1. Introduction
As previously reported [
1], evidence from both in vitro and in vivo data has demonstrated that drugs such as statins and bisphosphonates targeting the mevalonic acid pathway and consequently the synthesis of isoprenoids and cholesterol exert, beyond their lipid-lowering effects, pleiotropic actions, including immune regulation [
1,
2] and cancer prevention [
3,
4] as well as epigenetic effects [
5]. However, observed differences in the anti-cancer potency of these drugs might be related to cell type specific inhibitory activities from these drugs on uptake of glucose and other nutrients such as essential amino acids [
6,
7,
8,
9].
The anti-tumorigenic effects of statins vary between different types of cancer: the amelioration of breast cancer prognosis was extensively reviewed [
10]; survival or recurrence by statin was documented in one study with 146,326 participants [
11] and other studies with 75,684 [
12] or 124,669 [
13] women. There is also data available on the beneficial effect of bisphosphonates for the treatment of breast cancer [
14]. The curative effect of bisphosphonates on breast cancer is also mentioned in recent publications [
15] discussing potential options for the treatment of lysyl oxidase positive, estrogen receptor negative (LOX+, ER−) breast cancer patients. In prostate cancer patients, a statin-associated reduction of mortality has been documented in more than 100,000 cases [
16]. However, the treatment success also appears to be influenced by mitochondrial DNA mutations and associated metabolic consequences [
17] and non-responders to statin-therapy with persistent high serum cholesterol still have a higher cancer risk [
18].
Epidemiological evidence projecting statins and/or bisphosphonates as anticancer agents is conflicting, which largely depends on the type of cancer in question [
19,
20] and, to the best of our knowledge, no epidemiological data exist on the application of these drugs in osteosarcomas. Based on the working hypothesis, that statins and bisphosphonate-responses could be linked with the basic proliferation rate of respective tumor cells, the aim of this study was to elucidate underlying mechanisms by combining metabolic analyses with transcriptomic and complementary immune blot analyses.
3. Discussion
In our previous studies, we demonstrated that statins and bisphosphonates suppress the one-carbon metabolism and induce epigenetic alterations in tumor cells [
5]. Statins act through inhibition of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR) by contrast to the bisphosphonates, which act several steps downstream in this pathway as inhibitors of farnesyltransferase. This leads to an accumulation of isopentenyl-pyrophosphate (IPP), which also acts as a genoprotective agent and thus could be responsible for a weaker effect on gene-regulation [
56], as shown in
Table 6.
Based on these results, in this study we studied the effect of such mevalonate pathway inhibitors on metabolic processes. It was a challenge to find out the background of drug-responses of cancer cell-types, which are characterized by remarkable differences not only in their growth rates but also in their epithelial or mesenchymal background. A key metabolite related to cell proliferation rate is NADPH, which is essential for the synthesis of dTMP and therefore for DNA synthesis and replication during cell division and replication. For synthesis of dTMP, it is generated by an enzymatic reaction involving three enzymes (thymidylate synthetase (TYMS), dihydrofolate reductase (DHFR), serine hydroxymethyltransferase 1 and 2 (SHMT1/2) where DHFR needs NADPH to provide 5,10-methylene-THF. Importantly, further NADPH producing reactions and pathways like glycolysis include the NADPH-producing pentose-phosphate cycle as well as the KREBS or tricarboxylic acid cycle [
57,
58].
The differences in NADP(H) levels measured in this study after treatment with simvastatin or ibandronate could be linked to a general downregulation of the energy metabolism. Indeed, ibandronate had just a minor effect on NADPH, which was significantly reduced by simvastatin in PC-3 and MDA-MB-231 cells (
Figure 2A,B). Thus, we could not confirm a previous report indicating that simvastatin might protect MG-63 osteosarcoma cells from oxidative stress [
59], which could be a result of an induction of the anti-apoptotic BCL2 apoptosis regulator by the runt related transcription factor 2 (RUNX2) in response to the treatment with hydrogen peroxide in that study [
60]. As both genes had about the same basal expressions in the cell lines that were investigated in this study and were not significantly regulated, it rather appears possible that reduction of NADPH is a general sign for reduced energy metabolism in treated cells, also because data on the influence of simvastatin and ibandronate on the production of reactive oxygen species (ROS) appear contradictory (
Supplementary Table S2 showing respective literature citations). Additionally, a comparative analysis of our transcriptomic data from Venn diagrams showed that simvastatin had a stronger impact on gene regulation than ibandronate in all investigated cell lines with the exception of MG-63 osteosarcoma cells (
Table 6). However, it remains to be established if this is related to the (relatively) slow proliferation rate or to the osteogenic lineage of MG-63 cells or their different cell cycle response, which showed an arrest in the G2 phase.
Statins and bisphosphonates associated starvation affect mTOR-signaling resulting in an impaired uptake of nutrients such as essential amino acids including methionine [
61,
62], which is responsible for previously observed epigenetic alterations [
5] and glucose metabolism [
63,
64,
65]. This results in the stimulation of sestrin (SESN2) and inhibition of the mTOR [
64,
66,
67] and RHOB pathway leading to stimulation of autophagy [
68,
69,
70,
71], as well as the downregulation of fatty acid synthase FASN [
72]. The latter is partially associated with the upregulation of RHOB and sestrin [
63,
64,
69] in mevalonate-dependent or independent manner.
Regarding prostate cancer, it has been postulated that ibandronate exerts its anti-proliferative effect through a reduction in the prenylation of RAC and via disruption of the NADPH oxidase complex [
73]. The expression of RAC (gene name AKT1, not regulated in our study) at the protein level and the associated NADPH oxidases is cell type dependent and mirrors the mechanism of how bisphosphonates attenuate osteoclasts [
74,
75,
76]. As simvastatin downregulates the DNA methyltransferase DNMT1 to a higher extent than the demethylating agent decitabine, a cell-line specific epigenetic reaction, which appears to be present in all investigated cell lines except U-2 OS, cannot be excluded [
5].
The divergent responses induced by the tested drugs may lead to the hypothesis that the ectodermal (epithelial) origin of MDA-MB-231 and PC-3 cells versus the mesothelial origin of the osteosarcoma cell lines U2-OS and MG-63 might play a role in the observed differences. Furthermore, mutations in the retinoblastoma gene (RB1) as found in MDA-MB-231 and PC-3 cells [
25,
26] but not in MG-63 and U-2OS cells might also be partially responsible for the observed effects [
77,
78]. However, as the fast proliferating osteosarcoma cell line Saos-2 is more sensitive towards bisphosphonate-treatments [
79] than U-2 OS or MG-63 cells, it appears that the critical parameter is just the slower growth rate of MG-63 and U-2 OS cell lines and not the mesothelial background.
Recently, it has been shown that the above-mentioned downregulation of energy metabolism could play a role in the maintenance of a stem-cell-like status, where an increased autophagy plays a decisive role [
80]. Inhibitors of the mevalonate pathway are known to induce lysosomal activity and associated effects on autophagy [
81].
SESN2, which is a key molecule of the autophagy pathway [
36], was significantly upregulated in most of our treated cell lines (
Figure 3). A concordant stimulation of the small GTPase
RHOB could indicate an increase in protein degradation through an endolysosomal pathway [
38], especially in simvastatin-treated PC-3 and MDA-MB-231 cells. SESN2 belongs to the highly conserved gene family, playing a key role in processes of adaptation to extreme climatic conditions in Antarctica [
68]. Their primary function of SESN2 is to sensor lysine availability for further transport to mTOR via the GATOR complex that consists of a series of GTPases. SESN2 is upregulated upon stoppage of lysine import, as in situations of nutrient deprivation, starvation, or intoxication [
67,
82,
83]. The u-regulation of SESN2 in statin- or ibandronate-treated cells regulates the activity of AMP-activated protein kinase (AMPK) [
66,
84,
85] via liver kinase B (LKB1) mediated phosphorylation, thus promoting a status of quiescence [
86,
87,
88,
89]. Interestingly, in osteoblasts there is also a link between sestrin cell cycle attenuating activity of vitamin D (VD). In fact, VD induces the production of sestrins and thus leads to a cell cycle arrest [
90].
This has been associated with metabolic processes that are turned down in starving or quiescent cells, which do not proliferate but appear to be protected against necrosis or apoptosis: it is upregulated by simvastatin and ibandronate and is known to interact with a complex that interacts with GTPases of the RAG family to promote mTORC1 translocation to the lysosomal surface named GATOR2 in an amino-acid-sensitive manner. Thus, it functions as a negative regulator of this pathway by preventing proper mTORC1 localization to the lysosome in response to essential amino acids [
82] in all investigated cell lines. SESN2 attenuates the import of essential amino acids such as methionine by inhibiting the NPRL2 (nitrogen permease regulator like 2) gene, which is also responsible for the uptake of transcobalamin 2 (TCN2) and cobalamin (vitamin B12) [
62]. This could be responsible for the downregulation of the one carbon metabolism (folate cycle) and associated inhibition of the thymidylate synthase and downregulation of epigenetic regulators such as DNA-methyl-transferases [
5]. Data also indicate SESN2 protects cells from glucose starvation-induced necroptosis [
64]. SESN2 regulation has been demonstrated to occur via TP53 dependent and independent mechanisms [
91,
92,
93,
94].
The SESN2 [
95] gene is closely associated and deacetylated by the NAD–dependent histone deacetylase SIRT1 by a similar mechanism as described for the retinoblastoma gene RB1, which is known for its role in the transition from the G1 to the S-phase and mutated in PC-3 and MDA-MB-231 but not in MG-63 and U2-OS cell lines [
96].
SESN2 cooperates with the hypoxia-inducible gene REDD1/RTP801 (gene name:
DDIT4), which is part of a pathway, where mTOR inhibition is induced by hypoxia AMPK [
97]. This
DDIT4 gene was significantly stimulated in U-2 OS cells both by ibandronate and simvastatin. Inductions of REDD1/RTP801 together with SESN2 by DNA damage are required for phosphorylation of the controlling 4E-BP1 (gene name:
EIF4EBP1) elongation factor in situations of DNA damage [
98].
This gene and its metabolic background is tightly regulated by microRNAs [
46], which confers also to some microRNAs, where the extent of regulation is associated with cell type
MIR21 microRNA (
Table 4) which is known to be RAS-activated [
47]. Thus, a minus 3-fold ibandronate-induced downregulation in U-2 OS osteosarcoma cells might be associated with RAS-inactivation in this cell line.
As topoisomerase-inhibitors are known to act via the p-JUN-SESN2/AMPK pathway [
99], the observed (both here and in a further study [
100]) statin-mediated downregulation of topoisomerase (DNA) II α (TOP2A) could mimic this effect (see
Table S1 in the
supplementary materials).
Like
SESN2,
ANLN is also a Wnt/β-catenin responsive gene [
101] and the relation of sestrin to Wnt/β-catenin and AMPK-signaling and histone deacetylase 5 is well documented [
102].
The impact of ANLN on persistence of estrogen receptor positive breast cancer was shown by respective experiments showing a cell-cycle arrest in G2/M, lowered expression of cyclins D1, A2, and B1, as well as altered cell morphology [
103].
Furthermore, an upregulation of
SESN2 (with concomitant downregulation of cell cycle regulators such as
DDIT4,
CCNA2, and
CCNB1) and silencing of
ANLN are known to facilitate but not necessarily induce apoptosis [
104,
105].
In addition, a downregulation of methyl-histones appears to be associated with the above-mentioned mechanisms of growth arrest and starvation. This induces a downregulation of developmental genes, which is also known to be a sign of quiescence [
106]. This could also provide some explanation for the induction of growth arrest in cancer cells upon treatment with statins or bisphosphonates.