Impact of autophagy inducers on the NF-κB activation pathway. (A) Identification of IKK inhibitors as autophagy inhibitors in an unbiased screen. Levels of autophagy inhibition of ICCB compounds on nutrient deprivation in U2OS cells stably expressing GFP–LC3 and RFP–H2B were measured at 12 h after compound addition as described in Material and methods section (B–D). Effects of (B) starvation, (C) rapamycin or (D) pifithrin-α on IκBα levels and the phosphorylation status of the IKK complex. HeLa cells were exposed to autophagy inducers for the indicated time, followed by immunoblotting for the determination of the indicated proteins. The percentage values illustrate the fraction of cells that manifested a punctuated distribution of GFP–LC3 (% GFP–LC3^vac cells) after addition of the indicated agents, as assessed by immunofluorescence microscopy. (E–G) Determination of the enzymatic activity of the IKK complex. HeLa cells treated as in (B–D) were subjected to lysis and immunoprecipitation with an anti-NEMO antibody. Immunoprecipitates were tested for their capacity to phosphorylate in vitro a semi-synthetic derivative of IκBα, GST–IκBα (1–54). Lysates from cells treated with TNFα for 15 min were used as positive controls for IKK (and NF-κB activation). (H–J) NF-κB activation status. EMSA was used to determine the DNA-binding ability of NF-κB in cells treated as in (B–D). Results are representative of three independent experiments.

Autophagy induction by IKK can be uncoupled from NF-κB activation. (A–C) Effects of IκSR on NF-κB activation and autophagy induction by IKK, as determined by immunofluorescence microscopy. Cells were transfected simultaneously with a construct for the expression of GFP–LC3 plus the pcDNA3.1 vector or wild type (WT) IKKβ-, CA-IKKβ-, WT NEMO- or MN-NEMO-encoding plasmids alone or in combination with a plasmid that encodes IκSR, followed by immunofluorescence staining for the detection of p65. Representative microphotographs are shown in (A), whereas the percentage of transfected cells showing nuclear p65 translocation (% p65^nuc cells) or cytoplasmic GFP–LC3 aggregation (% GFP-LC3^vac cells) are shown in (B, C), respectively (mean±s.e.m., n=3, ^*P<0.05). (D) Effects of the IκB ‘superrepressor' (IκSR) on NF-κB activation and autophagy induction by IKK, as assessed by immunoblotting. HeLa cells were transfected with the pcDNA3.1 vector or with WT IKKβ-, CA-IKKβ-, WT NEMO- or MN-NEMO-encoding plasmids alone or in association with a construct for the expression of IκSR for 18 h, followed by lysis and processing for immunoblotting determinations. The maturation of LC3 was assessed in whole-cell extracts. Alternatively, lysates were subjected to subcellular fractionation to evaluate the presence of p65/RelA in the cytoplasm versus the nucleus, using GAPDH and TFIIA-α to control equal loading of cytoplasmic and nuclear fractions, respectively. (E, F) Knockout of p65/RelA fails to affect autophagy. WT and p65^−/− murine NIH-3T3 fibroblasts were exposed to starvation conditions for the indicated time, followed by immunoblotting assessment of LC3 maturation (E). Alternatively, cells were co-transfected with a GFP–LC3-encoding construct plus the pcDNA3.1 vector or the indicated IKK-relevant plasmids for 18 h, followed by immunofluorescence microscopy for the quantification of GFP–LC3^vac cells (mean±s.e.m., n=3, ^*P<0.05) (F).

Implication and functional impact of the IKK complex in the induction of autophagy in vitro. (A, B) Contribution of IKK subunits to autophagy in cultured cells. MEFs with the indicated genotype were transfected for 12 h with a GFP–LC3-encoding plasmid, cultured for 4 h in the presence of the indicated autophagy inducers, and eventually analysed by immunofluorescence microscopy for GFP–LC3 puncta. (B) Alternatively, HeLa cells were transfected with a control siRNA (Co.) or with siRNAs targeting the indicated proteins for 24 h, followed by transfection with the GFP–LC3-encoding plasmid, autophagy induction and immunofluorescence microscopy as in (A). In (A, B), columns depict the percentage of cells exhibiting GFP–LC3 in cytoplasmic dots (% GFP–LC3^vac cells, mean±s.e.m., n=3, ^*P<0.05), whereas inserts illustrate the efficacy of knockout or knockdown, as assessed by immunoblotting. (C, D) Impact of IKK constituents and autophagy-relevant proteins on cell viability in conditions of autophagy induction and metabolic stress. (C) MEFs of the indicated genotype were transfected for 24 h with a control siRNA (Co.) or with siRNAs targeting the indicated autophagy-relevant proteins, then subjected to metabolic stress (glucose deprivation, 0.1% oxygen) for 48 h and stained for cytofluorometric determination of apoptosis-associated parameters. Black and white or grey columns illustrate the percentage of cells (mean±s.e.m., n=3, ^*P<0.05) characterized by plasma membrane breakdown (PI⁺) or loss of ΔΨ_m (DiOC₆(3)^low/PI⁻), respectively. (D) MEFs characterized by the indicated genetic deficiency were subjected to metabolic stress and plated to determine clonogenic survival. Column depict mean clonogenic survival±s.e.m. (n=3, ^*P<0.05).

Activation of autophagy by IKK. (A, B) Induction of GFP–LC3 aggregation by CA-IKKβ or MN-NEMO. Human cervical carcinoma HeLa cells were co-transfected for 12 h with a GFP–LC3-encoding plasmid plus the pcDNA3.1 vector or constructs for the overexpression of wild type (WT) IKKβ, CA-IKKβ, WT NEMO or MN-NEMO, followed by immunofluorescence microscopy assessments of p62 expression. When indicated, the IKK inhibitor BAY11-7082 was added during the last 6 h of the experiment. In (A), representative immunofluorescence microphotographs are shown. The percentage of cells with GFP–LC3 puncta (% GFP–LC3^vac cells) is reported in (B) (mean±s.e.m., n=3, ^*P<0.05). (C) Maturation of LC3 and depletion of p62 determined by immunoblotting. HeLa cells were transfected with the pcDNA3.1 vector or with IKKβ-, CA-IKKβ-, NEMO- or MN-NEMO-encoding plasmids (with a transfection efficiency of ∼70%) for 12 h, then processed for immunoblotting as described in Materials and methods section. (D, E) Transmission electron microscopy detection of autophagic vacuoles. HeLa cells were transfected as in (C) and then processed for transmission electron microscopy. Representative images are shown in (D) and the number of autophagosomes (white columns) and autophagolysosomes (black columns) per cell (mean±s.e.m., n=30–40 cells, ^*P<0.05) is depicted in (E). (F) Requirement of autophagy-relevant proteins for GFP–LC3 aggregation induced by CA-IKKβ or MN-NEMO. HeLa cells were transfected with a control siRNA (Co.) or with siRNAs targeting the indicated autophagy-relevant proteins for 24 h, and then with a GFP–LC3-endoding plasmid plus the indicated IKK-relevant constructs for additional 12 h, followed by immunofluorescence microscopy to quantify the percentage of cells exhibiting GFP–LC3 aggregation (% GFP–LC3^vac cells, mean±s.e.m., n=3, ^*P<0.05).

Implication of AMPK, mTOR, p53 and JNK1 in autophagy induction by IKK activation. (A) Activation of AMPK, inhibition of mTOR and dissociation of the Beclin-1–Bcl-2 complex in response to IKK activation. HeLa cells were subjected to starvation conditions for 1 h or transfected with the pcDNA3.1 vector or wild type (WT) IKKβ-, CA-IKKβ-, WT NEMO- or MN-NEMO-encoding plasmids for 18 h, followed by immunoblotting detection of the indicated proteins. Alternatively, lysates were immunoprecipitated with a Beclin-1-specific antibody and then subjected to immunoblotting detection of Beclin-1 or Bcl-2. (B) Involvement of AMPK and mTOR in IKK-induced autophagy. Before transfection with a GFP–LC3-encoding construct plus the indicated plasmids for 12 h, HeLa cells were transfected for 48 h with a control siRNA (Co.) or an AMPKα-specific siRNA. Alternatively, 12 h after plasmid transfection alone, cells were treated with rapamycin or with an equivalent volume of DMSO (solvent control) for 6 h. Columns depict the percentage of cells exhibiting GFP–LC3 aggregation (% GFP–LC3^vac cells, mean±s.e.m., n=3, ^*P<0.05). (C) Involvement of p53 degradation in IKK-induced autophagy. MEFs were transfected with a control (Co.) or an MDM2-specific siRNA for 24 h, then transfected again with a plasmid encoding GFP–LC3 construct plus the indicated constructs for 6 h. As an alternative, 6 h after plasmid transfection alone cells were treated with Nutlin-3 or RITA for additional 12 h. Finally, cells were subjected to immunofluorescence microscopy for the quantification of GFP–LC3 aggregation. Columns report the % of GFP–LC3^vac cells (mean±s.e.m., n=3, ^*P<0.05). (D, E) Kinome analysis for the detection of serine/threonine kinase substrates phosphorylation triggered by autophagic stimuli. CelluSpot serine/threonine-kinase substrate microarrays were incubated for with extracts from cells that were transfected with the pcDNA3.1 vector or with CA-IKKβ- or MN-NEMO-encoding plasmids for 18 h. As an alternative, extracts were obtained from cells subjected to starvation conditions or treated with rapamycin or pifithrin-α for 2 h. Phosphorylation was revealed by means of fluorescence-labelled anti-phosphoserine/threonine antibodies. The dotted line in (D) highlights the position of a synthetic JNK1 substrate which is phosphorylated by extracts from cells undergoing autophagy. Panel (E) depicts the fold-induction of this signal (mean±s.e.m., n=2, ^*P<0.05). (F) CA-IKKβ and MN-NEMO induce JNK1 phosphorylation. Extracts from HeLa cells transfected or treated as in (A) were subjected to immunoblotting detection of phosphorylated (P-) and total JNK1. GAPDH levels were monitored to ensure equal loading. (G, H) Effect of JNK1 depletion/inhibition on IKK-induced autophagy. Cells were transfected with a control (Co.) or a JNK1-specific siRNA for 24 h, followed by transfection with the indicated constructs for additional 12 h and immunoblotting assessment of the maturation of LC3 and of the depletion of p62 (G). Cells were left untreated or transfected with the indicated siRNAs for 24 h, followed by further transfection with a construct encoding GFP–LC3 plus the indicated plasmids for 12 h. Alternatively, cell were pre-treated for 30 min with the indicated JNK1 inhibitors (or with an equal volume of DMSO) and then co-transfected with a GFP–LC3-encoding construct plus the indicated plasmids for additional 12 h (H). Columns depict the % of GFP–LC3^vac cells (mean±s.e.m., n=3, ^*P<0.05).

Implication of the IKK complex in the induction of autophagy in vivo. (A–D) Mice with the indicated genotype were subjected to food deprivation for 24 h or injected i.p. with rapamycin, followed by necropsy. Livers (A, C) and other organs (C) were subjected to the immunohistochemical detection of LC3 dots in the cytoplasm of parenchymatous cells. Columns illustrate the number of LC3 dots per mm² (mean±s.e.m., n=3, ^*P<0.05) (C). Alternatively, tissue sections were analysed by transmission electron microscopy for the detection of autophagosomes and autolysosomes (B), and parenchimal cells were processed for the detection of LC3 maturation, p53 and IKKβ (D). (E) Blockade of upstream autophagic signals in IKKβ-deficient hepatocytes. Livers and spleens from control or IKKβ^Δhep mice that had been housed or not in starvation conditions for 24 h were processed for the immunoblotting assessment of the indicated proteins. Results are representative of three independent determinations.