Abstract

   The DNA damage response is essential for preserving genome integrity
   and eliminating damaged cells. Although cellular metabolism plays a
   central role in cell fate decision between proliferation, survival, or
   death, the metabolic response to DNA damage remains largely obscure.
   Here, this work shows that DNA damage induces fatty acid oxidation
   (FAO), which is required for DNA damage‐induced cell death.
   Mechanistically, FAO induction increases cellular acetyl‐CoA levels and
   promotes N‐alpha‐acetylation of caspase‐2, leading to cell death.
   Whereas chemotherapy increases FAO related genes through peroxisome
   proliferator‐activated receptor α (PPARα), accelerated
   hypoxia‐inducible factor‐1α stabilization by tumor cells in obese mice
   impedes the upregulation of FAO, which contributes to its
   chemoresistance. Finally, this work finds that improving FAO by PPARα
   activation ameliorates obesity‐driven chemoresistance and enhances the
   outcomes of chemotherapy in obese mice. These findings reveal the shift
   toward FAO induction is an important metabolic response to DNA damage
   and may provide effective therapeutic strategies for cancer patients
   with obesity.

   Keywords: cell death, chemoresistance, DNA damage, fatty acid
   oxidation, obesity
     __________________________________________________________________

   The study identifies a novel metabolic response to DNA damage involving
   the induction of mitochondrial fatty acid oxidation (FAO). Obesity
   impedes the upregulation of FAO, causing resistance to chemotherapy.
   Restoration of FAO overcomes obesity‐driven chemoresistance, enhancing
   chemotherapy outcomes. These findings underscore FAO induction as a
   crucial metabolic response to DNA damage, offering potential
   therapeutic strategies for obese cancer patients.

   graphic file with name ADVS-11-2304702-g002.jpg

1. Introduction

   Obesity, a physical condition defined as abnormal or excessive
   accumulation of fat, is increasing worldwide.^[ [46]^1 ^]
   Epidemiological studies reveal strong links between obesity and at
   least 13 types of cancer, including colorectal, liver, breast and
   prostate cancers.^[ [47]^2 ^] Moreover, accruing evidence suggests that
   obesity is closely associated with inferior clinical outcomes of cancer
   therapy.^[ [48]^3 ^] Indeed, many cancer patients with obesity suffer
   from chemotherapy failure, cancer recurrence, and poor prognosis.^[
   [49]^4 ^] Thus, considering the global prevalence of overweight or
   obesity and the increase of cancer incidence in obese patients, it is
   crucial to elucidate mechanisms by which obesity influences cancer
   progression.

   The cancer's lethal nature stems from its profound resistance to
   therapy. Although chemotherapy and radiation therapy, the most commonly
   used tools for cancer treatment, have been developed to kill cancer
   cells by inducing DNA damage, some cancer cells escape DNA
   damage‐induced cell death and develop resistance to the treatments.^[
   [50]^5 ^] Whether innate or acquired, such therapeutic resistance is a
   massive obstacle to successful cancer treatment.^[ [51]^6 ^] Recently,
   systemic perturbations secondary to obesity have been implicated in
   enabling resistance to chemotherapy. For example, elevated circulating
   levels of insulin, insulin‐like growth factor (IGF‐1) and adipokines,
   associated with obesity‐mediated metabolic dysregulations,^[ [52]^7 ^]
   have been implicated in enabling resistance to chemotherapeutic
   agents.^[ [53]^4a ^] It is noteworthy that obese patients recapitulate
   a state of persistent low‐grade inflammation characterized by increased
   cytokines and chemokines that induce alterations in tumor
   microenvironment.^[ [54]^8 ^] Thus, enhanced fibrosis or reactive
   stroma caused by increased inflammation could impair drug delivery.^[
   [55]^4 , [56]^8 ^] However, it remains poorly understood whether
   overweight or obesity adversely affects therapeutic outcomes directly
   through intrinsic changes of cancer cells.

   When cells are exposed to genotoxic stresses, they respond by
   initiating tightly coordinated signaling responses that can promote
   survival by inducing cell cycle arrest and DNA repair or trigger cell
   death to limit the accumulation of damaged cells.^[ [57]^9 ^] Given the
   indispensability of cellular metabolism for survival and growth, cells
   exposed to DNA damage need to make fundamental metabolic decisions of
   whether to upregulate biosynthesis for promoting DNA repair or not, and
   whether to restrict catabolic pathways for limiting proliferation or
   not. Surprisingly, the metabolic response to DNA damage has been
   largely unknown. Recent studies have shown that there is a link between
   metabolism and DNA damage response, by repression of glutamine
   metabolism to promote proper DNA repair and cell cycle arrest, and to
   suppress tumorigenesis,^[ [58]^10 ^] implying that there is a metabolic
   area of cellular stress response with significant therapeutic
   potential.

   Here, we investigated how obesity modulates metabolic stress response
   of cancer cells to promote cell survival and chemoresistance. Using a
   murine model of obesity, we performed transcriptomic analyses of
   metabolic pathways in engrafted cancer cells after chemotherapy. We
   found that cancer cells in lean mice, but not in obese mice,
   dynamically respond to chemotherapy by upregulating fatty acid
   oxidation (FAO) pathway. The FAO induction promotes N‐alpha‐acetylation
   of caspase‐2 (CASP2), leading to apoptotic cell death. Importantly, we
   show that pharmacological induction of FAO alleviates obesity‐driven
   drug resistance and enhances therapeutic outcomes after chemotherapy.
   These findings may have important implications for future therapeutic
   approaches treating cancer patients with overweight or obesity.

2. Results

2.1. Obesity Impairs FAO Induction in Tumors after Chemotherapy

   In order to investigate how obesity induces chemoresistance of tumor
   cells, we performed allograft tumor formation assays in wild‐type (WT)
   and genetic leptin‐deficient ob/ob mice using syngeneic B16F10 melanoma
   cells. When the tumor volume reached ≈100 mm^3, mice were administrated
   with etoposide (ETS), which is one of the most common chemotherapeutic
   agents and has been shown to induce cell death across multiple
   cancers.^[ [59]^11 ^] As observed previously, whereas ETS treatment
   induced cell death in tumors from lean control mice, tumors from obese
   mice exhibited profound resistance to ETS (Figure [60]1a).

Figure 1.

   Figure 1
   [61]Open in a new tab

   Impairment of FAO induction is related to chemoresistance in obesity.
   a) Immunohistochemistry of cleaved CASP3 expression pre‐ and
   post‐chemotherapy in B16F10 tumors from lean and obese mice. Scale bar
   represents 50 µm. b) Number of genes related to top 10 KEGG pathways.
   c) Enrichment of KEGG metabolic signature scores that are altered by
   chemotherapy in RNA‐seq from lean (yellow‐green bars) and obese tumors
   (blue bars). d) Volcano plots of differentially expressed fatty acid
   oxidation genes in tumors from lean (upper panel) and obese mice (lower
   panel) post‐ and pre‐chemotherapy (>1.5‐fold, p < 0.05). e) Schematic
   representation of mitochondrial fatty acid oxidation. f) Relative
   expression of FAO related genes in B16F10 tumor cells from lean (upper
   panel) and obese mice (lower panel); n = 4. Statistical analysis was
   based on two‐tailed Student's t‐test. All error bars ±SEM. *p < 0.05,
   **p < 0.01, ***p < 0.001, and ****p < 0.0001.

   To explore mechanism(s) underlying the obesity‐induced drug resistance,
   we performed a comparative transcriptomic analysis of tumor cells from
   lean and obese mice after chemotherapy. We analyzed the following four
   groups: lean tumor, lean tumor after ETS treatment, obese tumor, obese
   tumor after ETS treatment. As KEGG pathway enrichment analysis,
   performed on differentially expressed genes (DEGs) in lean tumors
   before and after chemotherapy (2418 genes; |Fc| ≥ 2 & p‐value cutoff <
   0.05), suggested that metabolic pathways are the most significantly
   changed in lean tumors after chemotherapy (Figure [62]1b). Subsequent
   pathway mapping revealed that 16 KEGG metabolic signatures
   significantly changed in lean tumors after chemotherapy
   (Figure [63]1c). Importantly, tumors from obese mice also underwent
   significant changes in these pathways after chemotherapy except 3
   pathways including inositol phosphate metabolism, glycine, serine and
   threonine metabolism, and fatty acid degradation (Figure [64]1c). Among
   these differentially regulated pathways, we reasoned that fatty acid
   metabolism could be most relevant to obesity given its connection to
   abnormalities of lipid metabolism.

   To confirm that obesity abrogates the induction of fatty acid
   degradation after chemotherapy, we examined gene score enrichment
   analysis (GSEA) of DEGs in tumors from lean and obese mice pre‐ and
   post‐chemotherapy. The transcripts involved in fatty acid metabolism
   were particularly increased in lean tumors after chemotherapy, whereas
   these changes were not significant in obese tumors (Figure [65]S1a,
   Supporting Information). Additionally, we observed that the expression
   of genes involved in FAO pathway was more significantly changed in lean
   tumors than in obese tumors after chemotherapy (Figure [66]1d; Figure
   [67]S1b, Supporting Information). Furthermore, consistent with our
   transcriptomic analysis, quantitative RT‐PCR results confirmed that ETS
   treatment significantly increased the expression of FAO‐related genes
   in lean tumors, whereas the induction of these genes was markedly
   attenuated in obese tumors (Figure [68]1e,f). Taken together, our
   findings demonstrate that obesity impairs the induction of FAO gene
   signatures in tumors after chemotherapy.

2.2. Induction of FAO is Required for DNA Damage‐Induced Cell Death

   Chemotherapies target rapidly proliferating cancer cells by directly or
   indirectly inducing DNA damage.^[ [69]^12 ^] Because the expression of
   FAO related genes was highly induced after ETS treatment in lean tumors
   but not in obese tumors, we hypothesized that FAO could serve as an
   important regulator of DNA damage‐induced cell death after
   chemotherapy. First, the relevance of FAO induction was further tested
   using another chemotherapeutic agent, camptothecin (CPT). In line with
   our in vivo results, genes involved in fatty acid degradation were
   significantly increased upon ETS or CPT treatment. On the contrary,
   genes involved in fatty acid synthesis were decreased under these
   conditions, suggesting that the induction of FAO related genes may be a
   specific metabolic response to chemotherapy and not reflective of a
   nonspecific metabolic crisis (Figure [70]2a). When we directly measured
   FAO rates, the oxidation of palmitate, a saturated long‐chain fatty
   acid, was significantly increased after ETS or CPT treatment
   (Figure [71]2b).

Figure 2.

   Figure 2
   [72]Open in a new tab

   FAO induction is required for proper DNA damage‐induced cell death. a)
   Quantitative real time PCR of HepG2 cells showing indicated mRNA
   expression with CPT or ETS (n = 3). Statistical analysis was performed
   using two‐way ANOVA with Dunnett's multiple comparisons test. b) FAO
   was measured in HepG2 cells treated with CPT or ETS (n = 3).
   Statistical analysis was performed using one‐way ANOVA with Dunnett's
   multiple comparisons test. c) Cell death of B16F10 cells treated with
   ETS (25 µm, 40 h), ETO (50 or 200 µm, 40 h) or both (n = 3). Cell death
   was measured by propidium iodide exclusion assay. d) CASP3/7 GLO assay
   of B16F10 cells treated with ETS (25 µm, 24 h), ETO (50 µm, 24 h) or
   both (n = 4). e) Cell death of B16F10 cells transfected with
   nontargeting siRNA (siControl) or with siRNA against CPT1A (siCPT1A).
   Cells were treated with or without ETO (50 µm, 40 h) in the presence of
   ETS (25 µm, 40 h) (n = 3). Cell death was measured by propidium iodide
   exclusion assay. f) Relative Cd36 mRNA level in B16F10 cells treated
   with ETS (25 µm, 24 h); n = 4. Statistical analysis was based on
   two‐tailed Student's t‐test. g) FAO‐dependent ΔOCR was measured in
   B16F10 cells pre‐treated with or without palmitate (100 µm, 6 h); n =
   5. Statistical analysis was based on two‐tailed Student's t‐test. h)
   Cell death of B16F10 cells treated with ETS (25 µm, 40 h) in the
   presence of palmitate (100 µm, 40 h); n = 3. Statistical analysis was
   performed using one‐way ANOVA with Tukey's multiple comparisons test
   (c–e and h). All error bars ±SD. *p < 0.05, **p < 0.01, ***p < 0.001,
   and ****p < 0.0001.

   Next to assess whether FAO affects cell death after DNA damage, we
   inhibited FAO using etomoxir (ETO), which can inhibit mitochondrial
   import of fatty acids by inhibiting carnitine palmitoyltransferase 1
   family proteins (CPT1A, B and C). Notably, inhibition of FAO reduced
   apoptotic cell death after ETS treatment (Figure [73]2c,d). To further
   confirm the role of FAO in cell death, we suppressed CPT1A expression
   by using short interfering RNA (siRNA) and examined the effect on cell
   survival followed by ETS treatment. Knockdown of CPT1A significantly
   inhibited DNA damage‐induced cell death, and ETO treatment did not have
   addictive effects on cell death (Figure [74]2e; Figure [75]S2a,b,
   Supporting Information). This effect was not specific to B16F10
   melanoma cells. We observed similar results in immortalized embryonic
   fibroblasts (MEFs) (Figure [76]S2c–f, Supporting Information).
   Additionally, comparable results were obtained in cells treated with
   other DNA damaging agents, including cisplatin or irradiation (Figure
   [77]S2g,h, Supporting Information). These results demonstrate that
   blocking FAO protects cells from DNA damage‐induced cell death.

   Next, we tested whether increasing FAO potentiates cell death in
   response to DNA damage. First, we found that the expression of CD36,
   fatty acid translocase/receptor, was markedly induced following DNA
   damage (Figure [78]2f), in line with the induction of FAO. After
   palmitate treatment, cells exhibited an increased oxygen consumption
   rate (OCR), indicating that cells were oxidizing more fatty acids
   (Figure [79]2g; Figure [80]S2i, Supporting Information). Interestingly,
   cells treated with palmitate exhibited significantly elevated levels of
   cell death following ETS treatment (Figure [81]2h). We obtained
   comparable results in immortalized MEFs treated with palmitate or oleic
   acid (Figure [82]S2j,k, Supporting Information). Finally, to test
   whether the relationship between FAO and cell death is unique to DNA
   damage or a generalizable feature of other cell death triggers, we
   treated cells with tumor necrosis factor α (TNFα) and cycloheximide,
   which induces apoptotic cell death through the engagement of death
   receptors. Unlike DNA damage‐induce cell death, cell death induced by
   activation of death receptors was not reduced by FAO inhibition (Figure
   [83]S2l, Supporting Information). Taken together, these data illustrate
   that genotoxic stress induces FAO, which is required for DNA
   damage‐induced cell death.

2.3. DNA Damage Induces FAO through PPARα

   As our results show that DNA damaging drugs induce FAO, we sought to
   identify the molecular mechanism. Peroxisome proliferator‐activated
   receptor α (PPARα) is a critical regulator of lipid metabolism and
   promotes transcription of genes crucial for FAO,^[ [84]^13 ^] thus
   suggesting that DNA damage might potentially induce this metabolic
   response via PPARα. To test this concept, we first assessed the
   expression of PPARα in response to DNA damage. We observed that ETS
   treatment resulted in a robust increase in PPARα at mRNA level (Figure
   [85]3a), as well as the protein level (Figure [86]3b) in B16F10 cells
   and multiple cell lines (Figure [87]S3a–d, Supporting Information).
   Consistent with these results, the transcriptional activity of PPARα
   was significantly increased following DNA damage (Figure [88]3c).

Figure 3.

   Figure 3
   [89]Open in a new tab

   PPARα is an upstream target for DNA damage‐induced FAO induction. a–c)
   Relative Pparα mRNA (a), protein (b), and transcriptional activity (c)
   levels in B16F10 cells treated with ETS (25 µm, 24 h); n = 3.
   Statistical analysis was performed using two‐tailed Student's t‐test.
   d) Relative expression of FAO related genes in B16F10 cells treated
   with GW6471 (25 µm, 24 h) or WY14643 (300 µm, 24 h) in the presence of
   ETS (25 µm, 24 h); n = 4. Statistical analysis was performed using
   one‐way ANOVA with Tukey's multiple comparisons test. e) FAO‐dependent
   ΔOCR of B16F10 cells treated with GW6471 (25 µm) or WY14643 (300 µm)
   and/or ETS (25 µm); n = 5. The cell medium was replaced with assay
   media and the cells were incubated in a CO[2] free incubator at 37 °C
   for 1 h before the assay. FAO was measured using Seahorse XF Palmitate
   Oxidation Stress Test Kit. ETO (40 µm) was treated 15 min prior to
   running assay, and then Palmitate‐BSA was added to the wells
   immediately before running assay. Statistical analysis was performed
   using one‐way ANOVA with Tukey's multiple comparisons test. f,g)
   Relative PPARα mRNA (f) and protein (g) levels in allograft B16F10
   tumors from lean and obese tumors pre‐ and post‐chemotherapy (n = 4).
   Statistical analysis was performed using two‐way ANOVA with Sidak's
   multiple comparisons test. h) Relative mRNA levels of FAO related genes
   in B16F10 tumors from lean (up) and obese (down) mice treated with or
   without WY14643 (40 mg kg^−1 five times via intraperitoneal (i.p.)
   injection) in the presence of ETS (20 mg kg^−1 five times via i.p.
   injection). n = 4 (DMSO, ETS, ETS+WY14643) in WT mice; n = 3 (DMSO,
   ETS+WY14643) or n = 4 (ETS) in ob/ob mice. Statistical analysis was
   performed using one‐way ANOVA with Dunnett's multiple comparisons test.
   i) Cell death of B16F10 cells treated with GW6471 (25 µm) or WY14643
   (300 µm) and/or ETS (25 µm); n = 3. Statistical analysis was performed
   using one‐way ANOVA with Tukey's multiple comparisons test. j) Tumor
   volume of B16F10 cells expressing a control shRNA or two
   independent shRNAs to PPARα in lean mice treated with or without ETS
   (20 mg kg^−1 five times via i.p. injection). Statistical analysis was
   performed using one‐way ANOVA with Tukey's multiple comparisons test.
   Error bars indicate ±SD (a‐e, and i) or ±SEM (f and h). *p < 0.05, **p
   < 0.01, ***p < 0.001, and ****p < 0.0001.

   Next, to evaluate the role of PPARα on FAO induction after DNA damage,
   we treated cells with either GW6471, an antagonist of PPARα, or
   WY14643, a synthetic agonist of PPARα. Importantly, we observed that
   PPARα inhibition by GW6471 abrogated the induction of FAO gene
   signatures upon DNA damage (Figure [90]3d; Figure [91]S3e, Supporting
   Information). We obtained similar results in cells in which PPARα
   expression was reduced by siRNAs against PPARα (Figure [92]S3f,g,
   Supporting Information). As the PPAR family comprises three subtypes,
   such as PPARα, PPARβ, and PPARγ,^[ [93]^14 ^] we investigated the
   involvement of PPARβ or PPARγ in the induction of FAO‐related genes
   upon DNA damage. While PPARγ knockdown had no effect on the expression
   of these genes, PPARβ knockdown partially reduced the induction of some
   genes (Figure [94]S3f,h, Supporting Information). However, we observed
   that PPARβ expression remained unaffected after ETS treatment (Figure
   [95]S3i, Supporting Information), implying that PPARα may play a major
   role in the induction of FAO upon DNA damage. Consistent with this
   observation, inhibition of PPARα suppressed the induction of FAO upon
   DNA damage (Figure [96]3e). Moreover, to further confirm the in vivo
   relevance of PPARα, we assessed whether PPARα is induced in tumors
   after exposure to chemotherapy. Importantly, PPARα was significantly
   induced in lean tumors after ETS treatment, whereas not in obese tumors
   (Figure [97]3f, g; Figure [98]S3j, Supporting Information). These
   findings prompted us to further investigate whether enhancing PPARα
   activity rescues the defective FAO induction in obese tumors. We found
   that administration of WY14643 was able to rescue the impaired
   induction of FAO related genes in obese tumors after chemotherapy,
   whereas we did not detect additional effect in lean tumors
   (Figure [99]3h).

   Given the function of PPARα in FAO regulation, we speculated that the
   PPARα‐mediated transcriptional activation of FAO‐related genes promotes
   cell death after chemotherapy. To test this hypothesis, we examined the
   effect of PPARα inhibition on cell survival after DNA damage. Notably,
   GW6471 treatment markedly reduced DNA damage‐induced apoptotic cell
   death, whereas WY14643 had minimal effect on cell death (Fig [100]3i;
   Figure [101]S4a–d, Supporting Information). Additionally, these effects
   were diminished in CPT1A knockdown cells (Figure [102]S4e, Supporting
   Information). Similarly, knockdown of PPARα resulted in decreased cell
   death following ETS treatment compared to control cells (Figure
   [103]S4f,g, Supporting Information). To further confirm the importance
   of this pathway in DNA damage‐induced cell death, we performed
   allograft assays with PPARα knockdown melanoma cells to assess changes
   in drug‐sensitivity. Consistent with our in vitro results, PPARα
   knockdown led tumors to be more resistant to ETS treatment
   (Figure [104]3j). Together, these data support the idea that DNA damage
   induces FAO via PPARα, which promotes cell death.

2.4. Obesity‐Induced Hypoxia Suppresses FAO Induction after Chemotherapy

   Because obese tumors are more hypoxic than lean tumors,^[ [105]^15 ^]
   and hypoxia‐inducible factor‐1α (HIF1α) suppresses FAO by inhibiting
   PPARα,^[ [106]^16 ^] we hypothesized that hypoxic condition occurring
   in obese tumors would contribute to the defect in FAO induction after
   chemotherapy. To test this idea, we first investigated whether obesity
   augments hypoxia in tumors. As has previously shown,^[ [107]^15a ^]
   histological analysis showed that tumors from obese mice exhibited
   elevated levels of HIF1α relative to tumors from lean mice (Figure
   [108]4a; Figure [109]S5a, Supporting Information). To test whether
   hypoxia directly represses FAO induction after DNA damage, we examined
   the levels of FAO related genes under hypoxic conditions. ETS treatment
   clearly and reproducibly increased the expression of PPARα and its
   target genes in normoxia (Figure [110]4b–d; Figure [111]S5b, Supporting
   Information). However, the induction of these genes was markedly
   blunted by hypoxia (Figure [112]4b–d; Figure [113]S5b, Supporting
   Information). We also obtained similar results in cells treated with
   dimethyloxalglycine (DMOG), which stabilizes HIF1α by inhibiting prolyl
   hydroxylase (Figure [114]S5c, Supporting Information). Next, to further
   investigate whether HIF1α directly represses FAO induction after DNA
   damage, we reduced HIF1α expression by siRNA targeting HIF1α and
   examined the expression of FAO‐related genes under hypoxic conditions.
   Notably, knockdown of HIF1α significantly restored the impaired
   induction of PPARα and FAO‐related genes after ETS treatment
   (Figure [115]4e,f; Figure [116]S5d, Supporting Information). In
   conclusion, these data suggest that obesity‐mediated hypoxia opposed
   the FAO induction after chemotherapy.

Figure 4.

   Figure 4
   [117]Open in a new tab

   Obesity‐mediated hypoxia inhibits the FAO induction after chemotherapy.
   a) Immunohistochemistry of HIF1α expression in lean and obese B16F10
   tumors. Scale bars represent 50 µm. Statistical analysis was performed
   using two‐tailed Student's t‐test. Error bars represent ±SEM b)
   Relative Pparα mRNA levels in B16F10 cells treated with ETS (25 µm)
   under normoxic or hypoxic conditions for 24 h (n = 3). Statistical
   analysis was performed using one‐way ANOVA with Tukey's multiple
   comparisons test. c,d) Relative mRNA (c) and protein (d) levels of FAO
   related genes in B16F10 cells treated with ETS (25 µm) under normoxic
   or hypoxic conditions for 24 h (n = 3). Statistical analysis was
   performed using two‐way ANOVA with Sidak's multiple comparisons test.
   e) HIF1α and PPARα expression of B16F10 cells transfected with
   nontargeting siRNA (siControl) or with siRNA against HIF1α (siHIF1α)
   under hypoxic condition. β‐actin serves as a loading control. f)
   Relative expression of FAO related genes in B16F10 cells transfected
   with nontargeting siRNA (siControl) or with siRNA against HIF1α
   (siHIF1α) under hypoxic condition (n = 3). Statistical analysis was
   performed using one‐way ANOVA with Tukey's multiple comparisons test.
   Error bars ±SD (b–f). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p <
   0.0001.

2.5. FAO Regulates DNA Damage‐Induced Cell Death by Promoting
N‐Alpha‐Acetylation of Caspase‐2

   We asked how FAO regulates cell death after chemotherapy. First, to
   examine the involvement of p53 in FAO‐mediated cell death, we first
   reduced p53 expression using siRNA against p53. Notably, ETO suppressed
   cell death in p53 knockdown cells after ETS treatment (Figure [118]5a;
   Figure [119]S6a, Supporting Information), implying that FAO might
   regulate DNA damage‐induced cell death in a p53‐independent manner.
   Consistent with this observation, we observed similar results in p53
   null PC3 human prostate cancer cells (Figure [120]S6b, Supporting
   Information). Previous studies have shown that caspase‐2 (CASP2) is
   required for apoptosis in p53‐deficient cells.^[ [121]^17 ^] Thus, to
   investigate the contribution of CASP2 to the FAO‐mediated cell death,
   we examined CASP2 activity under these conditions. Interestingly, FAO
   inhibition markedly reduced levels of its cleaved forms after DNA
   damage (Figure [122]5b; Figure [123]S6c,d, Supporting Information).
   Consistent with this observation, ETO was not able to further reduce
   cell death in CASP2 knockdown cells (Figure [124]5c; Figure [125]S6e,f,
   Supporting Information), indicating that FAO might induce cell death by
   activating CASP2 after chemotherapy.

Figure 5.

   Figure 5
   [126]Open in a new tab

   FAO adjusts DNA damage‐induced cell death by N‐alpha‐acetylation of
   CASP2. a) Cell death of B16F10 cells transfected with nontargeting
   siRNA (siControl) or with siRNA against P53 (siP53). Cells were treated
   with ETS (25 µm, 40 h), ETO (50 or 200 µm, 40 h) or both (n = 3). b)
   Cleaved CASP2 expression of 293T cells treated with or without ETO in
   the presence of ETS. β‐actin serves as a loading control. c) Cell death
   of B16F10 cells transfected with nontargeting siRNA (siControl) or with
   siRNA against CASP2 (siCASP2). Cells were treated with or without ETO
   (50 µm, 40 h) in the presence of ETS (25 µm, 40 h); n = 3. Cell death
   was measured by propidium iodide exclusion assay. d) Relative
   acetyl‐CoA levels were measured in immortalized MEFs treated with ETO
   (200 µm, overnight) and/or citrate (10 mm, overnight) in the presence
   of ETS (25 µm, overnight); n = 4. e) Cell death of immortalized MEFs
   treated with the indicated drugs (n = 3). Cell death was measured by
   propidium iodide exclusion assay. f) Cell death of immortalized MEFs
   transfected with nontargeting siRNA (siControl) or with siRNA against
   ARD1 (siARD1). Cells were treated with or without ETO (200 µm,
   overnight) in the presence of ETS (25 µm, overnight); n = 3. Cell death
   was measured by propidium iodide exclusion assay. g) 293T cells were
   transfected with FLAG‐tagged CASP2 C320G (active cysteine mutant).
   Transfected cells were treated with or without ETO (200 µm, 24 h) in
   the presence of ETS (25 µm); n = 4. Lysates were immunoprecipitated
   with FLAG magnetic beads. The bound proteins were eluted with FLAG
   peptide and the eluent was performed mass spectrometry. Ratios for
   levels of N‐terminal acetylation were determined by normalizing to
   total detection peptides. h) Relative CASP3/7 activity of the indicated
   cells. HeLa cells were infected with mock vector, hCASP2 WT, or hCASP2
   A3P and selected by puromycin (2 µg mL^−1). Infected cells were
   transfected with siControl or with siRNA against CASP2 (siCASP2). Both
   cells were treated with or without ETO in the presence of ETS (n = 4).
   All error bars ±SD. Statistical analysis was performed using one‐way
   ANOVA with Tukey's multiple comparisons test. *p < 0.05, **p < 0.01,
   ***p < 0.001, and ****p < 0.0001.

   Previous studies have found that N‐alpha‐acetylation of CASP2 is
   required for its activation, which is dependent on cellular levels of
   acetyl‐CoA.^[ [127]^18 ^] Because FAO is an important source of
   cellular acetyl‐CoA pools,^[ [128]^19 ^] we hypothesized that FAO
   regulates DNA damage‐induced cell death by promoting CASP2 activity
   through acetyl‐CoA. To test this idea, we first examined whether FAO
   induction induces cell death via acetyl‐CoA upon DNA damage.
   Importantly, we found that ETS treatment elevated intracellular
   acetyl‐CoA levels (Figure [129]5d; Figure [130]S6g, Supporting
   Information). The induction of acetyl‐CoA levels required FAO, as ETO
   treatment blocked this (Figure [131]5d; Figure [132]S6g, Supporting
   Information). As citrate can be converted to acetyl‐CoA by ATP citrate
   lyase, we suspected that the decreased cell death by ETO could be
   restored by citrate. As expected, addition of citrate increased
   cellular acetyl‐CoA levels (Figure [133]5d; Figure [134]S6g, Supporting
   Information). Notably, citrate treatment rescued the reduced DNA
   damage‐induced cell death by FAO inhibition (Figure [135]5e; Figure
   [136]S6h, Supporting Information), suggesting that FAO regulates cell
   death by providing acetyl‐CoA.

   Next, to examine the relevance of N‐alpha‐acetylation of CASP2 in the
   FAO‐mediated cell death, we performed a series of experiments. Arrest
   defective 1 (ARD1) functions as a catalytic subunit of
   N‐acetyltransferase protein complexes and it was reported that
   knockdown of ARD1 disrupts N‐alpha‐acetylation of CASP2 induced by DNA
   damage.^[ [137]^18 ^] We found that knockdown of ARD1 reduced cell
   death upon DNA damage, consistent with previous work (Figure [138]5f;
   Figure [139]S6i, Supporting Information). Importantly, FAO inhibition
   could no longer affect cell death in ARD1 knockdown cells
   (Figure [140]5f). We next explored the direct effects of FAO inhibition
   on N‐alpha‐acetylation of CASP2 after DNA damage by performing mass
   spectrometry analysis. Indeed, we observed that the induction of
   N‐alpha‐acetylation of CASP2 after ETS treatment was significantly
   impaired by FAO inhibition (Figure [141]5g; Figure [142]S6j, Supporting
   Information). Finally, we conducted mutagenesis of CASP2 to disrupt its
   N‐alpha‐acetylation by replacing the third residue of CASP2 to Pro
   (A3P)^[ [143]^18 ^] and then expressed WT or mutant CASP2 in cells
   after knockdown of endogenous CASP2 (Figure [144]S6k, Supporting
   Information). Consistent with our model, ETO treatment suppressed DNA
   damage‐induced CASP3/7 activities in WT CASP2 expressed cells, whereas
   it had minimal effects in A3P CASP2 expressed cells (Figure [145]5h;
   Figure [146]S6l, Supporting Information). Collectively, these data
   support the idea that the induction of FAO increased intracellular
   acetyl‐CoA levels after DNA damage, which promotes cell death by
   enhancing N‐alpha‐acetylation of CASP2.

2.6. Pharmacological Intervention for Improving FAO Rescues Obesity‐Induced
Chemoresistance

   Our results support a model wherein FAO induction is required for DNA
   damage‐induced cell death, which is impaired by obesity. These results
   led us to hypothesized that restoring FAO in tumors of obese mice might
   inhibit their chemoresistance. We used a diet‐induced obesity (DIO)
   mouse model to investigate this issue. After 8 weeks on 60% high‐fat
   diet (HFD), mice weighed 39.1% more than lean, chow diet (CD) control
   mice (Figure [147]S7a, Supporting Information). To improve FAO
   induction after chemotherapy, mice subcutaneously injected B16F10 cells
   were treated with or without WY14643 in the presence of ETS (Figure
   [148]6a). First, we tested whether our cellular findings are relevant
   in this model. In line with our previous results, ETS treatment induced
   PPARα expression in tumors from control mice, whereas HFD abrogated
   this induction (Figure [149]6b). In addition, WY14643 administration
   strongly induced the expression of PPARα in obese tumors after
   chemotherapy, whereas it had no further effect on PPARα expression in
   lean tumors (Figure [150]6b). Consistent with this, PPARα activation
   restored the impaired induction of FAO‐related genes in obese tumors
   after chemotherapy (Figure [151]S7b, Supporting Information). We
   observed comparable results in ob/ob mice (Figure [152]S7c, Supporting
   Information). Moreover, we demonstrated that ETS treatment increased
   acetyl‐CoA levels in lean tumors but not in obese tumors and PPARα
   activation could rescue the defect in obese tumors (Figure [153]6c;
   Figure [154]S7d, Supporting Information). Thus, consistent with
   previous results, these data indicate that chemotherapy induces FAO in
   tumors, leading to increased acetyl‐CoA levels, and that obesity
   impairs this process.

Figure 6.

   Figure 6
   [155]Open in a new tab

   PPARα agonist improves the obesity‐induced chemoresistance. a)
   Schematic depicting in vivo experimental setup. b,c) b) Relative Pparα
   mRNA (n = 7 for CD and n = 10 for HFD) and c) acetyl‐CoA (n = 5) levels
   in allograft B16F10 tumors from CD‐ and HFD‐fed mice treated with or
   without WY14643 (40 mg kg^−1 five times via i.p. injection) in the
   presence of ETS (20 mg kg^−1 five times via i.p. injection).
   Statistical analysis was performed using two‐way ANOVA with Tukey's
   multiple comparisons test. d) Immunohistochemistry of cleaved CASP3
   expression in lean and obese tumors treated with or without WY14643 in
   the presence of ETS. Scale bar represents 50 µm. Statistical analysis
   was performed using one‐way ANOVA with Tukey's multiple comparisons
   test. e) Tumor volume of B16F10 tumors in CD‐ and HFD‐fed mice treated
   with or without WY14643 in the presence of ETS. f) Immunoblot for
   cleaved CASP2 in lean and obese tumors treated with or without WY14643
   in the presence of ETS. All error bars ±SEM. **p < 0.01, ***p < 0.001,
   and ****p < 0.0001.

   To validate the effect of FAO induction on the sensitivity of tumor
   cells after chemotherapy, we examined tumors from CD‐ or HFD‐fed mice.
   As expected, chemotherapy‐induced cell death was significantly reduced
   in tumors from HFD‐fed mice compared with those from CD‐fed mice
   (Figure [156]6d). Importantly, the treatment of WY14643 potently
   synergized with ETS to induce a strong increase of cleaved CASP3 in
   tumors from HFD‐fed mice, resulting in reduced tumor volume in these
   mice (Figure [157]6d,e). Consistent with our previous results, WY14643
   was able to increase CASP2 activity in obese tumors (Figure [158]6f;
   Figure [159]S7e, Supporting Information). However, we found that the
   combination did not induce the synergistic effects in tumors from
   control mice (Figure [160]6d–f; Figure [161]S7e, Supporting
   Information). Collectively, these results demonstrate that the
   treatment aimed at improving FAO synergizes with chemotherapy to kill
   tumor cells in obese mice.

3. Discussion

   In this study, we demonstrate that FAO regulates DNA damage‐induced
   cell death with important implications for obesity‐induced
   chemoresistance. Our study shows that FAO is induced after
   chemotherapy, which is abrogated by obesity (Figure [162]1). We find
   that the induction of FAO is required for apoptotic cell death in
   response to DNA damage (Figure [163]2). Molecularly, DNA damage induces
   FAO via PPARα, which is suppressed by obesity‐mediated hypoxia
   (Figure [164]3, [165]4). We demonstrate that increased acetyl‐CoA
   production by FAO contributes to DNA damage‐induced cell death by
   promoting N‐alpha‐acetylation of CASP2, highlighting the potential
   importance of this pathway in the regulation of chemoresistance of
   cancer cells (Figure [166]5). This idea is further validated by the
   finding that improving FAO reduces obesity‐induced chemoresistance
   (Figure [167]6). Taken together, our studies provide important insight
   into the role of FAO as a regulator of DNA damage‐induced cell death
   and uncover a mechanism whereby obesity promotes chemoresistance of
   tumor cells by inhibiting this critical metabolic response to DNA
   damage.

   Our studies reveal the significant impact of hypoxia on obesity‐driven
   chemoresistance. Hypoxia is a defining feature of solid tumors.^[
   [168]^20 ^] Previously, it has been shown that hypoxia leads to
   resistance to cancer therapies and a higher HIF1α expression is closely
   associated with poor prognosis in many cancers, including ovarian,
   colorectal and breast cancers.^[ [169]^21 ^] As inhibiting HIF1α shows
   efficacy in restoring the decreased response to cancer therapy, it has
   been proposed that targeting HIF1α has therapeutic potential for
   efficient cancer treatment.^[ [170]^22 ^] In this study, we demonstrate
   that obesity aggravates HIF1α stabilization in tumors, inhibiting
   PPARα‐dependent FAO induction after chemotherapy and, thus, suppressing
   cell death. Importantly, we suggest that the increased FAO is not
   simply compensative response for maintaining energy homeostasis.
   Rather, FAO actively functions as an important metabolic regulator of
   cell death by activating the CASP2‐dependent apoptosis pathway. Thus,
   taken together, these data provide a molecular mechanism whereby
   hypoxia leads to resistance of cancers to DNA damaging therapies, such
   as chemotherapy and radiation.

   Chemotherapy is designed to kill cancer cells.^[ [171]^23 ^] However,
   when cancer cells experience non‐lethal doses, they can remain to
   survive or proliferate, or drug‐resistant subpopulations can be emerged
   by clonal expansion.^[ [172]^24 ^] Thus, appropriate chemotherapy doses
   are crucial for treatment efficacy. Because chemotherapy entails
   systemic treatment with cytotoxic agents, optimal doses for
   normal‐weight patients are often not enough for overweight or obese
   patients.^[ [173]^25 ^] Although doses of drugs are determined by using
   the body surface area, many obese patients received suboptimal
   chemotherapy dosing, such as due to concerns regarding toxicity, which
   leads to limited treatment efficacy in obese patients.^[ [174]^26 ^]
   However, given that obese patients are more likely to develop several
   complications including peripheral neuropathy and that chemotherapy can
   cause neurotoxicity, selecting optimal chemotherapy doses for obese
   patients is challenging.^[ [175]^26 , [176]^27 ^] We propose that
   enhanced FAO significantly attenuates obesity‐induced drug resistance.
   As a proof of concept, we show that PPARα activation synergizes with
   chemotherapy to induce robust cell death in obese tumors without
   increasing chemotherapy doses. Thus, our findings may help to develop
   chemotherapeutic approaches for cancer patients with overweight or
   obesity.

   Although our current work highlights the potential role of FAO
   induction by PPARα in ameliorating obesity‐driven chemoresistance, it
   does not preclude the involvement of other PPARα‐dependent pathways. As
   PPARα plays crucial roles in cellular lipid homeostasis,^[ [177]^28 ^]
   it may regulate DNA damage‐induced cell death via other branches of
   lipid metabolism or other targets. In this regard, fatty acid
   degradation is closely associated with lipolysis process and lipophagy
   formation. As these lipid‐related pathways lead to the regulation of
   cell death through lipotoxicity or lipid peroxidation‐induced
   ferroptosis,^[ [178]^29 ^] it will be important for future work to
   examine how these pathways coordinately regulates cell survival and
   death upon genotoxic stress. Additionally, it has been reported that
   DNA damage induces mitochondrial biogenesis through the activation of
   AMP‐activated protein kinase (AMPK).^[ [179]^30 ^] Thus, an increased
   mitochondrial mass may contribute to the enhanced FAO after DNA damage.
   On the other hand, we demonstrate that FAO induction in cancer cells
   contributes to cell death after chemotherapy. However, it is important
   to note that in vivo cancer tissues consist of several types of cells,
   including cancer cells, cancer stem cells, inflammatory cells,
   adipocytes, and various related tissue cells.^[ [180]^31 ^] This
   heterogeneity of the cancer microenvironment could affect the
   FAO‐mediated cell death. In addition, metabolic alteration is a key
   feature of many cancer cells, which may influence this pathway. For
   example, metastatic cancer cells exhibit an aberrant dependency on
   FAO.^[ [181]^32 ^] Previously, it has been shown that FAO inhibition
   re‐sensitize radiation‐resistant nasopharyngeal carcinoma cells to
   radiation‐therapy.^[ [182]^33 ^] As therapy‐resistant cancer cells
   acquire resistance, they often undergo metabolic shifts or remodeling,
   potentially affecting the FAO‐dependent DNA damage‐induced cell death.
   Last, there may be cancer‐type‐specific differences, as different
   cancer cells have different metabolic dependencies.

   This study identifies an essential role of FAO in the DNA damage
   response by regulating cell death. As the DNA damage response is
   essential for maintaining cellular and organismal homeostasis,
   dysregulations of this pathway cause stress‐induced premature
   senescence and are frequently reported in accelerated aging
   syndromes.^[ [183]^34 ^] We found that cells with the defective FAO
   induction upon DNA damage resist to cell death. As senescence is an
   alternative decision for cells when they bypass death,^[ [184]^35 ^]
   surviving cells can enter a senescence‐like state. Intriguingly, recent
   studies have reported that defects or dysregulations of FAO are
   observed in senescent cells or in aged tissues.^[ [185]^36 ^] Thus, it
   is plausible that defects in this metabolic response involving FAO
   regulation could be one of the causes of the undesirable accumulation
   of senescent cells during aging.

   In sum, our work illustrates that FAO is intimately involved in
   cellular response to genotoxic stress, in part by regulating DNA
   damage‐induced cell death. Moreover, we found that improving this
   metabolic response could alleviate obesity‐driven chemoresistance.
   Given the essentiality of metabolism in cell fate decision and cancer
   progression, these findings may hold profound implications for
   developing novel therapeutic strategies for cancer as well as for
   elucidating the cellular metabolic response to stress.

4. Experimental Section

Cell Lines

   Primary MEFs were immortalized using SV40 large T antigen at passage 3.
   MEFs, B16F10, 293T, HeLa, HepG2 and PC3 cells were cultured in
   Dulbecco's modified Eagle's medium [DMEM high Glucose (with
   4500 mg L^−1 D‐glucose and L‐glutamine and without sodium pyruvate)]
   (WelGENE, LM001‐07) supplemented with 10% FBS (Gibco, 16000–044) and
   penicillin/streptomycin (Biowest, L0022‐100). The PPARα knockdown cell
   lines were generated via lentivirus. shRNAs for PPARα (TRCN0000026028
   and TRCN0000026046) were purchased from Sigma‐Aldrich. The sequences
   for shRNAs were: GCTCCTTTGATATGATACTTT for shPPARα #1 (TRCN0000026028),
   GCTATGAAGTTCAATGCCTTA for shPPARα #2 (TRCN0000026046). Lentiviral
   supernatants were produced by transfecting 293T cells using
   Lipofectamine 2000 (Invitrogen). B16F10 cells were infected with viral
   supernatant supplemented with 8 mg mL^−1 polybrene. Puromycin
   (2 µg mL^−1) was used to select shRNA or control expressing cells.
   Cells were cultured at 37 °C and 5% CO[2] in a humidified incubator.
   Hypoxic cultures were performed in Thermo Fisher Forma Anaerobic system
   (Thermo fisher scientific, 1029), 5% CO[2] and less than 1% O[2].

Mice

   All animal experiments were performed according to the National
   Institutes of Health's Guide for the Care and Use of Laboratory
   Animals, with approval of the Animal Experiment Ethics Committee of the
   Catholic University of Korea, College of Medicine. 4‐week‐old C57BL/6J
   and ob/ob mice were purchased from Central Lab Animal Inc (Korea).
   4‐week‐old C57BL/6 mice were purchased from Orient Bio (Korea). Mice
   were maintained in a specific pathogen‐free facility with 12 h light
   and 12 h dark cycles.

Mouse Tumor Models

   6‐week old C57BL/6J or ob/ob mice were injected subcutaneously in the
   abdominal flank with 1 × 10^5 B16F10 cells. Tumor volumes were measured
   using the following ellipsoid formula: volume (mm^3) = 0.5 × length ×
   (width^2), where length indicates the long diameter and width is the
   short diameter. When the tumor volumes reached ≈100 mm^3, ETS
   (20 mg kg^−1) and WY14643 (40 mg kg^−1) were intraperitoneally
   administered five times everyday. For a diet‐induced obesity mice
   model, 5‐week old C57BL/6 mice were fed normal CD or 60% HFD (Research
   Diets, D12492) for 8 weeks. After 8 weeks of CD or HFD feeding, mice
   were injected subcutaneously in the abdominal flank with 1 × 10^5
   B16F10 cells. When the tumor volumes reached ≈100 mm^3, ETS
   (20 mg kg^−1) and WY14643 (40 mg kg^−1) were intraperitoneally
   administered five times everyday. A day after a course of treatment,
   mice were sacrificed for collection of tumor tissue.

Antibody

   β‐actin (Genetex, GTX109639), CPT1A (Abcam, ab234111; Cell signaling,
   12252S), ACADL (Proteintech, 17526‐1‐AP), PPARα (Abcam, ab215270;
   Abcam, ab24509; Milipore, MAB3890), CASP2 (Enzo, 11B4), Cleaved CASP3
   (Cell signaling, 9664S), and HIF1α (Cayman, 10006421).

Reagent

   ETS (E1383), ETO (E1905), GW6471 (G5045), WY14643 (C7081), DMOG
   (D3695), CPT (C9911), Cisplatin (PHR1624), Cycloheximide (C4859),
   Palmitic acid (P0500), Oleic acid (O1008), Sodium citrate (PHR1416),
   Actinomycin D (A1410), Anti‐FLAG M2 Magnetic Beads (M8823), and FLAG
   Peptide (F3290) were purchased from Sigma‐Aldrich. Lipofectamine
   RNAiMAX (13778150) and Lipofectamine 2000 (11668019) were purchased
   from Invitrogen.

siRNAs and Transfection

   20 nm siRNAs were transfected in cells right after being seeded at a
   density of 30–50% confluency depending on experiments using
   Lipofectamine RNAiMAX (Invitrogen, 13778150) according to the
   manufacturer's protocols. Cells were harvested ≈24–40 h
   post‐transfection as described in the figure legends. Human CASP2 siRNA
   (hs.Ri.CASP2.13.1), mouse HIF1α siRNA #2 (mm.Ri.Hif1a.13.1), and mouse
   PPARα siRNA (mm.Ri.Ppara.13.1, mm.Ri.Ppara.13.3) were purchased from
   Integrated DNA Technologies. The oligomer sequences were:
   AAGAAGGCAAGCCAACCUUAGCUAC for human siCASP2 (hs.Ri.CASP2.13.1);
   CAGGGTCACTTGGAAGACTTA for mouse siCASP2; GGAGCGACUCUUCAAUACUUCCCGCAUCC
   for mouse siCPT1A; CCAACACCCUCAACUUUCAGAUCAG for mouse siARD1;
   GUGGAUAGCGAUAUGGUCAUU for mouse HIF1α siRNA #1;
   GAUAUGUUUACUAAAGGACAAGUCA for mouse HIF1α siRNA #2 (mm.Ri.Hif1a.13.1);
   GAAAGUCCCUUAUCUGAAGAAUUCT for mouse PPARα siRNA #1 (mm.Ri.Ppara.13.1);
   GACCAAGUCACCUUGCUAAAGUACG for mouse PPARα siRNA #2 (mm.Ri.Ppara.13.3);
   CACCUCACUGCAUGGACGAUCUGUU for mouse P53 siRNA; GGAGCAUCCUCACCGGCAA for
   mouse PPARβ siRNA; GCCCUUUACCACAGUUGAU for mouse PPARγ siRNA.

RNA Extraction and Quantitative RT‐PCR

   RNA was isolated from cell lines or tumor tissues using Trizol
   (Invitrogen, 15596026) according to the manufacturer's instructions.
   0.5 µg of total RNA was reverse‐transcribed using the PrimeScript RT
   Master Mix (Takara, RR036A). Quantitative RT‐PCR was performed using
   SYBR Green I Master (Roche, 04887352001) on a Lightcycler 480 (Roche).
   The level of gene expression was normalized to β‐actin. The primer
   sequences used for qPCR are listed in Supplementary Table (Supporting
   Information).

Western Blotting

   Cells were lysed with EZ‐RIPA lysis buffer (ATTO, WSE‐7420)
   supplemented with protease inhibitor cocktail (Roche, 11873580001).
   Cell lysates were separated by SDS‐PAGE and transferred onto
   nitrocellulose membranes (GE Healthcare) or polyvinylidene fluoride
   membranes (GE Healthcare). The membranes were blocked for 1 h in 5%
   nonfat milk or 3% BSA and then incubated with primary antibodies
   overnight at 4 °C. After washing five times in 1X TBST, the membranes
   were incubated with species‐specific secondary antibodies and detected
   using LAS 4000.

Flow Cytometric Measurement of Cell Death

   Cells at less than 70% confluence were treated with drugs. After
   treatment, cells were harvested by trypsinization, pelleted by
   centrifugation, and resuspended in PBS. The measurement of cell death
   was performed by flow cytometry using propidium iodide (PI) staining
   and/or annexin V staining (BD Biosciences, 556547).

Caspase 3/7 GLO

   Cells were plated into 96‐well plates at 2000 cells per well in 100 µL
   of growth media. Casp2 WT and CASP2 mutant (A3P) stable cell lines were
   transfected with siCASP2 for 24 h and plated into 96‐well plates at
   2000 cells per well in 100 µL of growth media. The following day, cells
   were treated with ETS, ETO, or both. 100 µL of Caspase‐GLO 3/7 reagent
   (Promega, G8091) were added to each well of white‐walled 96‐well plate
   and incubated at room temperature for 30 min. After incubation, the
   luminescent signal was detected using plate‐reading luminometer.

Acetyl‐CoA Measurement

   Cellular acetyl CoA level was detected by PicoProbe Acetyl CoA assay
   kit (Abcam, ab87546) according to the manufacturer's instructions.
   Briefly, 2 × 10^6 cells were harvested, resuspended in 500 µL of assay
   buffer on ice, and then homogenized using dounce homogenizer on ice.
   After centrifugation at 10000 g for 10 min at 4 °C, perchloric acid was
   added to a final concentration of 1 m in the supernatants. After
   centrifugation, the supernatants were neutralized with KOH. The samples
   were added CoA Quencher and then added Quencher Remover. The samples
   were supplied with reaction mix (assay buffer, substrate mix,
   conversion enzyme, enzyme mix, and PicoProbe) and measured using
   microplate reader (Ex/Em = 535/587 nm). The acetyl‐CoA standard was
   used in the range of 0–100 pm.

Cloning

   Human FLAG‐CASP2 WT overexpression vector was constructed by
   PCR‐amplifying human CASP2 with cloning primers containing NgoMIV
   restriction site using PfuUltra II Fusion HS DNA polymerase (Agilent)
   from human cDNAs. FLAG‐hCASP2 A3P was generated by PCR amplifying using
   the same reverse cloning primer of FLAG‐CASP2 WT cloning and the
   forward primer replaced the third residue of CASP2 with proline. The
   digested insert was ligated into pBabe vector using ligase (Takara,
   2011A). FLAG‐CASP2 C320G mutant was generated using mutagenesis kit
   (Agilent, 200523) according to the manufacturer's instructions. Primer
   sequences for cloning were: GCCGGCATGGCGGCGCCGAGCGCGGGG and
   GCCGGCTCACTTGTCGTCATCGTCTTTGTAGTCTGTGGGAGGGTGTCCTGG for human CASP2 WT
   with C‐terminal FLAG; GCCGGCATGGCGCCGCCGAGCGCGGGGTCTT and
   GCCGGCTCACTTGTCGTCATCGTCTTTGTAGTCTGTGGGAGGGTGTCCTGG for human CASP2 A3P
   with C‐terminal FLAG; GTTCTTCATCCAGGCCGGCCGTGGAGATGAGACTG and
   CAGTCTCATCTCCACGGCCGGCCTGGATGAAGAAC for human CASP2 C320G mutagenesis.

Mass Spectrometry

   293T cells were transiently transfected with FLAG‐tagged CASP2
   overexpression vector containing active cysteine mutation (C320G) to
   reduce cell death. Cells were treated with or without ETO in the
   presence of ETS for 24 h and then lysed with 0.2% Tween 20 and 0.2%
   Triton X‐100 buffer with protease inhibitors and phosphatase
   inhibitors. FLAG‐tagged CASP2 was immunoprecipitated with FLAG M2
   magnetic beads (Sigma‐Aldrich, M8823) overnight at 4 °C. The beads were
   washed five times with lysis buffer and additionally washed three times
   with PBS. The bound proteins were eluted with FLAG peptide. The Eluants
   were analyzed by mass spectrometry performed by the Proteomics Core
   Facility at the School of Biological Sciences in Seoul National
   University.

Immunohistochemistry

   Tumors were fixed in 4% paraformaldehyde and embedded in paraffin.
   Paraffin‐embedded tumors were sectioned at 3 µm thickness and section
   slides were deparaffinized, rehydrated, and incubated with citrate
   buffer (ImmunoBioScience, AR‐6544‐05) at 95 °C for antigen retrieval.
   The slides were incubated in 1.4% hydrogen peroxide/methanol for
   quenching endogenous peroxidase activity and blocked with 2.5% normal
   horse serum (Vector Laboratories, MP‐7401) for 1 h. The slides were
   incubated with anti‐cleaved Caspase 3 (Cell signaling, 9664S) or
   anti‐HIF1α antibodies (Cayman, 10006421) at 4 °C overnight. The
   following day, the slides were incubated with a micropolymer
   HRP‐conjugated secondary antibody (Vector Laboratories, MP‐7401) for
   1 h at room temperature. Antigens were revealed by DAB (Vector
   Laboratories, SK‐4100) and then the slides were counterstained with
   EASYSTAIN Harris Hematocylin (YD diagnostics, S2‐5). Finally, the
   slides were dehydrated in increasing concentrations of ethanol, mounted
   with Permount mounting medium (Fisher Scientific, SP 15–100), and
   imaged using Axio Scan.Z1 (ZEISS).

Fatty Acid Oxidation

   HepG2 cells were cultured overnight in media containing 100 µm
   palmitate and 1 mm carnitine. Cells were pulsed with 1.7 µCi
   [9,10(n)−3H]palmitic acid (GE Healthcare) for 2 h. To analyze the
   released 3H2O, the medium was collected and eluted using ion exchange
   columns packed with DOWEX 1 × 2‐400 ion exchange resin (Sigma).

Oxygen Consumption Rate Measurement

   OCR was measured using the Seahorse XFe96 Extracellular Flux Analyzer
   (Seahorse Bioscience). Briefly, cells were seeded in XFe96 cell culture
   microplates and cultured in DMEM supplemented with 10% fetal bovine
   serum, 100 U mL^−1 of penicillin, and 100 µg mL^−1 of streptomycin at
   37 °C in a humidified 5% CO[2] atmosphere. Thereafter, the cells were
   pre‐treated with BSA, PA, GW, WY, and/or ETS for 6 h. Afterward the
   cell medium was replaced with assay medium and the cells were
   transferred to a CO[2] free incubator and maintained at 37 °C for 1 h
   before the assay. FAO was measured using Seahorse XF Palmitate
   Oxidation Stress Test Kit (Agilent, 103693‐100) according to the
   manufacturer's protocols. ETO (40 µm) was added to the respective cells
   15 min prior to running the assay in the Seahorse analyzer. Finally,
   BSA or Palmitate‐BSA was added to the appropriate wells immediately
   before running the assay. The injection ports were loaded with 1.5 µm
   oligomycin A, 2 µm carbonyl cyanide‐4‐(trifluoromethoxy)
   phenylhydrazone (FCCP), and 1 µm rotenone/antimycine A. Following
   calibration, OCR measurements were recorded at every cycle of mixing
   (3 min) and measurement (3 min) for 75 min. Results were analyzed using
   Seahorse XFe96 Wave software (Agilent). FAO‐dependent OCR (ΔOCR) was
   determined as the (OCR untreated with ETO – OCR treated with ETO).

PPARα Transcription Factor Assay

   PPARα’s transcriptional activity was measured using the PPARα
   Transcription Factor Assay Kit (ab133107, Abcam) according to the
   manufacturer's instructions. Nuclear lysate was added to a 96‐well
   plate that was coated with a specific double‐stranded DNA (dsDNA)
   containing the peroxisome proliferator response element (PPRE)
   sequence. Samples were incubated overnight at 4 °C, and after five
   times washing, the primary antibody targeting PPAR‐α was added for 1 h.
   HRP‐conjugated secondary antibodies were added following five times
   washes. After adding the developing buffer for 45 min, stop solution
   added the samples. The samples were measured the absorbance at 450 nm.

RNA Sequencing

   RNA was isolated from tumor tissues of C57BL/6J and ob/ob mice pre‐ and
   post‐chemotherapy using Trizol reagent according to the manufacturer's
   instructions. RNA sequencing libraries were prepared using TruSeq
   Stranded mRNA LT Sample Prep Kit and sequenced with 101 bp paired‐ends
   on an Illumina platform at MacrogenTM (Seoul, South Korea). Reads were
   aligned to the mm10 genome using HISAT v2.1.0. HISAT utilizes two types
   of indexes for alignment (a global, whole‐genome index and tens of
   thousands of small local indexes). These two types of indexes were
   constructed using the same BWT (Burrows–Wheeler transform)/ a graph FM
   index (GFM) as Bowtie2. Because of its use of these efficient data
   structures and algorithms, HISAT generates spliced alignments several
   times faster than Bowtie and BWA widely used. Aligned reads were
   performed transcription assembly using StringTie v2.1.3b. Expression
   profiles were extracted using FPKM and TPM values and differentially
   expressed genes (DEGs) were identified based on |fold change| > 2 and
   the p‐value < 0.05. Gene set enrichment analysis was performed based on
   KEGG database. RNA seq data were deposited in the Gene Expression
   Omnibus under accession code [186]GSE208632.

RNA‐Sequencing Analysis

   The relative abundances of genes were measured in Read Count using
   StringTie. The statistical analysis was performed to find
   differentially expressed genes using the estimates of abundances for
   each gene in samples. Genes with one more than zeroed Read Count values
   in the samples were excluded. To facilitate log2 transformation, 1 was
   added to each Read Count value of filtered genes. Filtered data were
   log2‐transformed and subjected to TMM normalization. Statistical
   significance of the differential expression data was determined using
   exactTest using edgeR and fold change in which the null hypothesis was
   that no difference exists among groups. False discovery rate (FDR) was
   controlled by adjusting p value using Benjamini–Hochberg algorithm. For
   DEG set, hierarchical clustering analysis was performed using complete
   linkage and Euclidean distance as a measure of similarity.
   Gene‐enrichment and functional annotation analysis and pathway analysis
   for significant gene lists were performed based on gProfiler and KEGG
   pathway.

Statistical Analyses

   Sample numbers and sizes were indicated by dots. Statistical
   significance was calculated using two‐tailed Student's t‐test to
   compare two groups and one‐way ANOVA or two‐way ANOVA to compare
   multiple comparisons. All experiments were performed at least two
   independent experiments and the sample number of each experiment were
   more than 3. All statistical analyses and p values are described in the
   figure legends. Statistical analyses were performed using GraphPad
   Prism 6 software. Error bars represented as mean ±SD or ±SEM. *p <
   0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Data Availability

   RNA‐seq data were deposited in the Gene Expression Omnibus under
   accession code [187]GSE208632. All other data supporting the findings
   of this study are available by request to the corresponding author.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   S.M.J. designed the study and wrote the main manuscript text. S.H.
   performed the experiments, analyzed the data, and wrote the manuscript.
   S.Y., K.P., B.K., M.K., S.S., A.Y., J.A., and J. J. performed the
   experiments and reviewed the manuscript. Y.S.Y. and R.H.S. reviewed the
   manuscript.

Supporting information

   Supporting Information
   [188]ADVS-11-2304702-s001.pdf^ (2.2MB, pdf)

Acknowledgements