased significantly, and treatment with DHA by itself also raised NQO1 expression. The exact function of NQO1 in cancer genesis is not yet determined, but it is known that it activates the apoptotic protein TP53 and it is a priority target of glioblastoma chemotherapy [52, 53].
In our study combined treatment of 50 μM GLA and irradiation reduced significantly the over-expression of NOTCH1 which could be recorded when cells were subjected only to GLA or they were only irradiated (Figure 8). The main setback in radiotherapy is the radioresistance of cancer stem cells, which may be attributed to the Notch signaling pathway [5, 29]. Altered Notch activity was detected in several types of tumors; it mediates self-renewal of glioblastoma and influences the response to radiotherapy [5, 29, 54].
Endoplasmic reticulum (ER) stress response may be an indicator of the efficiency of glioma treatment [55–57]. We evaluated both elements of the ER stress response: the prosurvival arm (unfolded protein response (UPR) pathway) which is responsible for the alleviation of ER stress, and the proapoptotic arm, which is activated in case of intensive stress, when the UPR pathway is overwhelmed. The UPR pathway is represented by GRP78, while DDIT3 (GADD153) stands for the proapoptotic arm of the ER stress response [55–58]. Under our conditions significant over-expression of GRP78 could be recorded when U87 MG cells were treated with AA or DHA alone, or when cells were irradiated. Similar up-regulation could be observed when cells were treated in combination with irradiation and AA, DHA or GLA (Figure 8).
GRP78 silencing delays glioma cell growth and sensitizes human glioblastoma cell lines to chemotherapy [56, 58]. GRP78 is a prognostic marker; overexpression of GRP78 increases radioresistance of glioblastomas . Combination of PUFA treatment with irradiation did not decrease the overexpression of GRP78 or of DDIT3 (Figure 8), thus it seems that PUFAs radio sensitize U87 MG cells through other pathways than the ER stress response.
We examined the expression of EGR1, TNF-α, c-FOS and FOSL1 that were proven to be early-response genes and were up-regulated due to ionizing radiation: . c-FOS, EGR1 and FOSL1 contains a region with a serum response element (SRE) as promoter, which is responsible for the sensitivity of these genes to ionizing radiation [59, 60]. Ionizing radiation induces reactive oxygen species and up-regulates EGR1, a zinc-finger protein with six CArG elements, which regulates the transcription of genes involved in differentiation and cell growth [59–61]. AA, DHA and GLA up-regulated EGR1 and treatment with GLA enhanced the effect of irradiation (Figure 8). In contrast, co-exposure with AA and 10 Gy increased EGR1 expression in a significantly lower manner than application of 10 Gy by itself (Figure 8).
TNF-α is a growth promoting cytokine, which determines the outcome of glioblastoma . At low TNF-α concentration glioma cells have a higher survival rate, while overexpression of TNF-α induces neuronal cell death . Irradiation alone, and combined with DHA and GLA increased TNF- α expression significantly (Figure 8). AA treatment of irradiated U87 MG cells significantly decreased the overexpression of TNF- α compared to cells that were only irradiated (Figure 8). Thus, it seems that AA diminishes the harmful effect caused by irradiation induced TNF- α over-expression. Thus, this ω-6 fatty acid may have therapeutic effect when it is combined with irradiation, reducing possible side-effects.
In our study treatment of U87 MG cells with 25 μM AA, 50 μM GLA; irradiation or exposure to 10 Gy and PUFAs significantly increased the expression of FOSL1 compared to control cells (Figure 9). Overexpression of FOSL1 may cause carcinogenesis, and is a typical characteristic of glioma . FOSL1 over-expression induced differentiation, inhibited proliferation, growth and reduced tumorogenicity of C6 glioma cell line, so it may be a potential target for glioma treatment .
c-FOS also contains serum response elements in its promoter [59, 60]. c-FOS, is an oncogenic transcription factor, which regulates PKC-mediated signaling pathways , and it can induce carcinogenesis . Just as in case of TNF-α, when we treated U87 MG cells with 25 μM AA, it significantly decreased the overexpression of c-FOS, which is otherwise induced by irradiation (Figure 9).
GADD45A is a target for therapeutic interventions in cancer . Exposure to 25 μM AA or 10 Gy significantly increased the expression of GADD45A (Figure 9). As a consequence of GADD45A overexpression TP53 is phosphorylated and it stabilizes TP53 after DNA damage .
While different tumor types present specific microRNA signatures, several microRNAs are deregulated in glioblastoma, suggesting their involvement in the basic processes of tumorigenesis and response to therapy . To further analyze the mechanism of action of AA, DHA and GLA in combination with irradiation, miRNA expression levels were evaluated. Irradiation with 10 Gy and PUFA treatment did not alter significantly the expression of miR-34a, miR-96, miR-148a, miR-148b and miR-152. However, when cells were treated with DHA miR-146a was significantly up-regulated. Interestingly, its expression decreased when it was exposed to GLA. In case of combined exposure to irradiation and GLA the expression of miR-146a increased significantly compared with GLA alone or with irradiation. Recently it was shown that miR-146a suppresses gastric cancer cell invasion and metastasis in vitro and in vivo. From our data it would be interesting to investigate the differential effects of DHA and GLA on miR-146a in relevance with metastatic potential of glioblastoma, especially that GLA was the only PUFA which, in combination with radiation, could induce its expression suggesting potent antimetastatic effects.