Rictor/mTORC2 deficiency enhances keratinocyte stress tolerance via mitohormesis

Abstract

How metabolic pathways required for epidermal tissue growth and remodeling influence the ability of keratinocytes to survive stressful conditions is still largely unknown. The mechanistic target of rapamycin complex 2 (mTORC2) regulates growth and metabolism of several tissues, but its functions in epidermal cells are poorly defined. Rictor is an adaptor protein essential for mTORC2 activity. To explore the roles of mTORC2 in the epidermis, we have conditionally deleted rictor in mice via K14-Cre-mediated homologous recombination and found that its deficiency causes moderate tissue hypoplasia, reduced keratinocyte proliferation and attenuated hyperplastic response to TPA. Noteworthy, rictor-deficient keratinocytes displayed increased lifespan, protection from senescence, and enhanced tolerance to cellular stressors such as growth factors deprivation, epirubicin and X-ray in vitro and radioresistance in vivo. Rictor-deficient keratinocytes exhibited changes in global gene expression profiles consistent with metabolic alterations and enhanced stress tolerance, a shift in cell catabolic processes from glycids and lipids to glutamine consumption and increased production of mitochondrial reactive oxygen species (ROS). Mechanistically, the resiliency of rictor-deficient epidermal cells relies on these ROS increases, indicating stress resistance via mitohormesis. Thus, our findings reveal a new link between metabolic changes and stress adaptation of keratinocytes centered on mTORC2 activity, with potential implications in skin aging and therapeutic resistance of epithelial tumors.

Figure 1

Loss of rictor/mTORC2 in the epidermis results in tissue hypoplasia and impaired TPA response. (a) Diagram of the breeding strategy used to obtain K14-cre-mediated homozygous deletion of rictor (E-RiKO). WT: wild-type mouse, CT: WT mouse with exon 3 flanked by two LoxP sites. White tile: hexon; Black arrowhead: LoxP site. (b and c) Representative western blotting analysis with the indicated antibodies of (b) epidermal extracts of CT and E-RiKO newborn (P3) littermates (n=3); (c) cell extracts of CT and E-RiKO MPKs subjected to starvation/stimulation treatment for the indicated times. (d and e) Skins of E-RiKO and CT littermates (P3): (d) representative H&E staining (left panels) and histograms of the epidermal thickness; Epi: epidermis, Der: dermis; scale bar: 30 μm; (e) PCNA staining (left panels) and histograms of the percentage of PCNA⁺ epidermal area. Arrows: brightest PCNA⁺ cells in epidermal basal layers; scale bar: 50 μm. (f) Representative H&E images of skin section obtained from E-RiKO and CT littermates (7-week old) on 48 h of TPA or vehicle treatment, (n=3); scale bar: 30 μm. Histograms represent epidermal thickness. (g) Representative confocal images of IF performed on skin sections, obtained from mice treated as in f and injected with BrdU for the final hour of the experiment, with anti-BrdU antibody (red) and counterstained with anti-LaminA antibody (green). Arrowheads: BrdU⁺ cells; scale bar: 50 μm. The percentage of BrdU⁺ cells/area (μm²) was quantified (right histograms). Quantification of epidermal thickness (mean±SEM) for three mice/genotype; the standard error and the significance of the differences between groups were determined with linear regression using GEE. ***P<0.0005. PCNA and BrdU histograms represent the mean±S.D. of 30 fields/genotype (n=3). ***P<0.0005

Figure 2

Rictor-deficient keratinocytes display reduced proliferation and delayed senescence in vitro. (a) Quantification of MPKs derived from at least 10 skins of CT and E-RiKO newborn mice (P3). Histograms represent mean±S.D. ***P<0.0005. E-RiKO and CT MPKs were isolated from littermates P3 (n=5) and analyzed as follows. (b) Growth curve: MPKs (both adherent and suspended/differentiated cells) were detached and counted at the indicated times upon plating. Data are mean±S.D. obtained from triplicate samples/genotype of three independent experiments. CompareGrowthCurve function was used (see the ‘Materials and Methods' section); P=0.029. (c) Confluent MPKs, were subjected to a single BrdU pulse (3 h) under basal conditions. Histograms represent mean±S.D. of the percentage of BrdU⁺ cells determined from at least 200 cells/genotype of three independent experiments; ***P<0.0005. (d) Representative phase contrast images of MPKs cultured in LCM for the indicated times. Note that E-RiKO cells show at 3 days reduced density compared with CT counterparts; at 5–10 days, they display a comparable confluency; while at 25 days, they maintained an undifferentiated proliferative morphology; whereas CT cells show a flattened, senescent appearance; scale bar: 100 μm. (e) Western blotting analysis for the indicated markers of MPKs extracts of 25-day cultures. (f) Representative images of 25-day-old MPKs stained for SA-βGal (left panel). The percentage of SA-βGal⁺ cells is quantified in the histograms (right) as mean±S.D. of at least 200 cells/genotype of three independent experiments; scale bar: 50 μm; *P<0.05

Figure 3

Rictor-deficient keratinocytes display resistance to death-inducing stimuli both in vitro and in vivo. (a–h) E-RiKO and CT MPKs (P3) were analyzed at confluency (5–6 days in culture) as follows. (a) Representative western blotting analysis of extracts obtained from E-RiKO and CT MPKs on starvation (st) for the apoptosis marker cleaved caspase-3 (Cl-casp3) and Rictor. Loading control: Tubulin. (b) Representative AnnexinV/DAPI flow cytometry plots of E-RiKO and CT MPKs kept in complete LCM (−) or starvation (St) for 72 h. Histograms are mean±S.D. of dead cells expressed as variation of the fraction of dead cells in response to starvation (72 h) (see the ‘Materials and Methods' section) *P<0.05. (c and d) MPKs grown in LCM for 24 h were kept untreated (−) or treated with epirubicin (10 μM) for 18 h: (c) Representative Western blotting analysis for the indicated proteins. (d) Representative AnnexinV/DAPI flow cytometry plots of E-RiKO and CT MPKs kept untreated (−) or treated with Epirubicin for 15 h. Histograms are mean±S.D. of dead cells determined as in b; *P<0.05. (e) MPKs maintained in LCM or starved (St) for 18 h were treated with a single BrdU pulse. BrdU⁺ cells were determined out of 200 cells (DAPI⁺)/genotype in three independent experiments. Histograms represent mean±S.D. of BrdU⁺ cells. Note that on St 18 h, cells display comparable BrdU uptake; *P<0.05; ***P<0.0005. (f) MPKs were starved as in e and then treated with the indicated doses of epirubicin for 4 h. Cell extracts were analyzed by western blotting for Rictor, Cl-casp3 and Tubulin, loading control. Ratio between Cl-casp3 and Tubulin is reported. (g) Histograms represent mean colony number±S.D. from duplicate plates determined in untreated (−) and treated (X-ray) MPKs colony-forming efficiency assay, relative to the value obtained with CT (−) cells set as unitary. Data are representative of at least two independent experiments; ***P<0.0005, **P<0.005. (h) Representative AnnexinV/DAPI flow cytometry plots of E-RiKO and CT MPKs kept untreated (−) or X-ray treated (60 Gy) analyzed after 96 h. Histograms are mean±S.D. of dead cells determined as in b; *P<0.05. (i) Representative confocal images for TUNEL (red) and DAPI (blue) stainings of skin sections obtained from E-RiKO and CT littermates (P3) subjected to a full-body single dose of X-ray radiation (8 Gy) and sacrificed 24 h later; scale bar: 100 μm; n=5 mice/genotype. (j) Quantification of TUNEL staining of X-ray-treated skins with 4 Gy and 8 Gy as in i. Histograms represent mean±S.D. of TUNEL⁺ cells/area (10³ μm²) of at least 30 hair follicles of four mice/genotype. Note that the TUNEL⁺ area was limited to the hair follicle matrix, with similar cellular density and sensitive to radiations; ***P<0.0005

Figure 4

RNA-Seq analysis of rictor-deficient keratinocytes suggests a link between metabolic changes and stress adaptation. E-RiKO and CT MPKs were isolated from newborn littermates (n=6/genotype), grown in LCM and analyzed by RNA-Seq analysis under basal conditions (24 h LCM) and upon X-ray exposure (8 Gy) for 1 h and 24 h. (a) Upregulated and downregulated genes detected in E-RiKO versus CT cells under basal conditions. (b) Selected gene ontology (GO) categories enriched in differentially expressed genes as in a. The number of genes belonging to each GO category is indicated. Rictor deficiency was coupled with downregulation of genes involved in lipid metabolism (P=1.08E−16), keratinocyte differentiation (P=1.68E−9), oxidation–reduction process (P=1.89E−6), lipid catabolic process (P=2.41E−6), lipid biosynthetic process (P=2.7E−6) and upregulation of genes involved in cell motility (P=2.62E−18), signal transduction (P=3.15E−15), inflammatory response (P=1.66E−12), response to stress (P=3.11E−7) and defense response (P=2.28E−6) as in b. (c) Upregulated and downregulated genes of X-ray treated MPKs of the indicated genotype relative to basal conditions. (d) Differentially regulated genes overlap between CT MPKs X-ray treated for 1 h and E-RiKO MPKs under basal conditions (27.5% gene overlap; P=5.44E−57, Fisher's exact test). (e) Selected GO categories enriched in overlapping differentially regulated genes (DEGs) subgroup as in d

Figure 5

Rictor-deficient keratinocytes display metabolic reprogramming. E-RiKO and CT MPKs, isolated and cultured in LCM (n=5 littermates P3/genotype), were analyzed at confluency in LCM 24 h as follows: (a) total cellular ROS (nmol/mg prot); (b) mitochondrial ROS (nmol/mg mit prot); (c) mitochondrial/total ROS (%) determined by compairing mitochondrial and total ROS, each normalized for total protein content; (d) aldose reductase (nmol NADP⁺/min/mg prot); (e) NADPH oxidase (RLU/mg prot); (f) electron transport chain (nmol red cit c/min/mg mit prot); (g) ATP (nmol/mg mit prot); (h) lipid β-oxidation (pmol/h/mg prot); (i) triglycerides (nmol/mg prot); (j) glucose uptake (pmol glucose/mg prot); (k) hexokinase (nmol NADH/min/mg prot); (l) phosphofructose kinase 1 (nmol NAD⁺/min/mg prot); (m) glutaminase (μmol NADH/min/mg prot); (n) glutamic dehydrogenase (μmol NADH/min/mg prot); (o) TCA cycle with [¹⁴C] glutamine (pmol CO₂/h/mg mit prot); (p) TCA cycle with [¹⁴C] glutamate (pmol CO₂/h/mg mit prot); (q) intracellular [¹⁴C]glutamine (μmol/mg prot); (r) intracellular [¹⁴C]glutamate (μmol/mg prot); (s) glutamate/glutamate ratio; (t) total SOD (μmol/min/mg prot); (u) catalase (nmol/min/mg prot); (v) mitochondrial SOD (μmol/min/mg prot). All histograms represent mean±S.D. of the indicated metabolic parameters determined in at least three independent experiments; *P<0.05; **P<0.005; ***P<0.0005

Figure 6

ROS scavenging and glutaminase inhibition sensitize rictor-deficient keratinocytes to epirubicin-induced cell death. E-RiKO and CT MPKs were isolated (n=5/genotype), grown in LCM and analyzed at confluency for the following. (a) Total cellular ROS was measured on LCM (−) or NAC 10 mM for 48 h. Histograms represent mean±S.D. of ROS levels measured in at least three independent experiments; ***P<0.0005; *P<0.05. (b) Histograms represent mean±S.D. of mitochondrial ROS determined in cells treated as in a. (c) Representative western blotting analysis for the indicated antibodies of cell extracts obtained from MPKs maintained as in a and subsequently treated with DMSO (−) or epirubicin 10 μM for 10 h. Ratio between Cl-casp3 and keratin5 (K5) is reported. (d) Representative AnnexinV/DAPI flow cytometry plots of E-RiKO and CT MPKs maintained untreated (−) or pre-treated with NAC and subsequently treated with epirubicin 10 μM for 15 h. Histograms are mean±S.D. of dead cells determined as described in the ‘Materials and Methods' section; *P<0.05. (e) Histograms represent the mean±S.D. of the glutaminase activity (μmol NADH/min/mg prot) evaluated in CT and E-RiKO MPKs treated with BPTES (10 μM, 2 h) in at least three independent experiments; **P<0.005, *P<0.05. (f) Histograms represent mean±S.D. of total ROS levels (nmol/mg prot) measured in cells treated as in e in at least three independent experiments; *P<0.05. (g) Representative western blotting analysis with the indicated antibodies of cell extracts derived from E-RiKO and CT cells pre-treated with BPTES at the indicated doses for 48 h and exposed to epirubicin in the presence of BPTES for 15 h. Ratio between Cl-casp3 and keratin5 (K5) is reported. (h) Representative AnnexinV/DAPI flow cytometry plots of E-RiKO and CT MPKs maintained untreated (−) or pre-treated with BPTES and subsequently treated with epirubicin 10 μM for 15 h. Histograms are mean±S.D. of dead cells as in d; **P<0.005, *P<0.05

Figure 7

Rictor-deficient epidermis displays a metabolic rewiring and ROS-dependent lower sensitivity to stress. (a) Metabolic analysis of CT and E-RiKO skins obtained from at least three mice/genotype. Histograms represent mean±S.D. of the following metabolic parameters: total cellular ROS (nmol/mg prot); mitochondrial ROS (nmol/mg mit prot); electron transport chain (nmol red cit c/min/mg mit prot); ATP (nmol/mg mit prot); lipid β-oxidation (pmol/h/mg prot); hexokinase (nmol NADH/min/mg prot); glutaminase (μmol NADH/min/mg prot); glutamic dehydrogenase (μmol NADH/min/mg prot); **P<0.005, *P<0.05. (b) RNA was extracted from CT and E-RiKO epidermis of at least six mice/genotype and RT-PCR analysis for the following genes was performed: Gadd45-α; Gadd45-β; Gadd45-γ; Glul; Il1-α; Acsl1; Faah; Fabp4. Data are represented as mean±S.D. of the fold change relative to unitary value assigned to CT; **P<0.005, *P<0.05 (c) Representative confocal images of IF for TUNEL (red) and DAPI (blue) stainings on E-RiKO and CT skin sections pre-treated with NAC (or vehicle) for 48 h (i.p. injection), subjected to a single dose of X-ray radiation (8 Gy) and collected 24 h later; scale bar: 100 μm; n=10 mice/genotype. (d) Histograms represent mean±S.D. of ROS levels (nmol/mg prot) in CT and E-RiKO skins treated with NAC (or vehicle) for 48 h, of three mice/genotype; *P<0.05, **P<0.005. (e) Histograms represent mean±S.D. of TUNEL⁺ cells/area (10³ μm²) of at least 30 hair follicles of 10 mice/genotype treated as in c; ***P<0.0005.