脂肪细胞氧脂素与乳腺癌铁死亡

Fig. 1. Lean mammary adipocytes inhibit breast cancer cell growth. (A to F) Orthotopic tumor growth is accelerated in obese mice. (A) Growth kinetics of E0771 mammary tumors in female C57BL/6J mice fed a high-fat (obese; n = 16) or standard chow (lean; n = 16) diet. (B) End-point tumor burden. (C) Growth kinetics of Py230 mammary tumors in female C57BL/6J mice fed a high-fat (obese; n = 10) or standard chow (lean; n = 6) diet. (D) End-point tumor burden. (E) Growth kinetics of MCF7 mammary tumors in NRG mice fed a high-fat (obese; n = 16) or standard chow (lean; n = 11) diet. (F) End-point tumor burden. (G) Representative growth curves demonstrating that the secretomes of lean, but not obese, mammary adipocytes suppress breast cancer cell growth. (H) Secretome from lean inguinal adipocytes decreases the growth constant across seven breast cancer cell lines. Secretomes from obese inguinal or perigonadal (lean or obese) adipocytes do not inhibit breast cancer cell growth (n = 3 independent adipocyte isolations). (I) Coinjection of lean inguinal adipocytes with MCF7 breast cancer cells reduces tumor growth compared with coinjection with obese adipocytes or a saline control (n = 6). [(A) to (F) and (I)] Data are means ± SEMs. [(G) and (H)] Data are means ± SDs. P values calculated using two-way ANOVA followed by Šídák’s multiple comparisons test [(A), (C), and (E)] or followed by Tukey’s multiple comparison test [(G) and (I)], calculated using unpaired t test [(B), (D), and (F)], or calculated using one-way ANOVA followed by Dunnett’s multiple comparisons test (H). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 2. Lean adipocytes induce ferroptosis in breast cancer cells. (A) Gene set enrichment analysis for ferroptosis-related genes in E0771 cells treated with the lean inguinal adipocyte secretome compared with DMEM control. (B to E) Lean inguinal secretome and the ferroptosis inducer ML-162 increase cell death in E0771 cells (B), SkBr3 cells (C), MDA-MB-468 cells (D), and MCF7 cells (E), measured by Cytotox or SYTOX dye incorporation (n = 3). (F and G) Pharmacological inhibition of ferroptosis (F), but not other cell death pathways (G), rescues the growth of breast cancer cells treated with the lean inguinal secretome (n = 3 to 5). (H) Representative C11-BODIPY staining in E0771 and MCF7 breast cancer cells shows increased lipid peroxidation after treatment with lean inguinal secretome. (I) E0771 cells up-regulate iron-related proteins CD71/TFRC and HMOX1 in response to lean versus obese inguinal secretome (n = 3). (J) Intracellular Fe2+ levels are increased in E0771 cells treated with lean inguinal secretome compared with DMEM control medium or obese inguinal secretome, quantified by mean cellular RhoNox-1 fluorescence intensity (n = 2). (K) Immunob- lot analysis of 4-HNE in E0771 tumor lysates shows increased 4-HNE abundance in tumors from lean versus obese mice (n = 3 mice). (L to N) (L) Schematic of experimental design. (M) Daily injection of the ferroptosis inhibitor liproxstatin-1 accelerates mammary E0771 tumor growth in lean mice (n = 8) compared with vehicle control (n = 6), with no effect observed in obese mice with (n = 10) or without (n = 10) liproxstatin-1. (N) End-point tumor weights. [(B) to (G) and (J)] Data are means ± SDs. [(M) and (N)] Data are means ± SEMs. P values calculated using one-way ANOVA followed by Dunnett’s multiple comparison test [(B) to (E)], Šídák’s multiple comparisons test [(F) and (G)], or Tukey’s multiple comparisons test [(J) and (N)] or calculated using two-way ANOVA followed by Šídák’s multiple comparisons test (M). ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 3. Lean adipocytes secrete ferroptosis inducer 9S-HODE. (A) The growth-inhibitory effect of lean inguinal adipocyte secretome is not mediated by a heat-labile protein. Lipid depletion using charcoal dextran or Bio-Beads SM-2 Adsorbents, but not Cleanascite, restores E0771 growth (n = 3). (B) Schematic of lipidomics workflow. (C) Volcano plot of lipid features comparing lean and obese inguinal adipocyte secretomes (mean of three independent secretomes per group). Features enriched in lean versus obese secretome and depleted by charcoal dextran and Bio-Beads SM-2 Adsorbents, but not Cleanascite, are highlighted in red. (D and E) Quantification of candidate oxylipins in lean versus obese inguinal adipocyte secretome and DMEM control (n = 3 independent secretomes) (D) or lean versus obese female mammary white adipose tissue (n = 5) (E). (F) Treatment of E0771 cells with candidate lipids identifies 9S-HODE and 9-OxoODE as sufficient to inhibit E0771 growth (n = 4). (G to I) 9S-HODE and ML-162 induce cell death in E0771 cells (G), MCF7 cells (H), and SkBr3 cells (I) compared with DMEM or 13S-HODE controls (n = 3). (J and K) Representative C11-BODIPY staining in E0771 cells (J) and MCF7 cells (K) showed increased lipid peroxidation after 9S-HODE treatment (20 or 60 μM, respectively) compared with DMEM control. (L) Intracellular Fe2+ levels are increased in E0771 cells treated with 9S-HODE compared with 13S-HODE or DMEM control (n = 3). (M) Relative concentration of 9-HODE in the secretome of lean inguinal adipocytes was reduced after treatment of adipocytes with NDGA or zileuton (n = 3 to 7). (N) The growth-inhibitory effect of lean inguinal adipocyte secretome on E0771 cells is lost when adipocytes were treated with NDGA or zileuton. (O) Daily administration of zileuton accelerates E0771 mammary tumor growth in lean mice (n = 10) compared with vehicle control (n = 5). [(A), (D) to (I), (L), and (N)] Data are means ± SDs. [(M) and (O)] Data are means ± SEMs. P values calculated using two-way ANOVA followed by Dunnett’s multiple comparison test [(A) and (D)] or Šídák’s multiple comparisons test [(E) and (O)] or calculated using one-way ANOVA followed by Dunnett’s multiple comparison test [(F) to (I), (M), and (N)] or Tukey’s multiple comparisons test (L). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 4. Normal mammary epithelial cells are insensitive to lean adipocytes. (A and B) Representative growth kinetics of RL95-2 (A) and growth constant of RL95-2 (endome- trial), A375M (melanoma), and HCT116 (colorectal) cancer cell lines (B) show little to no sensitivity to lean inguinal adipocyte secretome or 9S-HODE at concentrations that strongly suppress breast cancer cell growth (n = 3). (C) Lean inguinal adipocyte secretome does not alter the growth constant of normal mammary epithelial cell lines NMuMG and EpH4 (n = 3). (D) Representative growth curve showing a transient growth arrest in NMuMG cells after treatment with lean inguinal adipocyte secretome. (E and F) Lean inguinal adipocyte secretome increases lipid droplet accumulation. (E) Representative time course of intracellular BODIPY fluorescence intensity. (F) Quantification of BODIPY fluorescence intensity 8 hours after treatment with lean inguinal adipocyte secretome or DMEM control (n = 2 to 3). (G to I) DGAT1/2-mediated lipid droplet accumulation is required for mammary epithelial cell protection from the lean inguinal adipocyte secretome. (G) Representative BODIPY fluorescence intensity over time with or without DGAT inhibition (iDGAT1/2). (H) Representa- tive images at 8 hours posttreatment. Scale bars, 100 μm. (I) Representative growth curves showing increased sensitivity of NMuMG cells to lean inguinal adipocyte secretome with DGAT1/2 inhibition. (J to L) E0771 cells stably expressing DGAT2 are resistant to the lean inguinal adipocyte secretome. (J) Representative growth curves. (K) Quantification of growth rate constants (n = 3 to 4). (L) Representative images showing increased lipid droplet accumulation labeled with BODIPY in DGAT2-GFP–expressing E0771 cells. Scale bars, 20 μm. DAPI, 4′,6-diamidino-2-phenylindole. [(A), (D), (E), (G), (I), and (J)] Data are means ± SDs of technical replicates. [(B), (C), (F), and (K)] Data are means ± SDs of biological replicates. P values calculated using two-way ANOVA followed by Dunnett’s multiple comparison test (B) or followed by Šídák’s multiple comparisons test (K) or calculated using one- way ANOVA followed by Dunnett’s multiple comparison test (C) or unpaired t test (F). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 5. 9S-HODE suppresses breast cancer growth. (A to H) Intratumoral administration of 9S-HODE suppresses breast cancer tumor growth. Orthotopic mammary tumors were treated three times per week with intratumoral injections of 9S-HODE or vehicle. Mice were randomized when tumors reached 80 to 150 mm3; lean mice were left untreated. E0771 mammary tumor growth (A) and end-point tumor weights (B) (n = 14 obese+vehicle, n = 14 obese+ 9S-HODE, n = 10 lean). Py230 mammary tumor growth (C) and end-point tumor weights (D) (n = 14 obese+vehicle, n = 13 obese+ 9S-HODE, n = 10 lean). MDA-MB-468 mammary tumor growth (E) (“#” indicates obese compared with lean, ****P < 0.0001; “@” indicates obese compared with obese 9S-HODE, ****P < 0.0001) and end-point tumor weights (F) (n = 11 obese+vehicle, n = 9 obese+ 9S-HODE, n = 8 lean). MCF7 mammary tumor growth (G) and end-point tumor weights (H) (n = 19 obese+vehicle, n = 15 obese+ 9S-HODE, n = 8 lean). (I and J) Human breast white adipose tissue from lean female donors (BMI < 25) have increased levels of 9-HODE compared with obese female donors (BMI > 30). (I) Abundance of select oxylipins in donor breast tissue, with left and right breast oxylipin levels averaged per donor. (J) 9-HODE concentrations as a function of donor BMI shows a significant inverse correlation between BMI and 9-HODE levels [coefficient of determination (R2) = 0.5708, P = 0.0011]; each data point is average of left and right breast 9-HODE levels per donor. (K) Breast cancer patient–derived xenograft organoids are sensitive to 9S-HODE treatment, quantified by Celltiter-Glo 3D for three TNBC patient–derived organoid lines. ATP, adenosine 5′-triphosphate. [(I) and (K)] Data are means ± SDs. [(A) to (H)] Data are means ± SEMs. P values calculated using one-way ANOVA followed by Dunnett’s multiple comparison test [(A), (C), (E), and (G)] or calculated using two-way ANOVA followed by Tukey’s multiple comparison test [(B), (D), (F), and (H)], followed by Šídák’s multiple comparisons test (I), or followed by Dunnett’s multiple comparisons test (K). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.