It is evident that the insulin/insulin-like growth factor-1 signalling (IIS) pathway plays a conserved role in lifespan determination (Gems & Partridge, 2013). Precisely how the IIS elicits its effects is unclear, although stress resistance in longevity may be important. Several long-lived IIS mutant worms and flies are stress resistant compared with wild-type (WT) controls [see (Miller, 2009)], and stress resistance correlates with longevity across species of mammals (Kapahi et al., 1999; Harper et al., 2007) and birds (Harper et al., 2011). An increased survival following paraquat or diquat injection has been reported in long-lived growth hormone (GH)-deficient Ames mice (Bokov et al., 2009), which have a secondary defect in IIS, and in female mice lacking a single copy of the IGF-1 receptor (Igf1R+/−) (Holzenberger et al., 2003; Bokov et al., 2011). A significant body of work has described broad-spectrum cellular resistance to various oxidative and nonoxidative stressors in GH dwarfs (Murakami et al., 2003; Salmon et al., 2005; Miller, 2009). While the precise mechanism underlying cellular stress resistance in GH dwarfs is unclear, it is associated with increased basal and arsenite-induced Nrf2 levels (Leiser & Miller, 2010), increased autophagy and reduced mTOR activity (Wang & Miller, 2012).
We previously reported that Irs1−/− mice were long-lived and had improved late-life health (Selman et al., 2008, 2011). Consequently, we predicted that Irs1−/− mice would demonstrate increased cellular stress resistance. To address this, we isolated and cultured dermal fibroblasts from young Irs1−/− and WT mice using published protocols (Murakami et al., 2003; Salmon et al., 2005; Leiser & Miller, 2010). Cells were exposed to hydrogen peroxide (H2O2), paraquat, cadmium chloride, arsenite and the DNA-alkylating agent methyl methanesulfonate (MMS) under both standard (21% oxygen) and physiological (3% oxygen) conditions in parallel experiments. Cytotoxicity was determined by calculating LD50 using probit analysis (Leiser & Miller, 2010). In contrast to our prediction, fibroblasts from Irs1−/− mice were not more resistant to cytotoxic stress at 21% oxygen (Fig. 1A–E) or 3% oxygen (Fig. S1A–E). Indeed, Irs1−/− cells were significantly more sensitive when treated with arsenite at 21% O2 (Fig. 1D). We also exposed primary myoblasts cultured at 3% oxygen (21% induced differentiation to myotubes) to H2O2 and paraquat and observed no genotype effect on cellular resistance (Fig. 2A,B). Additionally, no differences in basal Nrf2 or Keap1 protein levels (Fig. 2C,D), or in basal levels of Nrf1, Nrf2 and various antioxidant response element-dependent mRNA transcripts (Fig. 2E) were detected between WT and Irs1−/− fibroblasts. Cell size, proliferation rates and cell cycle distribution in cultured fibroblasts were similarly unaffected by genotype (Fig. S2A–C; P > 0.05 in all cases).
While both male and female Irs1−/− mice are long-lived (Selman et al., 2008, 2011), we did not observe a cellular stress resistance phenotype as predicted and as reported in long-lived GH-deficient mice (Murakami et al., 2003; Salmon et al., 2005; Leiser & Miller, 2010). However, a similar absence of cellular stress resistance has been reported in long-lived methionine-restricted and caloric-restricted (CR) mice (Harper et al., 2006). A potential explanation given for the different cellular stress phenotypes in GH dwarfs, methionine-restricted and CR mice was that epigenetic changes leading to cellular stress resistance may be manifested during development, rather than postnatally (Harper et al., 2006). However, we can discount this explanation in our study as the deletion event in Irs1−/− mice occurs during development. Long-lived homozygous Chico flies, which lack the single Drosophila IRS protein (Clancy et al., 2001), and long-lived brain-specific Igf1R+/− mice (Kappeler et al., 2008) are also not resistant to oxidative stress. It is currently unclear why different long-lived mouse models show such differences in cellular stress resistance. Irs1−/− mice, unlike GH dwarfs, appear to have relatively normal somatotropic function and GH levels (Selman et al., 2008) and this may help explain the absence of a stress-resistant phenotype. In Ames mice, treatment with GH shortened lifespan, decreased cellular stress resistance (Panici et al., 2010) and attenuated components of the cellular detoxification pathway, including glutathione transferases (Rojanathammanee et al., 2013). As discussed elsewhere (Harper et al., 2006), it may be that the precise cell type demonstrating cellular stress resistance, and hence critical to longevity, may vary across different long-lived mouse models. For example, while fibroblasts from methionine-restricted and CR mice are not stress resistant, both models, unlike Snell and GHR–KO mice, are resistant to acetaminophen (Harper et al., 2006). In conclusion, our findings demonstrate that fibroblasts and myoblasts from Irs1−/− mice are no more stress resistant than those derived from WT mice, suggesting that cellular stress resistance does not underlie the extended healthy lifespan of Irs1−/− mice.