Context-Specific Stress Causes Compartmentalized SARM1 Activation and Local Degeneration in Cortical Neurons

Abstract

Sterile alpha and TIR motif containing 1 (SARM1) is an inducible NADase that localizes to mitochondria throughout neurons and senses metabolic changes that occur after injury. Minimal proteomic changes are observed upon either SARM1 depletion or activation, suggesting that SARM1 does not exert broad effects on neuronal protein homeostasis. However, whether SARM1 activation occurs throughout the neuron in response to injury and cell stress remains largely unknown. Using a semiautomated imaging pipeline and a custom-built deep learning scoring algorithm, we studied degeneration in both mixed-sex mouse primary cortical neurons and male human-induced pluripotent stem cell-derived cortical neurons in response to a number of different stressors. We show that SARM1 activation is differentially restricted to specific neuronal compartments depending on the stressor. Cortical neurons undergo SARM1-dependent axon degeneration after mechanical transection, and SARM1 activation is limited to the axonal compartment distal to the injury site. However, global SARM1 activation following vacor treatment causes both cell body and axon degeneration. Context-specific stressors, such as microtubule dysfunction and mitochondrial stress, induce axonal SARM1 activation leading to SARM1-dependent axon degeneration and SARM1-independent cell body death. Our data reveal that compartment-specific SARM1-mediated death signaling is dependent on the type of injury and cellular stressor.

Keywords: SARM1; axon; cell death; neurodegeneration.

Figure 1.

SARM1-dependent axon degeneration in cortical neurons. a, Schematic of spot culture used for axon degeneration imaging after injury. b, Sample images and (c) quantification of axon degeneration in WT and SARM1 KO mouse cortical neurons after injury (n = 49–79 images/condition; Mann–Whitney test; ***p < 0.001 for all time points after 2 h). See Extended Data Figures 1-1–1-3 for additional information on the axon degeneration scoring algorithm. Quantification of axon degeneration after treatment with 10 µM SARM1 inhibitor or DMSO in either mouse (n = 14–20 images/condition; Mann–Whitney test, **p < 0.01, ***p < 0.001; d) or human (n = 10–16 images/condition; Mann–Whitney test; ***p < 0.001; e) cortical neurons. Sample images (f) and quantification (g) of neurite regrowth after injury in WT or SARM1 KO cortical neurons treated with DMSO or 10 nM taxol (n = 50–62 images/condition; two-way ANOVA, multiple comparisons; ***p < 0.001). Scale bars, 100 µm. See Extended Data Table 1-1 for further statistical analysis.

Figure 2.

SARM1 is activated in axons distal, but not proximal, to injury site. a, Schematic of hydrolysis of NAD⁺ to NAM and ADPR/cADPR or base-exchange reaction of NAD⁺ and 1a to NAM and fluorescent probe AD-1a catalyzed by SARM1. b, AD-1a autofluorescence with pH in different buffer conditions. c, Sample images and (d) quantification of AD-1a fluorescent in axons distal to the injury site 4 h after injury in WT or SARM1 KO mouse cortical neurons (n = 44–75 images/condition; Tukey's multiple-comparison test; ****p < 0.0001). Scale bar, 20 µm. Sample images (e) and quantification (n = 30 images/condition; Mann–Whitney test; ***p < 0.001 for all time points after 60 min; f) of time course of AD-1a fluorescence in axons after injury (top) or control (bottom) in WT mouse cortical neurons. Scale bar, 20 µm. g, AD-1a fluorescence proximal and distal to the injury site in WT mouse cortical neurons 4 h after injury. Scale bar, 50 µm. See Extended Data Figure 2-1 for uncropped images.

Figure 3.

Vacor selectively activates SARM1 to induce axon degeneration in mouse cortical neurons. a, Schematic of SARM1 activation and downstream events. Quantification of axon degeneration after treatment of WT or SARM1 KO mouse cortical neurons with (b) 100 µM CZ-48 and 2.5 nM FK866 (n = 21–22 images/condition; Mann–Whitney test; **p < 0.01; ***p < 0.001) or (c) 5 µM vacor (n = 18 images/condition; Mann–Whitney test; ***p < 0.001 for all time points after 6 h). See Extended Data Figure 3-1 for data on cotreatment of mouse cortical cultures with vacor and FK866. d, Quantification of axon degeneration after treatment of WT mouse cortical neurons with 5 µM vacor in the presence of 10 µM SARM1 inhibitor (n = 12–18 images/condition; Mann–Whitney test; ***p < 0.001 for all time points after 6 h). See Extended Data Figure 3-1 for data on vacor treatment of human iPSC-derived neurons. e, Sample images of AD-1a fluorescence after 4 h of 5 µM vacor treatment in WT or SARM1 KO mouse cortical neurons in either the cell body (left) or axon (right) compartment. Scale bar, 20 µm. See Extended Data Figure 3-2 for characterization of mouse cortical culture cell-type composition and glial culture AD-1a fluorescence after vacor treatment. f, Quantification of AD-1a fluorescence after 4 h of 5 µM vacor treatment in WT or SARM1 KO mouse cortical neurons in the axon compartment (n = 44–75 images/condition; Tukey's multiple-comparison test; ****p < 0.0001). Experiments were run in parallel with axotomy conditions (Fig. 2d), and the same control was used for both conditions. g, Quantification of cytotox dye-labeled puncta in dissociated WT or SARM1 KO cortical neurons after treatment with 5 µM vacor (n = 20 images/condition; Mann–Whitney test; ***p < 0.001 for all time points after 4 h). Quantification (h) and sample images (i) of TMRM fluorescence in axons after 5 µM vacor treatment (n = 5–8 images/condition; Mann–Whitney test; ***p < 0.001 for all time points after 4 h for WT + vacor compared with SARM1 KO + vacor). Scale bar, 100 µm. Time course of normalized NAD⁺ (j), ATP (k), and NfL (l) levels after 5 µM vacor treatment in WT mouse cortical neurons (n = 4–8, 6–20, and 4 measurements/condition, respectively; Šídák's multiple-comparison test; ****p < 0.0001). See Extended Data Figure 3-3 for data on axon degeneration and cell body death after stress in mouse cortical cultures treated with NAD⁺, nicotinamide, and 8-Br-cADPR or in mouse cortical cultures from TRPM2 KO animals. See Extended Data Table 1-1 for further statistical analysis.

Figure 4.

SARM1 KO or activation in mouse cortical neurons results in only minor proteomic changes. a, MS analysis of peptide fragment abundance in lysates of WT or SARM1 KO mouse cortical neurons. b, MS analysis of peptide fragment abundance in lysates of WT mouse cortical neurons treated for 4 h with 5 µM vacor or DMSO. Quantification of peptide abundance as measured by MS of proteins altered in SARM1 KO (c–i), or with SARM1 activation in WT (j,k), or related to SARM1 or SARM1 proteomic hit biology (l–q) in WT or SARM1 KO+/− 5 µM vacor treatment. n = 4 replicates/condition. See Extended Data Figure 4-1 for immunoblot quantification of RNF167 levels in WT and SARM1 KO cortical cultures and quantification of peptide abundance as measured by MS of other proteins of interest.

Figure 5.

Microtubule dysregulation and mitochondrial stress cause SARM1-dependent axon degeneration and SARM1-independent cell body death. Sample images and quantification of axon degeneration in WT or SARM1 KO mouse cortical neurons after treatment with 100 nM vincristine (a,b), 100 nM taxol (c,d), or 250 nM rotenone (e,f; scale bar, 100 µm; n = 40–60, 51–58, and 80–86 images/condition, respectively; Mann–Whitney test; ***p < 0.001). Quantification of axon degeneration in WT mouse cortical neurons treated with 10 µM SARM1 inhibitor and 100 nM vincristine (g), 100 nM taxol (h), or 250 nM rotenone (i; n = 12–15, 16–32, 20 images/condition, respectively; Mann–Whitney test; ***p < 0.001). Quantification of cytotox dye puncta in WT mouse cortical neurons treated with 10 µM SARM1 inhibitor and 100 nM vincristine (j), 100 nM taxol (k), or 250 nM rotenone (l). n = 32, 32, and 10 images/condition, respectively.

Figure 6.

Localization of SARM1 activation is dependent on the type of cellular stressor. Sample images (a–c) and quantification (d) of AD-1a fluorescence in WT or SARM1 KO mouse cortical neurons after 4 h treatment with 100 nM vincristine, 100 nM taxol, or 250 nM rotenone. Scale bar, 20 µm; n = 44–75 images/condition. Sample images (e) and quantification (f) of AD-1a fluorescence in cell bodies of WT mouse cortical neurons after stress treatment. Arrowheads denote cell bodies; scale bar, 20 µm; n = 80–88 images/condition. Two-way ANOVA, multiple comparisons; **p < 0.01; ****p < 0.0001. See Extended Data Figure 6-1 for sample images adjusted to equal brightness and contrast and AD-1a fluorescence in human iPSC-derived cortical neurons after stress.