TFEB ameliorates the impairment of the autophagy-lysosome pathway in neurons induced by doxorubicin

Abstract

Doxorubicin, a commonly used chemotherapy agent, induces severe cardio- and neurotoxicity. Molecular mechanisms of cardiotoxicity have been extensively studied, but mechanisms by which doxorubicin exhibits its neurotoxic properties remain unclear. Here, we show that doxorubicin impairs neuronal autophagy, leading to the accumulation of an autophagy substrate p62. Neurons treated with doxorubicin contained autophagosomes, damaged mitochondria, and lipid droplets. The brains from mice treated with pegylated liposomal doxorubicin exhibited autophagosomes, often with mitochondria, lipofuscin, and lipid droplets. Interestingly, lysosomes were less acidic in doxorubicin-treated neurons. Overexpression of the transcription factor EB (TFEB), which controls the autophagy-lysosome axis, increased survival of doxorubicin-treated neurons. 2-Hydroxypropyl-β-cyclodextrin (HPβCD), an activator of TFEB, also promoted neuronal survival, decreased the levels of p62, and lowered the pH in lysosomes. Taken together, substantial changes induced by doxorubicin contribute to neurotoxicity, cognitive disturbances in cancer patients and survivors, and accelerated brain aging. The TFEB pathway might be a new approach for mitigating damage of neuronal autophagy caused by doxorubicin.

Keywords: TFEB; autophagy; brain aging; chemotherapy; doxorubicin.

Figure 1. Doxorubicin induces autophagy in primary cortical neurons

(A) Autophagy is induced in cultured primary cortical neurons by 50 nM doxorubicin (overnight) as reflected by the increased levels of LC3-II. Tubulin was used as a loading control. LC3-II accumulated in neurons treated overnight with 50 nM doxorubicin or 5 μM 10-NCP (an autophagy enhancer as positive control) with or without 10 mM NH₄Cl, 4 h. LC3-II increased in doxorubicin-treated cells when NH₄Cl was added reflecting enhanced autophagic flux. (B) Measurements of the LC3-II bands from (A). The LC3-II intensities were normalized to the tubulin loading control. *p<0.01, **p<0.001, ***p<0.0001 (ANOVA). Results were pooled from three independent experiments. (C) Doxorubicin promotes the formation of pre-autophagosomal complexes as reflected by beclin1-GFP-positive puncta. Cortical neurons were transfected with mApple (a morphology and viability marker) and beclin1-GFP, and treated with a vehicle or 50 nM doxorubicin (overnight). Note changes in beclin1-GFP localization, consistent with beclin1 relocalization to pre-autophagosomal structures. White arrow points a beclin1-GFP-positive structure in a neurite. Bar, 20 μm. (D) To score autophagy induction, the redistribution of beclin1-GFP into puncta, is reflected by the fold-increase of puncta index, which is the standard deviation among pixels within the cellular region of interest. The puncta index significantly increased in doxorubicin-treated neurons. **p<0.001 (t-test). Two hundred neurons were analyzed from two independent experiments.

Figure 2. Levels of p62 increase in cultured cortical neurons treated with doxorubicin

(A) Autophagy was impaired by doxorubicin. Cortical neurons were transfected with mApple (a morphology and viability marker) and GFP-p62. The first neuronal cohort was treated with a vehicle, and the second cohort was treated with 50 nM doxorubicin (overnight). Neurons were imaged before and after the treatments. Small aggresomes are sometimes formed in neurons. Note large inclusion bodies formed by GFP-p62 in doxorubicin-treated neurons. Bar, 10 μm. (B) Quantification of fluorescent images from (A). The fold-increase of the puncta index in neurons, which express GFP-p62, treated with a vehicle or with 50 nM doxorubicin (overnight). *** p<0.0001 (t-test). Two hundred neurons were analyzed from two independent experiments. (C) Endogenous p62 accumulated in cultured cortical neurons treated with 50 nM doxorubicin (overnight). Actin was used as a loading control. (D) Quantification of western blots from (C). The levels of p62 were normalized to actin. *** p<0.0001 (t-test). Results were pooled from four independent experiments.

Figure 3. Accumulation of autophagosomes and organelles in primary cultured neurons and mouse brains induced by doxorubicin and doxil, respectively

(A) An example of an electron micrograph of cultured cortical neurons treated with a vehicle (overnight). M, mitochondria. Bar, 200 nm. (B) Electron micrographs of cultured cortical neurons treated with 50 nM doxorubicin (overnight). Left panel: A, autophagosomes. Bar, 200 nm. Right upper panel: A, an autophagosome. Bar, 100 nm. Right lower panel: L, lipid droplets. Bar, 500 nm. (C) An example of an electron micrograph of cultured cortical neurons treated with 50 nM doxorubicin (overnight). Note an abnormally large cluster of mitochondria. Arrows note atypical mitochondria. Bar, 200 nm. (D) Electron micrographs of mouse brain exposed to doxil. Left panels: A, autophagosomes. M, a mitochondrion being engulfed by an autophagosome. Bar, 200 nm. Right panels: L, lipid droplets; Lp, lipofuscin. Bar, 500 nm.

Figure 4. Doxorubicin raises the pH in lysosomes in cultured cortical neurons

(A) Primary cortical neurons were treated with a vehicle or 50 nM doxorubicin (overnight), and stained with a green lysotracker dye, which stains acidic organelles in live cells. Cultures were imaged, fixed, and stained with an antibody for MAP2c and with the nuclear dye, and imaged again to confirm that observed lysosomes were located in neurons. Bar, 20 μm. (B) Quantification of fluorescent images with the green lysotracker dye from (A). The green signal was less in cultures treated with doxorubicin than in controls. *** p<0.0001 (t-test). Three hundred neurons were analyzed. Results were pooled from three independent experiments.

Figure 5. TFEB increases survival of doxorubicin-treated neurons

(A) An example of longitudinal imaging and survival analysis. Primary cortical neurons transfected with mApple were followed with an automated microscope. Images collected every 24 h demonstrate the ability to return to the same field of neurons and track them over time. Each image is a montage of images captured in the center of one well of a 24- or 96-well plate. Scale bar is 400 μm. In the lower panel, images are zoomed in. Arrows represent two neurons that degenerate over time. Scale bar is 50 μm. (B) Two cohorts of primary cortical neurons transfected with mApple + GFP were treated either with a vehicle or 10 nM doxorubicin. Two cohorts transfected with mApple + TFEB-GFP were treated either with a vehicle or 10 nM doxorubicin. 16 hours after treatment, the four cohorts of neurons were imaged and tracked over 5 days. Risk of death associated with doxorubicin was calculated with JMP software. *p<0.01, ***p < 0.0001 (Log-Rank test), n.s., non-significant. One hundred fifty neurons were analyzed per condition. Results were pooled from two independent experiments.

Figure 6. Cyclodextrin reduces neuronal damage induced by exposure to doxorubicin

(A) Survival analysis of five neuronal cohorts transfected with mApple (a morphology and viability marker) were treated with a vehicle or with doxorubicin alone or in combination with HPβCD. The first neuronal cohort was treated with a vehicle. The second neuronal cohort was treated with 10 nM doxorubicin. The remaining three neuronal cohorts were treated with 10 nM doxorubicin in combination with 0.1 mM HPβCD (third neuronal cohort), 10 nM doxorubicin and 0.5 mM HPβCD (fourth neuronal cohort), and 10 nM doxorubicin and 1 mM HPβCD (fifth neuronal cohort). Neurons were tracked and imaged over 6 days. Risk of death associated with a treatment for each cohort was calculated with JMP software. *p<0.01, **p < 0.001 (Log-Rank test), n.s., non-significant. One hundred fifty neurons were analyzed per condition. Results were pooled from two independent experiments. (B) HPβCD does not prevent doxorubicin from binding to DNA. Cortical neurons were treated with 50 nM doxorubicin or with 50 nM doxorubicin and 0.5 mM HPβCD overnight, fixed, and stained for MAP2c (green) and with the nuclear Hoechst dye (blue), and imaged. Scale bar is 20 μm. (C) Images of fixed neurons from (B) were analyzed. ***p<0.0001 (t-test), n.s., non-significant. Three hundred neurons were analyzed per condition. Results were pooled from two independent experiments. (D) Three cohorts of cortical neurons were transfected with mApple and GFP-p62. The first neuronal cohort was treated overnight with a vehicle, the second cohort was treated with 50 nM doxorubicin (overnight), and the third cohort was co-treated with doxorubicin and 0.5 mM HPβCD (overnight). Quantification of fluorescent images revealed that HPβCD lowers the levels of GFP-p62 in doxorubicin-treated neurons. ***p < 0.0001 (t-test), n.s., non-significant. Two hundred neurons were analyzed. Results were from three independent experiments. (E) Cortical neurons were treated overnight with a vehicle or 50 nM doxorubicin or with a combination of doxorubicin and 0.5 mM HPβCD, and stained with a green lysotracker dye. HPβCD lowers the pH in in doxorubicin-treated neurons. ***p < 0.0001 (t-test), n.s., non-significant. Two hundred neurons were analyzed. Results were pooled from three independent experiments. (F) Electron micrographs of cultured cortical neurons co-treated with doxorubicin and 0.5 mM HPβCD (overnight). Left panel: Note that the cytoplasm lack the damage observed in neurons treated with doxorubicin alone. M, mitochondria. Bar, 1 μm. Right panel: Ls, lysosome; M, mitochondria. Bar, 500 nm.