TFE3 promotes ferroptosis in melanoma

cell


L E T T E R T O T H E E D I T O R TFE3 promotes ferroptosis in melanoma
Dear Editor, Melanoma is a highly heterogeneous malignancy that can present through multiple states of dedifferentiation, each with its own set of phenotypic and metabolic adaptations and vulnerabilities.
Phenotypic identity is in part regulated by the lineage-specific microphthalmia-associated transcription factor (MITF), a master regulator of melanocyte development that can control multiple facets of melanoma biology (Goding & Arnheiter, 2019).MITF is also a useful marker of distinct phenotypic states: high levels of MITF are associated with differentiation, while intermediate activity promotes proliferation and low levels are linked to invasion.Under stressful conditions, such as nutrient limitation, melanoma cells shutdown MITF expression and undergo a switch from a proliferative to invasive state (Falletta et al., 2017).
Efforts have been focused on understanding the key metabolic, and potentially therapeutic, vulnerabilities that cancer cells acquire while undergoing phenotypic transitions that enable cells to cope with and adapt to microenvironmental stresses.Invasive melanoma cells have been linked to a higher susceptibility to oxidative stress and ferroptosis (Ubellacker et al., 2020), a form of nonapoptotic cell death.In ferroptosis, Fenton chemistry via iron causes reactive oxygen species to target membrane phospholipids leading to lipid peroxidation (Dixon et al., 2012).This cell death pathway highlights an essential role of iron metabolism in cell survival that might represent a potential targetable vulnerability of invasive cells.
Lysosomes are key drivers of iron homeostasis and lysosomal dysfunction can contribute to ferroptosis by sequestering and accumulating iron, leading to oxidative stress and ultimately lysosomal membrane permeabilization (Mai et al., 2017).Interestingly, the master regulators of lysosomal biogenesis are TFEB and TFE3, two members of the MITF family that are retained in the cytoplasm when phosphorylated by mTORC1 under nutrient-replete conditions.Upon mTORC1 inhibition, via amino acid depletion for example, TFEB/3 translocate to the nucleus and drive adaptation to acute cellular stress through the upregulation of autophagic and lysosomal target genes (Settembre et al., 2011).Deprivation of the amino acid cyst(e)ine impacts mTORC1 function and activates ferroptosis (Zhang et al., 2021).In cases where stress is prolonged and unresolved, TFEB/3 can also control cell death via the induction of ATF4 and other genes that comprise part of the Integrated Stress Response (ISR), a pathway that can also regulate cyst(e)ine metabolism (Kreß et al., 2023;Martina et al., 2016).
While MITF High melanoma cells can achieve resistance to the induction of lipid peroxidation through the lineage-restricted expression of enzymes like the long chain fatty acid desaturase SCD (Kreß et al., 2023;Vivas-García et al., 2019), it is currently not well understood how MITF Low invasive cells might transcriptionally regulate ferroptosis.Therefore, we sought to understand the potential role of MIT/TFE family members in driving an adaptive response to ferroptosis in melanoma.
As previous work from others (Tsoi et al., 2018) showed that melanoma cells can undergo ferroptotic cell death, we first assessed the proliferation of melanoma cells under treatment with ferroptosis inducers (FINs) via crystal violet staining (Materials and Methods are provided in the supporting information).We observed that MITF Low IGR39 and A375M cells were highly sensitive to drugs directly inhibiting the main lipid peroxide detoxifying enzyme GPX4, such as RSL-3 (Figure 1a) and ML162 (Figure 1b) while growth of MITF High 501mel and IGR37 cells was much less affected with much higher levels of drug required to blunt their growth.
We confirmed that treatment with GPX4 inhibitors led to an increase in lipid peroxidation by BODIPY C11 staining (Figure 1c).To (Figure 1d).To rule out any contribution of other cell death pathways like apoptosis or necroptosis, 501mel cells were deprived of cystine and treated at the same time with ZVAD or Nec1s, which inhibit apoptosis and necroptosis respectively, or Ferrostatin1 and Deferoxamine (DFO) that block ferroptosis.Cell death was then measured by Propidium Iodide (PI) staining (Figure 1e).The results confirmed that cell death was inhibited using a ferroptosis inhibitor, but not those blocking apoptosis or necroptosis.We then starved the MITF Low IGR39 cells of cystine and showed using the PI staining assay that cell death was again prevented using the ferroptosis inhibitors (Figure 1f).
Gene expression analysis by qRT-qPCR confirmed that cystine deprivation led to the upregulation of targets associated with a ferroptotic response, especially CHAC1, PTGS2, ATF4, and the cystine transporter SLC7A11 (Figure 2a) (Upadhyayula et al., 2023).Next, we sought to understand whether the behavior of MIT/TFE factors in melanoma was altered upon induction of ferroptosis and whether their canonical target genes involved in lysosomal biogenesis would be affected.Firstly, we determined the protein levels of MIT/TFE members in both MITF High and MITF Low melanoma cells and observed that while TFEB is not detected in the A375M cells, TFE3 was expressed in all four melanoma cell lines examined, with the highest expression in the MITF Low IGR39 cell line (Figure 2b).
Second, quantification by qRT-PCR showed upregulation of lysosomal genes after 12 h of cystine starvation, with TFE3 mRNA, but not TFEB mRNA, showing an increase close to 4-fold (Figure 2c).Surprisingly, data from the Cancer Therapeutics Response Portal which analyses correlations between sensitivity to 481 drugs and gene expression revealed a strong positive correlation between sensitivity to FINs and expression of TFE3 in melanoma, meaning that high expression of TFE3 in melanoma correlates with increased sensitivity to FINs (Figure 2d).Note that of the 481 drugs tested, only those lying in the outlier 1.5× interquartile range are shown.
We therefore focused on the role of TFE3 as a potential mediator of ferroptosis in melanoma.
Treatment with FINs (Figure 2e, left panel) or cystine starvation (Figure 2e, right panel) revealed a shift in mobility of TFE3, consistent with its dephosphorylation and nuclear translocation.The moderately elevated TFE3 protein levels also observed following cystine starvation may be explained by the almost 4-fold increase in TFE3 mRNA levels (Figure 2c).Nuclear localization was confirmed by immunofluorescence of cells stably expressing physiological levels of HA-tagged TFE3 (Figure 2f).To test whether the induction of TFE3 mRNA following cystine starvation might be mediated by ATF4, a key effector of the integrated stress response whose translation is increased in response to cystine starvation (Kreß et al., 2023), we used RNA-seq from previously published 501mel cells engineered to induce ATF4 expression using doxycycline (Falletta et al., 2017).
The results (Figure 2g) revealed after induction with doxycycline, ATF4 was able to induce TFE3 mRNA levels around 2-fold.No significant effect was observed for TFEB mRNA, consistent with the lack of induction of TFEB mRNA by cystine deprivation (Figure 2c), while VEGF, a known ATF4 target, was also induced.
After establishing that melanoma cells can drive a transcriptional response to ferroptosis by upregulating lysosomal gene expression, potentially via regulating TFE3 levels and activity, we next asked how these cells might behave in the absence of TFE3.Using CRISPR we generated TFE3 knockout (KO) cells in both 501mel and IGR39 melanoma lines (Figure 2h).qRT-PCR revealed a dramatic reduction in the expression of CHAC1 mRNA, encoding a critical regulator of ferroptosis, in both 501mel and IGR39 TFE3 KO cells (Figure 2i).
ChIP-seq data revealed that TFE3 directly binds upstream from the CHAC1 gene and can therefore potentially directly influence its transcriptional expression (Figure 2j), although we cannot be certain that the binding of TFE3 detected at the CHAC1 locus is productive.
Lastly, PI staining revealed the high levels of ferroptotic cell death induced by overnight cystine deprivation were significantly suppressed in both the 501mel and IGR39 TFE3 KO cell lines (Figure 2k).
No effect on cell survival of the TFE3 KO was seen under standard culture conditions.
In summary, we have shown that in melanoma TFE3 is required for the expression of genes such as CHAC1 that promote ferroptosis.Consequently, loss or inactivation of TFE3, for example under nutrient-rich conditions may render cells less susceptible

Significance
Melanoma is characterized by high degrees of phenotypic heterogeneity originating from the impact of the microenvironment on epigenetically plastic cancer cells, with different phenotypes exhibiting different sensitivities to both targeted and immunotherapies.For example, MITF Low invasive melanoma cells are more resistant to BRAFi but exhibit an increased sensitivity to ferroptosis.What determines sensitivity to ferroptosis is not well understood.
Here, we reveal that TFE3, an MITF-related transcription factor expressed in melanomas, plays a key role in sensitizing cells to ferroptotic cell death.The results highlight a novel mechanism underpinning a potential therapeutic vulnerability in melanoma.
assess the specificity of ferroptosis induction under more physiologically relevant conditions, melanoma cells grown overnight in cystine-free media also exhibited elevated lipid peroxidation This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.© 2023 The Authors.Pigment Cell & Melanoma Research published by John Wiley & Sons Ltd.

F
I G U R E 1 MITF Low cells are sensitive to Ferroptosis.(a, b) Melanoma cell number detected via crystal violet staining after treatment with increasing amounts of RSL-3 (a) or ML162 (b) for 48 h.(c) Lipid peroxidation quantification using BODIPY C11 in 501mel cells after 1 μM RSL-3 treatment for 6 h.(d) Lipid peroxidation quantification using BODIPY C11 in 501mel cells after cystine starvation for 12 h.(e) Cell death assay using propidium iodide staining of 501mel cells starved of cystine for 12 h and treated simultaneously with indicated cell death inhibitors at 10 μM, 2 μM, 1 μM and 100 μM for ZVAD, Nec1s, Ferr1 and DFO respectively.(f) Quantification of cell death via PI staining of IGR39 after cystine starvation for 12 h.Statistical comparisons between two groups were made using the Student t-test, and ANOVA was used for more than two groups.*p < .05;**p < .01;***p < .001;****p < .0001.Error bars represent mean ± SD.Quantification of cell death and lipid peroxidation assays was achieved by acquiring data from at least 5000 cells.to ferroptosis.These findings shine a light on the regulation of ferroptosis by TFE3 and provide a new target to treat metastatic melanoma by inducing TFE3 activity and ferroptotic cell death in invasive cells.FU N D I N G I N FO R M ATI O N This work was funded by FCT PD/BD/114127/2015 as part of the Graduate Program in Areas of Basic and Applied Biology (DD) and Ludwig Cancer Research (CRG, PL and DD).

F
I G U R E 2 TFE3 promotes ferroptosis in melanoma.(a) qRT-PCR of indicated ferroptosis-associated genes after cystine starvation of 501mel cells for 12 h.(b) Western blot of MIT/TFE family members in melanoma cells with ACTIN used as a loading control.(c) qRT-PCR of lysosomal/autophagy associated genes after cystine deprivation of 501mel 12 h.(d) Cancer Therapeutics Response Portal data from melanoma showing correlation between TFE3 expression and sensitivity to FINs.Data is presented as a box-and-whisker plot representing z-scored Pearson expression-sensitivity correlation coefficients.The data was generated from sensitivity to 481 small molecules, expressed as area under the curve (AUC) (Low AUC = High sensitivity to drug), correlated with genome-wide gene expression of TFE3 across 52 melanoma cell lines.Red dots shown indicate drugs representing Tukey outliers of the 1.5× interquartile range.(e) Western blot of TFE3 in 501mel treated with indicated 1 μM FINs for 6 h or starved of cystine for 12 h with ACTIN as loading control.(f) Immunofluorescence of HA-tagged TFE3 in 501mel cells treated with indicated FINs at 1 μM for 6 h.(g) RNA-seq of doxycycline-inducible ATF4-expressing 501mel cells comparing induction of indicated genes 24 h after doxycycline induction of ATF4 expression.3′ RNA-seq was performed in triplicate to obtain statistical significance.(h) Western blot of TFE3 KO cells in parental 501mel and IGR39 cells.ACTIN was used as a loading control.(i) qRT-PCR showing CHAC1 mRNA expression in WT vs TFE3 KO 501mel and IGR39 cells.(j) UCSC genome browser screenshot of duplicate TFE3 ChIP-seq and input control at the CHAC1 locus.(k) Propidium iodide staining for dead cells in WT or TFE3 KO 501mel and IGR39 cells after cystine starvation for 12 h.Statistical comparisons between two groups were made using Student t-test, and ANOVA was used for more than two groups.*p < .05;**p < .01;***p < .001;****p < .0001.Error bars represent mean ± SD.Quantification of the immunofluorescence assays was achieved by imaging a minimum of 30 cells per replicate.