GPX4 and vitamin E cooperatively protect hematopoietic stem and progenitor cells from lipid peroxidation and ferroptosis

Abstract

Ferroptosis, a newly defined mode of regulated cell death caused by unbalanced lipid redox metabolism, is implicated in various tissue injuries and tumorigenesis. However, the role of ferroptosis in stem cells has not yet been investigated. Glutathione peroxidase 4 (GPX4) is a critical suppressor of lipid peroxidation and ferroptosis. Here, we study the function of GPX4 and ferroptosis in hematopoietic stem and progenitor cells (HSPCs) in mice with Gpx4 deficiency in the hematopoietic system. We find that Gpx4 deletion solely in the hematopoietic system has no significant effect on the number and function of HSPCs in mice. Notably, hematopoietic stem cells (HSCs) and hematopoietic progenitor cells lacking Gpx4 accumulated lipid peroxidation and underwent ferroptosis in vitro. α-Tocopherol, the main component of vitamin E, was shown to rescue the Gpx4-deficient HSPCs from ferroptosis in vitro. When Gpx4 knockout mice were fed a vitamin E-depleted diet, a reduced number of HSPCs and impaired function of HSCs were found. Furthermore, increased levels of lipid peroxidation and cell death indicated that HSPCs undergo ferroptosis. Collectively, we demonstrate that GPX4 and vitamin E cooperatively maintain lipid redox balance and prevent ferroptosis in HSPCs.

Fig. 1. GPX4 deficiency induces HSPC ferroptosis in vitro.

a LT-HSC single cells were sorted and cultured in a medium with or without RSL3 for 14 days. Colony numbers were counted (n = 4 mice). Scale bars = 0.5 mm. LT-HSCs derived from the Gpx4^flox/flox mice and Gpx4^flox/flox Vav-Cre mice (b) or the pIpC-treated Gpx4^flox/flox Mx-Cre mice (c) were tested with single-cell colony-forming assay as in (a) (n = 2 Vav-Cre mice or 4 Mx-Cre mice). d The viability of LSK cells isolated from the wild-type mice after 48 h of culture with the indicated drugs (n = 3 mice). e LSK cells isolated from the wild-type mice were cultured with the indicated drugs for 24 h, and lipid ROS were measured with C11-BODIPY and detected by flow cytometry (n = 3 mice). f The viability of LSK cells isolated from the Gpx4^flox/flox mice or the Gpx4^flox/flox Vav-Cre mice after 48 h of culture with the indicated drugs (n = 3 mice). g LSK cells isolated from the Gpx4^flox/flox mice or the Gpx4^flox/flox Vav-Cre mice were cultured with the indicated drugs for 24 h, and lipid ROS were measured with C11-BODIPY and detected by flow cytometry (n = 3 mice). Data are the mean ± SD. (ns not significant, *P < 0.05, **P < 0.01, ***P < 0.001 vs. the RSL3 group in (d, e), vs. the Gpx4^−/− group in (f, g).

Fig. 2. Gpx4 deletion does not affect hematopoietic system homeostasis in the Gpx4^flox/flox Vav-Cre mice.

a Gpx4 deletion in the blood cells of the Gpx4^flox/flox Vav-Cre was verified by Western blots. b The BM cell numbers were counted. c, d The numbers of HSPCs were measured by flow cytometry. Left panels: representative flow cytometric plots. Right panels: statistical data. e The relative MFI of ROS in LK cells, LSK cells, and LT-HSCs (CD34⁻Flt3⁻LSK cells). f The relative MFI of lipid peroxidation in BM cells was detected by C11-BODIPY. N = at least five mice in each group. Data are the mean ± SD. (ns not significant, **P < 0.01).

Fig. 3. Gpx4 deficiency does not affect HSC function in vivo.

a The number of HSPCs, b the relative MFI of ROS in LSK cells and LT-HSCs, and c the cell cycle of LSK cells and LT-HSCs from the pIpC-treated Gpx4^flox/flox Mx-Cre mice 7 days after 5-FU treatment were measured. b The experimental schematic for serial competitive transplantation with LT-HSCs. e The chimerism of peripheral blood cells, f the chimerism of HSPCs at the 4th month after primary transplantation, g the chimerism of HSPCs at the 6th months after secondary transplantation in the recipient mice receiving LT-HSCs from the Gpx4^flox/flox Vav-Cre mice. h The chimerism of peripheral blood cells, i the chimerism of HSPCs at the 7th month after primary transplantation, j the chimerism of HSPCs at the 4th month after secondary transplantation receiving LT-HSCs from the Gpx4^flox/flox Mx-Cre mice. N = at least three recipient mice in each group in (e−j). Data are the mean ± SD.

Fig. 4. α-Toc rescues HSPCs from ferroptosis ex vivo.

a Schematic illustration of the ferroptosis pathway and the targets of different drugs. Ferroptosis was triggered by iron-dependent accumulation of lipid peroxidation. GPX4 makes use of GSH to reduce lipid peroxidation and inhibit ferroptosis. Ferroptosis can be induced by the GPX4 inhibitor RSL3 and inhibited by the iron chelator DFO or lipophilic antioxidants such as Fer-1 and α-Toc. N-Acetyl-L-cysteine (NAC) is the precursor of cysteine, which promotes GSH synthesis. LSK cells (b, c) or GMPs (d, e) from the Gpx4^flox/flox Vav-Cre mice were cultured with NAC or α-Toc. The viability and lipid ROS levels of these cells were measured (n = 3 mice). f LT-HSCs derived from the Gpx4^flox/flox Vav-Cre mice were tested with single-cell colony-forming assay in a medium containing NAC or α-Toc (n = 3 mice). # indicates no colony was formed. Data are the mean ± SD. (ns not significant, ***P < 0.001).

Fig. 5. Deficiency of both GPX4 and vitamin E impairs hematopoietic system homeostasis.

a The body weight, b spleen weight, and c total BM cell counts of the Gpx4^flox/flox Vav-Cre mice were measured after 3 weeks of a V_E-depleted diet. d Left panel: representative flow cytometric plots of the T cell, B cell, and myeloid cell groups in peripheral blood. Right panel: statistical data. e The number of HSPCs in the Gpx4^flox/flox Vav-Cre mice after 3 weeks of a V_E-depleted diet. N = at least four mice in each group. Data are the mean ± SD. (*P < 0.05, **P < 0.01, ***P < 0.001).

Fig. 6. Deficiency of both GPX4 and vitamin E results in HSC ferroptosis in vivo.

a The cell cycle of LK cells, LSK cells, and LT-HSCs in the Gpx4^flox/flox Vav-Cre mice after 3 weeks of a V_E-depleted diet. b The HSPCs in the Gpx4^flox/flox Vav-Cre mice after 3 weeks of a V_E-depleted diet were determined to be DAPI negative. c Left panel: representative flow cytometric plots of lipid ROS in c-kit⁺ cells of the Gpx4^flox/flox Vav-Cre mice after 3 weeks of a V_E-depleted diet. Right panel: statistical data of the relative MFI of lipid ROS. d Left panel: representative flow cytometric plots of ROS in LK cells, LSK cells, and LT-HSCs of the Gpx4^flox/flox Vav-Cre mice after 3 weeks of a V_E-depleted diet. Right panel: statistical data of the relative MFI of ROS. N = at least three mice in each group. Data are the mean ± SD. (ns not significant, *P < 0.05, **P < 0.01, ***P < 0.001).