**3. Discussion**

OS is the most common type of primary bone cancer, often associated with a high degree of malignancy, early metastasis, and rapid progression. There are different molecular types of OS; however, the tumor suppressor *TP53* is the most frequently altered gene in OS [6,26]. Although the survival rate in the case of non-metastatic patients is quite high, distant metastases (mainly to lungs) are found in about 20% of OS patients and the prognosis for these patients is still poor due to strong resistance of OS to chemotherapy [34]. Given the high rates of recurrence after tumor resection, resistance to chemotherapy and stagnation in the survival rates of OS patients during the last years, an intensive search for novel agents and alternative strategies to combat this malignancy is suggested [35].

1

Evidence from recent studies suggests that exogenous supplementation of AKG may exert anti-cancer effects against colorectal carcinoma or breast cancer [19,22]. Moreover, our recent study has revealed that exogenous AKG can inhibit cell proliferation and stimulate differentiation of normal osteoblasts [36]. However, the influence of AKG on osteosarcoma cell lines has not been studied so far. In the present study, to assess the anti-osteosarcoma potential of exogenous AKG, two primary osteosarcoma cell lines harboring *TP53* mutations were used, i.e., Saos-2 (p53-null cell line) and HOS (p53 mutant). The AKG treatment resulted in significantly reduced OS cell proliferation with IC<sup>50</sup> values of approx. 35 mM for both cell lines (in the BrdU assay). In the study conducted by other authors [22], AKG inhibited DNA synthesis in colon carcinoma cell lines such as Caco-2, HT-29, and LS-150 with IC<sup>50</sup> values of approx. 55 mM, 64 mM, and 67 mM, respectively. In turn, AKG was able to inhibit slightly the growth of MDA-MB-231 breast cancer cells at a concentration of 1 mM [19], while OS cell proliferation was inhibited by AKG at a concentration of 2.5 and 5 mM (in the Saos-2 and HOS cell lines, respectively). These data suggest a cell type-specific effect of AKG, probably related to different oncogenic pathways in the tested tumor cell lines. It is worth mentioning that exogenous AKG does not easily penetrate into the cell (although this occurs via simple diffusion), thus the intracellular level of this metabolite depends on the extracellular concentration [8] in an in vitro study and probably on the time required for its consumption by the cell. Nevertheless, even cell-permeable AKG derivatives (e.g., dimethyl alpha-ketoglutarate), which are often used to increase the intracellular AKG level, were applied at a concentration of 4 mM [25].

Previous studies conducted by other authors have shown that exogenous AKG (25 and 50 mM) can modulate the expression of cell cycle-related proteins such as cyclin D1 and the inhibitor of cyclin-dependent kinases p21Waf1/Cip1 [22]. Cyclin D1 is well known for its role in the response to mitogenic signals and regulation of the G<sup>1</sup> to S phase transition in the cell cycle. This protein activates cyclin-dependent kinases CDK4 and CDK6, which form active complexes with cyclin D1 and phosphorylate the RB protein, leading to transcriptional activation of genes required for cell division [37]. Cyclin D1 and CDK4 have also been reported to be overexpressed in osteosarcoma and related to its occurrence and development [38,39]. In turn, the p21 protein can function as a regulator of cell cycle progression at the G<sup>1</sup> checkpoint through binding to cyclin/CDK2 complexes and inhibition of RB phosphorylation or direct interaction with PCNA (proliferating cell nuclear antigen), which both trigger inhibition of DNA replication [40]. P21 has been shown to be involved in both p53-dependent and p53-independent control of cell proliferation, differentiation, and cell death [40]. Recently, p21 has been found to be significantly downregulated in osteosarcoma tissue, compared to their matched adjacent non-tumor tissues [41], and upregulation of this protein has been shown to inhibit proliferation of OS cells [42,43]. In our study, AKG (25 and 50 mM) was able to decrease cyclin D1 expression in the HOS cells but not in the Saos-2 cells, which do not express cyclin D1 [32]. On the other hand, AKG remarkably upregulated p21 only in the p53 null OS cells, but not in HOS cells harboring a *TP53* mutation. The results of our study suggest that AKG can disturb cell cycle progression through different mechanisms depending on the distinct genetic characteristics of OS cells.

A successful OS therapy requires, among others, effective agents that promote apoptotic cell death [44]. Drug-induced apoptosis can often occur through extrinsic or intrinsic pathways involving the activation of initiator caspase-8 and -9, respectively. The initiation of the mitochondrial pathway is under the control of the Bcl-2 family members, such as pro-apoptotic Bax and anti-apoptotic Bcl-2, and the Bax/Bcl-2 ratio is a critical determinant of the cell's apoptotic threshold. Bax insertion into the outer mitochondrial membrane leads to its permeabilization and release of various apoptotic proteins, which trigger the activation of caspase 9/3 signaling cascade [45]. In the present study, the AKG treatment induced apoptosis in both OS cell lines through activation of caspase 3. The further study revealed that the AKG-treatment of the Saos-2 cells resulted mainly in the upregulation of Bax, suggesting that changes in the ratio of Bax/Bcl-2 proteins could contribute to the subsequent activation of caspases-9 and -3. Surprisingly, a similar mechanism associated with the induction of apoptotic death was mediated by upregulation of IDH1 in osteosarcoma cell lines [24].

Mitogen-activated protein kinases (MAPKs), such as extracellular signal-regulated kinases (ERK1/2), c-Jun N-terminal protein kinases (JNKs), and p38, have been identified as key proteins of the signaling pathways that transmit mitogenic signals into the nucleus in response to various extracellular stimuli [46]. The ERK pathway is usually identified as a key mediator of cell proliferation and survival [47]. In turn, JNK may play a dual role in cancer cell survival, but preferentially exerts a pro-apoptotic effect [33,46,48]. Mounting evidence indicates that activation of JNK kinase in response to various anti-cancer agents may contribute to OS cell death [49,50]. In our study, AKG induced phosphorylation of JNK in the Saos-2 cells. Moreover, the pretreatment of the OS cells with the JNK inhibitor abolished the AKG-induced increase in the phosphorylation of this kinase and partially inhibited the AKG treatment-induced apoptosis. These data suggest an essential role of the JNK signaling pathway in AKG-induced apoptotic death of OS cells. It is well known that activated JNK promotes an intrinsic apoptotic pathway and cytochrome c release from the mitochondrion via multiple mechanisms, including Bax and Bcl-2 regulation [33]. More importantly, it phosphorylates cytoplasmic Bax-anchor proteins, which triggers dissociation of Bax from the complexes, its translocation to mitochondria, and induction of outer mitochondrial membrane permeabilization [51]. Moreover, JNK may increase the expression of Bax through transcriptional activation of c-Jun [52]. Furthermore, it may induce apoptosis through direct Bcl-2 phosphorylation and inhibition of its anti-apoptotic activity [33]. Since the AKG treatment of the Saos-2 cells resulted in both JNK activation and reduction of the Bax/Bcl-2 ratio, we can suppose that the caspase-9 and caspase-3 activation observed was linked with the mechanisms above mentioned.

The present study also revealed that AKG decreased ERK1/2 phosphorylation in the Saos-2 cells. The ERK pathway mediates several upstream signals from growth factors (e.g., vascular endothelial growth factor, VEGF) or proinflammatory stimulants, and regulates cell proliferation, migration, and metastasis in most cancers, including osteosarcoma [53]. It has been shown that overexpression and abnormal activation of the ERK signaling pathway are implicated in the pathogenesis of OS; therefore, this pathway is an attractive molecular target in OS [53–56]. Many studies have shown that suppression of this pathway by anticancer agents results in increased apoptosis and decreased metastasis in OS [53,57,58]. Therefore, besides the JNK pathway, inhibition of ERK1/2 activation may also be implicated in the AKG-mediated inhibition of cell cycle progression and programmed OS cell death observed in our study. However, this issue needs further investigations, to confirm these suggestions. On the other hand, the decrease in the ERK1/2 activation may be also implicated in the anti-migratory and anti-invasive effects of AKG observed in the OS cells.

Earlier studies have shown that exogenous AKG has the ability to reduce the level of the HIF-1α subunit, resulting in downregulation of HIF-1 downstream targets, including the production of VEGF, and inhibition of angiogenesis [20,21]. Moreover, a recent study has shown that exogenous supplementation of AKG prevented tumor growth and metastasis of breast cancer cells through stabilization of PHD2 and decreasing HIF-1α [19]. Furthermore, other strategies associated with intracellular AKG accumulation resulted in anti-metastatic effects [17,24]. In the present study, the AKG supplementation also markedly inhibited cell motility and invasion of both OS cell lines in a concentration-dependent manner, which confirms the anti-metastatic potential of this compound. Moreover, AKG was able to decrease the production of TGF-β and VEGF by the OS cells, i.e., growth factors that are implicated in osteosarcoma progression and metastasis.

Although TGF-β acts in most cancers both as a tumor suppressor in premalignant stages and a tumor promoter in advanced stages of the disease, in OS it exerts only pro-tumoral effects through the promotion of metastasis [59]. It has been shown that the level of TGF-β in sera of OS patients is higher compared to those of healthy donors, which is correlated with a high grade of disease and associated with chemoresistance and presence of metastases in lungs and other sites [60]. Moreover, in vitro studies have revealed that TGF-β is implicated in the EMT-like phenomenon, stimulates proliferation of OS cells, and exerts pro-angiogenic properties in OS [60,61]. In addition, the secretion of TGF-β by OS cells or stromal cells can regulate the phenotype and function of the microenvironment in order

to stimulate switching its function to pro-tumoral [59]. Since TGF-β plays a pro-tumoral role in OS, the downregulation of TGF-β by AKG seems to be an important feature of this compound in terms of anti-cancer activity.

In the present study, AKG was also able to inhibit the VEGF production in both OS cell lines, which is in agreement with the results of previous studies conducted by Matsumoto et al. [20,21] in the Hep3B hepatocellular carcinoma cell line and LCC (Lewis lung carcinoma) cell line. In their study, AKG decreased VEGF production at a concentration of 7.5 mM and 5 mM, respectively. Similarly, in our study, the lowest concentrations inducing significant inhibition of VEGF production were 5 mM and 10 mM of AKG for HOS and Saos-2 cells, respectively. VEGF is a key potent tumor-derived pro-angiogenic factor influencing both the tumor microenvironment and cancer cells. It acts in a paracrine manner on endothelial cells which leads to the promotion of angiogenesis [62]. Moreover, VEGF can act in an autocrine manner on several cancer cells, including aggressive osteosarcoma phenotypes, which leads to activation of various signaling pathways in these cells (e.g., PI3K/Akt and MEK/ERK), ultimately supporting tumor growth [63]. VEGF plays an important role in the pathogenesis of OS [64]. VEGF serum levels in OS patients are elevated and associated with poor prognosis [65]. Moreover, overexpression of VEGF is a predictor of pulmonary OS metastasis [66,67]. Recently, it has been shown that silencing of VEGF in Saos-2 cells inhibited cell proliferation and promoted apoptosis in vitro [68]. Therefore, the inhibition of VEGF production in osteosarcoma cells by AKG treatment may have therapeutic value.

As mentioned earlier, we used cell lines with p53 function deficiency in our study; however, the *TP53* gene mutation does not occur in all OS cases. Early studies reported that the rate of the *TP53* gene mutation in OS is around 20% [69]; however, recent studies have shown that more than 90% of osteosarcomas have either missense mutations in this gene or structural variation in p53 [70]. Nevertheless, some percentages of osteosarcomas have functional p53, which raises a question of whether AKG would act in these cases in a similar way as in cells harboring the *TP53* mutation. A recent study has shown that AKG is an effector molecule of p53-mediated tumor suppression, and its accumulation in p53-deficient tumors can partially recapitulate the p53 action linked with the remodeling of cancer cell metabolism through epigenetic modifications and alterations of gene expression [25]. It has also been shown that exogenous AKG can switch metabolism from glycolytic to oxidative, which can prevent the growth and metastasis of breast cancer cells [19]. Based on the analysis of cell lines in which AKG exhibited anti-cancer activity [19–22], we identified that all these cell lines, except the LS-180 colon cancer cell line, have *TP53* mutations; nevertheless, AKG also inhibited the proliferation of LS-180 cells with the wild type of p53. Therefore, we can speculate that AKG would similarly affect OS lines with functional p53, and that more than one mechanism of AKG activity may operate in cancer cells, depending on the distinct genetic/epigenetic characteristics of these cells. However, further studies of the AKG influence on OS cell lines with functional p53 and their metabolism are needed to confirm this hypothesis.

#### **4. Materials and Methods**
