br G Glycolysis rates of mTN cells transduced with a
(G) Glycolysis rates of mTN Calcipotriol transduced with a control plasmid or a plasmid encoding the catalytic subunit of G6PC3.
(J) Western blot showing levels of HK2 in mTC-SAM-sgTom and mTC-SAM-sgBach1 after retroviral transduction of two different Hk2 shRNAs.
(K) Glycolysis rates of cells shown in (J).
(L) Left, Transwell migration assay of mTC-SAM-sgTom and mTC-SAM-sgBach1 after retroviral transduction of Hk2 shRNAs. Right, representative photos of migrated cells.
(M) Glycolysis-derived ATP production rate of mTC and mTN cells (n = 3) as judged by Seahorse analyses.
(N) Total ATP production rate of mTC and mTN cells (n = 3) as judged by Seahorse analyses.
(P) Steady-state ATP levels in mTC and mTN cells with and without BACH1 expression (n = 2).
Figure 7. Targeting Glycolysis Inhibits Antioxidant- and BACH1-Induced Migration and Metastasis
(A) Schematic of glycolysis and inhibitors (red) and an activator (green) used to define the role of an enzymatic step in the ability of antioxidants to increase cell migration. Blue ovals, proteins whose genes were increased in mTN versus mTC cells. Red oval, gene reduced in mTN cells.
(B) Survival of mTC and mTN cells (n = 3) incubated for 48 h with the glycolysis inhibitor 2-DG. Values are the percentage of untreated mTC cells.
(legend continued on next page)
Both inhibitors reduced lung metastases of mTN cells, but 3-BP was more effective (Figure 7J) and also reduced lung metastases of mTC-SAM-Bach1 cells (Figure 7K).
This study identifies BACH1-induced glycolysis as a new mechanism that drives lung cancer metastasis, which can be stimulated by antioxidants and inhibited by oxidative stress. The results are relevant to the debate on the promises and perils of antioxidant supplementation, especially for lung cancer patients and for the 30% of them who harbor KEAP1/NRF2 mu-tations (Berger et al., 2016).
Metastasis requires at least six steps: detachment, intravasa-tion, immune evasion, survival in circulation, extravasation, and colonization in a distant organ. We find that antioxidant-induced BACH1 stabilization increases the efficiency of several of these steps. BACH1 was involved in antioxidant-induced metastasis in KP mice and metastasis from subcutaneously transplanted tumor cells, which involves all six steps. BACH1 was both required and sufficient for metastasis in the i.v. transplantation experiments, which involves the last three steps. Immune evasion is likely a less important factor because the effect was cell autonomous and observed in immune-competent and immune-deficient mice.
Antioxidants stimulated metastasis in the presence and absence of p53. However, in early lung tumors, antioxidants accel-erate progression only when p53 is present (Sayin et al., 2014). Indeed, p53 appears to underlie the effects observed in those studies. Thus, the mechanisms underlying ROS-mediated barriers to tumor progression differ depending on the stage of tumor devel-opment. Previous studies concluded that the K mouse model is metastasis resistant in the absence of cooperating mutations— although advanced-grade adenocarcinoma are observed (Fisher et al., 2001; Jackson et al., 2001, 2005; Meuwissen et al., 2001; Sutherland et al., 2014). Therefore, metastasis in the presence of p53 was surprising. But similarly, antioxidants activated BACH1, glycolysis, and migration in human NSCLC cells with and without known TP53 mutations. Thus, BACH1 enables antioxidants to overcome p53’s potent metastasis-inhibiting activity.
We find that BACH1 is stabilized by antioxidants and that this effect is driven by reduced levels of ROS and free heme. The latter result confirms reports that free heme stimulates nuclear export and proteasomal degradation of BACH1 (Gozzelino et al., 2010; Pamplona et al., 2007; Suzuki et al., 2004; Zenke-Kawasaki et al., 2007). In a companion paper, Lignitto et al.
show that heme induces BACH1 degradation by stimulating an interaction with the ubiquitin ligase FBXO22. They also show that NRF2 activation stimulates lung cancer metastasis by inducing HO-1, leading to heme degradation, FBXO22 uncou-pling, and BACH1 stabilization. Consequently, in that study, high BACH1 levels stimulated metastasis in the setting of high NRF2 activity. In contrast, antioxidant supplementation sup-pressed the activity of NRF2 in lung cancer cells—and thus reduced the expression of NRF2-target genes—without altering Hmox1 transcripts or HO-1 protein levels. Thus, in our study, high BACH1 levels stimulated metastasis in the setting of low NRF2 activity. Our two studies provide complementary evidence that BACH1 stabilization stimulates lung cancer metastasis and that BACH1 can be stabilized by both exogenous and endoge-nous antioxidants. But importantly, we also find that BACH1 is pro-metastatic in the absence of antioxidants.