Automated measurement of fast mitochondrial transport in neurons.

Kyle E Miller,Sathyanarayanan V Puthanveettil,Xin-An Liu

doi:10.3389/fncel.2015.00435

Kyle E Miller, Sathyanarayanan V Puthanveettil + Show 1 more

Open Access

https://doi.org/10.3389/fncel.2015.00435

Copy DOI

Abstract

There is growing recognition that fast mitochondrial transport in neurons is disrupted in multiple neurological diseases and psychiatric disorders. However, a major constraint in identifying novel therapeutics based on mitochondrial transport is that the large-scale analysis of fast transport is time consuming. Here we describe methodologies for the automated analysis of fast mitochondrial transport from data acquired using a robotic microscope. We focused on addressing questions of measurement precision, speed, reliably, workflow ease, statistical processing, and presentation. We used optical flow and particle tracking algorithms, implemented in ImageJ, to measure mitochondrial movement in primary cultured cortical and hippocampal neurons. With it, we are able to generate complete descriptions of movement profiles in an automated fashion of hundreds of thousands of mitochondria with a processing time of approximately one hour. We describe the calibration of the parameters of the tracking algorithms and demonstrate that they are capable of measuring the fast transport of a single mitochondrion. We then show that the methods are capable of reliably measuring the inhibition of fast mitochondria transport induced by the disruption of microtubules with the drug nocodazole in both hippocampal and cortical neurons. This work lays the foundation for future large-scale screens designed to identify compounds that modulate mitochondrial motility.

Highlights

Considering the abnormalities in mitochondrial transport observed in neurological disease (De Vos et al, 2008), novel compounds that manipulate transport could have therapeutic value (Kadakkuzha et al, 2014)
To confirm that the Lucas Kanade Motion Tracking Algorithm (LKMTA) as implemented in FIJI as the FlowJ plugin (Abràmoff et al, 2000, 2004) is capable of detecting fast mitochondrial transport (Gerencser and Nicholls, 2008), we analyzed a region of a larger movie where only one out of a total of approximately 200 mitochondria were undergoing fast transport
Because the difference between the movies was the removal of one mitochondrion, we could be reasonably confident that differences in the output of the LKMTA could be attributed to the movement of that mitochondrion

Summary

Introduction

Considering the abnormalities in mitochondrial transport observed in neurological disease (De Vos et al, 2008), novel compounds that manipulate transport could have therapeutic value (Kadakkuzha et al, 2014). With the development of automated microscopes (Mei et al, 2012) and the availability of libraries of small molecules (Kadakkuzha et al, 2014) it is possible to generate in a single day raw datasets where the effects of hundreds of compounds can be analyzed. The analysis of mitochondrial transport in tens to hundreds of neurons is straightforward using manual tracking and kymographs (Miller and Sheetz, 2004), methodologies have not been developed for the large-scale analysis of mitochondrial transport data generated by robotic microscopes. With the long-term goal of screening large libraries of compounds, our focus here is to test the feasibility of applying automated methods to track mitochondrial motion using data generated by a robotic microscope.

Objectives

Methods

Results

Conclusion