Autophagy promotes cell survival by maintaining NAD levels

Abstract

Autophagy is an essential catabolic process that promotes the clearance of surplus or damaged intracellular components. Loss of autophagy in age-related human pathologies contributes to tissue degeneration through a poorly understood mechanism. Here, we identify an evolutionarily conserved role of autophagy from yeast to humans in the preservation of nicotinamide adenine dinucleotide (NAD) levels, which are critical for cell survival. In respiring mouse fibroblasts with autophagy deficiency, loss of mitochondrial quality control was found to trigger hyperactivation of stress responses mediated by NADases of PARP and Sirtuin families. Uncontrolled depletion of the NAD(H) pool by these enzymes ultimately contributed to mitochondrial membrane depolarization and cell death. Pharmacological and genetic interventions targeting several key elements of this cascade improved the survival of autophagy-deficient yeast, mouse fibroblasts, and human neurons. Our study provides a mechanistic link between autophagy and NAD metabolism and identifies targets for interventions in human diseases associated with autophagic, lysosomal, and mitochondrial dysfunction.

Keywords: DNA damage; NAD; PARP; Sirtuins; ageing; autophagy; metabolism; mitochondria; mitophagy.

Figure 1.. Metabolic deficit and apoptotic cell death in respiring autophagy-deficient cells

(A–C) Apoptotic cell death in respiring Atg5^−/− MEFs. Phase contrast images (A), cytotoxicity assays (B), and immunoblot analyses for caspase-3 cleavage (C) of Atg5^+/+ or Atg5^−/− MEFs cultured in glucose (glu) or galactose (gal) medium for 24 h (A and C) or 40 h (B). (D and E) Apoptotic cell death in respiring cells with the deletion of autophagy-related genes. Phase contrast images (D) and immunoblotting analyses for caspase-3 cleavage (E) of wild-type (WT) or isogenic Atg5^−/−, Atg7^−/−, and Rbcc1^−/− cell lines generated by the CRISPR-Cas9 system; and Npc1^+/+ and Npc1^−/− MEFs grown in gal medium for 110 h (Atg5, Atg7, and Rb1cc1 CRISPR-Cas9 generated cell lines) or 72 h (Npc1 cell lines). (F) Volcano plot representation of all analyzed metabolites in a pairwise comparison of Atg5^−/− to Atg5^+/+ MEFs after 16 h in gal medium. Thresholds are shown as dashed red lines. (G) List of discovered depleted metabolites in Atg5^−/− MEFs. Highlighted are metabolites that change significantly (−1 ≥ log₂(FC) ≥ 1, log₁₀[p adjusted > 1.3, highlighted in red]) and correlate with cell death/survival (bold). (H) Metabolite profiling in Atg5^+/+ and Atg5^−/− MEFs is depicted as a heatmap of log₂(FC) of Atg5^−/− to Atg5^+/+ MEFs based on their association to glucose oxidation pathways of glycolysis, pentose phosphate pathway (PPP), and tricarboxylic acid (TCA) cycle. (I) NAD(H) pool is depleted in respiring Atg5^−/− MEFs. Measurement of NAD⁺ and NADH levels in Atg5^+/+ and Atg5^−/− MEFs cultured in glu or gal medium for 20 h. Graphical data are mean ± SEM of n = 3 biological replicates (B, C, E, and I). p values were calculated by one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (B, C, and I), unpaired two-tailed Student’s t test (E) or multiple t test with the original FDR method of Benjamini and Hochberg (F and H) on three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant). Scale bars, 200 μm in (A) and (D). See also Figure S1.

Figure 2.. NAD(H) depletion contributes to cell death due to autophagy deficiency

(A) Graphical representation of the NAD⁺ synthesis pathways. Enzyme inhibitors are highlighted in red. (B–D) Boosting NAD(H) levels via salvage pathway rescues cell death in respiring Atg5^−/− MEFs. Measurement of NAD⁺ and NADH (B), cytotoxicity assays (C), and phase contrast images (D) in Atg5^+/+ and Atg5^−/− MEFs cultured for 20 h (B), 40 h (C), or 24 h (D) in gal medium supplemented with NAM or NR in the presence or absence of FK866. (E) Volcano plot representation of all analyzed metabolites in a pairwise comparison of Atg5^−/− + NAM to Atg5^−/− MEFs cultured in gal medium for 16 h. Thresholds are shown as dashed red lines. (F–H) Boosting NAD(H) levels via de novo pathway rescues cell death in respiring Atg5^−/− MEFs. Measurement of NAD⁺ and NADH (F), cytotoxicity assays (G), and phase contrast images (H) in Atg5^+/+ and Atg5^−/− MEFs cultured for 20 h (F), 40 h (G), or 24 h (H) in gal medium supplemented with L-tryptophan (Trp). (I–K) NAD supplementation rescues cell death in various cell lines with autophagy deficiency. Measurement of NAD⁺ and NADH (I), cytotoxicity assays (J), and phase contrast images (K) in isogenic (CRISPR KO) non-targeted control (WT), Atg5^−/−, Atg7^−/−, and Rbcc1^−/− cell lines generated by the CRISPR-Cas9 system, cultured for 96 h (I), 144 h (J), or 110 h (K) in gal medium supplemented with NAM. Graphical data are mean ± SEM of n = 3–4 biological replicates as indicated (B, C, F, G, I, and J). p values were calculated by one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (B, C, F, G, I, and J) or multiple t test with the original FDR method of Benjamini and Hochberg (E) on three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant) with respect to untreated Atg5^−/− MEFs or between the indicated groups. Scale bars, 200 μm in (D), (H), and (K). See also Figures S2 and S3.

Figure 3.. NAD(H) depletion due to the hyperactivation of NADases in autophagy-deficient cells

(A and B) Kinetic hyperactivation of NADases in respiring Atg5^−/− MEFs. Immunoblot analyses for poly(ADP-ribose) (PAR), acetylated lysine (AcK), acetylated p53 (Ac-p53), and total p53 as indicators of PARP and SIRT activities in Atg5^+/+ and Atg5^−/− MEFs cultured in gal medium for 14 h (A) or 20 h (B). (C–H) Chemical or genetic inhibition of NADases activity rescues NAD(H) levels and cell death in respiring Atg5^−/− MEFs. Measurement of NAD⁺ and NADH levels (C and F), cytotoxicity assays (D and G), and phase contrast images (E and H) in Atg5^+/+ and Atg5^−/− MEFs treated with sirtinol (SIRTi), olaparib, (PARPi) or solvent (DMSO); and in Atg5^+/+ and Atg5^−/− MEFs transfected with Control, Sirt1, or Parp1 siRNA, after 20 h (C and F), 40 h (D and G), or 24 h (E and H) culture in gal medium. Graphical data are mean ± SEM of n = 3–4 biological replicates as indicated (A–D, F, and G). p values were calculated by unpaired two-tailed Student’s t test (A and B) or one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (C, D, F, and G) on three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001 with respect to untreated Atg5^−/− MEFs or between the indicated groups. Scale bars, 200 μm in (E) and (H). See also Figure S3.

Figure 4.. Loss of mitochondrial quality control contributes to the depletion of NAD(H) and cell death

(A and B) Mitochondrial ROS and DNA damage are induced in respiring Atg5^−/− MEFs. Fluorescence microscopy images and quantification of MitoSOX staining (A) or γH2AX immunostaining (B) of Atg5^+/+ or Atg5^−/− MEFs cultured in glu or gal medium for 24 h. (C) Transmission electron micrographs of Atg5^+/+ or Atg5^−/− MEFs cultured in glu or gal medium for 24 h. White arrows: healthy mitochondria; red arrows: mitochondria with aberrant morphology. (D and E) Mitophagy defect in Atg5^−/− MEFs and PentaKO HeLa cells. Fluorescence microscopy images and quantification of mitophagy of Atg5^+/+ or Atg5^−/− MEFs (D) and WT or PentaKO HeLa cells (E), expressing mt-mKeima and cultured in glu or gal medium for 24 h (D) or 96 h (E). (F) Autophagy is activated in respiring PentaKO cells. Confocal fluorescence microscopy images and quantification of autophagosomes and autolysosomes in WT or PentaKO HeLa cells expressing mRFP-GFP tandem fluorescent-tagged LC3 (tfLC3) cultured in glu or gal medium for 96 h. (G–I) Loss of mitophagy contributes to NAD(H) depletion and cell death in respiring PentaKO HeLa cells. Measurement of NAD⁺ and NADH (G), cytotoxicity assays (H), and phase contrast images (I). (G) in WT or Penta KO HeLa cells cultured for 144 h (H), 120 h (I), or 110 h (G) in gal medium supplemented with 1 mM NAM. (J) Mitophagy defect is upstream of NAD(H) depletion. Fluorescence microscopy images and quantification of mitophagy in WT or PentaKO HeLa cells expressing mt-mKeima and cultured for 96 h in gal medium supplemented with 1 mM NAM. Data are mean ± SEM (A, C–H, and J) or displayed as cell popular violin plots (B, D–F, and J). p values were calculated by one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli on three independent experiments (A–H and J). *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant). Scale bars: 10 μm in (A); 20 μm in (B), (D)–(F), and (J); 500 nm in (C); and 200 μm in (I). See also Figure S4.

Figure 5.. Loss of PINK1 supresses mitophagy activated by galactose media culture and recapitulates NAD(H) depletion and cell death

(A) Fluorescence microscopy images and quantification of mitophagy of isogenic non-targeted control (CRISPR WT) or Pink1^−/− (CRISPR Pink1^−/−) MEFs generated by the CRISPR-Cas9 system, expressing mt-mKeima and YFP-Parkin, cultured in glu or gal medium for 96 h. (B) Confocal fluorescence microscopy images and quantification of autophagosomes and autolysosomes in CRISPR WT or Pink1^−/− MEFs expressing tfLC3 cultured in glu or gal medium for 96 h. (C–E) Measurement of NAD⁺ and NADH (C), cytotoxicity assays (D), and phase contrast images (E) of CRISPR WT or Pink1^−/− MEFs cultured for 134 h (C), 144 h (D), or 168 h (E) in gal medium supplemented with 5 mM NAM. (F) Fluorescence microscopy images and quantification of mitophagy in CRISPR WT or Pink1^−/− MEFs expressing mt-mKeima and YFP-Parkin cultured for 96 h in gal medium supplemented with 5 mM NAM. Data are mean ± SEM (A–D and F) or displayed as cell popular violin plots (A, B, and F). p values were calculated by one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli on three independent experiments (A–D and F). *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant). Scale bars: 20 μm in (A), (B), and (F) and 200 μm in (E). See also Figure S5.

Figure 6.. Respiration via dysfunctional mitochondrial CI is upstream of NAD(H) depletion and loss of viability in autophagy-deficient cells

(A–D) Mitochondrial CI is dysfunctional in Atg5^−/− MEFs. (A–C) CI- and CII-linked respiration were assessed by measuring oxygen consumption rate (OCR) in permeabilized Atg5^+/+ and Atg5^−/− MEFs in assay buffer supplemented with either CI substrates, pyruvate, and malate (P+M) (A) or CII substrate and succinate (S) (B). Oligomycin, FCCP, and animycin A were added at the indicated times during OCR measurements (A and B). Respiratory control ratios (RCRs) indicate a capacity for substrate oxidation and ATP turnover (C). (D) Respirometry analysis of Atg5^+/+ and Atg5^−/− MEFs cultured in gal medium for 20 h was performed by sequential additions of CI substrates P+M, ADP, and S into suspension of 1 million permeabilized cells with digitonin. (E–J) Rotenone causes NAD(H) depletion and cell death in respiring wild-type cells. Measurement of NAD⁺ and NADH (E and H), cytotoxicity assays (F and I), and phase contrast images (G and J) in Atg5^+/+ MEFs cultured for 6 h (E and H), 12 h (F and I), and 9 h (G and J) in glu or gal medium supplemented with 1 μM rotenone (RTN) in the presence or absence of 5 mM NAM. Cells were pre-treated with NAM for 16 h prior to RTN treatment (H–J). (K–P) Chemical interventions in CI dysfunction, mitochondrial ROS, and boosting CII activity rescue NAD(H) and cell death phenotypes. Measurement of NAD⁺ and NADH (K and N), cytotoxicity assays (L and O), and phase contrast images (M and P) in Atg5^+/+ and Atg5^−/− MEFs cultured for 20 h (K and N), 40 h (L and O), or 24 h (M and P) in gal medium supplemented with CP2, S1QEL2.2 (S1QEL), MitoQ, or N-acetyl-L-cysteine (NAC) (K–M) or with dimethyl succinate (DS) (N–P). Graphical data are mean ± SEM of biological replicates (n = 3) (A and B) or more replicates as indicated (C–F, H, I, K, L, N, and O). p values were calculated by multiple t test (C and D) or one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (E, F, H, I, K, L, N, and O) on three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant) with respect to untreated Atg5^−/− MEFs or between the indicated groups. Scale bars, 200 μm in (G), (J), (M), and (P). See also Figure S6.

Figure 7.. Depletion of mitochondrial NADH and membrane potential mediates cell death due to autophagy deficiency

(A) Mitochondrial NADH is depleted in respiring Atg5^−/− MEFs. Measurement of NAD⁺ and NADH levels in cytoplasmic (cyto) and mitochondrial (mito) fractions of Atg5^+/+ and Atg5^−/− MEFs after 20 h culture in gal medium. (B and C) Mitochondrial depolarization in respiring Atg5^−/− MEFs is rescued by NAD supplementation. Confocal fluorescence images of live cells after 20 h culture in gal medium supplemented with NAM or oligomycin (Oligo) and co-stained with TMRM and MTG (B). ΔΨm quantified as a ratio of TMRM to MTG (C). (D–F) Mitochondrial hyperpolarization by oligomycin rescues cell death with NADH but not NAD⁺ restoration in respiring Atg5^−/− MEFs. Measurement of NAD⁺ and NADH levels (D), cytotoxicity assays (E), and phase contrast images (F) in Atg5^+/+ and Atg5^−/− MEFs treated with Oligo or solvent (DMSO) after 20 h (D), 40 h (E), or 24 h (F) culture in gal medium. (G–K) Boosting NADH consumption by alternative NADH dehydrogenase NDI1 augments mitochondrial depolarization and cell death phenotypes in respiring Atg5^−/− MEFs. Confocal fluorescence images of live cells stained with TMRM (G), ΔΨm quantified as the relative intensity of TMRM in GFP positive cells (H), measurement of NAD⁺ and NADH levels (I), cytotoxicity assays (J), and phase contrast images (K) in Atg5^+/+ and Atg5^−/− MEFs stably expressing NDI1-IRES-GFP (NDI1) or GFP after 20 h (G–I), 40 h (J), or 24 h (K) culture in gal medium. Graphical data are mean ± SEM of n = 3–4 biological replicates as indicated (A, C–E, and H–J). p values were calculated by unpaired two-tailed Student’s t test (A) or one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (C–E and H–J) on three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant) with respect to untreated Atg5^−/− MEFs or between the indicated groups. Scale bars: 20 μm in (B) and (G) and 200 μm in (F) and (K). See also Figures S6 and S7.

Figure 8.. Evolutionarily conserved role of autophagy in the maintenance of NAD(H)

(A and B) Spot-testing (5-fold serial dilutions) of S288C WT and atg5Δ yeast strains on nutrient-rich agar media (A) and measurement of NAD⁺ and NADH levels (B) following 5 days (A) or 3 days (B) of nitrogen starvation in the presence of 10 mM NAM or solvent (H₂O). Image is a representative of n = 4 experiments. (C and D) Representative images (overlay of DIC and fluorescence images) (C) and quantification of cell viability (D) of BY4741 (WT), ScPPS2 (atg5Δ), and sSUN99 (atg1Δ) yeast strains cultured in SD-N medium supplemented with 10 mM NAM for 6 days (C) or the indicated times (D) followed by Phloxine B staining. (E–G) Fluorescence images (E), quantification of TUNEL⁺ apoptotic nuclei in TUJ1⁺ cells (F), and measurement of NAD⁺ and NADH levels (G) in control (Ctrl_#13) and NPC1 (NPC1–2_#26) iPSC-derived neurons after 3 (F and G) or 4 (F) weeks of neuronal differentiation, where NPC1 neurons were treated with or without NAM for the last 6 days. (H) Schematic representation of the mechanism of NAD(H) depletion leading to mitochondrial depolarization and apoptosis in respiring autophagy-deficient cells. SARs, selective autophagy receptors. Graphical data are mean ± SEM of n = 3 (B, D, and F) or 4 (G) biological replicates. p values were calculated by unpaired two-tailed Student’s t test (B and D within dataset on day 6) or one-way ANOVA followed by multiple comparisons with the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli (F and G) on three or four independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001; ns (non-significant) with respect to untreated atg5Δ yeast or between the indicated groups. ^§p < 0.05 with respect to atg1Δ yeast. Scale bars: 5 μm in (C) and 100 μm in (E). See also Figure S8.