br RNA seq data and qPCR analyses confirmed regulation
RNA-seq data and qPCR analyses confirmed regulation of BACH1’s previously known—and new Transfac-identified—target
genes. For example, Nqo1 expression was lower, and Pkp3 and Mmp1a expression was higher, in mTN than mTC DAPT (Figure 2C). Pkp3 and Mmp1a have been linked to cell invasion (Furukawa et al., 2005; Toruner et al., 2004). These gene expression changes would be consistent with increased BACH1. Analyses of data in The Cancer Genome Atlas (TCGA) revealed that high BACH1 expression in LUAD patients is associated with poor survival (Fig-ure 2C), as are its targets PKP3 and MMP1 (Figures S3G and S3H). BACH1 also predicts poor survival in patients with kidney clear cell carcinoma and pancreatic adenocarcinoma (Figures S3I and S3J).
BACH1 is degraded by the proteasome when levels of intra-cellular free heme are high (Suzuki et al., 2004; Zenke-Kawasaki et al., 2007), and ROS stimulates the release of free heme by oxidizing heme-containing proteins (Gozzelino et al., 2010; Pam-plona et al., 2007). We therefore hypothesized that antioxidants, by lowering ROS levels, inhibit the release of free heme and thereby prevent BACH1 degradation (Figure 2D).
To test this possibility, we first confirmed that BACH1 protein levels were higher in mTN than mTC cells (Figure 2E); Bach1 mRNA levels were not altered, which suggests stabilization of the protein (Figure 2B). NAC and Trolox also increased BACH1 protein levels in mTC and A549 cells (Figures 2F, S3K, and S3L). Furthermore, activating the expression of endogenous an-tioxidants with the NRF2-activating peptide KI-696 (Sayin et al., 2017) increased BACH1 levels in mTC and mTN cells and increased the migration of mTC cells (Figures S3M and S3N). Conversely, increasing oxidative stress by adding H2O2 reduced BACH1 (Figure 2G).
Next, we measured free heme under conditions of low and high oxidative stress. Free heme levels were reduced by NAC and Tro-lox and increased by H2O2 and diamide, as judged by enzymatic assays and a fluorescent heme sensor (Figures 2H and 2I). More-over, BACH1 levels in mTC and mTN cells were reduced after addition of hemin (a heme derivative) to the culture medium (Fig-ure 2J). Hemin-induced BACH1 degradation was blocked by the proteasome inhibitors MG132 and bortezomib and the neddyla-tion inhibitor MLN4924 (Figure 2K). BACH1 also disappeared at a slower rate in antioxidant-treated lung cancer cells after protein synthesis was inhibited with cycloheximide (Figures 2L and Fig-ure S3O); incubation with MG132, bortezomib, and MLN4924 blocked BACH1 degradation in the cycloheximide-treated cells.
BACH1 Is Required for Antioxidants to Induce Metastasis and Can Induce Metastasis in the Absence of Antioxidants Next, we evaluated the functional role of BACH1 for the migra-tory and metastatic phenotype of antioxidant-treated cells. The
(G) Top, western blots of BACH1 in mTC and mTN cells incubated for 24 h with H2O2 (200 mM). Bottom, amounts of BACH1 were quantified from two experiments. (H) Steady-state levels of free heme in A549 cells incubated with 1 mM NAC for 2 weeks. Red bar, positive control; cells were incubated with 200 mM H2O2 overnight to stimulate release of free heme.
(I) Free heme levels measured with the pCDNA-HS1 sensor in A549 cells incubated for 2 weeks with 1 mM NAC or 100 mM Trolox. Cells were incubated with 5 mM diamide for 48 h as a positive control.
(J) Left, western blot showing amounts of BACH1 and ACTIN in cells incubated with 10 mM hemin for 24 h. Right, amounts of BACH1 determined with densi-tometry. Values are the mean of 3 cell lines/condition from 2 experiments and are normalized to mTC.
(K) Western blot showing amounts of BACH1 and ACTIN in cells incubated with 10 mM hemin. Cells were incubated with 10 mM MG132, 1 mM bortezomib, or 2 mM MLN4924 for 30 min before the addition of hemin.
Figure 3. BACH1 Is Required for Antioxidant-Induced Metastasis and Can Induce Metastasis in the Absence of Antioxidants
(A) Left, Transwell migration assay of mTC and mTN cells incubated with 10 mM hemin for 24 h. Right, representative photos of migrated cells.
(B) Western blots showing amounts of BACH1 in mTC and mTN cells transduced with CAS9 and sgRNAs targeting Bach1; sgRNA targeting dTomato (Tom) was used as control. HISTONE 3 (H3) was the loading control.
(C) Left, Transwell migration assay of mTN-sgBach1, mTN-sgTom, and control mTC cells (n = 2 biological replicates/condition). Right, representative photos of migrated cells.
(D) Left, lung tumor burden in NSG recipient mice 3 weeks after i.v. injection of control and Bach1-deficient mTC and mTN cells (0.5 3 105 cells/mouse; n = 5–8 mice/condition). Right, representative lung sections.