Discussion on the meta-analysis
Discovery anti-tumor drugs is a luxurious and time-consuming process, also the percentage of drugs meets the clinic is quite a little. Thus, the development of new functions of existing drugs has become a hot study topic. It is known to all that obesity and T2DM relate with poor prognosis of breast cancer closely. Metformin, the most widely used antihyperglycemic drug to treat T2DM, could maintain weight loss as well [5]. It has been verified that most tumors are sensitive to metformin [6, 8]. Consistently, metformin synergizes with conventional anticancer therapy to kill tumors and repress migration [32]. But the outcomes of clinical studies of metformin in the treatment of breast cancer were not entirely positive [15, 29, 33], the metformin effect in breast cancer treatment was still under discussion. In this study, we intended to estimate the connection between metformin and distant breast cancer distant metastasis. With the forest plot (Figs. 2 and 3), we demonstrated that adjuvant metformin might contribute to suppress breast cancer metastasis.
The adverse effects brought by metformin should not be neglected, while gastrointestinal distress, including transient mild nausea and moderate diarrhea, was the toxicity that people usually met during metformin treatment [31, 34]. There was no grade 3 and 4 treatment-related adverse event (TRAE) reported in the included publications.
This meta-analysis had several limitations. First, the small sample size limited to obtain firm conclusions. Second, the dose of metformin could not adjust to consistent because most of the trails were retrospective and the individual distinctions among the patients were unavoidable. It has been proved that adjuvant metformin therapy repressed HER 2 + breast cancer cells [35, 36] while ER- breast cancer cells resisted to this drug [37]. Thus, the effects of metformin might depend on the molecular types of breast cancer. Here we recommended that subgroups for different hormone receptors status should be designed in the future clinical trials.
The mechanisms of metformin to suppress tumor metastasis
Inhibiting tumor metastasis by metformin was a complex process of multiple pathways, including: AMPK activation, epithelial mesenchymal transition (EMT) inversion, DNA methylation modulation, interfering with TGF-β pathway and tumor microenvironment. Moreover, N-cadherin, vimentin, β-catenin, snail, Rac1 and MMP-2/9 were downregulated while E-cadherin and phosphorylated AMPK increased, which resulted in tighter intercellular connections, weaker migration and movement of tumor cells. Metformin affected insulin-like growth factor (IGF) pathway by repressing IGF-1 receptor and IGF-2 molecule [38]. As shown in Fig. 5, metformin mediated different pathways to prevent metastasis.
Intercellular reaction
AMPK, a widely known metformin effector, was activated in two pathways. One way was that metformin activated liver kinase B1 (LKB1), the upstream of AMPK, and the other pathway was mitochondrial complex I activation decreased ATP/AMP ratio under metformin stimulation, which triggered AMPK activation indirectly [39, 40]. Phosphorylated AMPK functioned as an inhibitor to repress a series of moleculars activation, containing: STAT3, smad-2/3, Akt, ERK, mTOR, PKCγ and Twist. Snail, which was the downstream of ERK and Y-box binding protein-1 (YB-1) [41, 42], a oncogenic transcription/translation factor, and also the direct target gene of metastasis-related gene YAP [43]. This protein regulated E-cad expression with Twist. Additionally, it participated in E-cad promoter hypomethylation modulation with Slug [41]. Metformin induced Snail ubiquitination after LKB1 phosphorylation, which helped Snail’s interaction with E3 ligase FBXL14 [44]. Furthermore, metformin decreased Twist with obliterating interaction between GSK-3β and Twist via reducing Akt/GSK-3β pathway [45]. mTOR, of which inhibition under metformin therapy mediated suppression of HIF-1α/VEGF-A and p70s6k [46, 47], was another downstream molecule of Akt. In addition, metformin augmented Foxo3a nuclear localization and protein stabilization to active Foxo3a with IKKβ repression and MDM2 phosphorylation involvement, leading to the level of E-cad increase [42]. For protein kinase Cγ (PKCγ), which was attenuated after AMPK-α1 phosphorylation, it modulated Hs90α activation [48].
The Rac1 and RhoA GTP downstream migratory protein took charge of the migration of cell, thus they were required in cancer migration and metastasis. Metformin downregulates Rac1 in different pathways. Firstly, metformin suppressed FAK/Akt signaling pathway [49] or CXCL12/CXCR4 [50] to decrease downstream factors Rac1 and RhoA GTP expression. Secondly, metformin elevated the level of phosphatase and tensin (PTEN), a protein controls tumor metastasis, to inhibit Akt/Rac1 axis [51]. Thirdly, Rac1 GTP was reduced by metformin-mediated cAMP increase [50]. Moreover, CD24, a mucin-like adhesion molecule, enhanced the metastasis potential of malignant cells. Distant metastasis in patients with refractory breast cancer was mainly composed of CD24 positive cells, which was thought as a marker indicating poor prognosis of breast cancer. Recently, CD24 has been confirmed to significantly be downregulated by metformin [52].
Micro-RNA, modulators of many cellular signaling pathways, have been verified to take parts in metformin treatment. MiR-26a was enhanced under metformin stimulation to inhibit Akt phosphorylation [53]. Similarly, miR-381, upregulated by metformin, significantly interfered with YAP transcription to affect Snail [43]. MiR-30a, another member upregulated by metformin, attenuated SOX4, which was a oncogenic transcription factor and epithelial mesenchymal transformation (EMT) regulator, reversing the process of EMT [54]. Because of the DNA methylation, metformin therapy increased the level of mir-570-3p, while decreased lncRNA H19 by metformin-induced DNA methylation. The former was shown to reduce the invasion of tumor cells through inhibiting LCMR1 and ATG12 [55], while the latter was also been implicated to control tumor metastasis by not only reducing MMP-9, but also elevating AMPK phosphorylation and let-7, a potent tumor suppressor microRNA [56, 57].
Effects on tumor microenvironment cells
The word “tumor microenvironment” was created in 2011, and defined to include endothelial cells, pericytes and immune inflammatory cells [58]. Endothelial cells were shown to benefit to developmental and tumor-associated angiogenesis, while pericytes wrapped around the endothelial tubing of blood vessels and prevented the tumors from entering the circulatory system, making it possible to reduce subsequent hematogenous dissemination [59, 60]. Metformin downregulated micro-vessel density (MVD) by inhibiting platelet-derived growth factor B (PDGF-B), ending with reducing the ratio of endothelial cells/ pericytes, leakage and hypoxia, namely “vessel normalization” [61, 62]. Moreover, Angiopoietin-like protein 4 (ANGPIL4) was decreased during metformin treatment after HIF-1α suppression [63]. Another modulation of metformin to block tumor metastasis was the attenuation of M2-like polarization of tumor associated macrophages (TAM) with phosphorylated AMPK α1 [64].