A small-molecule scaffold induces autophagy in primary neurons and protects against toxicity in a Huntington disease model

Abstract

Autophagy is an intracellular turnover pathway. It has special relevance for neurodegenerative proteinopathies, such as Alzheimer disease, Parkinson disease, and Huntington disease (HD), which are characterized by the accumulation of misfolded proteins. Although induction of autophagy enhances clearance of misfolded protein and has therefore been suggested as a therapy for proteinopathies, neurons appear to be less responsive to classic autophagy inducers than nonneuronal cells. Searching for improved inducers of neuronal autophagy, we discovered an N(10)-substituted phenoxazine that, at proper doses, potently and safely up-regulated autophagy in neurons in an Akt- and mTOR-independent fashion. In a neuron model of HD, this compound was neuroprotective and decreased the accumulation of diffuse and aggregated misfolded protein. A structure/activity analysis with structurally similar compounds approved by the US Food and Drug Administration revealed a defined pharmacophore for inducing neuronal autophagy. This pharmacophore should prove useful in studying autophagy in neurons and in developing therapies for neurodegenerative proteinopathies.

Fig. 1.

N¹⁰-substituted phenoxazine (10-NCP) induces autophagy in primary neurons. (A) Autophagy was efficiently induced in striatal, cortical, and hippocampal neurons by 10 μM 10-NCP as reflected by the increased levels of LC3-II. Actin was used as a loading control throughout this figure. *P < 0.001 (pairwise t test). norm., normalized. (B) Autophagy was efficiently induced in striatal neurons by 0.5, 1, 5, 10, and 25 μM 10-NCP (overnight). *P < 0.01 (ANOVA). The slight increase in LC3-I levels represents up-regulation of the LC3 gene. N.S., not significant. (C) LC3-II accumulation is noticeable 1 h after treatment with 10 μM 10-NCP. LC3-II levels peak at ∼4 h in cultured striatal neurons. *P < 0.01 (ANOVA). (D) LC3-II accumulation in striatal neurons treated with 5 μM 10-NCP with or without 1 nM bafilomycin A (10-NCP, overnight; bafilomycin A was then added for 4 h). Bafilomycin A reached a ceiling effect, and higher concentrations (e.g., 10 nM) did not further increase LC3-II levels. LC3-II increased in 10-NCP–treated cells when bafilomycin A was added. The film was exposed to encompass all signals fully but especially to demonstrate the differences in LC3-II levels. *P < 0.001 (ANOVA). (E) Striatal neuron from GFP-LC3 transgenic mice expressing mCherry before (Left) and after (Right) treatment with 10-NCP (10 μM, 4 h). Note the changes in GFP-LC3 localization, which were consistent with GFP-LC3 relocalization to autophagosomes. (Scale bar, 50 μm.) (F) Electron micrographs of striatal neurons treated with 10-NCP (10 μM, overnight) (0- and 1-μM conditions are shown in Fig. S2). A, autophagosomes; Al, autolysosome; C, curving phagophores; E, extending phagophore; L, lysosome with degraded contents. (Scale bar, 0.6 μm.) Quantification of autophagic structures in neurons treated with 0, 1, and 10 μM 10-NCP is shown in Fig. S2.

Fig. 2.

10-NCP is not toxic to neurons, decreases diffuse and aggregated forms of mutant Htt^ex1 protein, and is protective in a neuron model of HD. (A) Cumulative risk for death associated with 10-NCP treatment of GFP-transfected striatal neurons. Treatment with low-dose (1 μM, weak stimulation of autophagy) 10-NCP led to a small decrease in baseline risk for death. A higher dose (10 μM, overstimulated autophagy) led to increased cell death. Note differences in LC3-II levels in neurons treated with 1 and 10 μM 10-NCP in Fig. 1B. *P < 0.001; **P < 0.0001 (Mantel–Cox test). (B) Striatal neurons transfected with mCherry and mutant Htt^ex1-GFP were treated with 1 μM 10-NCP or vehicle. Cumulative risk for death was calculated from Kaplan–Meier curves. 10-NCP reduced the risk for death (i.e., improved survival) of neurons expressing mutant Htt^ex1. *P < 0.001 (Mantel–Cox test). (C) Fraction of neurons that formed IBs and subsequently cleared IBs in the presence of vehicle, 1 μM 10-NCP, and 1 nM bafilomycin A (bafA). Neurons were transfected as in B and were followed with an automated microscope for several days. *P < 0.001 (ANOVA). (D) IB (arrow) that spontaneously disappeared during the experiments in C. Numbers reflect hours after transfection (Fig. S3). (Scale bar, 15 μm.) (E) Two cohorts of neurons transfected and treated as in A were monitored longitudinally. The fraction of surviving neurons with IBs was greater after treatment with vehicle than with 10-NCP. *P < 0.001 (t test). (F) Cumulative risk for IB formation in neurons transfected and treated as in A shows that IBs form more readily in untreated neurons. *P < 0.0001 (Mantel–Cox test). (G) Levels of diffuse mutant Htt and mCherry from neurons transfected and treated as in B. After exclusion of neurons with IBs, GFP or mCherry levels were normalized (Norm.) to fluorescence at the first time point. The levels of cotransfected mCherry did not change. *P < 0.01 (ANOVA). N.S., not significant.

Fig. 3.

Structural scaffold that induces neuronal autophagy. Chemical structures of compounds that induced autophagy in striatal, cortical, and hippocampal neurons. Note the base ring structure (or a biphenyl ring system for pimozide, niguldipine, loperamide, and fluspirilene) with a 2–5 carbon aliphatic substituent linker that contains a secondary or tertiary amine. Phenoxazine, phenothiazine, and quinacrine failed to induce autophagy and variably lacked the amino-capped substituent, had a 4-aminopyridine central ring, or had a strictly planar tricyclic ring system. Potency of the drugs was assayed in striatal neurons using the levels of LC3-II induction for each compound at 0.5, 1, and 5 μM as a semiquantitative readout (Fig. S6). Ranked potencies fell into five groups: +++++, 10-NCP and trifluoperazine (strongest stimulators); ++++, promazine, promethazine, chlorpromazine, and triflupromazine; +++, mesoridazine and thioridazine (medium stimulators); ++, niguldipine, loperamide, pimozide, and fluspirilene; +, nortriptyline (weakest stimulator). −, Drugs that do not induce autophagy.