Histone deacetylase inhibition expands cellular proteostasis repertoires to enhance neuronal stress resilience

Abstract

Neurons are long-lived, terminally differentiated cells with limited regenerative capacity. Cellular stressors such as endoplasmic reticulum (ER) protein folding stress and membrane trafficking stress accumulate as neurons age and accompany age-dependent neurodegeneration. Current strategies to improve neuronal resilience are focused on using factors to reprogram neurons or targeting specific proteostasis pathways. We discovered a different approach. In an unbiased screen for modifiers of neuronal membrane trafficking defects, we unexpectedly identified a role for histone deacetylases (HDACs) in limiting cellular flexibility in choosing cellular pathways to respond to diverse types of stress. Genetic or pharmacological inactivation of HDACs resulted in improved neuronal health in response to ER protein folding stress and endosomal membrane trafficking stress in C. elegans and mammalian neurons. Surprisingly, HDAC inhibition enabled neurons to activate latent proteostasis pathways tailored to the nature of the individual stress, instead of generalized transcriptional upregulation. These findings shape our understanding of neuronal stress responses and suggest new therapeutic strategies to enhance neuronal resilience.

Figure 1.. Deletion of sin-3 protects against rab-10 recycling stress and ire-1 folding stress.

(A) Schematic of a wild-type PVD dendrite (top) and routes of DMA-1 trafficking in the soma and primary dendrite (bottom). (B) Representative images of PVD morphology in wild-type, sin-3, rab-10, and ire-1 single mutants, and rab-10, sin-3 and ire-1, sin-3 double mutants. Scale bars are 50 µm. (C) Quantification of total dendritic length per 100 µm of primary dendrite for genotypes shown in B. (D) Representative images of endogenous DMA-1::GFP and endogenous mScarlet::RAB-7 in PVD somatodendrite region. Scale bars are 5 µm. (E) Representative images of endogenous DMA-1::GFP and endogenous mScarlet::SP12 in PVD soma. Scale bars are 5 µm. (F) Quantification of Pearson’s correlation coefficient, indicating degree of co-localization between DMA-1::GFP and mScarlet::RAB-7. (G) Quantification of Pearson’s correlation coefficient, indicating degree of co-localization between DMA-1::GFP and mScarlet::SP12. All statistical comparisons were performed using one-way ANOVA with Tukey’s correction for multiple comparisons.

Figure 2.. HDAC inhibition protects against ER stress across contexts.

(A) Representative images of PVD morphology in wild-type, ire-1 mutants, and double mutants of ire-1 with hda-3, arid-1, or athp-1. Scale bars are 50 µm. (B) Quantification of total dendritic length per 100µm of primary dendrite for genotypes shown in A and additional hda- mutants. Statistical comparison was performed using one-way ANOVA with Tukey’s correction for multiple comparisons. Wild-type and ire-1 quantification which first appear in Figure 1C are reproduced here as controls. (C) Schematic of tunicamycin and HDACi treatment of mouse cortical neurons. (D) Representative images of tubulin-βIII staining in mouse cortical neurons treated with or without tunicamycin and with or without apicidin or TSA. Axonal microtubules (MTs) were assessed after treatment. (E) Quantification of microtubule depolymerization index determined from tubulin-βIII staining. Scale bar, 100 µm.

Figure 3.. HDAC inhibition regulates transcriptional programs that protect against ER stress.

(A) Schematic of RNA sequencing experiment performed in mouse cortical neurons. (B) Principal component analysis of all drug treatment groups, displaying PC1 and PC2. (C) K-means clustering, k=10. Highlighted cluster indicates genes upregulated by Tunicamycin and further enhanced by HDAC inhibition. (D) Altered expression of ER stress genes in Tunicamycin vs. control neurons; red indicates ER stress genes with altered expression; dark grey indicates unchanged ER stress genes. (E) ER stress genes in neurons treated with either Tunicamycin and TSA or Tunicamycin alone.

Figure 4.. ER-associated degradation (ERAD) is required for the protective effect of HDAC inhibition.

(A) Representative images of PVD morphology in triple mutants between ire-1, sin-3 and the ERAD genes sel-1 or sel-11. Scale bars are 50 µm. (B) Quantification of total dendritic length per 100 µm of primary dendrite for genotypes shown in A. Quantification of wild-type, ire-1 single mutant, and ire-1, sin-3 double mutant, which first appear in Figure 1C, are reproduced here as controls. Quantification of the triple mutants is also summarized in the candidate screen shown in S4B. (C) Representative images of endogenous DMA-1::GFP in PVD soma. Scale bars are 5 µm. (D) Quantification of DMA-1::GFP intensity. All genotypes were normalized to the median of wild-type and data are plotted on a log-scale. (E) Schematic of siRNA experiment in mouse cortical neurons. (F) Cortical primary neurons treated with both tunicamycin and TSA after Syvn1 or non-targeting siRNA were added. Axonal microtubules (MTs) were assessed after treatment. (G) Quantification of microtubule depolymerization index determined from tubulin-βIII staining. Scale bar, 100 µm. All statistical comparisons were performed using one-way ANOVA with Tukey’s correction for multiple comparisons.

Figure 5.. Protein folding components are required for the protective effect of HDAC inhibition.

(A) Representative images of PVD morphology in triple mutants between ire-1, sin-3 and the folding-related genes manf-1, atf-6, and pek-1. Scale bars are 50 µm. (B) Quantification of total dendritic length per 100µm of primary dendrite for genotypes shown in A. Quantification of wild-type, ire-1 single mutant, and ire-1, sin-3 double mutant which first appear in Figure 1C are reproduced here as controls. Quantification of the triple mutants is also summarized in the candidate screen shown in S4B. (C) Representative images of endogenous DMA-1::GFP in PVD soma. Scale bars are 5 µm. (D) Quantification of DMA-1::GFP intensity. All genotypes were normalized to the median of wild-type and data are plotted on a log-scale. Quantification of DMA-1::GFP in wild-type, ire-1 mutant, and ire-1, sin-3 double mutant which first appear in Figure 4B are repeated here as controls. (E) Schematic of siRNA experiment in mouse cortical neurons. (F) Cortical primary neurons treated with both tunicamycin and TSA after Manf or non-targeting siRNA were added. Axonal microtubules (MTs) were assessed after treatment. (G) Quantification of microtubule depolymerization index determined from tubulin-βIII staining. Scale bar, 100 µm. All statistical comparisons were performed using one-way ANOVA with Tukey’s correction for multiple comparisons.

Figure 6.. RAB-11.1 recycling endosomes are required for the protective effect of HDAC inhibition.

(A) Representative images of endogenous GFP::RAB-5 (top), endogenous GFP::RAB-7 (middle), and endogenous GFP::RAB-11.1 (bottom) in wild-type, sin-3 and rab-10 mutants, and rab-10, sin-3 double mutants. Scale bars are 5 µm. (B) Quantification of GFP::RAB-5 intensity in PVD soma and dendrite for genotypes in A. (C) Quantification of GFP::RAB-7 intensity in the PVD soma and dendrite for genotypes in A. (D) Quantification of GFP::RAB-11.1 intensity in the PVD soma and dendrite for genotypes in A. (E) Representative images of PVD morphology with and without expression of RAB-11.1(DN). Scale bars are 50 µm. (F) Quantification of total dendritic length per 100 µm of primary dendrite for genotypes in D. Quantification of wild-type, rab-10 mutant, and rab-10, sin-3 double mutant that first appear in Figure 1C are repeated here as controls. (G) Representative images of endogenous DMA-1::GFP. Scale bars are 5 µm. (H) Quantification of DMA-1::GFP intensity for genotypes in H. (I) Representative images of endogenous GFP::RAB-11.1 localization in higher order dendrites. All statistical comparisons were performed using one-way ANOVA with Tukey’s correction for multiple comparisons.