Sulforaphane Activates a lysosome-dependent transcriptional program to mitigate oxidative stress

Abstract

Oxidative stress underlies a number of pathological conditions, including cancer, neurodegeneration, and aging. Antioxidant-rich foods help maintain cellular redox homeostasis and mitigate oxidative stress, but the underlying mechanisms are not clear. For example, sulforaphane (SFN), an electrophilic compound that is enriched in cruciferous vegetables such as broccoli, is a potent inducer of cellular antioxidant responses. NFE2L2/NRF2 (nuclear factor, erythroid 2 like 2), a transcriptional factor that controls the expression of multiple detoxifying enzymes through antioxidant response elements (AREs), is a proposed target of SFN. NFE2L2/NRF2 is a target gene of TFEB (transcription factor EB), a master regulator of autophagic and lysosomal functions, which we show here to be potently activated by SFN. SFN induces TFEB nuclear translocation via a Ca²⁺-dependent but MTOR (mechanistic target of rapamycin kinase)-independent mechanism through a moderate increase in reactive oxygen species (ROS). Activated TFEB then boosts the expression of genes required for autophagosome and lysosome biogenesis, which are known to facilitate the clearance of damaged mitochondria. Notably, TFEB activity is required for SFN-induced protection against both acute oxidant bursts and chronic oxidative stress. Hence, by simultaneously activating macroautophagy/autophagy and detoxifying pathways, natural compound SFN may trigger a self-defense cellular mechanism that can effectively mitigate oxidative stress commonly associated with many metabolic and age-related diseases.Abbreviations: ANOVA: analyzes of variance; AREs: antioxidant response elements; Baf-A1: bafilomycin A₁; BHA: butylhydroxyanisole; CAT: catechin hydrate; CCCP: carbonyl cyanide m- chlorophenylhydrazone; CLEAR: coordinated lysosomal expression and regulation; DCFH-DA: 2',7'-dichlorofluorescin diacetate; FBS: fetal bovine serum; GFP: green fluorescent protein; HMOX1/HO-1: heme oxygenase 1; KD: knockdown; KEAP1: kelch like ECH associated protein 1; KO: knockout; LAMP1: lysosomal associated membrane protein 1; MCOLN1/TRPML1: mucolipin 1; ML-SA1: mucolipin-specific synthetic agonist 1; ML-SI3: mucolipin-specific synthetic inhibitor 3; MTOR: mechanistic target of rapamycin kinase; MTORC1: mechanistic target of rapamycin kinase complex 1; NAC: N-acetylcysteine; NFE2L2/NRF2: nuclear factor: erythroid 2 like 2; NPC: Niemann-Pick type C; PBS: phosphate-buffered saline; PPP2/PP2A: protein phosphatase 2; Q-PCR: real time polymerase chain reaction; ROS: reactive oxygen species; RPS6KB1/S6K1/p70S6K: ribosomal protein S6 kinase B1; SFN: sulforaphane; TFEB: transcription factor EB; WT, wild-type.

Keywords: Autophagy; NFE2L2/NRF2; ROS; TFEB; lysosome; sulforaphane.

Figure 1.

SFN promotes autophagic flux. (A) Chemical structure of SFN. (B) SFN treatment increased the number of autophagosomes. HeLa cells stably expressing mRFP-GFP-LC3 were treated with SFN (5–15 μM) for 9 h. Scale bar: 10 μm or 2 μm (for zoom-in images). (C) Quantification of autophagosome formation shown in B. N = 30–40 randomly selected cells from at least 4 independent experiments. Experimenters were blind to the treatment conditions. (D) Western blot analysis of LC3-II protein expression in HeLa cells that were treated with SFN (15 μM, 3–9 h). Torin1 (1 μM) was used as a positive control to induce autophagosome formation. (E) Quantitative analysis of LC3-II levels shown in D. from n = 4 independent experiments. (F) Western blot analysis of LC3-II protein expression with or without SFN (15 μM, 9 h) treatment, and in the presence and absence of Baf-A1 (0.5 μM; used as an inhibitor of autophagosome-lysosome fusion [30]) in HeLa cells. (G) Quantification of results shown in F. (n = 3 independent experiments). For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 1.

SFN promotes autophagic flux. (A) Chemical structure of SFN. (B) SFN treatment increased the number of autophagosomes. HeLa cells stably expressing mRFP-GFP-LC3 were treated with SFN (5–15 μM) for 9 h. Scale bar: 10 μm or 2 μm (for zoom-in images). (C) Quantification of autophagosome formation shown in B. N = 30–40 randomly selected cells from at least 4 independent experiments. Experimenters were blind to the treatment conditions. (D) Western blot analysis of LC3-II protein expression in HeLa cells that were treated with SFN (15 μM, 3–9 h). Torin1 (1 μM) was used as a positive control to induce autophagosome formation. (E) Quantitative analysis of LC3-II levels shown in D. from n = 4 independent experiments. (F) Western blot analysis of LC3-II protein expression with or without SFN (15 μM, 9 h) treatment, and in the presence and absence of Baf-A1 (0.5 μM; used as an inhibitor of autophagosome-lysosome fusion [30]) in HeLa cells. (G) Quantification of results shown in F. (n = 3 independent experiments). For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 2.

SFN promotes lysosomal biogenesis. (A) LAMP1 staining in SFN-treated (15 μM, 3–9 h) HeLa cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm or 2 μm (for zoom-in images). (B) Quantification of LAMP1 immunofluorescence shown in A. N = 30–40 randomly selected cells from at least 3 independent experiments. (C) Western blot analysis of LAMP1 protein expression in HeLa cells that were treated with SFN (15 μM, 3–9 h). GAPDH served as a loading control. (D) Quantitative analysis of LAMP1 levels, as shown in C. n = 3 independent experiments. (E) Effects of SFN on lysosome activation. HeLa cells were treated with SFN (10 or 15 μM) for 9 h, and then stained with LysoTracker Red DND-99 (50 nM; labeling acidified lysosomes [24]) for 15 min. Scale bar: 10 μm or 2 μm (for zoom-in images). (F) Quantification of LysoTracker intensity, as shown in E. N = 30–40 randomly selected cells from at least 3 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 2.

SFN promotes lysosomal biogenesis. (A) LAMP1 staining in SFN-treated (15 μM, 3–9 h) HeLa cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm or 2 μm (for zoom-in images). (B) Quantification of LAMP1 immunofluorescence shown in A. N = 30–40 randomly selected cells from at least 3 independent experiments. (C) Western blot analysis of LAMP1 protein expression in HeLa cells that were treated with SFN (15 μM, 3–9 h). GAPDH served as a loading control. (D) Quantitative analysis of LAMP1 levels, as shown in C. n = 3 independent experiments. (E) Effects of SFN on lysosome activation. HeLa cells were treated with SFN (10 or 15 μM) for 9 h, and then stained with LysoTracker Red DND-99 (50 nM; labeling acidified lysosomes [24]) for 15 min. Scale bar: 10 μm or 2 μm (for zoom-in images). (F) Quantification of LysoTracker intensity, as shown in E. N = 30–40 randomly selected cells from at least 3 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 3.

SFN induces TFEB nuclear translocation and expression of TFEB target genes. (A) SFN (10 or 15 μM, 4 h) induced TFEB nuclear translocation in HeLa cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm. (B) Average ratios of nuclear vs. cytosolic TFEB immunoreactivity shown in A. N = 40–50 randomly selected cells from at least 5 independent experiments. (C) SFN (15 μM for 9 h) treatment increased mRNA expression of TFEB target genes in 1321N1 cells analyzed by Q-PCR (n = 3 independent experiments). (D) Western blot analysis of MTOR and RPS6KB1 activity, assayed by ratios of p-MTOR vs. total MTOR and p-RPS6KB1 vs. total RPS6KB1 in SFN (15 μM) -treated HeLa cells for 6–12 h. (E) Quantification of the results shown in D. from n = 3 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 3.

SFN induces TFEB nuclear translocation and expression of TFEB target genes. (A) SFN (10 or 15 μM, 4 h) induced TFEB nuclear translocation in HeLa cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm. (B) Average ratios of nuclear vs. cytosolic TFEB immunoreactivity shown in A. N = 40–50 randomly selected cells from at least 5 independent experiments. (C) SFN (15 μM for 9 h) treatment increased mRNA expression of TFEB target genes in 1321N1 cells analyzed by Q-PCR (n = 3 independent experiments). (D) Western blot analysis of MTOR and RPS6KB1 activity, assayed by ratios of p-MTOR vs. total MTOR and p-RPS6KB1 vs. total RPS6KB1 in SFN (15 μM) -treated HeLa cells for 6–12 h. (E) Quantification of the results shown in D. from n = 3 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 4.

SFN regulates TFEB activation and autophagic induction via ROS elevation. (A) SFN (15 μM for 2 h) treatment increased cellular ROS levels, as measured by the fluorescence intensity of DCFH-DA (green; a ROS-sensitive dye [24]) in 1321N1 cells. The increase was prevented by co-treatment of NAC (5 mM), a commonly used antioxidant compound [24]. Scale bar: 40 μm. (B) Quantification of results shown in A. N = 30–40 randomly selected cells from 3 independent experiments. (C) NAC (5 mM) pretreatment blocked SFN (15 μM, 4 h)-induced, but not torin1 (1 μM)-induced TFEB translocation in HeLa GFP-TFEB stable cells. Nuclei were labeled with DAPI (blue). Scale bar: 10 μm or 2 μm (for zoom-in images). (D) Ratios of nuclear vs. cytosolic TFEB shown in C. N = 30–40 randomly selected cells from at least 4 independent experiments. (E) Effects of NAC (5 mM) co-treatment on SFN (15 μM, 9 h) -induced expression of TFEB target autophagic genes in 1321N1 cells (n = 3 independent experiments). (F) Effects of NAC (5 mM) co-treatment on SFN-induced autophagosome accumulation in HeLa cells stably expressing mRFP-GFP-LC3. Scale bar: 10 μm. (G) Quantitative analysis of results shown in F. (N = 30–40 randomly selected cells from 3 independent experiments). For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 4.

SFN regulates TFEB activation and autophagic induction via ROS elevation. (A) SFN (15 μM for 2 h) treatment increased cellular ROS levels, as measured by the fluorescence intensity of DCFH-DA (green; a ROS-sensitive dye [24]) in 1321N1 cells. The increase was prevented by co-treatment of NAC (5 mM), a commonly used antioxidant compound [24]. Scale bar: 40 μm. (B) Quantification of results shown in A. N = 30–40 randomly selected cells from 3 independent experiments. (C) NAC (5 mM) pretreatment blocked SFN (15 μM, 4 h)-induced, but not torin1 (1 μM)-induced TFEB translocation in HeLa GFP-TFEB stable cells. Nuclei were labeled with DAPI (blue). Scale bar: 10 μm or 2 μm (for zoom-in images). (D) Ratios of nuclear vs. cytosolic TFEB shown in C. N = 30–40 randomly selected cells from at least 4 independent experiments. (E) Effects of NAC (5 mM) co-treatment on SFN (15 μM, 9 h) -induced expression of TFEB target autophagic genes in 1321N1 cells (n = 3 independent experiments). (F) Effects of NAC (5 mM) co-treatment on SFN-induced autophagosome accumulation in HeLa cells stably expressing mRFP-GFP-LC3. Scale bar: 10 μm. (G) Quantitative analysis of results shown in F. (N = 30–40 randomly selected cells from 3 independent experiments). For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 5.

SFN induces TFEB translocation through Ca²⁺-calcineurin-mediated de-phosphorylation of TFEB. (A) BAPTA-AM (10 μM, 1 h; a membrane-permeable Ca²⁺ chelator [24]) pretreatment blocked SFN (15 μM, 4 h)-induced TFEB nuclear translocation in HeLa cells. Scale bar: 10 μm (B) Quantification of results shown in A. N = 30–40 randomly selected cells from at least 4 independent experiments. (C) The effects of SFN (15 μM) and NAC (5 mM) co-treatment for 4 h on ML-SA1 (20 μM; synthetic agonist of MCOLN1 [40])-induced Ca²⁺ release, measured with Fura-2 (F₃₄₀/F₃₈₀) imaging, in MCOLN1-expressing HEK293 cells. Mean ± s.e.m. is shown from at least 30 cells per coverslip. Ionomycin was used to achieve the maximal Ca²⁺ response. (D) Quantification of results shown in C. from n = 3 independent experiments. (E) Acute application of SFN did not evoke lysosomal Ca²⁺ release. SFN (15 μM) did not evoke Ca²⁺ release, measured with the GCaMP3 signal (F₄₇₀), in GCaMP3-MCOLN1-expressing HEK293 cells. ML-SA1 (20 μM) readily elicited rapid and robust Ca²⁺ release in the same cells. (F) FK506 (5 μM; a calcineurin inhibitor [21]), but not okadaic acid (50 nM; a PP2A inhibitor [21]) inhibited SFN (15 μM)-mediated TFEB nuclear localization in HeLa GFP-TFEB stable cells. Scale bar: 10 μm. (G) Quantification of results shown in F. N = 30–40 randomly selected cells from 3 independent experiments. (H) SFN (15 μM for 4 h) induced downshift of endogenous TFEB electrophoretic mobility. TFEB was detected using a specific antibody against TFEB, and TFEB KO HeLa cells were used as a negative control. Experiments were repeated multiple times and a representative blot was shown. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 5.

SFN induces TFEB translocation through Ca²⁺-calcineurin-mediated de-phosphorylation of TFEB. (A) BAPTA-AM (10 μM, 1 h; a membrane-permeable Ca²⁺ chelator [24]) pretreatment blocked SFN (15 μM, 4 h)-induced TFEB nuclear translocation in HeLa cells. Scale bar: 10 μm (B) Quantification of results shown in A. N = 30–40 randomly selected cells from at least 4 independent experiments. (C) The effects of SFN (15 μM) and NAC (5 mM) co-treatment for 4 h on ML-SA1 (20 μM; synthetic agonist of MCOLN1 [40])-induced Ca²⁺ release, measured with Fura-2 (F₃₄₀/F₃₈₀) imaging, in MCOLN1-expressing HEK293 cells. Mean ± s.e.m. is shown from at least 30 cells per coverslip. Ionomycin was used to achieve the maximal Ca²⁺ response. (D) Quantification of results shown in C. from n = 3 independent experiments. (E) Acute application of SFN did not evoke lysosomal Ca²⁺ release. SFN (15 μM) did not evoke Ca²⁺ release, measured with the GCaMP3 signal (F₄₇₀), in GCaMP3-MCOLN1-expressing HEK293 cells. ML-SA1 (20 μM) readily elicited rapid and robust Ca²⁺ release in the same cells. (F) FK506 (5 μM; a calcineurin inhibitor [21]), but not okadaic acid (50 nM; a PP2A inhibitor [21]) inhibited SFN (15 μM)-mediated TFEB nuclear localization in HeLa GFP-TFEB stable cells. Scale bar: 10 μm. (G) Quantification of results shown in F. N = 30–40 randomly selected cells from 3 independent experiments. (H) SFN (15 μM for 4 h) induced downshift of endogenous TFEB electrophoretic mobility. TFEB was detected using a specific antibody against TFEB, and TFEB KO HeLa cells were used as a negative control. Experiments were repeated multiple times and a representative blot was shown. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 6.

TFEB is required for SFN-induced autophagic flux. (A) Western blot analysis of LC3-II protein expression in SFN (15 μM, 9 h)-treated WT and TFEB KO HeLa cells in the presence and absence of Baf-A1 (0.5 μM). Torin1 (1 μM) was used as a control to induce autophagy. (B) Quantitative analysis of LC3-II levels under various experimental conditions shown in A. from n = 5 independent experiments. (C) The effects of TFEB KO on SFN (15 μM, 9 h)-induced expression of autophagic genes in HeLa cells (n = 3 independent experiments). For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 6.

TFEB is required for SFN-induced autophagic flux. (A) Western blot analysis of LC3-II protein expression in SFN (15 μM, 9 h)-treated WT and TFEB KO HeLa cells in the presence and absence of Baf-A1 (0.5 μM). Torin1 (1 μM) was used as a control to induce autophagy. (B) Quantitative analysis of LC3-II levels under various experimental conditions shown in A. from n = 5 independent experiments. (C) The effects of TFEB KO on SFN (15 μM, 9 h)-induced expression of autophagic genes in HeLa cells (n = 3 independent experiments). For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 7.

TFEB is required for SFN-induced expression of NFE2L2. (A) Effects of SFN (15 μM for 9 h) treatment on mRNA expression of NFE2L2 gene in WT and TFEB KO HeLa cells (n = 3 independent experiments). (B) Effects of SFN (15 μM, 4 h) treatment on NFE2L2 subcellular localization, detected by an anti-human NFE2L2 antibody in HepG2 cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm. (C) Ratios of nuclear vs. cytosolic NFE2L2 shown in B. N = 40–50 randomly selected cells from at least 4 independent experiments. (D) Western blot analysis of the KD efficiency of a specific TFEB-targeting siRNA in HepG2 cells (n = 3 independent repeats). (E) Effects of TFEB KD on SFN (15 μM, 4 h)-induced NFE2L2 nuclear translocation in HepG2 cells. Scale bar: 10 μm. (F) Average ratios of nuclear vs. cytosolic NFE2L2 immunoreactivity shown in E. N = 30–40 randomly selected cells from 3 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 7.

TFEB is required for SFN-induced expression of NFE2L2. (A) Effects of SFN (15 μM for 9 h) treatment on mRNA expression of NFE2L2 gene in WT and TFEB KO HeLa cells (n = 3 independent experiments). (B) Effects of SFN (15 μM, 4 h) treatment on NFE2L2 subcellular localization, detected by an anti-human NFE2L2 antibody in HepG2 cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm. (C) Ratios of nuclear vs. cytosolic NFE2L2 shown in B. N = 40–50 randomly selected cells from at least 4 independent experiments. (D) Western blot analysis of the KD efficiency of a specific TFEB-targeting siRNA in HepG2 cells (n = 3 independent repeats). (E) Effects of TFEB KD on SFN (15 μM, 4 h)-induced NFE2L2 nuclear translocation in HepG2 cells. Scale bar: 10 μm. (F) Average ratios of nuclear vs. cytosolic NFE2L2 immunoreactivity shown in E. N = 30–40 randomly selected cells from 3 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 8.

TFEB is required for SFN-induced removal of excessive ROS. (A) Western blot analysis of the KD efficiency of TFEB-targeting siRNA in 1321N1 cells (n = 3 independent experiments). (B) The effects of SFN (15 μM, 9 h) treatment and TFEB KD on CCCP (10 μM) -induced ROS increases, measured by DCFH-DA (green) imaging, in 1321N1 cells. Scale bar, 40 μm. (C) Quantification of results shown in B. from 40–50 randomly selected cells (n = 4). (D) Western blot analysis of the TFEB KD efficiency in NPC fibroblast cells (n = 3 independent experiments). (E) The effects of SFN (15 μM, 9 h) treatment and TFEB KD on ROS levels in NPC cells. Scale bar: 20 μm. (F) Quantitative analysis of ROS levels shown in E. N = 40–50 randomly selected cells from at least 4 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 8.

TFEB is required for SFN-induced removal of excessive ROS. (A) Western blot analysis of the KD efficiency of TFEB-targeting siRNA in 1321N1 cells (n = 3 independent experiments). (B) The effects of SFN (15 μM, 9 h) treatment and TFEB KD on CCCP (10 μM) -induced ROS increases, measured by DCFH-DA (green) imaging, in 1321N1 cells. Scale bar, 40 μm. (C) Quantification of results shown in B. from 40–50 randomly selected cells (n = 4). (D) Western blot analysis of the TFEB KD efficiency in NPC fibroblast cells (n = 3 independent experiments). (E) The effects of SFN (15 μM, 9 h) treatment and TFEB KD on ROS levels in NPC cells. Scale bar: 20 μm. (F) Quantitative analysis of ROS levels shown in E. N = 40–50 randomly selected cells from at least 4 independent experiments. For all panels, data are presented as mean ± s.e.m.; *P < 0.05, ANOVA

Figure 9.

SFN induces a mild increase of ROS to activate TFEB-dependent lysosome biogenesis and autophagy, facilitating removal of excessive ROS. A working model to illustrate the role of TFEB in SFN-mediated enhancement of autophagic and lysosomal functions. A mild elevation in intracellular ROS levels by SFN, e.g., through mitochondrial and other sources, increases release of Ca²⁺, e.g., through lysosomal MCOLN1 channels and other unidentified Ca²⁺ release channels. Ca²⁺-bound calcineurin dephosphorylates TFEB to cause nuclear translocation of TFEB. Nuclear TFEB then promotes the transcription of a unique set of genes related to autophagy induction, autophagosome biogenesis, lysosome biogenesis, and detoxification (e.g., NFE2L2 and HMOX1). Subsequently, cells are pre-conditioned to promote the clearance of damaged mitochondria, removal of excessive ROS, and activation of detoxifying pathways

Figure 9.

SFN induces a mild increase of ROS to activate TFEB-dependent lysosome biogenesis and autophagy, facilitating removal of excessive ROS. A working model to illustrate the role of TFEB in SFN-mediated enhancement of autophagic and lysosomal functions. A mild elevation in intracellular ROS levels by SFN, e.g., through mitochondrial and other sources, increases release of Ca²⁺, e.g., through lysosomal MCOLN1 channels and other unidentified Ca²⁺ release channels. Ca²⁺-bound calcineurin dephosphorylates TFEB to cause nuclear translocation of TFEB. Nuclear TFEB then promotes the transcription of a unique set of genes related to autophagy induction, autophagosome biogenesis, lysosome biogenesis, and detoxification (e.g., NFE2L2 and HMOX1). Subsequently, cells are pre-conditioned to promote the clearance of damaged mitochondria, removal of excessive ROS, and activation of detoxifying pathways