Aging Differentially Affects Axonal Autophagosome Formation and Maturation

ABSTRACT

Misregulation of neuronal macroautophagy/autophagy has been implicated in age-related neurodegenerative diseases. We compared autophagosome formation and maturation in primary murine neurons during development and through aging to elucidate how aging affects neuronal autophagy. We observed an age-related decrease in the rate of autophagosome formation leading to a significant decrease in the density of autophagosomes along the axon. Next, we identified a surprising increase in the maturation of autophagic vesicles in neurons from aged mice. While we did not detect notable changes in endolysosomal content in the distal axon during early aging, we did observe a significant loss of acidified vesicles in the distal axon during late aging. Interestingly, we found that autophagic vesicles were transported more efficiently in neurons from adult mice than in neurons from young mice. This efficient transport of autophagic vesicles in both the distal and proximal axon is maintained in neurons during early aging, but is lost during late aging. Our data indicate that early aging does not negatively impact autophagic vesicle transport nor the later stages of autophagy. However, alterations in autophagic vesicle transport efficiency during late aging reveal that aging differentially impacts distinct aspects of neuronal autophagy.

Abbreviations: ACAP3: ArfGAP with coiled-coil, ankyrin repeat and PH domains 3; ARF6: ADP-ribosylation factor 6; ATG: autophagy related; AVs: autophagic vesicles; DCTN1/p150^Glued: dynactin 1; DRG: dorsal root ganglia; GAP: GTPase activating protein; GEF: guanine nucleotide exchange factor; LAMP2: lysosomal-associated protein 2; LysoT: LysoTracker; MAP1LC3B/LC3B: microtubule-associated protein 1 light chain 3 beta; MAPK8IP1/JIP1: mitogen-activated protein kinase 8 interacting protein 1; MAPK8IP3/JIP3: mitogen-activated protein kinase 8 interacting protein 3; mCh: mCherry; PE: phosphatidylethanolamine

Introduction

Macroautophagy (hereafter autophagy) is a degradative process that is integral to maintaining cellular homeostasis in eukaryotic cells. Autophagy is especially critical in neurons, as neurons are post-mitotic, metabolically very active, and highly polarized cells with functional compartments such as synapses localized far away from the buffering capabilities of the cell soma. Furthermore, adult neurogenesis occurs infrequently and individual neurons must be retained throughout an animal’s lifespan. The importance of autophagy in sustaining neuronal function is demonstrated by early-onset neurodegeneration in mice with neuron-specific knockout of critical autophagy genes [1–3]. Additionally, dysregulation of autophagy has been implicated in age-related neurodegenerative diseases including Alzheimer disease, Parkinson disease, Huntington disease, and amyotrophic lateral sclerosis [4–6].

Autophagy is a complex but evolutionarily-conserved pathway originally delineated in yeast and predominately studied in yeast and mammalian cell culture in the context of acute stressors such as starvation [7–12]. The pathway begins with autophagosome formation – the de novo creation of a double membrane structure engulfing, in the case of bulk macroautophagy, nonspecific cytoplasmic cargo and organelle fragments. Over 40 autophagy related (ATG) genes have been identified in mammals, the majority of which coordinate autophagosome biogenesis. The initiation complex, the nucleation complex, the expansion complex, the ATG2 complex, and other individual ATG proteins work in concert to build and extend the double membrane structure. During biogenesis, the outer leaflet of the double membrane becomes decorated with phosphatidylethanolamine (PE)-conjugated MAP1LC3B/LC3B (microtubule-associated protein 1 light chain 3 beta) and its orthologs, members of the Atg8 family. LC3B and other Atg8-family proteins are the only known proteins that mark autophagosomes both during and after formation, as the rest of the autophagy machinery dissociates from the organelle prior to closure of the phagophore membrane to generate the hallmark double-membrane autophagosome. Following closure, PE-conjugated LC3B remains in the lumenal-facing leaflet, while the LC3B – PE present on the outer, cytosolic face is cleaved off the autophagosome [13].

Once autophagosome biogenesis is complete, the autophagosome fuses with a late endosome or lysosome to generate an autolysosome. The internal pH of the autolysosome drops as lysosomal V-type ATPases acidify the compartment, leading to the activation of degradative enzymes that break down internalized cargos to facilitate recycling of their components [14].

In neurons, the stages of autophagy are spatially and temporally regulated. Autophagosome biogenesis predominately occurs in the distal axon [15–19]. Once formed, autophagosomes are transported along microtubules. Initially, autophagosomes exhibit bidirectional movement in the distal axon, enabling initial fusion with distal late endosomes and lysosomes [15]. Subsequently, immature autophagosomes and maturing autolysosomes, jointly termed autophagic vesicles (AVs), undergo processive retrograde movement to the cell soma directed by cytoplasmic dynein. Cargo degradation initiates en route, while the final delivery of AVs to the soma may enhance the recycling of component macromolecules into biosynthetic pathways to generate new proteins and organelles [1,15,18,20–26].

Many studies have linked defects in autophagy to age-related neurodegenerative diseases [4–6]. Here, we examined autophagosome biogenesis, transport and maturation in murine dorsal root ganglia (DRG) neurons during development and aging using live imaging approaches that provide high temporal and spatial resolution. DRG neurons are a powerful model, as these cells are one of the few neuronal types that can be cultured from any age mouse, enabling us to interrogate the autophagy pathway in neurons across relevant timescales [27]. We used this model to interrogate the specific steps involved in autophagosome biogenesis, maturation, and trafficking to better understand which steps in the pathway are altered during aging. Together, our results indicate that while autophagosome biogenesis slows significantly, most of the subsequent steps in the pathway are not adversely affected by early aging. However, later aging results in a reversion of autophagosome transport and maturation to a less efficient state.

Results

Autophagosome biogenesis and transport flux decrease with age

Misregulation of autophagy has been implicated in neurodegeneration. Since age is the most common risk factor for neurodegenerative disease [28], we asked how aging affects each of the steps of basal autophagy in murine neurons. Using GFP-LC3B transgenic mice [29], we assessed the rate of autophagosome biogenesis in primary DRG neurons from four different ages of mice: 1-month-old young mice, 3-month-old young adult mice, 16–17-month-old aged mice, and 24-month-old advanced aged mice. First, we examined the recruitment of autophagy components required for autophagosome biogenesis, focusing on markers for the initiation complex (mCherry-ATG13) [30–32] and the nucleation complex (Halo-ATG14) [33–35]. We used live-cell imaging to observe the recruitment of each marker to developing phagophores in the distal axon (Figure 1A). We did not observe any significant changes with age in the appearance of either mCh-ATG13 (Figure 1B) or Halo-ATG14 (Figure 1C) puncta in the distal axon, demonstrating that these early stages of autophagosome biogenesis are not altered during aging in DRG neurons. Additionally, we did not observe any significant changes in the protein levels of ATG13 and ATG14 in immunoblot of whole brain lysates (Figure S1). In striking contrast, but consistent with our previous results [27], we observed a significant deficit in the rates of autophagosome formation corresponding to a 70% drop between neurons from young and advanced aged mice (Figure 1D), consistent with the dysregulation of autophagosome biogenesis at the axonal terminal of DRG neurons during aging. Therefore, while the rates of both initiation and nucleation of phagophore formation appear unchanged with age, the generation of LC3-II-positive autophagosomes is dramatically inhibited in neurons from aged mice.

Figure 1.

Distal autophagosome biogenesis and mid-axon autophagosome transport flux decrease with age. (A) Time series of merge micrographs of mCh-ATG13, Halo-ATG14, and GFP-LC3B from live cell imaging of the distal neurite of DRGs from young adult mice depicting an autophagosome biogenesis event. White arrowhead denotes colocalization of mCh-ATG13, GFP-LC3B and Halo-ATG14; green arrowheads denote a GFP-LC3B-positive punctum from which mCh-ATG13 and Halo-ATG14 have dissociated. Magnified views of denoted puncta are shown below full micrograph; border color represents channel (left 3 grayscale boxes per timepoint) or colocalization state in merge (right box per timepoint). Retrograde is to the right. Scale bar: 2 μm. (B-C) Quantification of the rate of mCh-ATG13 (B) and Halo-ATG14 (C) puncta formation in live-cell imaging of the distal axon of DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 18 neurons from three biological replicates). Comparisons not significant by Kruskal-Wallis test. (D) Quantification of the rate of autophagic vesicle (AV) biogenesis (assayed by GFP-LC3B puncta formation per minute) in DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 18 neurons from at least three biological replicates). Unmarked comparisons not significant; ****p < 0.0001; ***p < 0.0005; by Kruskal-Wallis test with Dunn’s multiple comparisons test. (E) Cartoon of DRG neuron depicting time-lapse imaging of AVs in the mid-axon. (F) Representative time series of micrographs of GFP-LC3B from live-cell imaging in the mid-axon of DRGs from a young adult mouse. Colored arrowheads denote three individual GFP-positive puncta moving retrograde. Retrograde is to the right. Scale bar: 10 μm. (G) Quantification of the number of GFP-positive puncta detected per minute per 100 μm in the mid-axon of DRG neurons from young, young adult, and aged mice (violin plot, median + first and third quartiles; n ≥ 90 neurons from three biological replicates). Unmarked comparisons not significant; **p < 0.005; ****p < 0.0001 by Kruskal-Wallis test with Dunn’s multiple comparisons test.

Once autophagosomes form in the distal axon, they undergo retrograde trafficking toward the cell soma [15]. We asked if transport of autophagosomes in the axon changes with age. First, we assessed transport flux of autophagosomes in the mid-axon using live-cell imaging of the GFP-LC3B probe in primary DRG neurons (Figure 1E). We counted the number of GFP-LC3B puncta observable during a three-minute imaging window (Figure 1F) and normalized for the length of the axon segment within the micrograph frame. We detected a significant decrease in the transport flux of AVs with age corresponding to a 28% drop between neurons from young and young adult mice and a 48% drop between neurons from young and aged mice (Figure 1G). Taken together, these data suggest that autophagosome biogenesis decreases in the distal axon with age leading to a significant decrease in the transport flux of autophagosomes in the mid-axon with age.

Autophagosome maturation in the distal axon increases with age

The marked decrease in autophagosome biogenesis with age is one factor contributing to the age-related decrease in the transport flux of AVs in the mid axon. Other factors may also contribute, including alterations in the retrograde transport of AVs in the axon, driven by the molecular motors cytoplasmic dynein and kinesin along the microtubule cytoskeleton [15]. Fusion of autophagosomes with lysosomes and late endosomes or the acidification of AVs following fusion may also be altered in aging.

The pH of AVs directly affects the ability to detect GFP-LC3B-marked AVs. Since LC3B remains only inside the autophagosome after closure [13], once the pH of the autolysosome falls below the pKa of GFP, the fluorophore is quenched. The pH-sensitive quenching of GFP-LC3B may have a significant impact on the interpretation of AV transport flux data in the mid axon. Thus, we subsequently monitored AV transport in the axon using a tandem mCherry-eGFP-LC3B probe [36]. Since the pKa of mCherry is 4.5 [37], lower than that of eGFP (pKa 6.0) [38], mCherry is more resistant to quenching in acidic environments. Thus, autophagosome maturation can be monitored using this mCh-eGFP-LC3B marker; newly formed, immature autophagosomes are marked with both red and green fluorescence, while mature autolysosomes with acidified lumens emit only mCh fluorescence (Figure 2A) [36].

Figure 2.

Autophagic vesicle maturation increases in the distal axon during aging. (A) Schematic of the tandem mCh-eGFP-LC3B probe in an immature autophagosome (top) and an acidified, mature autolysosome (bottom). (B) Quantification of the total number of autophagic vesicles present in the distal (left) and proximal (right) axon in DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 40 neurons from three biological replicates), normalized to axon length and imaging duration. Unmarked comparisons not significant; **p < 0.005; ***p < 0.0005 by Kruskal-Wallis test with Dunn’s multiple comparisons test. (C) Quantification of the percent of GFP⁺ mCh⁺ autophagic vesicles in the distal (left) and proximal (right) axon in DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 39 neurons from three biological replicates). Unmarked comparisons not significant; *p < 0.05; ****p < 0.0001 by Kruskal-Wallis test with Dunn’s multiple comparisons test. (D) Quantification of the number of GFP⁺ mCh⁺ AVs detected per minute per 100 μm in the distal axon of DRG neurons from young, young adult, and aged mice (violin plot, median + first and third quartiles; n ≥ 79 neurons from three biological replicates). Unmarked comparisons not significant; ****p < 0.0001 by Kruskal-Wallis test with Dunn’s multiple comparisons test. (E-H) Representative micrographs mCh-eGFP-LC3B in the distal axon of DRG neurons from young (E), young adult (F), aged (G), and advanced aged (H) mice. Top micrograph is the GFP channel; middle micrograph is the mCh channel; and bottom micrograph is the merge. In the merge panels, yellow arrowheads denote an immature, GFP⁺ mCh⁺ autophagosome and red arrowheads denote a mature, GFP quenched, mCh⁺ autolysosome. Retrograde is to the right. Scale bar: 5 μm. (I-L) Representative micrographs mCh-eGFP-LC3B in the proximal axon of DRG neurons from young (I), young adult (J), aged (K), and advanced aged (L) mice. Top micrograph is the GFP channel; middle micrograph is the mCh channel; and bottom micrograph is the merge. In the merge panels, yellow arrowheads denote an immature, GFP⁺ mCh⁺ autophagosome and red arrowheads denote a mature, GFP quenched, mCh⁺ autolysosome. Retrograde is to the right. Scale bar: 5 μm.

Using the mCh-eGFP-LC3B tandem probe transiently transfected into primary DRG neurons, we first determined the absolute number of AVs, independent of maturation state, in the distal (≤150 µm from the axonal tip) and proximal (≤150 µm from the soma) axon of DRG neurons during aging. Interestingly, we observed a significant decrease in the number of AVs in the distal axon during development and early aging, but no significant decrease between neurons from young and advanced aged mice. We also detected no age-related changes in the total number of AVs in the proximal axon (Figures 2B, S2). We next assessed the percentage of AVs that were positive for both mCh and eGFP in the distal and proximal axon. At all ages examined, we saw a significantly higher fraction of immature AVs, positive for both mCh and eGFP fluorescence, in the distal axon as compared to the proximal axon (Figure 2C,E–L). These data indicate that aging does not disrupt the overall pathway for maturation of axonal autophagosomes. However, in neurons from aged mice, we detected a decrease in the fraction of immature AVs in the distal axon, but no age-related changes in AV maturation in the proximal axon (Figure 2C). Interestingly, when we compared the fraction of immature AVs between neurons from young and advanced aged mice, we did not detect a statistically significant difference, suggesting that the maturation of AVs in neurons from 24-mo mice reverts to a less efficient state.

We next used the transiently transfected tandem probe to ask whether we could replicate the age-related decrease in autophagosome transport flux we observed using DRG neurons from GFP-LC3B transgenic mice (Figure 1G). Indeed, we observed significant decreases in immature AV transport flux in the distal axon between DRG neurons from young, young adult and aged mice with the ectopically expressed mCh-eGFP-LC3B marker (Figure 2D). It is important to note that the GFP-LC3B probe was expressed stably at a low level in neurons isolated from B6.Cg-Tg(CAG-EGFP/LC3)53 Nmi/NmiRbrc transgenic mice [15,16,27,29], while the mCh-eGFP-LC3B tandem probe was transiently transfected into DRG neurons after isolation. Higher expression levels of the dual reporter construct relative to the GFP-LC3B probe likely drive the higher number of total AVs observed in these experiments. Collectively, these results explain our observations of AV transport flux using the GFP-LC3B probe (Figure 1G), showing a decrease in total AV transport flux in neurons from young to young adult mice with a subsequent decrease in the percentage of distal immature AVs in neurons from young adult to aged mice.

Lysosomal content of the distal axon does not change until late aging

We next investigated the molecular and cellular underpinnings of the decrease in the fraction of immature AVs in the distal axon during aging. The age-related change in AV maturation that we observed in the distal axon could be due to alterations in lysosomal content or in AV transport in the distal axon. To examine the first possibility, we assessed multiple parameters of lysosomes in the distal axon of neurons from four ages of mice. We first used LysoTracker (LysoT) Red to label acidic vesicles and CellMask Deep Red to label the plasma membrane (Figure 3A). We acquired z-stacks of the distal tip of axons and then calculated the number of LysoT-positive puncta (Figure S3A) and LysoT-positive area (Figure 3C), both normalized to the area of the distal axon. We observed no significant changes with age in either normalized number or area of LysoT+ puncta in the distal axon until the 24-mo time point. Second, we used Cresyl Violet (CresylV [39]) as an alternative label of acidic vesicles, again in parallel with CellMask Deep Red labeling of the plasma membrane (Figure 3B,D). Again, we observed no age-related alterations in normalized number of CresylV-positive puncta in the distal axon (Figure S3B) or in the normalized area of CresylV-positive vesicles in the distal axon (Figure 3F) until the 24-mo time point. With advanced aging, we observed a significant loss of acidified vesicles in the distal axon (Figures 3C–D, S3A-B). Thus, these data do not explain the increase in AV maturation we observed in neurons from aged mice, but the loss of distal acidified vesicles during advanced aging may underpin the reversion of AV maturation (Figure 2C).

Figure 3.

Lysosomal content in the distal axon does not change with age. (A-B) Representative maximal projection micrographs of LysoTracker Red (“LysoT”, green) and Cell Mask Deep Red (“Cell Mask”, magenta) (A) and of Cresyl Violet (“CresylV”, green) and Cell Mask Deep Red (“Cell Mask”, magenta) (B) from live cell imaging of the distal neurite of DRG neurons from young (top), young adult (top middle), aged (bottom middle), and advanced aged (bottom) mice. Scale bars: 5 μm. (C-D) Quantification of the area of LysoTracker Red-positive puncta (C) and the area of Cresyl Violet-positive puncta (D), normalized to distal tip area in DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 83 neurons from three biological replicates. Unmarked comparisons not significant; *p < 0.05; ***p < 0.0005; ****p < 0.0001 by Kruskal-Wallis ANOVA test with Dunn’s multiple comparisons test. (E-K) Representative maximal projection micrographs of the distal neurites of fixed DRG neurons from young (G), young adult (H), aged (I), and advanced aged (K) mice. White arrowheads indicate colocalization of anti-LGMN/AEP (red), anti-CTSB (blue), and anti-LAMP2 (green). Borders of magnifications of indicated puncta denote channel (left 3 grayscale boxes) or colocalization state in merge (right box). Scale bar: 2 μm. (F, H, J, L) Line scans of puncta indicated by arrows. Intensity of each channel was normalized to facilitate ease of viewing.

Degradatively competent lysosomes are present in the distal axon of primary hippocampal, cortical, and DRG neurons [40–43]. Thus, we asked whether we could detect any age-related changes in the presence of lysosomal enzymes in the distal axon. We used multi-color immunocytochemistry to examine the localization of endogenous lysosomal enzymes LAMP2 (lysosomal-associated membrane protein 2), CTSB (cathepsin B), and LGMN/AEP (legumain) in DRG neurons from young (Figure 3E–F), young adult (Figure 3G–H), aged (Figure 3I–J), and advanced aged (Figure 3K–L) mice. We observed colocalization of all three lysosomal enzymes in the distal axon of neurons from all four ages of mice, indicating that lysosomes containing degradative enzymes are present in the distal axon during development and aging. Next, we examined the endogenous levels of lysosomal proteins by immunoblot of whole brain lysates. We did not detect age-related changes in the levels of lysosomal proteins: ATP6V1A (ATPase, H+ transporting, lysosomal V1 subunit A), ATP6V1B (ATPase, H+ transporting, lysosomal V1 subunit B), CTSD (cathepsin D), CTSL (cathepsin L), LAMP1 (lysosomal-associated membrane protein 1), SCARB2/LIMP2 (scavenger receptor class B, member 2), and PLA2G15/LYPLA3 (phospholipase A2, group XV) in mouse whole brain lysate (Figure S3C-D). Taken together, our data suggest that neither lysosomal residency nor degradative capacity changes in the distal axon between development and early aging, but does decline during advanced aging, implying that age-related alterations in lysosomal capacity in the distal axon do not account for the observed increase in distal AP maturation in neurons from aged mice, but may underlie the reversion of distal AP maturation in neurons from advanced aged mice.

Parameters of AV transport in the distal axon do not change with age

We next asked whether AV transport changed with age. To assess AV transport, we used multi-color, live-cell microscopy to acquire 3-minute time-lapse videos of DRG neurons transiently expressing the tandem mCh-eGFP-LC3B probe. We chose the tandem probe so that we could analyze axonal transport of both immature, mCh⁺ eGFP⁺ autophagosomes and mature, mCh⁺ eGFP⁻ autolysosomes. By analyzing the kymographs (Figure 4A) generated from the time-lapse videos, we calculated several parameters of axonal AV transport. Each AV was categorized based on its net displacement during imaging: “anterograde” if the AV net displacement was greater than 10 μm toward the axonal tip, “bidirectional/stationary” if the AV net displacement was less than 10 μm, and “retrograde” if the AV net displacement was greater than 10 μm toward the cell soma (Figure S2).

Figure 4.

Transport dynamics of immature and mature AVs do not considerably change with age in the distal neurite. (A) Representative kymographs of the mCh channel of the distal neurite of DRG neurons ectopically expressing mCh-eGFP-LC3B from young, young adult, aged, and advanced aged mice. Annotations of individual AV tracks are below the grayscale kymographs. Retrograde is to the right. Horizontal scale bar: 5 μm; vertical scale bar: 1 min. (B) Cartoon of DRG neuron depicting time-lapse imaging of immature AVs in the distal neurite. (C-D) Quantification of the net displacement of immature autophagic vesicles that moved in a net bidirectional or stationary (C) or net retrograde (D) direction in the distal neurite of DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 31 puncta from three biological replicates). Comparisons are not significant by Kruskal-Wallis test. (E) Cartoon of DRG neuron depicting time-lapse imaging of mature AVs in the distal neurite. (F-G) Quantification of the net displacement of mature autophagic vesicles that moved in a net bidirectional or stationary (F) or net retrograde (G) direction in the distal neurite of DRG neurons from young, young adult, aged, and advanced aged mice (violin plot, median + first and third quartiles; n ≥ 59 puncta from three biological replicates). Comparisons are not significant by Kruskal-Wallis test. (H) Quantification of the percent of immature AVs in the distal axon that moved in a net anterograde (gray), bidirectional/stationary (cyan), or retrograde (magenta) direction from young, young adult, aged and advanced aged mice. **p = 0.0015 by Chi-square test (mean ±95% confidence intervals; n ≥ 11 puncta from three biological replicates). Unmarked comparisons not significant; **p < 0.005; ***p < 0.005 between neurons from young mice and indicated adult mice and *p < 0.05 between indicated groups by Fisher’s exact test. (I) Ratio of retrograde to anterograde immature AVs. Representation of data presented in H. (J) Quantification of the percent of mature AVs in the distal axon that moved in a net anterograde (gray), bidirectional/stationary (cyan), or retrograde (magenta) direction from young, young adult, aged, and advanced aged mice. **p = 0.0047 by Chi-square test (mean ±95% confidence intervals; n ≥ 14 puncta from three biological replicates). Unmarked comparisons not significant; *p < 0.05; **p < 0.005; ***p < 0.0005 between neurons from young mice and indicated adult mice and *p < 0.05 between indicated groups by Fisher’s exact test. (K) Ratio of retrograde to bidirectional mature AVs. Representation of data presented in J.

We first analyzed the transport dynamics of immature, mCh⁺ eGFP⁺ autophagosomes in the distal axon. Given the small number of distal AVs with a net anterograde displacement (Figure S2), we only analyzed bidirectional and retrograde AVs. We observed no age-related changes in net displacement (Figure 4B–D), average speed (Figure S4A-C), or pause time fraction (the time spent not moving/total time of AV run; Figure S5A-C) across conditions. We did detect a significant increase in AV switch count for bidirectional AVs between neurons from young adult and advanced aged mice (Figure S6A-C). Additionally, we did not observe significant age-related changes in total run length, pause duration, or pause number for bidirectional or retrograde immature AVs (data not shown). These data indicate that most transport parameters for immature autophagosomes in the distal axon are not significantly altered with age.

We next analyzed the transport dynamics of mature, mCh⁺ eGFP⁻ autolysosomes in the distal axon. We did not detect any age-related changes in AV net displacement for bidirectional or retrograde AVs (Figure 4E–G). Similarly, we did not observe consistent changes in average speed with age, although we did detect a slight decrease in the average speed of bidirectional and retrograde AVs in neurons from aged mice compared to young mice (Figure S4D-F). While we did observe an increase in the AV pause time fraction in bidirectional AVs in the distal axon in neurons from adult mice compared to young mice, we detected no age-related changes in AV pause time fraction in retrograde AVs in the distal axon (Figure S5D-F). Furthermore, we did not find age-related changes in AV switch count in bidirectional or retrograde mature AVs in the distal axon (Figure S6D-F). Finally, we did not observe significant age-related changes in total run length, pause duration, or pause number for anterograde or retrograde AVs (data not shown). Collectively, our data indicate that AV transport parameters for both immature and mature AVs are generally unchanged in the distal axon with age.

Transport of autophagic vesicles becomes more efficient with age in the distal axon

We also examined the fraction of AVs that moved in a net anterograde, bidirectional, or retrograde direction. For immature AVs in the distal axon, independent of age, the plurality of AVs moved bidirectionally, with a slightly lower fraction moving in a net retrograde direction (Figure 4H). Of interest, the fraction of immature AVs that moved in a net anterograde direction increased between neurons from young and young adult mice and was maintained during early aging, but was lost during late aging (Figure 4H,I). This increase in the anterograde fraction was accompanied by a concomitant slight decrease in the retrograde fraction (Figure 4H,I). These data suggest that there are modifications during development in the distal axon that result in immature AVs being held in the distal axon. These modulations are preserved during early aging, but are lost during advanced aging.

Similarly, we assessed the net direction of movement of mature AVs in the distal axon. Independent of age, the majority of mature AVs in the distal axon moved in a net retrograde direction (Figure 4J). In contrast to the immature AVs in the distal axon, we observed an increase in the fraction of mature AVs that moved in a net retrograde direction between neurons from young and young adult mice which was retained during early aging (Figure 4J,K). This age-related increase in the retrograde fraction was balanced by a concomitant decrease in the fraction of mature AVs that moved bidirectionally. Similar to the immature AVs, this increase in the retrograde fraction was lost during late aging (Figure 4J,K). Taken together, our data suggest that a sorting mechanism for AVs is refined in the distal axon during development that holds immature AVs in the distal axon while simultaneously enabling robust retrograde transport of mature AVs out of the distal compartment. Importantly, the efficacy of this distal sorting mechanism is sustained during early aging, but lost during late aging.

Retrograde transport of mature autolysosomes in the proximal axon becomes more efficient during early aging

We next assessed AV transport in the proximal axon. Approximately 80% of AVs in the proximal axon are mature, mCh⁺ eGFP⁻ autolysosomes (Figure 2C). Independent of age, in both the distal and proximal axon, the majority of mature AVs are transported in a net retrograde direction (Figure S2). We calculated the transport parameters for each category of AVs (immature and mature) in the proximal axon: anterograde, bidirectional, and retrograde. However, due to the small numbers of immature AVs moving in any direction and mature AVs moving net anterogradely or bidirectionally (Figure S2), we focused on the mature AVs moving retrogradely in the proximal axon (Figure 5A,B). Comparing neurons from young and young adult mice, we observed an increase in total run length (Figure 5C), net displacement (Figure 5D), and average speed (Figure 5E). Additionally, we observed decreases in pause time fraction (Figure 5F) and switch count (Figure 5G) between neurons from young and young adult mice. These changes were generally maintained during early aging (Figure 6C–G), with only average speed differing significantly between neurons from young adult and aged mice (Figure 5E). These data diverge from our observations of few age-related changes in AV transport parameters in the distal axon and suggest that age-related changes in the transport dynamics of mature, retrogradely moving AVs occur primarily in the proximal axon. Interestingly, some parameters are also maintained during advanced aging: average speed, pause time fraction, and switch count (Figure 5E–G). In contrast, total run length and net displacement show no significant differences between neurons from advanced aged and young mice (Figure 5C,D). However, our findings in both the distal and proximal axon are congruent with an axonal sorting mechanism for AVs that becomes more efficient during development, is sustained during early aging, but is lost during late aging.

Figure 5.

Transport dynamics of mature AVs do not considerably change with age in the proximal neurite. (A) Representative kymographs of the mCh channel of the proximal neurite of DRG neurons ectopically expressing mCh-eGFP-LC3B from young, young adult, aged, and advanced aged mice. Annotations of individual AV tracks are below the grayscale kymographs. Retrograde is to the right. Horizontal scale bar: 5 μm; vertical scale bar: 1 min. (B) Cartoon of DRG neuron depicting time-lapse imaging of immature AVs in the proximal neurite. (C-G) Quantification of the total run length (C), net displacement (D), average speed (E), pause time fraction (F), and switch count (G) for mature AVs that had a net retrograde displacement in the proximal axon of DRG neurons (violin plot, median + first and third quartiles; n ≥ 191 puncta from three biological replicates). Unmarked comparisons not significant; *p < 0.05; **p < 0.005; ***p < 0.0005; ****p < 0.0001 by Kruskal-Wallis test with Dunn’s multiple comparisons test. (H) Quantification of the percent of mature AVs in the proximal axon that moved in a net anterograde (gray), bidirectional/stationary (cyan), or retrograde (magenta) direction from young, young adult, aged and advanced aged mice. ****p < 0.0001 by Chi-square test (mean ±95% confidence intervals; n ≥ 13 puncta from three biological replicates). Unmarked comparisons not significant; *p < 0.05; **p < 0.005; ***p < 0.0005 between neurons from young mice and indicated adult mice and *p < 0.05 between indicated groups by Fisher’s exact test. (I) Ratio of retrograde to bidirectional mature AVs. Representation of data presented in H.

Figure 6.

Overexpression of ACAP3 does not individually alter AV maturation in the distal axon in early aging. (A) Immunoblots of brain lysates from young, young adult, aged, and advanced aged mice of one sex (n = 3 biological replicates shown for each age; blots were repeated on another set of 3 biological replicates of the other sex for each age). Total protein was used as a loading control (normalization factor indicated below each blot as a percentage). (B) Quantification of protein levels, normalized to total protein and represented as fold change relative to the mean for lysates from young (1 mo) mice. For individual data points, circles represent female, triangles represent male. Unmarked comparisons not significant; *p < 0.05 by One-way ANOVA with Tukey’s multiple comparisons test. (C) Representative micrographs mCh-eGFP-LC3B in the distal axon of a DRG neuron from a young adult mouse. Top micrograph is the GFP channel; second micrograph is the mCh channel; third micrograph is the Halo-ACAP3 channel; and bottom micrograph is the merge. In the merge panel, white arrowheads denote immature, GFP⁺ mCh⁺ autophagosomes colocalized with ACAP3, magenta arrowheads denote mature, GFP quenched, mCh⁺ autolysosomes colocalized with ACAP3, and the blue arrowhead denotes an ACAP3 punctum not colocalized with an AV. Retrograde is to the right. Scale bar: 5 μm. (D-G) Quantification of the percent of AVs colocalized with Halo-ACAP3 (D), the percent of GFP⁺ mCh⁺ autophagic vesicles (E), the flux of immature AVs (F), and total AV flux (G) in the distal axon of neurons from aged mice ectopically expressing Halo control or Halo-ACAP3 (violin plot, median + first and third quartiles; n ≥ 26 neurons from three biological replicates). Comparisons are not significantly different by two-tailed unpaired t-test.

Additionally, we analyzed the net direction of movement of mature AVs in the proximal axon. Similar to the distal axon, independent of age, the majority of mature AVs were transported in a net retrograde direction in the proximal axon (Figure 5H). Also similar to the distal axon, we observed an increase in the fraction of mature AVs that moved in a net retrograde direction with a simultaneous decrease in the fraction that moved bidirectionally between neurons from young and young adult mice. These alterations were partially retained during aging, but we did detect a significant decrease in the fraction moving retrograde between neurons from advanced aged and young adult mice (Figure 5H,I). Collectively, these data in the proximal axon mirror our data from the distal axon, in which a sorting mechanism is consolidated during development and preserved during early aging to efficiently move mature autolysosomes through the proximal axon into the cell soma to enable effective recycling of neuronal AV contents into new proteins and organelles. This sorting mechanism is lost in the distal axon, but is partially maintained in the proximal axon during late aging.

Regulation of AV-associated motor adaptor proteins may modulate axonal AV maturation during early aging

After formation, autophagosomes are transported from the distal axon to the soma along microtubules, with the microtubule motors cytoplasmic dynein and KIF5 (kinesin family member 5) detectable on neuronal autophagosomes [15,44–46]. To further understand the molecular mechanisms that underpin the age-related changes in AV transport we observed, we asked if any components of the transport machinery change with age in brain lysate. We detected no age-related changes in protein expression of tubulin or dynein cytoplasmic 1 intermediate chains. Surprisingly, we did observe a slight, yet significant decrease in KIF5B (kinesin family member 5B) expression between young and advanced aged mice. Furthermore, we found an increase in DCTN1/p150^Glued (dynactin 1) expression between young and aged mice (Figure S7). This increase in DCTN1 expression is subsequently lost during advanced aging, which mirrors the age-related changes we observed with AV maturation and transport efficiency in the distal axon (Figures 2C, 4H–K), suggesting that changes in DCTN1 expression might contribute to age-related changes in neuronal AV transport dynamics and AV maturation.

However, if DCTN1 was a significant factor underlying alterations in age-related AV transport, we would expect to detect significant changes in AV average speed, run length and/or net displacement. Given the robust age-related changes in AV transport efficiency, we next examined motor adaptors that mediate autophagosome interactions with motor complexes. MAPK8IP1/JIP1 (mitogen-activated protein kinase 8 interacting protein 1) is a motor adaptor that inactivates kinesin-1 on newly formed autophagosomes in the distal axon to promote their retrograde transport [20]. Molecularly unrelated MAPK8IP3/JIP3 (mitogen-activated protein kinase 8 interacting protein 3) modulates the retrograde transport of mature AVs in the axon [40]. However, we did not detect age-related changes in protein expression of either MAPK8IP1/JIP1 or MAPK8IP3/JIP3 (Figure S7).

ARF6 (ADP-ribosylation factor 6) was recently identified as a MAPK8IP3/JIP3 effector on neuronal autophagosomes [47]. ARF6 is a small GTPase whose nucleotide state regulates MAPK8IP3/JIP3 interaction with dynein, with ARF6-GTP promoting MAPK8IP3/JIP3-SPAG9/JIP4 interaction with DCTN1 and ARF6-GDP promoting MAPK8IP3/JIP3-SPAG9/JIP4 binding to kinesin light chain [47,48]. Small GTPases are regulated by guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs). ARF6 has 10 known GEFs and 20 identified GAPs. We narrowed this list to GEFs and GAPs significantly enriched or expressed in neural tissues and implicated in the autophagosomal and/or endolysosomal pathways [49]. ARF6 GEFs IQSEC2 (IQ motif and Sec7 domain 2) and IQSEC3 (IQ motif and Sec7 domain 3) are highly expressed in brain and found at the autophagosomal membrane [47]. We did not observe significant age-related changes in either IQSEC2 (Figure 6A,B) or IQSEC3 (data not shown) protein expression in brain lysate. ARF6 GAPs ACAP2 (ArfGAP with coiled-coil, ankyrin repeat and PH domains 2) and ACAP3 (ArfGAP with coiled-coil, ankyrin repeat and PH domains 3) are expressed in brain and associated with the endolysosomal pathway [49,50]. We did not observe age-related changes in ACAP2 expression in brain lysate (Figure 6A,B). Conversely, we did observe a significant decrease in ACAP3 expression in brain lysate between young and aged mice, which was not maintained in advanced aged mouse brain lysate (Figure 6A,B). An age-related decrease in ARF6 GAP expression could lead to an increase in ARF6-GTP at the autophagosome, promoting retrograde transport of autophagosomes via MAPK8IP3/JIP3 in neurons from aged mice.

We next asked if we could modulate autophagosome transport and maturation in the distal axon by ectopic expression of ACAP3 in neurons from aged mice. When we overexpressed Halo-ACAP3, we observed ACAP3 recruitment to both immature and mature autophagosomes in the distal axon (Figure 6C,D). However, Halo-ACAP3 overexpression did not significantly increase the fraction or flux of immature AVs in the distal axon (Figure 6E–G), suggesting that modulating the autophagosomal ARF6 GEF as well as the GAP might be necessary to ectopically modulate AV transport and maturation in neurons from aged mice.

Together, these data extend the current model for axonal transport of autophagic vesicles [51] to focus on changes during development and aging. During development, axonal compartments establish sorting mechanisms for immature and mature autophagic vesicles, restricting immature autophagosomes distally and releasing mature autolysosomes to move proximally (Figure 7). This sorting capability is retained in neurons during early aging, suggesting that aged neurons would have the capacity to effectively process AVs if autophagosome biogenesis was restored. However, during late aging, neurons lose the AV sorting capability established during development, indicating that neurons lose additional aspects of this critical homeostatic pathway as aging progresses.

Figure 7.

Model for autophagic vesicle transport and maturation during aging in neurons. Autophagic vesicles are sorted in the distal axon, with immature autophagosomes (yellow) being held in the distal axon and mature autolysosomes (red) released to undergo processive retrograde transport to the cell soma. The color and weight of the arrows show the net transport direction of the AVs (magenta = retrograde, gray = anterograde). This sorting mechanism and autophagic vesicle transport become more efficient in the distal and proximal axon during development. Efficient sorting and transport of AVs is preserved during early aging in both the distal and proximal axon. However, during advanced aging, AV sorting and transport revert to a less efficient state.

Discussion

We have examined the dynamics of axonal autophagy during aging and shown that while the rate of autophagosome biogenesis decreases with age, transport of autophagosomes and autolysosomes is not globally impaired during aging. Conversely, we have demonstrated that transport of mature AVs in the proximal axon actually becomes more efficient during development and that this efficiency is maintained during early aging. Interestingly, while we observed an age-related decrease in AV transport flux in the mid-axon with the GFP-LC3B autophagosomal reporter, we determined that this decrease was likely due to a significant decrease in the rate of autophagosome biogenesis combined with an increase in the maturation of AVs in the distal axon during early aging. Ultimately, we demonstrated that DRG neurons from aged mice do not have a significant reduction in AV maturation nor a diminished capacity to transport AVs in the distal or proximal axon, suggesting that the later stages of autophagy remain robust in neurons during aging. Interestingly, during later aging, DRG neurons lose their AV transport efficiency gained during development, reverting to an immature state reminiscent of DRG neurons from young mice.

Ectopic expression of autophagy pathway components has been shown to affect the rate of autophagosome biogenesis. Overexpression of LC3B can enhance autophagosome biogenesis [52–54], consistent with the higher flux measurements observed in experiments overexpressing the tandem reporter probe (Figure 2D) compared to data collected using the transgenic GFP-LC3B reporter (Figure 1G). We found that by ectopically expressing the tandem mCh-eGFP-LC3B AV reporter, we eliminated the variable of age-related changes in AV number per minute in the axon. With ectopic expression, we observed no age-related changes in AV number in the proximal axon and an age-related decrease of only one AV per minute in the distal axon (Figure 2B). Thus, in our experiments examining AV transport dynamics, the neurons from young, young adult, aged, and advanced aged mice were transporting approximately the same number of AVs in the distal and proximal axon, further suggesting that neurons preserve their capacity to transport AVs efficiently during aging.

Given our data indicating that autophagosomes mature into autolysosomes more rapidly in the distal axon in neurons from aged mice relative to neurons from young, young adult, and advanced aged mice (Figure 2C), we hypothesized that this age-related increase in autophagosome maturation could be due to an age-induced increase in endolysosomal content in the distal axon or to changes in AV transport dynamics. Lysosomes and late endosomes fuse with autophagosomes to generate autolysosomes, which eventually degrade and recycle engulfed AV cargo. Since the endolysosomal system directly regulates the maturity of AVs, we sought to identify any age-related alterations in lysosomal content in the distal axon. Using acid-sensitive dyes and interrogating endogenous endolysosomal proteins, we did not detect age-related changes in lysosomal content or the presence of degradative enzymes in the distal axon until late aging. These data suggest that age-related changes in the endolysosomal system contribute to age-related alterations in AV maturation during late aging.

Importantly, we did not examine the degradative capacity of the acidified endolysosomes in the distal axon. Others have suggested that there is a gradient of degradative capability of endolysosomes in the axon, with more degradatively competent lysosomes positioned in the proximal axon [55], although lysosomes containing degradative capability have been observed in the distal axon of primary neurons [41–43]. Since our initial results indicated age-related changes to AV maturation only in the distal axon, we did not investigate whether aging modulates the gradient of endolysosomal degradative capacity in the proximal axon of DRG neurons. Defects in the endolysosomal system have been implicated in age-related neurodegenerative diseases [56,57]. It is possible that our results suggesting that distal AVs acidify faster in neurons from aged mice could contribute to age-related defects in the endolysosomal pathway. However, acidification of endolysosomes is required for the activation of their degradative machinery; thus, we hypothesize that this larger fraction of acidified AVs indicates that aging alone does not inhibit progression through the later stages of autophagy in neurons until late aging. Conversely, we did observe significant loss in acidified vesicles in neurons from 24 mo mice. Thus, late loss of acidified endolysosomes in the distal axon could underpin the decrease in AV maturation we observed in the distal axon in neurons from advanced aged mice. In the future, it will be important to assess how age modulates the degradative capacity of endolysosomes in distinct axonal regions and in models of NDDs, particularly focusing on late aging.

Despite the majority of AV transport parameters not changing with age in the distal axon for both immature and mature AVs (Figure 4), we found that the fraction of immature autophagosomes moving anterograde increased with age, while the fraction of mature autolysosomes moving retrograde increased with age. However, these changes occurred during development, observed by comparing neurons from young (1 mo) and young adult (3 mo) mice. Furthermore, these increases in AV transport efficiency were lost during late aging, with some AV transport parameters in neurons from 24 mo mice resembling those from young mice, suggesting a deterioration to a less mature state during advanced aging.

We observed an increase in distal AV maturation in neurons from aged mice (Figure 2C). Since the ectopic expression of the mCh-eGFP-LC3B tandem marker normalized the number of distal autophagosomes between neurons from young adult and aged mice, age-related changes in distal AV maturation are not the result of age-related changes in the proportion of lysosomes to autophagosomes in the distal axon. Therefore, our data suggest that there are other mechanisms that are modulated during aging that alter AV maturation state. One possible mechanism is that the fusion of autophagosomes to lysosomes becomes more efficient during aging, yielding a higher fraction of mature AVs in the distal axon in neurons from aged mice. It will be interesting to examine age-related changes in autophagosome-endolysosome fusion in future studies.

Interestingly, while we observed few AV transport parameters changing substantially with age in the distal axon (Figure 4), several AV transport parameters changed during development in the proximal axon (Figure 5). These changes were maintained during early aging, with some parameters returning to a more immature state during late aging. These data indicate that the proximal axon is more resistant to regression than the distal axon during late aging, suggesting that the homeostatic and buffering capabilities of the cell soma may protect the proximal axon.

Together, our data from the distal and proximal axon suggest that the properties of the microtubule motors cytoplasmic dynein and kinesins are not altered with age. Furthermore, we did not observe global age-related changes in tubulin or microtubule motor machinery levels (Figure S7). Instead, our data suggest that motor adaptors may be modulated during development and aging. Motor adaptors not only link vesicles, including AVs, to the motor complexes, but also regulate the activation state of the cytoplasmic dynein and kinesin motors that are simultaneously bound to these organelles [20,51]. In embryonic primary neurons, different motor adaptors predominate in different axonal compartments, which could explain how aging broadly affects AV transport parameters for mature AVs in the proximal axon while only specifically altering AV switch count in the distal axon [40]. Remarkably, we observed that AV sorting in the distal axon becomes more efficient during development, with a higher fraction of immature autophagosomes moving anterograde and a higher fraction of mature autolysosomes moving retrograde in neurons from young adult mice compared to neurons from young mice. This increased efficiency is retained during early aging and suggests a robust sorting mechanism for the transport of immature and mature AVs. Since motor adaptors participate in cargo selection for microtubule motors [51], motor adaptors are particularly attractive candidates for this sorting mechanism.

While we did not observe age-related changes in known AV-associated motor adaptors MAPK8IP1/JIP1 or MAPK8IP3/JIP3, we investigated a MAPK8IP3/JIP3 effector, ARF6. ARF6 is a small GTPase, an established MAPK8IP3/JIP3 interactor [47,48], and present on the outside of AVs isolated from mouse brain lysate [47]. The nucleotide binding state of ARF6 modulates MAPK8IP3/JIP3-SPAG9/JIP4 binding to microtubule motor complexes [48], and we observed an age-related decrease in ARF6 GAP ACAP3 expression in brain lysate. In contrast, overexpression of tagged ACAP3 in neurons from aged mice did not significantly increase the fraction of immature AVs in the distal axon (Figure 6). It is possible that we would also need to modulate the ARF6 GEF that acts on distal AVs to significantly alter AV maturation. While we assessed likely ARF6 GEF candidates for neuronal AVs, we did not detect age-related changes in their protein expression (Figure 6). It will be interesting to identify which ARF6 GEF regulates ARF6 nucleotide state on AVs in the axon to better understand the molecular mechanisms that underlie age-related changes in axonal AV maturation and transport.

Misregulation of autophagy has been implicated in many age-related neurodegenerative diseases, with multiple stages of autophagy being affected in a specific disease and multiple diseases influencing a given stage of autophagy [4–6]. The most relevant shared risk factor for age-related neurodegenerative diseases is age [28]. Our data suggest that while autophagosome biogenesis dramatically decreases during aging (Figure 1), later stages of autophagy, including autophagosomal maturation, do not become less effective during early aging. Conversely, AV transport becomes more efficient during development and is maintained during early aging, while autophagosome maturation in the distal axon occurs more rapidly in neurons from aged mice. Thus, our results imply that if we can restore the rate of autophagosome biogenesis in aged neurons, those neurons will be able to process and transport autophagosomes to eventually recycle the contents in the cell soma. However, our data also suggest that additional mechanisms break down during later aging, including presence of acidified vesicles in the distal axon. For neurodegenerative diseases with known defects in the endolysosomal system and autophagosomal transport and/or maturation, it will be important to address the synergistic effects of both age and disease.

Our previous work indicates that increasing the proportion of dephosphorylated WIPI2B in neurons from aged mice restores the age-related decrease in neuronal autophagosome biogenesis [27], identifying a potential therapeutic avenue for treating neurodegenerative diseases during early aging. Ultimately, our data suggest that while axonal autophagosome biogenesis decreases substantially during aging, maturation and transport of AVs do not decline during early aging. However, AV maturation and transport efficiency regress to a less efficient state during late aging, indicating that neurons experience declines in specific cellular pathways at different points during aging.

Materials and methods

Reagents

GFP-LC3B transgenic mice (strain: B6.Cg-Tg[CAG-EGFP/LC3]53 Nmi/NmiRbrc) were generated by N. Mizushima (Tokyo Medical and Dental University, Tokyo, Japan [29]) and obtained from RIKEN BioResource Center in Japan (RBRC00806). These mice were bred with C57BL/6J mice obtained from The Jackson Laboratory (000664). Hemizygous and wild-type littermates were used in experiments. Constructs used include: mCherry-ATG13 (subcloned from Addgene, 22875; deposited by Noboru Mizushima), mCherry-eGFP-LC3B (gift from T. Johansen, University of Tromsø, Tromsø, Norway [58]), Halo-ATG14L (subcloned from Addgene, 21635; deposited by Tamotsu Yoshimori), and Halo-ACAP3 (cloned by PCR from whole mouse brain cDNA).

Primary neuron culture

Mice were euthanized prior to dissection. All animal protocols were approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania. DRG neurons were isolated as previously described [59] from mice of either sex in these postnatal ranges: P21–28 (1 mo, young), P90–120 (3 mo, young adult), P480–540 (16–17 mo, aged), or P730–761 (24 mo, advanced aged). DRG neurons were plated on glass-bottomed dishes (MatTek Corporation, P35G-1.5-14-C) and cultured in F-12 Ham’s media (Gibco, 11765–047) with 10% heat-inactivated fetal bovine serum (HyCLone, SH30071.03), 100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco, 15140122). DRG neurons were imaged or fixed after being maintained for 2 days at 37°C in a 5% CO₂ incubator.

Prior to plating, neurons were transfected with a maximum of 0.6 μg total plasmid DNA using a Nucleofector (Lonza, Basel, Switzerland) and following the manufacturer’s instructions. For Halo-ATG14- and Halo-ACAP3-transfected neurons, DRG neurons were incubated with 100 nM of JF646-Halo ligand (from Luke Levis, Janelia Farms, HHMI) for at least 30 min at 37°C in a 5% CO₂ incubator. After incubation, neurons were washed three times with complete equilibrated F-12 media, with the final wash remaining on the neurons for at least 15 min at 37°C in a 5% CO₂ incubator.

Immunofluorescence

DRGs were isolated, plated, and cultured as described above. At DIV2, DRG neurons were fixed in pre-warmed 50% Bouin’s solution (Millipore-Sigma, HT10132) with 4% sucrose (Fisher, S5–500) in PBS (137 mM NaCl [Fisher, S271–500], 2.7 mM KCl [Fisher, P217–500], 10 mM Na₂HPO₄ [Fisher, S374–500], 2 mM KH₂PO₄ [Fisher, P285–500], pH 7.4) for 30 min at room temperature. Neurons were washed three times with 1X PBS, then incubated in Cell Block (1X PBS with 1% BSA [Fisher, BP1605–100] and 5% normal goat serum [Gibco, 16210064]) with 0.4% saponin (Millipore-Sigma, 47036) for 1 h at room temperature. DRG neurons were then incubated in Cell Block with 0.1% saponin containing primary antibodies for one hour at room temperature. After three five-minute washes in 1X PBS, neurons were incubated in Cell Block containing secondary antibodies for 1 h at room temperature. DRG neurons were washed three additional times in 1X PBS and then once with ddH₂O. Neurons were then mounted in Prolong Gold (ThermoFisher, P36930), cured overnight in the dark at room temperature, and assessed by spinning disk confocal microscopy. See Table 1 for antibodies used.

Table 1.

Key resources.

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
genetic reagent (M. musculus)	C57BL/6J	Jackson Laboratory	000664; RRID:IMSR_JAX:000664
genetic reagent (M. musculus)	GFP-LC3B	RIKEN BioResource Center in Japan; PMID: 14699058	RBRC00806; RRID:IMSR_RBRC00806
transfected construct (H. sapiens)	mCherry-ATG13	PMID: 31309927		Subcloned from RRID:Addgene_22875
transfected construct (M. musculus)	Halo-ATG14L	This paper		Subcloned from RRID:Addgene_21635
transfected construct (M. musculus)	Halo-ACAP3	This paper		Cloned by PCR from brain cDNA
transfected construct (H. sapiens)	mCh-EGFP-LC3B	T. Johansen, University of Tromsø
transfected construct	pHTC	Promega
Antibody	Anti-ACAP2	Invitrogen	PA5–18209; RRID: AB_10977709	WB (1:750)
Antibody	Anti-ACAP3	Proteintech	17570–1-AP; RRID:AB_2257668	WB (1:1000)
Antibody	Anti-ATG13, rabbit polyclonal	Abcam	ab105392; RRID: AB_10892365	ICC (1:50), IHC (1:100)
Antibody	Anti-ATG14, rabbit Monoclonal	Cell Signaling Technology	96752; RRID:AB_2737056	WB (1:1000)
Antibody	Anti-ATP6V1A, mouse monoclonal	ThermoFisher	MA5–27730; RRID: AB_2735124	WB (1:500)
Antibody	Anti-ATP6V1B/ V-ATPase B1/2, mouse monoclonal	Santa Cruz Biotechnology	sc -55,544; RRID: AB_831844	WB (1:500)
Antibody	Anti-CTSB (cathepsin B)	Cell Signaling Technology	31718; RRID:AB_2687580	IF (1:50)
Antibody	Anti-CTSD (cathepsin D)	R&D Systems	MAB1029; RRID:AB_2292411	WB (1:500)
Antibody	Anti-CTSL (cathepsin L)	Abcam	ab6314; RRID:AB_305417	WB (1:500)
Antibody	Anti-DYNC1I1 (dynein cytoplasmic 1 intermediate chain 1)-DYNC1I2 (dynein cytoplasmic 1 intermediate chain 2)	EMD Millipore	MAB1618; RRID:AB_2246059	WB (1:500)
Antibody	Anti-IQSEC2	Invitrogen	PA5–72831; RRID: AB_2718685	WB (1:250)
Antibody	Anti- MAPK8IP1/JIP1	R&D Systems	AF4366; RRID: AB_2281680	WB (1:500)
Antibody	Anti- MAPK8IP3/JIP3	Invitrogen	PA5–38828; RRID:AB_2555421	WB (1:500)
Antibody	Anti- MAPK8IP3/JIP3	MyBioSource	MBS820556	WB (1:1000)
Antibody	Anti-KIF5B	EMD Millipore	MAB1614; RRID:AB_94284	WB (1:500)
Antibody	Anti-LGMN/ asparaginyl endopeptidase (legumain)	R&D Systems	AF2058; RRID:AB_2234536	IF (1:50)
Antibody	Anti-LAMP1	Developmental Studies Hybridoma Bank	1D4B; RRID:AB_2134500	WB (1:500)
Antibody	Anti-LAMP2	Developmental Studies Hybridoma Bank	GL2A7; RRID:AB_528182	IF (1:50)
Antibody	Anti-SCARB2/LIMP2	ThermoFisher	PA3–16802; RRID:AB_2182836	WB (1:500)
Antibody	Anti-PLA2G15/LYPLA3	Biorbyt	orb185108; RRID: AB_2904247	WB (1:1000)
Antibody	Anti-DCTN1	BD BioScience	610474; RRID:AB_397846	WB (1:500)
Antibody	Anti-TUBA/TUBB (α/β-tubulin)	Cell Signaling Technology	2148; RRID: AB_2288042	WB (1:1000)
Antibody	Anti-mouse IgG-Alexa Fluor Plus 680, donkey polyclonal	ThermoFisher	A32788; RRID:AB_2762831	WB (1:20,000)
Antibody	Anti-rabbit IgG-Alexa Fluor 647, donkey polyclonal	ThermoFisher	A31573; RRID:AB_2536183	ICC (1:200)
Antibody	Anti-rabbit IgG-IRDye 680RD, donkey polyclonal	LI-COR Biosciences	926–68073; RRID:AB_10954442	WB (1:10,000)
Antibody	Anti-rabbit IgG-Alexa Fluor Plus680, donkey polyclonal	ThermoFisher	A32802; RRID:AB_2762836	WB (1:20,000)
Antibody	Anti-rabbit IgG-Alexa Fluor Plus800, donkey polyclonal	ThermoFisher	A32808; RRID:AB_2762837	WB (1:20,000)
Antibody	Anti-rat IgG-Alexa Fluor 555, goat polyclonal	ThermoFisher	A-21434; RRID:AB_2535855	ICC (1:200)
Antibody	Anti-rat IgG-IRDye 800CW, goat polyclonal	LI-COR Biosciences	926–32219; RRID:AB_1850025	WB (1:20,000)
Antibody	Anti-sheep IgG-Alexa Fluor 488, donkey polyclonal	ThermoFisher	A-11015; RRID:AB_2534082	ICC (1:200)
Antibody	Anti-sheep IgG-Alexa Fluor 680, donkey polyclonal	ThermoFisher	A-21102; RRID:AB_2535755	WB (1:20,000)
Antibody	Anti-goat IgG-Alexa Fluor 680, donkey polyclonal	ThermoFisher	A-21084; AB_2535741	WB (1:10,000)
chemical compound, drug	CellMask Deep Red	ThermoFisher	C10046
chemical compound, drug	Cresyl Violet	Sigma-Aldrich	C5042	1 μM
chemical compound, drug	LysoTracker Red DND99	ThermoFisher	L7528	25 nM
chemical compound, drug	JF646-Halo ligand	Luke Levis, HHMI Janelia Farm
software, algorithm	Adobe Illustrator CS4	Adobe Systems
software, algorithm	FIJI	PMID: 22743772
software, algorithm	Prism Prism 9	GraphPad
software, algorithm	Volocity	PerkinElmer
Software, algorithm	Nikon Elements	Nikon Instruments

Live-cell imaging and image analysis

Microscopy was performed in Hibernate A low fluorescence nutrient media (BrainBits, HALF500) supplemented with 2% B27 (Gibco, A3582801) and 2 mM GlutaMAX (Gibco, 35050061). Confocal images were captured with a spinning-disk confocal (UltraVIEW VoX; PerkinElmer, Waltham, MA) microscope (Eclipse Ti; Nikon, Tokyo, Japan) with an Apochromat 100×, 1.49 NA oil immersion objective (Nikon, Tokyo, Japan) at 37°C in an environmental chamber. Digital micrographs were acquired with an EM charge-coupled device camera (C9100; Hammamatsu Photonics, Japan) using Volocity software (PerkinElmer, Waltham, MA). Confocal images for Figure 6 were captured with a spinning-disk confocal (W1 confocal system; Nikon Instruments, Tokyo, Japan) with an Apochromat Lambda 100×, 1.45 NA oil immersion objective (Nikon Instruments, Tokyo, Japan) at 37°C in an environmental chamber. Digital micrographs were acquired with a back-illuminated sCMOS camera (Teledyne Photometrics, Tucson, AZ) using Nikon Elements software (Nikon Instruments, Tokyo, Japan). The Perfect Focus System was used to maintain Z position during time-lapse acquisition on both confocal systems.

To capture autophagosome biogenesis, time-lapse videos were acquired for 10 min with a frame every 3 s. To capture autophagosome transport, time-lapse videos were acquired for 3 min with a frame every 3 s. For each imaged neuron, a time-lapse video was acquired in the distal axon (within 150 µm of the axonal tip) and then a subsequent time-lapse video was acquired in the proximal axon (as defined by being within 150 µm of the soma). For mid-axon autophagosome transport, the mid-axon was defined as at least 250 µm away from both the distal tip and the soma. Multiple channels were acquired consecutively, with the green (488 nm) channel captured first, followed by red (561 nm), and far-red (640 nm). DRG neurons were selected for imaging based on morphological criteria and low expression of transfected constructs. To minimize artifacts from overexpression, neurons within a narrow range of low fluorescence intensity were chosen for imaging, ensuring the analyzed neurons expressed low levels of the ectopic tagged proteins.

All image analysis was performed on raw data. Images were prepared in FIJI [60]; contrast and brightness were adjusted equally to all images within a series. Time-lapse micrographs were analyzed with FIJI [60]. To quantify AV biogenesis, GFP-LC3B puncta were tracked manually using FIJI. An AV biogenesis event was defined as the de novo appearance of a GFP-LC3B punctum based on changes in fluorescence intensity over time. For GFP-LC3B puncta that were present at the start of the time-lapse series, only those puncta that increased in fluorescence intensity and/or area with time were counted as AV biogenesis events. To quantify autophagosome transport, kymographs were generated in FIJI using the MultiKymograph plugin (line width 5) and analyzed in FIJI. Autophagic vesicles were classified as anterograde (net movement ≥10 μm toward the axon distal tip), retrograde (net movement ≥10 μm toward the cell body), or bidirectional/stationary (net movement <10 μm). Time-lapse videos were referenced during kymograph analysis to ensure that extraneous neurites were not included in the data. Transport parameters were calculated in Microsoft Excel from the coordinates identified in FIJI.

To visualize lysosomes, LysoTracker Red DND99 (Invitrogen, L7528) or Cresyl Violet (Millipore-Sigma, C5042) was added to DIV2, untransfected DRGs. LysoTracker was added at 25 nM and incubated for 30 min at 37°C in a 5% CO₂ incubator. CellMask^TM Deep Red (Invitrogen, C10046) was added for the last 5 min of the incubation. Cells were washed once with equilibrated complete F-12 prior to imaging. CellMask^TM Deep Red and Cresyl Violet (1 μM) were added for 5 min and washed three times prior to imaging.

Z-stack micrographs were initially processed in FIJI into maximal projections. Maximal projection micrographs were segmented, processed, and subsequently analyzed in FIJI using Analyze Particles.

Immunoblotting

Brains of non-transgenic mice were dissected and subsequently homogenized and lysed. Brains were homogenized individually in RIPA buffer [1× PBS (see above), 1% Triton X-100 (Fisher, BP151500), 0.5% deoxycholate (Fisher, BP349100), 0.1% SDS (Fisher, BP166500), 1× cOmplete^TM protease inhibitor cocktail (Roche, 11836170001), and 1× Halt^TM protease and phosphatase inhibitor cocktail (ThermoFisher Scientific, 78440)]. Total protein in each lysate was determined by BCA assay (ThermoFisher Scientific, 23225) to ensure equal protein loading during western blot analysis.

All supernatants were analyzed by SDS-PAGE western blot, transferred onto FL PVDF membranes (Millipore-Sigma, IPFL00010), and visualized with fluorescent secondary antibodies (Li-Cor, ThermoFisher) using an Odyssey® CLx imaging system (Li-Cor, Lincoln, Nebraska). See Key Resources Table for antibodies used. All western blots were analyzed with Image Studio (Li-Cor). Total protein was used as a loading control (REVERT^TM Total Protein Stain, Li-Cor, 926–11021). The normalization factor is listed below each blot as a percent.

Statistical Analysis

Prism 9 (GraphPad) was used to plot graphs and perform statistical tests. Specific statistical tests used are indicated in the text and figure legends.

Funding Statement

The work was supported by the National Institute of Neurological Disorders and Stroke [R01 NS060698]; National Institute of Neurological Disorders and Stroke [K99 NS109286]; National Institute of Neurological Disorders and Stroke [R00 NS109286].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2023.2236485