NRF2 regulates lipid droplet dynamics to prevent lipotoxicity

Abstract

Lipid droplets (LDs) are dynamic organelles comprising a neutral lipid core encapsulated by a phospholipid monolayer. LD structure and function are influenced by a variety of intrinsic and extrinsic signals, and cells alter LD content and distribution to adapt to their environment. Here, we show that LD content increases in response to stabilization of the transcription factor NRF2 under conditions of lipotoxic stress. Notably, NRF2 activity leads to increased expression of the G0S2, a protein that inhibits ATGL, the enzyme responsible for degradation of triacylglycerol and the release of fatty acids from LDs. Importantly, stabilization of NRF2 in the absence of stress is sufficient to increase LD content, and inhibition of ATGL partially rescues the impact of NRF2 deletion on stress-induced ferroptosis. These data support a model in which stress-induced NRF2 stabilization protects cells against lipotoxicity in part through the sequestration of fatty acids in lipid droplets.

Keywords: Biochemistry; Biological sciences; Cell biology.

Graphical abstract

Figure 1

NRF2 stability and activity influences LD accumulation in HEK-TtH cells (A) Immunoblot confirming NRF2 activation in sgKEAP1 cells. (B) Quantification of LD area per cell from (C). n = 4 biological replicates. Mean ± SD. (C) Representative images of LDs in sgKEAP1 HEK-TtH cells taken with 63× objective. Scale bars represent 20 μm. (D) Immunoblot confirming NRF2 activation in cells stimulated with 10 μM Sulforaphane for 24 h. (E) Quantification of LD area per cell from (F). n = 4 biological replicates. Mean ± SD. (F) Representative images of LDs in HEK-TtH cells treated with 10 μM sulforaphane for 24 h taken with 63× objective. Scale bars represent 20 μm. (G) Immunoblot confirming NRF2 depletion in sgNRF2 cells. (H) Quantification of LD area per cell from (I). n = 3 biological replicates. Mean ± SD. (I) Representative images of LDs in sgNRF2 HEK-TtH cells taken with 63× objective. Scale bars represent 20 μm. (J) Immunoblot confirming NRF2 depletion and inactivation of transcriptional activity in sgKEAP1 cells with addition of sgNRF2. (K) Quantification of LD area per cell from (L). n = 3 biological replicates. Mean ± SD. (L) Representative images of LDs in sgKEAP1/sgNRF2 double-KO HEK-TtH cells taken with 63× objective. Scale bars represent 20 μm.

Figure 2

NRF2-mediated LD accumulation occurs in specific lung cancer cell lines (A) Representative images of LDs in lung cancer cells taken with 63× objective. Scale bars represent 20 μm. (B) Quantification of LD area per cell from (A). n = 3 biological replicates. Mean ± SD. (C) Immunoblot of basal NRF2 and HMOX1 levels in lung cancer cell lines. (D) Relative basal mRNA expression (2-ΔΔCT relative to GAPDH) of NRF2 transcriptional targets in lung cancer cell lines. Data normalized to H1650 mRNA expression. Horizontal dotted line represents no change relative to H1650 mRNA levels. n ≥ 3 biological replicates. Mean ± SD. (E) Immunoblot confirming NRF2 depletion in sgNRF2 H1650 cells. (F) Immunoblot confirming NRF2 depletion in sgNRF2 H2009 cells. (G) Representative images of LDs in sgNRF2 H1650 cells taken with 63× objective. Scale bars represent 20 μm. (H) Representative images of LDs in sgNRF2 H2009 cells taken with 63× objective. Scale bars represent 20 μm. (I) Quantification of LD area per cell from (G). n = 3 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (J) Quantification of LD area per cell from (G). Horizontal dotted line represents no change relative to sgScr cells. n = 3 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, and paired). Mean ± SD. (K) Quantification of LD area per cell from (H). n = 3 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (L) Quantification of LD area per cell from (H). Horizontal dotted line represents no change relative to sgScr cells. n = 3 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, and paired). Mean ± SD.

Figure 3

The negative regulator of adipose triglyceride lipase, G0S2, acts downstream of NRF2 to regulate LD abundance in HEK-TtH cells (A) Fold change of mRNA expression (2-ΔΔCT relative to GAPDH) of regulators of LD structure, lipids, and lipid storage in sgKEAP1 cells vs. sgScr cells. Vertical dotted line represents no change relative to sgScr control mRNA levels. n = 3 biological replicates. Mean ± SD. (p < 0.05 = ∗, p < 0.01 = ∗∗). (B) Fold change of mRNA expression (2-ΔΔCT relative to GAPDH) of G0S2 in lung cancer cell lines. Data represented normalized to G0S2 expression of H1650 cells. Horizontal dotted line represents no change relative to H1650 mRNA levels. n = 3 biological replicates. Mean ± SD. (C) Representative images of LDs in sgKEAP1/sgG0S2 double-KO HEK-TtH cells taken with 63× objective. Scale bars represent 20 μm. (D) Quantification of LD area per cell from (C). n = 5 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (E) Representative images of LDs in sgNRF2/mG0S2-Flag-overexpression HEK-TtH cells taken with 63× objective. Scale bars represent 20 μm. (F) Quantification of LD area per cell from (E). n = 3 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD.

Figure 4

Specific fatty acids stabilize NRF2 and drive downstream signaling to protect cells from lipotoxic stress (A) Representative images of LDs in HEK-TtH cells treated with fatty acids conjugated to fatty acid-free BSA for 4 h taken with 63× objective. Scale bars represent 20 μm. (B) Quantification of LD area per cell from (A). n = 3 independent biological replicates. Mean ± SD. (p < 0.05 = ∗, p < 0.01 = ∗∗, p < 0.001 = ∗∗∗). (C) Quantification of cell viability based on luminescence from CellTiter-Glo Luminescent Cell Viability Assay for HEK-TtH cell treated with gradient of fatty acids conjugated to fatty acid-free BSA relative to vehicle controls for 24 h. Horizontal dotted line represents half maximal inhibitory concentration (IC50). n = 3 biological replicates of each. (D) IC50 and standard error calculated from (C) for each fatty acid treatment. (E) Immunoblot of NRF2 and downstream transcriptional target HMOX1 after 4 h of treatment of fatty acids conjugated to fatty acid-free BSA. (F) Quantification of cell viability based on luminescence from CellTiter-Glo Luminescent Cell Viability Assay for HEK-TtH sgNRF2 and sgKEAP1 cell lines treated with 200 μM arachidonic acid (AA) conjugated to fatty acid-free BSA for 24 h. n = 3 biological replicates of each. Mean ± SD. (G) Quantification of cell viability based on luminescence from CellTiter-Glo Luminescent Cell Viability Assay for H2009 sgNRF2 lung cancer cell lines treated with 200 μM arachidonic acid (AA) conjugated to fatty acid-free BSA for 24 h. n = 3 biological replicates of each. Mean ± SD. (H) Immunoblot comparing 4-hydroxynonenal in sgScr and sgNRF2 cells after 4 h treatment with gradient of arachidonic acid conjugated to fatty acid-free BSA. (I) Quantification of cell viability based on luminescence from CellTiter-Glo Luminescent Cell Viability Assay of WT HEK-TtH cells treated with 200 μM arachidonic acid (AA) conjugated to fatty acid-free BSA, co-treated with 10 μM Ferrostatin-1 (Fer-1) or 10 μM Z-VAK-FMK for 24 h. n = 3 biological replicates of each. Mean ± SD.

Figure 5

Inhibition of lipid droplet degradation protects cells from ferroptosis (A) Immunoblot of NRF2 stability and downstream transcriptional target HMOX1 after treatment with 200 nM RSL3 for 4 h. (B) Quantification of cell viability based on luminescence from CellTiter-Glo Luminescent Cell Viability Assay for HEK-TtH cell treated with gradient of RSL3 relative to DMSO controls for 24 h. Horizontal dotted line represents half maximal inhibitory concentration (IC50). n = 4 biological replicates of each. (C) IC50 and standard error calculated from (B) for each lung cancer cell line treated with RSL3 gradient. (D) Quantification of luminescence from CellTiter-Glo Luminescent Cell Viability Assay of sgNRF2 H2009 cell lines treated with 200 nM RSL3 relative to vehicle control for 24 h. n = 3 biological replicates of each. Mean ± SD. (E) Quantification of luminescence from CellTiter-Glo Luminescent Cell Viability Assay of sgNRF2 and sgKEAP1 HEK-TtH cell lines treated with 200 nM RSL3 relative to vehicle control for 24 h. n = 6 biological replicates of each. Mean ± SD. (F) Representative images of LDs of HEK-TtH cells treated with 50 nM RSL3 for 24 h taken with 63× objective. Scale bars represent 20 μm. (G) Quantification of LD area per cell from (F). n = 3 biological replicates. Mean ± SD. (H) Quantification of cell viability from CellTiter-Glo Glo Luminescent Cell Viability Assay of sgNRF2 HEK-TtH cells treated with DMSO or 10 μM ATGListatin (ATGLi) in combination with 25 nM RSL3. n = 4 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (I) Representation of data from (H) when paired within each individual biological replicate. Statistics represented calculated by Student’s t test (one-tailed, paired). Mean ± SD. (J) Quantification of cell viability from CellTiter-Glo Glo Luminescent Cell Viability Assay of sgNRF2 HEK-TtH cells treated with DMSO or 10 μM ATGListatin (ATGLi) in combination with 50 nM RSL3. n = 4 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (K) Representation of data from (J) when paired within each individual biological replicate. Statistics represented calculated by Student’s t test (one-tailed, paired). Mean ± SD. (L) Quantification of cell viability from CellTiter-Glo Glo Luminescent Cell Viability Assay of sgNRF2 HEK-TtH cells overexpressing mG0S2-Flag or empty vector (EV) treated with 25 nM RSL3. n = 5 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (M) Representation of data from (L) when paired within each individual biological replicate. Statistics represented calculated by Student’s t test (one-tailed, paired). Mean ± SD. (N) Quantification of cell viability from CellTiter-Glo Glo Luminescent Cell Viability Assay of sgNRF2 HEK-TtH cells overexpressing mG0S2-Flag or empty vector (EV) treated with 50 nM RSL3. n = 5 biological replicates. Statistics represented calculated by Student’s t test (one-tailed, unpaired with equal variance). Mean ± SD. (O) Representation of data from (N) when paired within each individual biological replicate. Statistics represented calculated by Student’s t test (one-tailed, paired). Mean ± SD.