AKT-mediated phosphorylation of TSC2 controls stimulus- and tissue-specific mTORC1 signaling and organ growth

Abstract

Mechanistic target of rapamycin (mTOR) complex 1 (mTORC1) integrates diverse intracellular and extracellular growth signals to regulate cell and tissue growth. How the molecular mechanisms regulating mTORC1 signaling established through biochemical and cell biological studies function under physiological states in specific mammalian tissues are unknown. Here, we characterize a genetic mouse model lacking the 5 phosphorylation sites on the tuberous sclerosis complex 2 (TSC2) protein through which the growth factor-stimulated protein kinase AKT can active mTORC1 signaling in cell culture models. These phospho-mutant mice (TSC2-5A) are developmentally normal but exhibit reduced body weight and the weight of specific organs, such as brain and skeletal muscle, associated with cell intrinsic decreases in growth factor-stimulated mTORC1 signaling. The TSC2-5A mouse model demonstrates that TSC2 phosphorylation is a primary mechanism of mTORC1 activation in some, but not all, tissues and provides a genetic tool to facilitate studies on the physiological regulation of mTORC1.

Figure 1:. A genetic mouse model to disconnect AKT signaling from TSC2 regulation

(A) Schematic of proposed mechanisms of mTORC1 regulation. (B) Schematic of the domain structure and phosphorylation sites on TSC2. (C) Schematic of the strategy for generating conditional TSC2-WT and −5A mice, involving transgene knock-ins at the Rosa26 locus and a Tsc2-floxed allele. (D) qPCR quantification of mRNA expression of the mouse and human TSC2 transcripts in Rosa26-LSL-TSC2-WT; Tsc2^fl/fl (WT Flox) or Rosa26-LSL-TSC2-5A; Tsc2^fl/fl (5A Flox) and following expression of CMV-Cre, Rosa26-TSC2-WT; Tsc2^−/− (WT Cre) or Rosa26-TSC2-5A; Tsc2^−/− (5A Cre) 15.5 dpc whole embryos. Data are graphed as mean ± SEM; N = 2 biological replicates per genotype. (E) Immunoblots of total protein (Input) and immunoprecipitations (IP) with TSC1 or IgG control antibodies from protein extracts from Rosa26-LSL-TSC2-WT or −5A; Tsc2^fl/fl (Cre −) and Rosa26-TSC2-WT or −5A; Tsc2^−/− (Cre +) 15.5 dpc whole embryos are shown. (F) Quantification of the TSC2/TSC1 ratio observed in (E), graphed as mean fold change from Cre- control ± SEM; N = 2 biological replicates per genotype. (G) Immunoblots of protein extracts from 15.5 dpc whole embryos obtained from 3 independent litters from heterozygous Rosa26-TSC2^WT/5A; Tsc2^−/− crosses are shown. (H) Quantification of pS6/S6 ratios from the data in (G), graphed as mean ± SEM; N = 7 TSC2-WT/WT, 11 TSC2-WT/5A, and 6 TSC2-5A/5A. Statistical analysis in (H) was performed using the Kruskal-Wallis test: ns (not significant), p > 0.5.

Figure 2:. TSC2-5A cells have decreased growth factor-stimulated mTORC1 signaling

(A) Immunoblots of lysates from primary TSC2-WT and TSC2-5A MEFs cultured for 16 h in full serum media (10% FBS). N = 3 biological replicates per genotype. (B) Quantification of pS6K1/S6K1 ratio from (A) expressed as percentage of TSC2-WT MEFs and graphed as mean ± SEM; N = 3 biological replicates per genotype. (C) Representative cell size (diameter) distribution between TSC2-WT and TSC2-5A MEFs cultured as in (A). (D) Mean volume of TSC2-WT and TSC2-5A MEFs cultured as in (A). Data are graphed as mean ± SEM ; N = 3 biological replicates per genotype (E) Proliferation curve of the TSC2-WT and TSC2-5A MEFs cultured in full serum media for 3 days. Data are expressed as fold change in cell number relative to day 0 and graphed as mean ± SEM; N = 3 biological replicates per genotype. (F-H) Representative immunoblots and pS6K1/S6K1 ratio quantifications (below) of lysates from primary TSC2-WT and TSC2-5A MEFs serum-starved and then stimulated with time courses of FBS (10%; F), EGF (100 ng/ml; G), or insulin (100 nM; H). Graphed as mean percentage of serum-starved TSC2-WT MEFs ± SEM; N = 3 biological replicates per genotype. (I) Representative immunoblots and pS6K1/S6K1 ratio quantifications (below) of lysates from primary TSC2-WT and TSC2-5A MEFs serum-starved and then stimulated with the indicated concentrations of insulin. Graphed as mean percentage of serum-starved TSC2-WT MEFs ± SEM; N = 3 biological replicates per genotype. (J) Representative immunoblots and pS6K1/S6K1 ratio quantifications (below) of lysates from primary TSC2-WT and TSC2-5A MEFs treated for 1 h with vehicle (−) or an AMPK activator (compound 991, 100 μM) with or without 30 min pretreatment with an AMPK inhibitor (BAY-3827, 1 μM). Graphed as mean percentage of vehicle-treated TSC2-WT MEFs ± SEM; N = 3 biological replicates per genotype. (K) Representative immunoblots and pS6K1/S6K1 ratio quantifications (below) of lysates from primary TSC2-WT and TSC2-5A MEFs deprived of amino acids for 1 h and then, where indicated (+), refed amino acids for 15 min. Graphed as mean percentage of amino acid deprived TSC2-WT MEFs ± SEM; N = 3 biological replicates per genotype. Statistical analysis: (B, D) by Mann-Whitney test; (E-I) repeated measurement two-way ANOVA with Greenhouse–Geisser correction; and (J, K) ordinary two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 3:. TSC2-5A mice display decreased body and organ-specific weight, including the brain with associated differences in mTORC1 signaling and neuron size

(A) Body weight measured over 12 months in a cohort of TSC2-WT and TSC2-5A littermate males, graphed as mean ± SEM; N = 12 TSC2-WT and 10 TSC2-5A. (B) Body weight measured over 12 months in a cohort of TSC2-WT and TSC2-5A littermate females, graphed as mean ± SEM; N = 10 TSC2-WT and 7 TSC2-5A. (C) Body length of male littermates measured at 11 weeks of age and graphed as mean ± SEM; N = 12 TSC2-WT and 10 TSC2-5A. (D) Body length of female littermates measured at 11 weeks of age and graphed as mean ± SEM; N = 10 TSC2-WT and 7 TSC2-5A. (E) Male organ weights normalized to body weight (BW) measured in 16-week-old TSC2-5A mice, graphed as mean percentage of TSC2-WT littermates ± SEM; N = 6 TSC2-WT and 8 TSC2-5A. (F) Brain weight measurement of male littermates at 9 weeks (N = 7 per genotype), 16 weeks (N = 6 TSC2-WT and 8 TSC2-5A), and 26 months (N = 7 TSC2-WT and 5 TSC2-5A), graphed as mean ± SEM. (G) Ad libitum pS6/S6 ratio measured in brain lysates of 9-week-old TSC2-5A mice (N = 7), graphed as mean percentage of TSC2-WT littermates (N = 7) ± SEM. Quantification of immunoblots in Figure S3D. (H) Soma size (area) of primary hippocampal neurons after 4 and 7 days in vitro, graphed as mean ± SEM for 2 independent experiments each. 4 days: N = 149 TSC2-WT and 191 TSC2-5A; 7 days: N = 103 TSC2-WT and 117 TSC2-5A. (I, J) Representative immunoblots of lysates from primary TSC2-WT and TSC2-5A cortical neurons following 4 h of growth factor and amino acid withdrawal, followed by a 15-min stimulation with BDNF (100 ng/mL; I) or IGF1 (100 ng/mL; J) after a 30-min pretreatment with vehicle (−), MK2206 (2 μM, AKTi), trametinib (5 μM, MEKi), or both, as indicated. Statistical analysis: (A, B) by mixed model effect with Šídák’s multiple comparisons; (C-E, G, H) by Mann-Whitney test; and (F) by ordinary two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 4:. Tissue-specific effects on feeding-induced mTORC1 signaling and organ size in skeletal muscle from TSC2-5A mice.

(A) Quantification of male skeletal muscle (gastrocnemius) and hepatic ad libitum pS6/S6 ratios in lysates from 9-week-old TSC2-5A mice (N = 7), graphed as mean percentage of TSC2-WT littermates (N = 7) ± SEM. Quantification of immunoblots in Figure S4A. (B-D) Immunoblots and pS6/S6 ratio quantifications of lysates from skeletal muscle (quadriceps (B) and gastrocnemius (C)) and liver (D) from mice subjected to a 12-h daytime fast, or similarly fasted then refed for 2 h. Graphed as mean percentage of fasted TSC2-WT littermates ± SEM; N = 3 fasted and 4 refed per genotype. (E) Echo-MRI lean mass measurements of male littermates at 3, 6, and 12 months of age, graphed as mean ± SEM; N = 12 TSC2-WT and 9 TSC2-5A. (F) Lean mass change of the measurements in (E) from 3 to 6 months of age, graphed as mean ± SEM. (G) Skeletal muscle weights in 16-week-old male TSC2-5A mice (N = 8), graphed as mean percentage of TSC2-WT littermates (N = 6) ± SEM. (H, I) Pooled (H) and binned (I) fiber cross-sectional area (CSA) distribution of the skeletal muscles of 16-week-old male TSC2-WT (N = 6) and TSC2-5A (N = 8) littermates. Data are represented as mean ± SEM. Statistical analysis by Mann-Whitney or ordinary two-way ANOVA with Tukey’s multiple comparisons tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 5:. TSC2-5A myotubes display a cell intrinsic defect in mTORC1 signaling and protein synthesis.

(A) Bright-field images of isolated satellite cells from TSC2-WT and TSC2-5A skeletal muscle before (D0) and after 6 days of differentiation into myotubes (D6). Scale bars: satellite cells, 100 μm; myotubes, 500 μm. (B) Representative immunoblots of lysates from primary TSC2-WT and TSC2-5A myotubes serum-starved for 6 h and then stimulated with time courses of insulin (100 nM). (C) Representative immunoblots of lysates from primary TSC2-WT and TSC2-5A myotubes serum-starved for 6 h followed by a 15-min stimulation with insulin (100 nM) after a 30-min pretreatment with vehicle (−), MK2206 (2 μM, AKTi), trametinib (5 μM, MEKi), or both, as indicated. (D) Representative immunoblots of lysates from primary TSC2-WT and TSC2-5A myotubes serum-starved for 6 hours and then stimulated with insulin (100 nM) for 1.5 h. Puromycin was added for the final 30 min of insulin stimulation to measure protein synthesis. N = 2 technical replicates. (E) Quantification of puromycin incorporation in TSC2-5A myotubes, graphed as percentage of TSC2-WT ± SEM. N = 2 independent biological experiments, each with 2 technical replicates.