Dynamics of Mitochondrial Transport in Axons

Abstract

The polarized structure and long neurites of neurons pose a unique challenge for proper mitochondrial distribution. It is widely accepted that mitochondria move from the cell body to axon ends and vice versa; however, we have found that mitochondria originating from the axon ends moving in the retrograde direction never reach to the cell body, and only a limited number of mitochondria moving in the anterograde direction from the cell body arrive at the axon ends of mouse hippocampal neurons. Furthermore, we have derived a mathematical formula using the Fokker-Planck equation to characterize features of mitochondrial transport, and the equation could determine altered mitochondrial transport in axons overexpressing parkin. Our analysis will provide new insights into the dynamics of mitochondrial transport in axons of normal and unhealthy neurons.

Keywords: anterograde transport; axonal transport; mitochondrial transport; models; retrograde transport; theoretical.

Figure 1

Mitochondrial transport in the entire axon. Mito-dendra2 was photoswitched at either the soma (anterograde) or an axonal endpoint (retrograde). (A) Kymograph for anterograde movement. Axon branch length: 1.064 mm. Scale bars: 50 μm for distance (green line), 15 min for time (red line). (B) Kymograph for retrograde movement. Axon branch length: 1.162 mm. Note that the dim red signal seen in the soma is due to the long emission tail of unconverted dendra2 and represent mitochondria already present in the soma prior to photo-activation. This signal is typically negligible; however, the cumulative brightness of hundreds of mitochondria in the soma can exhibit a red background. (C) Velocity distribution of anterograde and retrograde moving mitochondria. (D) Comparison of velocity distribution of mitochondria traveling within 200 μm of soma, and mitochondria traveling from 200 μm from the soma to the end. Anterograde forward movement proximal to the cell body was found to be slower than other areas of the axon. (E,F) Velocity distribution of mitochondrial movement in axon is compared between normal neuron and neuron overexpressing parkin. Anterograde movement is much slower in axon overexpressing parkin compared to that of normal axon (E). Retrograde movement shows similar velocity distribution between normal and parkin overexpression (F). N = 10 primary neurons for wild type anterograde, and 7 for wild type retrograde. N = 4 (anterograde) and 3 (retrograde) for neurons overexpressing parkin. ^*P < 0.03. Values shown are mean ± s.e.m., and tested for statistical significance by student's t-test.

Figure 2

The probability distribution function (PDF) derived from Fokker-Planck equation fits well into experimental data. (A) Curve fit of anterograde movement. We defined anterograde forward as movement from the cell body toward the distal axon point, and anterograde reverse as movement that reverses in direction toward the cell body. Anterograde reverse (right) was consistent regardless of distance from the cell body, but was significantly slower than anterograde forward. (B) Retrograde forward was defined as movement from the axonal endpoint to the cell body, and retrograde reverse was opposite of the retrograde forward. A retrogradely forward moving mitochondrion is much faster than that of reverse. (C,D) Comparison mitochondrial transport in normal axons with that of axons overexpressing parkin. (C–F) Anterograde forward movement is much slower in axon overexpressing parkin compared to that of normal axon. Anterograde reverse direction is also slower due to parkin overexpression. Retrograde forward movement shows similar velocity distribution between normal and parkin overexpression. Interestingly, however, retrograde reverse movement in axon with parkin overexpression exhibits extremely narrow velocity distribution than that of normal axon. Anterograde reverse direction is also a little slower due to parkin overexpression. Retrograde forward movement shows similar velocity distribution between normal and parkin overexpression.

Figure 3

Velocity distribution of mitochondrial transport in different experimental conditions. Different length of axon segments and observation time were tested to determine the minimum experimental conditions necessary to generate velocity distribution of mitochondrial transport using the derivative of Fokker-Planck equation. (A) Velocity distribution of moving mitochondria plotted for experimental conditions as indicated. Data from eight different segments of axons were combined to calculate probability of velocity distributions. (B) Modeled velocity distribution plotted for each condition using the derivative of Fokker-Planck equation. (C) Comparison of modeled velocity distribution for data taken 200 μm beyond the cell body to the distal axonal area with the indicated condition.

Figure 4

Representative kymographs for mitochondrial events in normal neurons and neurons overexpressing parkin. (A) Tagged mitochondria were tracked for dynamics events as they traveled either anterograde or retrograde through the axon over 2 h (i) full stop of mitochondria. (ii) fusion and fission. (iii) reversal of movement. (B) Quantitative analysis of anterograde mitochondrial transport in axons. (C) Dynamics of anterograde mitochondrial events in axons. For the fusion/fission event, a red moving mitochondrion is split into 1 stationary mitochondrion and 1 that continued its movement. (D) Number of mitochondria leaving the axonal endpoint over 2 h. (E) Dynamics of retrograde mitochondrial transport in axons. The number and destiny of mitochondria moving in the anterograde or retrograde direction were normalized to the length of the axon. N = 10 primary neurons for wild type anterograde, and 11 for wild type retrograde. N = 4 (anterograde) and 3 (retrograde) for neurons overexpressing parkin. ^*P < 0.01.

Figure 5

Distribution of stationary mitochondria in the entire axon. (A) Representative kymograph demonstrating analysis of stationary mitochondria. A typical kymograph (top) was thresholded to eliminate background (middle). Peaks that appeared constant over the course of 2 h were quantified as stationary (bottom). (B) Density of mitochondria throughout axon. Axonal lengths of 50 μm were binned, and the frequency of mitochondria in reference to distance from the cell body was plotted. Value = mean ± S.D. ^*P < 0.02 compared to density at 750 μm. 750 μm was selected for statistical comparison because it is roughly the middle of the average axon length monitored. A longest axon in a neuron was analyzed. N = the longest branch from 10 primary neurons.