Autophagic lipid metabolism sustains mTORC1 activity in TSC-deficient neural stem cells

FIP200 ablation in Tsc1^GFAP cKO mice reverses brain abnormalities driven by mTORC1 hyperactivation

Recent studies showed that mTORC1 hyperactivation⁷ and autophagy deficiency^{37, 38} both led to defective maintenance of neural stem/progenitor cells (NSCs). Autophagy inhibition by mTORC1 hyperactivation is well established^{1, 3, 39}, but it is not known if reduced autophagy is responsible for NSCs defects^7–9. To explore this question, we generated Tsc1^f/f;Fip200^f/f;hGFAP-Cre (designated as 2cKO), Tsc1^f/f;Fip200^f/+;hGFAP-Cre (Tsc1^GFAPcKO), and Tsc1^f/+;Fip200^f/+;hGFAP-Cre (Ctrl) mice by crossingTsc1^{f/f40, 41} and Fip200^f/f;hGFAP-Cre³⁸ mice. Mice were produced in expected Mendelian ratios. Despite NSC defects by either Tsc1 or Fip200 deletion alone, we found that, surprisingly, the 2cKO mice were rescued from aberrant growth in the subventricular zone (SVZ) and rostral migratory stream (RMS), and enlarged brains compared to Tsc1^GFAPcKO mice^{7, 9} (Figs. 1A, 1B, ED1A, ED1B). Autophagy was functional in Tsc1^GFAPcKO but reduced in 2cKO SVZ, indicated by p62 aggregates and autophagy flux, consistent with autophagy blockade with FIP200 loss^{37, 38, 42} (Figs. 1C, 1D, Fig. ED1C-1F). Staining of phosphorylated S6 ribosomal protein (pS6RP) and 4EBP1 (p4EBP1), downstream effectors of mTORC1, revealed diminished mTORC1 activity in the SVZ of 2cKO mice (Figs. 1E, 1F, ED1G, ED1H). These results suggested that abnormal Tsc1^GFAPcKO NSCs were associated with sustained autophagy combined with mTORC1 hyperactivation.

Autophagy inhibition in Tsc1^GFAP cKO mice rescues Tsc1-deficient NSCs and suppresses tumorigenesis

We next examined the cellular basis for the rescue of brain defects in 2cKO mice. At P0, Ki67⁺ cell frequency was increased in the VZ/SVZ of Tsc1^GFAPcKO mice compared to Ctrl (Fig. 1G), contributing to increased cellularity and RMS cell migration at P7 and P21 (see Figs. 1A, 1B)^{7, 9}. In Tsc1^GFAP cKO mice at P21, Ki67⁺ cell frequency was reduced and apoptosis increased compared to Ctrl (Figs. 1H, 1I). Fip200 deletion in 2cKO mice restored VZ/SVZ proliferation and apoptosis, and NSC proliferation, to Ctrl levels (Figs. 1G–1K, ED1I-1J).

We examined NSC self-renewal by in vitro neurosphere assay using P21 SVZ cells. Primary neurospheres from Tsc1^GFAPcKO and 2cKO SVZ cells were similar in number and size to those from Ctrl (Figs. 1L, ED1K, ED1L). Tsc1^GFAPcKO secondary neurosphere growth was significantly decreased compared to Ctrl and 2cKO (Figs. 1L, 1M, ED1M), suggesting that Fip200 deletion sustained the renewal potential of Tsc1-deficient NSCs. Autophagy in Tsc1^GFAPcKO and Ctrl neurospheres was similar, and was blocked in 2cKO neurospheres by Fip200 deletion (Figs. ED1N, ED1O). As previously observed^{7, 9}, increased DCX⁺ neuroblasts were found in Tsc1^GFAPcKO mice at P21 (Figs. 1N, ED2A). Mature neurons (NeuN⁺ cells) were found in Tsc1^GFAPcKO SVZ (Fig. 1O). Labeling proliferative SVZ cells with BrdU at P7 and tracing to P21, we found ~7% NeuN⁺BrdU⁺ cells in Tsc1^GFAPcKO SVZ but almost none in Ctrl SVZ (Figs. 1P, 1Q)^{37, 38}. This indicates premature differentiation of mutant neuroblasts consistent with either defective migration into RMS or prevention of migration out of the SVZ. Fip200 deletion reversed aberrantly increased DCX⁺ neuroblasts, NeuN⁺ cells, NeuN⁺BrdU⁺ cells, and astrocytes in 2cKO SVZ (Figs. 1N–1R, ED2B). Culturing neurospheres in vitro³⁸ to analyze Tsc1^GFAPcKO SVZ cell differentiation, we found aberrantly increased β-III tubulin⁺ cells (Fig. ED2C), GFAP⁺ astrocytes (Fig. ED2D), and reduced NeuN+ neurons (Fig. ED2E) and MBP+ cells (Fig. ED2F); all were rescued in 2cKO neurospheres, consistent with observations in vivo. These results suggest that autophagy is required for the defective maintenance and differentiation of Tsc1-deficient NSCs induced by hyperactivation of mTORC1 in Tsc1^GFAPcKO mice.

SEN/SEGAs are brain lesions found in TSC patients. SEN-like lesions in the SVZ of Tsc1^GFAPcKO mice (Fig. ED2G) were not found until P21. They developed along the lateral wall on the side of striatum with no regional preference (Figs. 1S, 1T). Some lesions appeared as “floating” spheres as tumors protruded and bent in the LV when cross-sectioned (Fig. 1S, arrow). Murine SEN-like cells contained hyperchromatic nuclei surrounded by scanty cytoplasm, resembling NSC, and were pS6RP⁺ (Fig. ED2H). As previously reported⁴³, proliferation in the SEN-like lesions was lower than in SVZ (~5% vs ~20% Ki67⁺ cells at P21). SEN-like cells were nestin⁺ (Fig. ED2I); most were medium Sox2⁺ compared to strong Sox2⁺ in Tsc1^GFAPcKO mice (Fig. ED2J); some were NG2⁺ (Fig. ED2K); and most expressed high levels of DCX (Fig. ED2L). Only a small fraction was GFAP⁺ (Fig. ED2J), suggesting mixed lineage with neuronal differentiation preference. We found only one SEN-like lesion in ten 2cKO mice at P21 and none out of six at P28 (Fig. 1T, ED2G). These results suggest that autophagy supports tumorigenesis with hyperactivated mTORC1 in Tsc1^GFAPcKO mice.

Many autophagy genes have functions independent of their roles in canonical autophagy. We studied autophagy-specific functions of FIP200 in vivo using our Fip200 4A knock-in mutant mouse, in which the autophagy function of FIP200 is inactivated⁴⁴. Removing the FIP200 autophagy function in Tsc1^f/f;Fip200^f/KI;hGFAP-Cre (designated 2cKI) mice in Tsc1-deficient NSC rescued elevated SVZ cell number, mTORC1 hyperactivation, defective NSC maintenance (Fig. ED2M-ED2R), and SEN-like lesion formation (Fig. 1T), similar to FIP200 ablation. These results support that autophagy is required for defective NSC maintenance, differentiation, and tumorigenesis in Tsc1^GFAPcKO mice.

Energy-stressed Tsc-deficient cells require autophagy for hyperactivation of mTORC1

To explore the mechanisms by which autophagy inhibition reverses NSC defects and tumorigenesis in Tsc1^GFAPcKO mice, we first examined mTORC1 activation in Tsc2 KO MEF cultured under various conditions. The accumulation of LC3-II and LC3 puncta in response to Bafilomycin A1 (BafA1) was comparable in Tsc2 KO and WT MEF under basal and glucose-free conditions (Figs. ED3A, ED3B). Both glucose and amino acid starvation reduced mTORC1 phosphorylation of S6K in WT, while Tsc2 KO MEF had sustained mTORC1 signaling (Fig. ED3C). Inhibition of autophagy by Spautin1⁴⁵ significantly reduced mTORC1 hyperactivation in glucose starved but not amino acid starved Tsc2 KO MEF (comparing lanes 7 and 8; 11 and 12). mTORC1 is activated on lysosomes⁴⁶, and mTOR was found on LAMP2⁺ (Lysosomal Associated Membrane Protein 2) lysosomes in Tsc2 KO MEF under glucose-free conditions. Spautin1 treatment abolished co-localization (Figs. ED3D, ED3E). Spautin1 autophagosome inhibition was verified by measuring LC3-II with or without BafA1 treatments (Fig. ED3F). To complement pharmacological inhibition of autophagy, we generated Atg7 KO in WT and Tsc2 KO MEF (Fig. ED3G). This genetic block of autophagy decreased hyperactivation of mTORC1 and abolished co-localization of mTOR and LAMP2 in glucose-free conditions (Figs. ED3G, ED3H, see Fig. ED3D bottom panels and Fig. ED3E). Consistent with Tsc-deficient MEF, mTORC1 hyperactivation in Tsc1^GFAPcKO neurospheres was reversed in 2cKO to Ctrl levels under glucose-free conditions (Figs. 2A, 2B). These results correspond to mTORC1 activation status in SVZ (see Figs. 1E, 1F), connecting in vitro mechanistic studies using MEF/SVZ cells and observations in vivo.

Interrupting autophagy by Spautin1 treatment or Atg7 KO induced AMPK phosphorylation of its targets Raptor at Ser⁷⁹² and ACC at Ser⁷⁹ in Tsc2 KO MEF in response to glucose starvation or glycolysis inhibitor 2-deoxy-d-glucose (2DG) (Figs. ED4A, S4B). Autophagy inhibition by either Spautin1, chloroquine (CQ), or Atg7 KO decreased ATP content (Figs. ED4C, ED4D), accounting for increased AMPK activity. In vivo, an increased fraction of pACC+ and pAMPK+ cells in SVZ of 2cKO compared to Tsc1^GFAPcKO mice was seen (Figs. 2C–2F). Likewise, ATP was increased in Tsc1^GFAPcKO neurospheres, and reversed in 2cKO neurospheres (Fig. 2G). Glucose-free conditions in SVZ cells and MEF may approximate the relatively low glucose concentration in SVZ⁴⁷.

We stably expressed scrambled (Ctrl), TSC1, or FIP200 shRNA singly or in combination (Figs. ED4E-ED4G) in HEK293 cells. Lack of increase in LC3-II with BafA1 indicated decreased autophagy after FIP200 knockdown (Figs. ED4H-ED4I). FIP200 knockdown decreased hyperactivation of mTORC1 induced by TSC1 knockdown in cells treated with 2DG (Fig. ED4J), but not normal culture conditions (see Fig. ED4E). These results suggested that FIP200-mediated autophagy is necessary to maintain mTORC1 hyperactivation with loss of either Tsc1 or Tsc2 under energy stress conditions.

Autophagy maintains OXPHOS to sustain ATP production for hyperactivated mTORC1

Autophagy is proposed to provide substrates for oxidative metabolism under energy stress^{48, 49}. We studied mitochondrial oxygen consumption rate (OCR) and found that Tsc2 KO MEF had increased ATP-related and maximum OCR (Figs. 2H, ED4K, ED4L), indicative of high mitochondrial oxidative phosphorylation (OXPHOS). After glucose deprivation (Fig. ED4K) or 2DG treatment (Fig. ED4L), both OCRs were significantly reduced in WT but maintained in Tsc2 KO MEF. Autophagy inhibition by Spautin1, CQ, or Atg7 KO reduced OCRs in Tsc2 KO MEF (Figs. 2I, 2J, ED4M, ED4N), consistent with reduced ATP (see Figs. ED4C, ED4D). Similarly, Tsc1 KD cells (Figs. 2K, 2M) and WT MEF expressing constitutively active Rheb (Rheb-CA MEF) (Fig. 2L, 2N) maintained OCR in glucose-free conditions in an autophagy-dependent manner. Lastly, increased ATP-related and maximum OCR in Tsc1^GFAPcKO neurospheres were reversed in 2cKO (Fig. 2O). Together, these results suggest a role of autophagy in maintaining elevated OXPHOS to sustain ATP production and mTORC1 hyperactivation under energy stress.

Lipophagy produces FFAs as an energy source for hyperactivation of mTORC1

We explored potential autophagy-dependent energy sources to fuel mTORC1 activity in energy stressed Tsc-deficient cells. Glutamine did not restore mTORC1 hyperactivation in Tsc2 KO MEF after autophagy inhibition (Fig. ED5A), in contrast to previous results in pancreatic cancers ⁵⁰. Amino acid carbon skeletons were unlikely energy sources, as similar ammonia concentrations were found in WT and Tsc2 KO MEF under normal or glucose-free conditions (Fig. ED5B). Glycogen content was comparable in WT and Tsc2 KO MEF after glucose starvation (Fig. ED5C), and BafA1 had no effect, suggesting autophagy did not regulate glycogen levels (Fig. ED5C).

We then focused on lipid catabolism. Previous studies showed that Tsc-deficient cells have higher triglyceride (TG) stores than WT cells⁵¹. Glucose deprivation significantly promoted glycerol release in Tsc2 KO but not WT MEF (Fig. 3A); Spautin1 reduced it. Glucose deprivation and 2DG decreased TG and increased intracellular FFAs in Tsc2 KO MEF; Spautin1 treatment counteracted these changes (Figs. 3B, 3C). FFAs were increased in TSC1-knockdown after glucose starvation, but not normal culture conditions, and this increase was abolished by autophagy inhibition by FIP200 knockdown (Fig. ED5D). In mammalian cells, TG is stored in organelles as lipid droplets (LDs). The baseline number of LDs in Tsc2 KO was higher than WT MEF. Glucose deprivation reduced LDs in Tsc2 KO but not WT MEF in an autophagy-dependent manner (Figs. 3D, 3E). Tsc1^GFAPcKO neurospheres maintained increased LDs; glucose deprivation significantly reduced LDs in Tsc1^GFAPcKO but not in 2cKO neurospheres (Figs. 3F, 3G).

LDs can be sequestered by autophagosomes and transported to lysosomes for degradation by lipophagy³². Under glucose-free conditions LDs co-localized with LC3⁺ autophagosomes in Tsc1^GFAPcKO but not Ctrl or 2cKO neurospheres (Figs. 3H, 3I). Co-localization of LC3 and LDs was slightly induced upon glucose starvation in both WT and Tsc2-deficient MEF; the co-localization of LDs with autophagosomes was significantly increased by BafA1 (Figs. 3J, ED5E) and abolished by Spautin1. Similar results were obtained for co-localization of LDs with lysosomes (Figs. 3K, ED5F). In vivo, we observed fewer LDs in Tsc1^GFAPcKO versus Ctrl SVZ (Fig. ED5H-ED5M), consistent with higher fatty acid oxidation. Both the mitochondrial β-oxidation inhibitor etomoxir (ETO) and genetic autophagy inhibition (2cKO) increased LDs, suggesting that the lipid content of SVZ cells is responsive to mitochondrial beta-oxidation, which requires functional autophagy (Fig ED5M). These results indicate that lipophagy facilitates FFA production from stored TGs as an energy source for TSC deficient cells under energy stress to maintain mTORC1 hyperactivation.

β-oxidation of FFAs supports bioenergetics in Tsc-deficient cells during energy stress

FFAs can be metabolized to produce ATP through catabolic β-oxidation. We analyzed β-oxidation under glucose-free conditions and found higher activity in Tsc2 KO MEF, reversed by Spautin 1 treatment (Fig. 3L). We then cultured Tsc2 KO MEF with medium and long chain FFA. After autophagy inhibition, we found that both FFAs abrogated ATP loss by Spautin1 and CQ under glucose-free (Fig. 3M) but not normal culture conditions (Fig. ED5G). Consistent with ATP, under energy stress long chain FFAs increased the OCR (Fig. 3N) and restored S6RP phosphorylation (Fig. 3O) in Spautin1-treated Tsc2 KO but not WT MEF. These results demonstrate a role for autophagy-mediated lipid catabolism in Tsc-deficient cells under energy stress conditions.

We then examined FFA β-oxidation directly and found that treatment with β-oxidation inhibitors ETO, trimetazidine (TMZ)⁵², and ranolazine (Rano)⁵³ triggered bioenergetic stress characterized by phosphorylation of AMPK, Raptor, and ACC (Figs.4A, ED6A, ED6C), and reduced phosphorylation of ED6K and ED6RP (Figs. 4B, ED6B, ED6D) in Tsc2 KO MEF under glucose-free and 2DG treatment. ETO inhibited mTORC1 activation under energy stress but not normal culture in Tsc1^GFAPcKO neurospheres (Fig. 4C). ETO and Rano dissociated mTOR from LAMP2+ structures in Tsc2 KO MEF under glucose deprivation (Figs. ED6E, bottom panels and ED6F, compare to Fig. ED3D, second row right 3 panels and S3E), but not normal culture conditions (Figs. ED6E, top panels, ED6F). Lastly, inhibition of AMPK by Compound C prevented mTORC1 suppression by ETO (Fig. ED6G) in Tsc2 KO MEF, providing further support that FAO inhibitors blocked mTORC1 via AMPK in TSC-deficient cells under energy stress.

All β-oxidation inhibitors reduced ATP levels in Tsc2 KO (Fig. 4D) but not WT (Fig. ED6H) MEF under energy stress. Similar results were obtained for TSC1 KD cells treated with ETO (Fig. ED6I). ETO and Rano also decreased ATP levels under energy stress in Tsc1^GFAPcKO neurospheres (Fig. 4E). We also observed corresponding decrease of maximum OCR in Tsc2 KO MEF by all inhibitors under energy stress (Figs. 4F, 4G). These results suggest that Tsc-deficient cells have an increased dependence on FFA β-oxidation fueled by FFAs to maintain hyperactivated mTORC1 under energy stress.

All three β-oxidation inhibitors also increased the number of LDs co-localized with lysosomes in Tsc2 KO but not in WT MEF under glucose deprivation (Figs. 4H, 4I). Co-localization of LDs with LC3+ autophagosomes increased together with FFAs (Fig. 4J, 4K), suggesting feedback to increase lipophagy when β-oxidation is inhibited. These results further support that FFAs generated by lipophagy of LDs are used to sustain mTORC1 hyperactivation in Tsc-deficient cells under energy stress.

Lysosome acid lipase (LAL) digests LDs to release FFAs in Tsc-deficient cells

We next explored potential lipase(s) for lipophagy of LDs. We used lipase inhibitors including Orlistat for lysosomal acidic lipase (LAL)⁵⁴, Atglistat for adipose triglyceride lipase (ATGL)⁵⁵, and JZL184 for monoacylglycerol lipase⁵⁶. Similar to a previous report suggesting a role for LAL in autophagy⁵⁷, we found that Orlistat, but not Atglistat or JZL184, significantly reduced glycerol release (Fig. 5A) and increased the number of LDs (Fig. 5B) in Tsc2 KO MEF with glucose starvation. No inhibitor affected glycerol release in normal culture conditions, nor did they affect glycerol release or LD numbers of WT MEF under either condition (Figs. ED7A, ED7B). Orlistat also blocked FFA release (Fig. 5C), reduced mitochondrial OCR (Fig. 5D), and suppressed ATP content (Fig. 5E) of Tsc2 KO MEF under both glucose-free and 2DG treatment conditions. Consequently, Orlistat increased the AMPK phosphorylation of Raptor at Ser⁷⁹² and ACC at Ser⁷⁹, reduced mTORC1-dependent phosphorylation of S6K and S6RP (Figs. 5F, 5G), and disrupted co-localization of mTOR with LAMP2 in Tsc2-deficient cells under glucose-free conditions (Fig. ED7C). Importantly and consistent with results in Tsc-deficient MEF, we found that Orlistat reduced mTORC1 activation under energy stress, but not normal culture conditions in Tsc1^GFAPcKO primary SVZ neurospheres (Fig. 5H). Lastly, blockade of LAL resulted in increased localization of LDs on lysosomes (marked by LAMP2) in Tsc2 KO MEF under glucose starvation condition (Fig. 5I). As LAL is the only lipase known to degrade lipids in lysosomes⁵⁸, these results suggest that under energy stress conditions, LDs are delivered to lysosomes via lipophagy, and LAL is used to hydrolyze TG to generate FFAs to sustain ATP and mTORC1 hyperactivation in Tsc-deficient cells.

Indeed, we observed increased LAL in Tsc2 KO MEF (Fig. ED7D). In glucose-starved Tsc2 KO MEF, expression of LAL shRNA (Fig. ED7E) reduced glycerol release (Fig.ED7F) and FFA content (Fig. ED7G), accompanied by corresponding increases in LDs (Fig. ED7H) and TG (Fig. ED7I). LAL shRNA also decreased mitochondrial OCR, ATP levels, and mTORC1 activation under energy stress (Figs. ED7J-ED7L). Supplementation of BSA-Palmitate restored ATP and mTORC1 hyperactivation in Tsc2 KO MEF lacking LAL in glucose limiting conditions (Fig. 5J, ED7K). Together, these results demonstrate LAL as the major lipase for the production of FFAs from lipophagy of LDs to sustain hyperactivated mTORC1 in Tsc deficient cells under energy stress.

Lipophagy inhibition restores function of Tsc1-deficient NSCs and inhibits tumorigenesis

To evaluate targeting lipophagy to counteract mTORC1 hyperactivation as new therapies, we examined pharmacological inhibition of autophagy in Tsc1^GFAPcKO mice. We injected autophagy inhibitor CQ (50 mg/kg) and glycolysis inhibitor 2DG (500 mg/kg) I.P. into P7 Ctrl and Tsc1^GFAPcKO mice every other day for 14 days, individually or in combination. Combined CQ+2DG inhibited murine SEN-like lesions, but each alone did not (Fig. 6A). Lesions were also eliminated by positive control rapamycin (2.5 mg/kg, daily I.P.). Combined CQ+2DG, but neither alone decreased S6RP phosphorylation (Figs. 6B, ED7M) and increased AMPK phosphorylation and ACC (Figs. 6C, 6D, ED7N, ED7O). mTORC2 was unaffected (Figs. ED7P, ED7Q). 2DG, CQ, or the combination all reduced ERK signaling comparably (Figs. ED7R, ED7S). The combination treatment restored the pool of GFAP⁺Nestin⁺ and GFAP⁺Sox2⁺ NSC (Figs. 6E, ED8A). Combination CQ+2DG decreased proliferation and apoptosis (Figs. 6F, 6G, ED8B), aberrant neurogenesis and premature neuronal differentiation, and increased astrogenesis (Figs. 6H–6J) of Tsc1^GFAPcKO but single treatment had no effect on mTORC1 activity, or the maintenance and differentiation of SVZ NSC in Ctrl mice (Figs. 6B–6J).

We next tested β-oxidation inhibitors ETO, Rano, and TMZ with or without 2DG in mice. The combination blocked tumorigenesis, but β-oxidation inhibitors alone did not (see Fig. 6A). Phosphorylation of AMPK and ACC increased (Figs. 6K–6L), and mTORC1 signaling was reduced (Fig. 6M) in Tsc1^GFAPcKO but not Ctrl mice. The combination rescued defective NSC maintenance, proliferation, and apoptosis (Figs.6N-6P, ED8C, ED8D), and neurogenesis and neuronal differentiation defects (Figs. ED8E, ED8F) in Tsc1^GFAPcKO but not Ctrl mice. β-oxidation inhibitors alone did not alter AMPK and ACC phosphorylation, mTORC1 activation, or proliferation and apoptosis in Tsc1^GFAPcKO mice (Figs. 6K–6P). Each suppressed aberrant neurogenesis and neuronal differentiation (Figs. ED8E, ED8F) without affecting SEN-like lesion formation (see Fig. 6A) in Tsc1^GFAPcKO mice. Together, these results indicate that combined inhibition of glycolysis and lipophagy, or their downstream catabolic processes, prevents tumorigenesis, exhaustion, and abnormal NSC differentiation, establishing the therapeutic potential of autophagy inhibition for Tsc-deficient conditions.

Animals

FIP200^f/f, FIP200^f/KI, Tsc1^f/fand hGFAP-Cre transgenic mice were described previously^{37, 38, 41, 44}. FIP200^f/f;Tsc1^f/f;hGFAP-Cre, FIP200^f/KI;Tsc1^f/f;hGFAP-Cre and control mice were maintained on a mixed C57Bl6 (87.5%) with FVB (12.5%) background. Age and littermate-matched control and mutant mice were used for analysis to minimize the impact of modifier genes. Mice were housed and handled according to local, state, and federal regulations. All experimental procedures were carried out according to the guidelines of Institutional Animal Care and Use Committee at the University of Cincinnati. Genotyping for FIP200, Tsc1 and Cre alleles was performed by PCR analysis of tail DNA essentially as described previously⁴¹.

Cells and cell culture

Human embryonic kidney 293 cells (HEK293) from ATCC is cultured in DMEM with 10% FBS under growth conditions. Human Tsc1, FIP200 shRNA, and scrambled shRNA infected HEK293 cells were selected with 1 μg/ml Puromycin in media. Wild type (WT) mouse embryonic fibroblast (MEF) and Tsc2 knockout MEF were kindly gifted by Dr. Diane Fingar from the Department of Cell and Developmental Biology, University of Michigan and cultured in DMEM with 10% FBS and 25 μg/ml Hygromycin B for selection. Atg5 knockout MEF and paired WT MEF were gifted by Dr. Noboru Mizushima from the University of Tokyo. Atg5 KO MEF and WT MEF were transduced with rat RhebS16H ⁶⁵ or empty plasmid and cultured in DMEM with 10% FBS and 1 μg/ml Puromycin for selection. Glucose-free DMEM was purchased from Gibco™ (Cat#11966025). 100 nM Bafilomycin A1 was used for 2 hr to inhibit autophagosome degradation.

CRISPR knockout of Atg7 and AMPKα1 in MEF cells

Two pairs of sgRNA targeting Atg7 (#1, 5’-TGGACACCAGGGAGAGCCGG-3’ and #2, 5’-GAACGAGTACCGCCTGGACG-3’) and two pairs of sgRNA targeting AMPKα1 (#1, 5’-TTATTGTCACAGGCATATGG-3’ and #2, 5’-GTGGGCAAGCACGAGTTGAC-3’) were subcloned into lentiCRISPR v2 (Addgene, #52961) according to the protocol in Dr. Feng Zhang’s lab. These two CRIPSR plasmids were packaged into lentiviruses with psPax2 and pMD2.G in HEK293 cells using the calcium phosphate method. MEF cells were infected with CRISPR lentiviruses and selected with Puromycin (2 – 4 μg/ml) for one week. Single clones were screened by limited dilution in 96-well plate. The selected clones were cultured in DMEM with 10% FBS.

Virus production and infection

Human scrambled shRNA, FIP200 shRNA, and Tsc1 shRNA encoded in pGIPZ lentiviral vector were purchased from Open Biosystems (Lafayette, CO). Mouse scrambled shRNA, LAL shRNA #1 and #2 were purchased from Sigma. Package vectors of pSPAX2 and pMD2.G were mixed with shRNA vectors and transfected in HK293 cells for virus production essentially as described before³⁷.

Neurosphere formation assay and in vitro differentiation

Neurospheres were cultured in neurobasal medium supplemented with B27 (Invitrogen, CA), 10 ng/ml basic fibroblast growth factor (bFGF) and 20 ng/ml epidermal growth factor (EGF) (Invitrogen, CA) in Ultra-Low Attachment dishes (Corning, MA), essentially as described in our previous reports ^{37, 38}. Neurospheres with diameter larger than 50 μm were counted 7–9 days after culturing. For their passage, primary neurospheres were collected by centrifugation, then incubated in 0.2% Trypsin at 37°C for 10 min with ~50 times pipetting. After addition of trypsin inhibitor supplemented with DNase I (Worthington, NJ) to stop digestion, dispersed cells were collected by centrifugation at 400 g for 10 minutes, counted, and cultured for secondary neurosphere formation and quantified as described before ^{37, 38}. In some experiments, the dispersed cells from neurospheres were cultured for differentiation in neurobasal medium supplemented with B27 (Invitrogen, CA), 10 ng/ml bFGF, and 5% (v/v) FBS in 48-well plates coated with 150 μg/ml poly-D-lysine (Biomedical Technologies, MA) and 20 μg/ml laminin (Invitrogen, CA) for another 7 or 14 days. Differentiated cells were fixed in 4% PFA for 30 min and then washed twice in cold PBS for 10 min. These cells were stained with β-III tubulin, NeuN, GFAP, MBP and DAPI for lineage analysis.

Antibodies and reagents

Mouse anti-Nestin (Cat: Rat-401, PMID: 23403289, DSHB, IA), IHC, 1:50; anti-GFAP (Cat# 3607, PMID: 28794441, Cell Signaling, MA), IHC, 1:100; anti-mTOR (Cat#: 14–2190-82, PMID: 18803507, ThermoFischer Scientific, CA), IF, 1:100; Alexa 647 conjugated anti-BrdU (Cat: B35133, PMID: 28754836, ThermoFischer Scientific, CA), IHC, 1:50; anti-beta-actin (Cat: A2228, PMID: 21498634, Sigma, MO), WB, 1:4,000; anti-GAPDH (Cat: G8795, PMID: 23222489, Sigma, MO), WB, 1:10,000; anti-vinculin (Cat:V9264, PMID: 23685627, Sigma, MO), WB, 1:5,000; anti-NeuN (Cat: MAB377, PMID: 26373451, Millipore, MA), IHC, 1:500. rabbit anti-Ki67 (Cat: M3060, PMID: 26402841, Spring Bioscience, CA), IHC, 1:200; anti-p62 (Cat: BML-PW9860, PMID: 22562096, Enzo, PA), WB, 1:1,000, IHC, 1:200; anti-GFAP (Cat: Z0334, PMID: 8229641, DAKO, CA), IHC, 1:200; anti-phosphorylated (Ser65) 4EBP1 (Cat: 9451, PMID: 28778941,Cell Signaling, MA), WB, 1:1,000; anti-phosphorylated (Thr37/46) 4EBP1 (Cat: 2855, PMID: 26913956,Cell Signaling, MA), IHC, 1:100, WB, 1:1,000; anti total 4EBP1 (Cat: 9644, PMID: 29720666, Cell Signaling, MA), WB, 1:1,000; anti-FIP200 (Cat: 12436, PMID: 27305347, Cell Signaling, MA), WB, 1:1,000; anti-beta-III tubulin (Cat: 5568, PMID: 25726526, Cell Signaling, MA), IF, 1:100; anti-NG2 (Cat: AB5320, PMID: 25902404, Millipore, MA), IHC, 1:100; anti-phosphorylated (Thr202/Tyr204) ERK1/2 (Cat: 4307, PMID: 22027685,Cell Signaling, MA), IHC, 1:100; anti-phosphorylated S6 kinase (Thr389) (Cat: 9234, PMID: 28928465, Cell Signaling, MA), WB, 1:1,000; anti total S6 kinase (Cat: 2708, PMID: 28835610, Cell Signaling, MA), WB, 1:1,000; anti-phosphorylated S6 ribosomal protein (Ser240/244) (Cat: 5364, PMID: 27026523, Cell Signaling, MA), IHC, 1:500, WB, 1:3,000; anti total S6 ribosomal protein (Cat: 2317, PMID: 29162813, Cell Signaling, MA), WB, 1:3,000; anti-LC3 (Cat: 2775, PMID: 28674081, Cell Signaling, MA), IF, 1:100, WB, 1:1,000; anti-phosphorylated (Ser473) pAKT (Cat:4060, PMID: 24670654, Cell Signaling, MA), IHC, 1:100; anti-phosphorylated-AMPK (Thr172)(Cat: 2535, PMID: 27409632, Cell Signaling, MA), IHC, 1:100, WB, 1:1,000; anti AMPK (Cat: 5831, PMID: 27172265, Cell Signaling, MA), WB, 1:1,000; anti-phosphorylated-Raptor (Ser792)(Cat: 2083, PMID: 23352126, Cell Signaling, MA), WB, 1:1,000; anti total Raptor (Cat: 2280, PMID: 28613987, Cell Signaling, MA), WB, 1:1,000; anti-phosphorylated-ACC (Ser79) (Cat: 3661, PMID: 28552492, Cell Signaling, MA), IHC, 1:100, WB, 1:1,000; anti total ACC (Cat: 3662, PMID: 28552492, Cell Signaling, MA), WB, 1:1,000; anti-pjosphorylated-Ulk1 (Ser757) (Cat: 6888, PMID: 30784596, Cell Signaling, MA), WB, 1:1,000; anti-pjosphorylated-Ulk1 (Ser555) (Cat: 5869, PMID: 30957628, Cell Signaling, MA), WB, 1:1,000; anti-Ulk1 (Cat: 8054, PMID: 30957628, Cell Signaling, MA), WB, 1:1,000; anti-SOX2 (Cat: AB5603, PMID: 25580734, Millipore, MA), IHC, 1:500; anti-LAL (Cat: ab154356, PMID: 26134519, Abcam, MA); WB, 1:2,000. Rat anti-Ki67 (Cat: 652402, PMID: 25675507, Biolegend, CA), IHC, 1:200; anti-MBP (Cat: MAB386, PMID: 25706475, Millipore, MA), IHC, 1:200; anti-LAMP2 (Cat: H4B4, PMID: 16874023, DSHB, IA), IF, 1:100. Guinee pig anti-DCX (Cat: Ab2253, PMID: 24880216, Millipore, MA), IHC, 1:500. Secondary antibodies: Anti-mouse IgG, HRP-linked Antibody (Cat: 7076, Cell Signaling MA); 1:10,000; Anti-rabbit IgG, HRP-linked Antibody (Cat: 7074, Cell Signaling MA); 1:10,000; AffiniPure Donkey Anti-Mouse IgG (H+L) (Jackson Immunology, PA); conjugated FITC Cat: 715–095-150; conjugated TRITC, Cat: 715–025-150; IF and IHC, 1:200; AffiniPure Donkey Anti-Rabbit IgG (H+L) (Jackson Immunology, PA); conjugated FITC Cat: 711–095-152; conjugated TRITC, Cat:711–025-152; IF and IHC, 1:200; AffiniPure Donkey Anti-Rat IgG (H+L) (Jackson Immunology, PA); conjugated FITC Cat: 712–095-153; conjugated TRITC, Cat: 712–025-153; IF and IHC, 1:200; AffiniPure Goat Anti-Guinea Pig IgG (H+L) (Jackson Immunology, PA); conjugated TRITC, Cat: 106–025-033; IHC, 1:200.

BODIPY 493/503 was purchased from Invitrogen (Invitrogen, CA). DAPI was purchased from Millipore (Millipore, CA). DMEM, glucose-free DMEM and HBSS were from Thermo Fischer Scientific. ATP, BSA, palmitate, octanoate, glucose, glutamine, sodium pyruvate, Bafilomycin A1, chloroquine and Spautin1 were purchased from Sigma (St Louis, MO).2DG, Orlistat, Atglistat, JZL, etomoxir, ranolazine, and trimetazidine were purchased from Cayman (Ann Arbor, MI); rapamycin was from LC laboratory.

Histology, immunofluorescence (IF) and BODIPY (493/503) staining

Mice were euthanized using CO₂, and a complete tissue set was harvested during necropsy. The weight of brains was measured before fixation. Fixation was carried out for 16 hr at 4°C using 4% (w/v) freshly made, pre-chilled PBS-buffered paraformaldehyde (PFA). The brain tissues were all and embedded in paraffin and sagittal sectioned at 5 μm sections. Slides from histologically comparable positions (triangular lateral ventricle with intact RMS) were stained with hematoxylin and eosin (H&E) for routine histological examination or left unstained for immunofluorescence (IF). H&E stained sections were examined under a BX41 light microscope (Olympus America, Inc., Center Valley, PA), and images were captured with an Olympus digital camera (model DP70) using DP Controller software (Version 1.2.1.10 8). For immunofluorescence, unstained tissues were first deparaffinized in 3 washes of xylene (3 min each) and then rehydrated in graded ethanol solutions (100, 95, 70, 50, and 30%). After heat-activated antigen retrieval (Retriever 2000, PickCell Laboratories B.V., Amsterdam, Holland) according to the manufacturer’s specifications, sections were treated with Protein Block Serum Free (Dako Corp, CA) at room temperature for 10 min. Slices were then incubated with primary antibodies at 4°C for 16 hr in a humidified chamber, washed in PBS 3 times (5 min each) and incubated with the 1:200 secondary antibodies for 1 hr at room temperature. After secondary antibody incubation, slides are washed in PBS 3 times (5 min each). Nuclei were stained with DAPI and mounted with Vectashield mounting medium (Vector Laboratories, CA). Digital photography was carried out as described previously⁶².

For BODIPY staining, a fresh solution of BODIPY was diluted in PBS at a concentration of 1μg/mL and applied to 4% PFA fixed samples. After PFA was removed by three washes with PBS, samples were incubated in 1 μg/ml BODIPY 493/503 for 20 min at room temperature. Some samples were also labeled with antibodies (LAMP2 and LC3) to identify lysosome and autophagosomes in cells. The samples were observed using a Zeiss LSM 700 confocal microscope (Carl Zeiss).

Colocalization of mTOR with lysosomal marker LAMP2 was quantitated using Pearson’s correlation coefficient (PCC). Confocal acquired imaging data were analyzed using the Coloc 2 plugin of the Fiji distribution of ImageJ software (http://imagej.net/Coloc_2). Cells were selected as regions of interest (ROIs) using the freehand selection tool. Pearson’s R value (no threshold) reported in the ImageJ Log window was recorded. Values shown represent mean ± SE (5 independent experiments from at least 30 cells).

Oil Red O staining

25 μm frozen brain sections were washed with 60% 2-propanol (Fisher), incubated in filtered 0.3% Oil Red O (Allied Chemical) for 10 min, then rinsed in 60% 2‐propanol followed by one rinse with distilled H₂O. All samples were mounted in Vectashield, covered with glass cover slips (No. 1, VWR), and mounted with Permount (Biomeda, CA).

BrdU incorporation assay and TUNEL assay

For in vivo studies, BrdU was administrated intra-peritoneally (I.P.) at 100 μg/gram for 3 times at 2 hr intervals. Mice were either euthanized 2 hr after the last injection (short term incorporation) or 2 weeks later (long term retention), and tissues were processed as described above. For BrdU detection, the samples were treated with 2 M HCL at room temperature for 20 min to denature nucleotides, then neutralized with 0.1 M sodium borate at room temperature for another 20 min. After 3 washes in PBS (5 min each), slides were incubated with mouse anti-BrdU antibody and secondary antibodies as described in immunofluorescent staining. Nuclei were stained with DAPI and mounted with Vectashield mounting medium (Vector Laboratories, CA). For IF and detection of BrdU in the same tissue, IF was carried out first and samples were post-fixed with 4% PFA at room temperature for 20 min before nucleotide denaturing with HCL. Histological examination and digital photography were carried out as described above.

Apoptotic cells were detected by TUNEL according to the protocol provided by the manufacture in the In Situ Cell Death Detection Kit-TMR Red (Roche, Germany).

Cellular ATP content measurement

Intracellular ATP level was measured by using CellTiter-Glo® Luminescent Cell Viability Assay Kit (Promega, WI). ATP dissolved in DMEM was used as standards of 5 μM, 500 nM, 50 nM, and 5 nM. The experimental procedure was provided by the manufacture in the kit. 10,000 cells were used for each well and 3 wells were used for each sample. The measurement was carried out on a Synergy™ 2 Multi-Mode Microplate Reader (BioTEK, VT).

Measurement of β-oxidation

β-oxidation was measured using a ³H-palmitate adapted from a previous publication (Djouadiet al., 2003 Mol Genet Metab). Briefly, 5×10⁵ MEF were washed with glucose-free DMEM, then cultured with 400 μl of (9,10-³H) Palmitate:albumin for 2 hrs under glucose-free conditions at 37°C. After incubation, 10% TCA was added to each tube and centrifuged at 3300 rpm for 10 min at 4°C, then mixed with 6N NaOH and applied to ion-exchange columns. The column was washed with 1 mL of water and the eluates were counted using a scintillation counter.

Measurement of triglyceride, FFA and glycerol release

MEFs (4×10⁶) were lysed in 0.2 ml lysis buffer (5% NP40 in H₂O) and assayed for triglyceride using a Triglyceride Quantification Colorimetric Kit (Biovision Inc, CA) according to the manufacturer’s instructions. Samples were twice heated to 90°C for 2 min and cooled to room temperature, then centrifuged to remove any insoluble material, and diluted twofold in Triglyceride Assay Buffer. Samples were then incubated with lipase for 20 min, followed by a 30-min incubation with Triglyceride Reaction Mix. The OD at 570 nm was measured and triglyceride amount was calculated using a standard curve.

MEF (2×10⁶) cells were extracted by homogenization with 200 μl of chloroform-Triton X-100 (1% Triton X-100 in pure chloroform) in a microhomogenizer and assayed for FFA using a Free Fatty Acid Quantification Colorimetric/Fluorometric Kit (Biovision Inc. CA). Extracts were spun for 5–10 minutes at top speed in a microcentrifuge. The organic phase (lower phase), was collected, air dried at 50°C to remove chloroform, and vacuum dried 30 min to remove trace chloroform. The dried lipids (in Triton X-100) are dissolved in 200 μl of Fatty Acid Assay Buffer by vortexing extensively for 5 min. The extraction procedure can be proportionally scaled up if a larger amount of sample is desired. 1– 50 μl of the extracted sample is used per assay.

Glycerol in MEF supernatants was measured using absorbance at 570 nm with a Free Glycerol Colorimetric/Fluorometric Assay Kit (Biovision Inc. CA) by comparison to a glycerol standard in according to the protocol provided by the manufacture.

2DG, CQ, ETO, Rano, TMZ and rapamycin administration

PBS, 2DG (500 mg/kg in PBS), CQ (50 mg/kg in PBS), 2DG+CQ, ETO (50 mg/kg in PBS), Rano (25 mg/kg in PBS), TMZ (5 mg/kg in PBS), 2DG+ETO, 2DG+Rano, 2DG+TMZ were administrated intra-peritoneal (IP) into Tsc1^GFAP cKO mice every other day from P7 for 14 days. Ctrl mice were injected with 2DG+CQ, 2DG+ETO, 2DG+Rano, 2DG+TMZ at same frequency from P7 to P21. For administration of rapamycin (LC lab), freshly prepared rapamycin (Sun et al., 2015) was IP injected into Ctrl and FIP200^hGFAP cKO mice at a dose of 2.5 mg/kg once a day from P5 to P21. After treatment, mice were euthanized as described previously and analyzed for histological changes.

Analysis of murine subependymal nodules (SEN)-like structure

We analyzed the SEN-like structures by H&E staining of brain samples from P7, P14, P21, and P28 mice. The slides from the start point to the end point of lateral ventricle were examined at an interval of 30–40 μm under both high and low magnifications. Each slide containing 3–4 slices of brain sections. The lateral ventricle and surrounding striatum, CC, hippocampus, and cortex from Ctrl, FIP200 cKO, Tsc1 cKO, and FIP200/Tsc1 2cKO mice at P21 were examined. Brain slides from P21 Tsc1 cKO mice receiving vehicles, 2DG, CQ, 2DG + CQ, beta-oxidation inhibitors, 2DG + beta-oxidation inhibitors and P21 Ctrl mice receiving 2DG + CQ, as well as 2DG + beta-oxidation inhibitors were examined.

Protein extraction, SDS-PAGE and immunoblotting

Cells were used for protein extraction by homogenization in modified radioimmune precipitation assay buffer (50 mm Tris-HCl, pH 7.4, 1% Triton X-100, 0.2% sodium deoxycholate, 0.2% SDS, 1 mm sodium EDTA) supplemented with protease inhibitors (5 μg/ml leupeptin, 5 μg/ml aprotinin, and 1 mm phenylmethylsulfonyl fluoride). After removing cell debris by centrifugation at 13,000 rpm for 10 min at 4°C, protein concentration was determined using Bio-Rad protein assay reagent. The lysates were boiled for 5 min in 1×SDS sample buffer (50 mm Tris-HCl, pH 6.8, 12.5% glycerol, 1% SDS, 0.01% bromphenol blue) containing 5% β-mercaptoethanol. They were then analyzed by SDS-PAGE followed by immunoblotting using various antibodies, as described previously (Wang et al., 2013, 2016).

Seahorse extracellular assay for oxygen consumption rate (OCR)

Measurement of OCR by XF24 microplate-based respirometry from intact cells was made using an XF24 Extracellular Flux Analyzer (Seahorse Bioscience, MA). 20,000 cells per well were seeded in Seahorse 24-well XF Cell Culture Microplates and incubated overnight at 37°C with 5% CO₂. Mitochondria stress tests were initiated according to the manufacturer’s instructions by replacing the growth medium in each well with 500 μl of Seahorse XF Base Medium (supplemented with 10 mM glucose, 2 mM L-glutamine, and 1 mM sodium pyruvate for normal conditions, or without glucose only for glucose deprivation conditions) pre-warmed at 37°C. Glycolysis inhibitor 2DG (10 mM) was included in Seahorse XF Base media for some experiments. Cells were then incubated at 37 °C in a non-CO2 incubator for 60 min to allow temperature and pH equilibration. Following the equilibration period, cells were loaded into the XF24 extracellular flux analyzer and further equilibrated for 20 min prior to the first measurement. After an OCR baseline measurement, the minimum oxygen consumption was determined by injection of 4 μM oligomycin (Olig) and sequentially the maximal respiration rate was assessed upon injection of 0.5 μM FCCP (Trifluoromethoxy carbonylcyanide phenylhydrazone). Finally, the non-mitochondrial oxygen consumption was evaluated after injection of both 0.5 μM rotenone and antimycin (R/A). The measurements were normalized to cell number, which were quantified using CyQuant DNA dye (Invitrogen). Analysis of data was carried out using Wave software and the XF Mito/Glycolysis stress test report generator (Seahorse Bioscience).

Spautin1 (10 μM) and chloroquine (20 μM) were incubated with cells overnight before experiments. ETO, Rano or TMZ prepared in assay medium (75 μl) were preloaded into reagent delivery chambers. The assay medium final working concentration (200 μM ETO, 400 μM Rano, or 1 mM TMZ) was injected after FCCP and OCR was measured 3 times before injection of rotenone and antimycin.

Statistical analysis

Lengths, areas, and the number of cells from comparable sections were quantified using the ImageJ software package. Statistical significance was evaluated by Student’s t-test, one way ANOVA, two way ANOVA, and Pearson correlation coefficient, and differences between groups for frequency of SEN-like structures were performed using ELDA, as described previously⁶⁶, with p<0.05 as indicative of statistical significance using Graph Pad Prism (Version 5.0). The number of animals used for quantification was indicated in the figure legends.