Identification of genes supporting cold resistance of mammalian cells: lessons from a hibernator

Abstract

Susceptibility of human cells to cold stress restricts the use of therapeutic hypothermia and long-term preservation of organs at low temperatures. In contrast, cells of mammalian hibernators possess remarkable cold resistance, but little is known about the molecular mechanisms underlying this phenomenon. In this study, we conducted a gain-of-function screening of genes that confer cold resistance to cold-vulnerable human cells using a cDNA library constructed from the Syrian hamster, a mammalian hibernator, and identified Gpx4 as a potent suppressor of cold-induced cell death. Additionally, genetic deletion of or pharmacological inhibition of Gpx4 revealed that Gpx4 is necessary for suppressing lipid peroxidation specifically under cold in hamster cell lines. Genetic disruption of other ferroptosis-suppressing pathways, namely biopterin synthesis and mitochondrial or plasma membrane CoQ reduction pathways, also accelerated cold-induced cell death under Gpx4 dysfunction. Collectively, ferroptosis-suppressing pathways protect the cells of a mammalian hibernator from cold-induced cell death and the augmentation of these pathways renders cold resistance to cells of non-hibernators, including humans.

Fig. 1. Gpx4 was identified by hamster cDNA expression screening as a gene rendering cold resistance to a human cancer cell line.

a Cold resistance of hamster HapT1 cells but not of human HT1080 cells. The proportion of dead cells was determined by LDH assay (one-way ANOVA with the Tukey’s multiple comparison test, p < 0.05). b Schematic illustration of hamster cDNA expression screening for genes that render the resistance to cold-rewarming culture to human HT1080 cells. c Venn diagram of the genes inserted in the genome of survived cells under cold-rewarming culture in four screening experiments. d The proportion of dead cells in HT1080 cells after 24 h of cold (4 °C) culture when each indicated gene was overexpressed (N = 3 wells, One-way ANOVA with the Dunnett’s multiple comparison test, ***p < 0.001).

Fig. 2. Gpx4 is necessary for preventing lipid peroxidation and cell death induced by long-term cold exposure in hamster HapT1 cells.

a Immunoblot of Gpx4 and β-Actin proteins in parental (wild type; WT) and clonal HapT1 cells (#19, #21, #35) in which Gpx4 gene is disrupted. b The proportion of dead cells in WT or Gpx4-KO HapT1 cells during cold (4 °C) culture. c The proportion of dead cells in HT1080 and parental or Gpx4-KO HapT1 cells after 5 days of cold (4 °C) culture in the absence (non-treated; NT) or presence of 100 µM DFO, 1 µM ferrostatin-1 (Fer-1), 10 µM idebenone (Ide), 10 µM mitoquinol (MitoQ), 2 µM BAPTA-AM (One-way ANOVA with the Dunnett’s multiple comparison test compared to NT within each cell line, ***p < 0.001). d Oxidized lipid detection by BODIPY C11 ratio imaging in HT1080 and HapT1 cells (parental or Gpx4-KO) during cold (4 °C) culture. Scale bar = 200 µm. e Time-course changes of oxidized lipid level determined with BODIPY C11 ratio imaging in the presence or absence of 2 µM RSL3 or 1 µM Fer-1 N = 3. Plot is represented as mean ± s.d. f Detection of lipid peroxidation by LC–MS/MS analysis. The amount of mono-(left), di-(middle), tri-(right) oxidized phosphatidylethanolamine (PE38:4) in WT or Gpx4-KO HapT1 cells are shown (One-way ANOVA with the Tukey’s multiple comparison test, p < 0.05). g The proportion of dead cells in HapT1 after 5 or 7 days of cold (4 °C) culture in the absence or presence of 2 µM RSL3. Each bar corresponds to the condition shown in the upper panel wherein RSL3 was added only during the period indicated by black during the culture (One-way ANOVA with the Dunnett’s multiple comparison test, *p < 0.05, **p < 0.01).

Fig. 3. Overexpression of either human or hamster Gpx4 prevents cold-induced cell death.

a Immunoblot of Gpx4 and β-Actin proteins in HT1080 and HapT1 (WT and Gpx4-KO clone #21), in which GFP, human (hs)Gpx4, or hamster (ma)Gpx4 were exogenously expressed by lentivirus vectors at 37 °C. The proportion of dead cells after 5 days of cold (4 °C) culture in HT1080 (b) and HapT1(WT or Gpx4-KO) (c), in which lentivirus vectors expressing GFP or hs/maGpx4 were infected (One-way ANOVA with the Tukey’s multiple comparison test, p < 0.05).

Fig. 4. Biopterin and CoQ reduction pathways are required for short-term cold resistance under Gpx4 dysfunction in hamster cells.

a Assessment of knock-out efficiency of ferroptosis-suppressors in HapT1 cells. Immunoblots of each protein in non-treated HapT1 (WT) or HapT1 cell populations, in which lentivirus vectors expressing SpCas9 and sgRNA (non-targeting control; sgNT, or sgRNA targeting each of three genes) were infected. b–d The proportion of dead cells during cold culture (4 °C) in the each HapT1 cell population infected with lentivirus vectors expressing SpCas9 and sgNT or sgRNA for FSP1 (b) or Dhodh (c) or Gch1 (d) in the presence or absence of 2 µM RSL3 (One-way ANOVA with the Tukey’s multiple comparison test, p < 0.05). e Biopterin synthesis and related pathways. f (upper panel) Experimental time course and (lower panel) the proportion of dead cells during cold culture in WT or Gch1 KO HapT1 cells, which were infected with lentivirus vectors expressing SpCas9 and sgNT or sgRNA for Gch1, in the presence of 6 µM ML210 with or without 100 µM BH₂ and 4 µM methotrexate (MTX). g Immunoblot of parental HT1080 and HT1080 cell populations in which FSP1 or Gch1 were exogenously overexpressed at 37 °C by lentivirus vector infection. h The proportion of dead cells during cold culture in parental HT1080 and HT1080 cell populations in which FSP1 or Gch1 were exogenously overexpressed (One-way ANOVA with the Dunnett’s multiple comparison test, ***p < 0.001).

Fig. 5. Gpx4 OE and RTAs additively augment cold resistance of mouse primary hepatocytes.

a Schematic illustration of experimental time course. b Overexpression of Gpx4 protein was confirmed by immunoblot of hepatocytes infected with the AAV vectors expressing GFP or maGpx4. c The proportion of dead cells in hepatocytes infected with the AAV vectors expressing GFP or maGpx4 after 24 h cold culture with DMSO or 200 µM BH₂ and/or 10 µM αTocopherol (αT) (one-way ANOVA with the Dunnett’s multiple comparison test compared to GFP-DMSO, *p < 0.05, **p < 0.01, ***p < 0.001).