Neuronal migration defects in the Loa dynein mutant mouse
© Ori-McKenney and Vallee; licensee BioMed Central Ltd. 2011
Received: 9 April 2011
Accepted: 25 May 2011
Published: 25 May 2011
Skip to main content
© Ori-McKenney and Vallee; licensee BioMed Central Ltd. 2011
Received: 9 April 2011
Accepted: 25 May 2011
Published: 25 May 2011
Cytoplasmic dynein and its regulatory proteins have been implicated in neuronal and non-neuronal cell migration. A genetic model for analyzing the role of cytoplasmic dynein specifically in these processes has, however, been lacking. The Loa (Legs at odd angles) mouse with a mutation in the dynein heavy chain has been the focus of an increasing number of studies for its role in neuron degeneration. Despite the location of this mutation in the tail domain of the dynein heavy chain, we previously found a striking effect on coordination between the two dynein motor domains, resulting in a defect in dynein run length in vitro and in vivo.
We have now tested for effects of the Loa mutation on neuronal migration in the developing neocortex. Loa homozygotes showed clear defects in neocortical lamination and neuronal migration resulting from a reduction in the rate of radial migration of bipolar neurons.
These results present a new genetic model for understanding the dynein pathway and its functions during neuronal migration. They also provide the first evidence for a link between dynein processivity and somal movement, which is essential for proper development of the brain.
Cytoplasmic dynein is a minus-end directed microtubule motor protein involved in a wide variety of functions. In order to perform such diverse activities in the cell, dynein makes use of numerous regulatory partners, including dynactin, LIS1, NudE and NudEL [1–3]. LIS1 and NudE/L, in particular, have received considerable attention for their role in brain development [4–7]. Sporadic mutations in LIS1 cause type I, or classical, lissencephaly, a severe brain developmental disease characterized by a smooth cortical surface . Patients with lissencephaly exhibit lamination defects of the cerebral cortex, consistent with a defect in neuronal migration. Developmental analysis of LIS1 mutant mouse lines has shown that a reduction of LIS1 protein leads to neocortical organization defects in a dose-dependent manner [5, 9–11]. Live analysis of embryonic rat brain subjected to LIS1 RNA interference (RNAi) has directly demonstrated defects in neural progenitor cell division, radial migration, and axon elongation [7, 12]. The genes encoding NudE and NudEL, Nde1 and Ndel1, are closely related and interact genetically with LIS1 and cytoplasmic dynein . Nde1 and Ndel1 mutant mice exhibit either microcephalic or lissencephalic brain phenotypes, consistent with a functional relationship to LIS1 [6, 14–16].
Several studies have focused on the role of genes in the dynein pathway in mammalian brain development. However, there are no genetic models with which to study the involvement of dynein directly. Mice homozygous null for the cytoplasmic dynein heavy chain gene exhibit early embryonic lethality, whereas heterozygotes show no obvious abnormalities . However, several dynein mutant mouse strains have since been identified in screens for genes involved in neurodegeneration [18, 19]. Of these dynein mutants, the Loa mutation has received particular attention [20, 21]. The Loa heterozygous mice are viable and develop early onset neurodegenerative disease, but the homozygous animals die perinatally due to an inability to feed . Interestingly, both the Loa/+ and the Loa/Loa mice exhibited defects in retrograde axonal transport [18, 22]. The Loa mutation resides in the amino-terminal tail region of the dynein heavy chain. Nonetheless, enzymatic, biochemical, and single molecule analysis revealed a decreased microtubule affinity for the Loa mutant dynein due to a disruption in the coordination of its two motor domains . This defect was manifested as a decrease in processivity for purified mutant dynein in in vitro single molecule assays. Particle tracking of lysosome/late endosome transport in cultured wild-type and Loa hippocampal neurons revealed a marked reduction in vesicular run lengths that could be accounted for quantitatively by the reduced processivity of the mutant dynein .
The Loa mouse represents an important model for the study of neurodegeneration. It should also, in principle, be useful in evaluating the role of dynein in other functions, such as neuronal migration. In view of its particular effects on dynein mechanochemical behavior, Loa also affords an opportunity to test whether dynein processivity contributes to this process. The Loa/Loa mice have been reported to display abnormal facial motor neuron organization, suggesting potential cellular migration defects . We report direct evidence that the Loa mutation disrupts this process, causing a delay in neocortical development, a result that confirms a more general value for this mutant mouse strain and has implications for the role of motor processivity in neuronal migration.
Many cell bodies of the migrating wild-type and Loa/Loa neurons were distorted into an elongated shape during migration. However, the wild-type cell bodies remained in this stretched conformation for 20.0 ± 11.5 minutes before returning to their normal rounded morphology after advancing, whereas Loa/Loa cell bodies remained in the stretched state for 38.0 ± 13.0 minutes (Figure 5A,B; P < 0.05; n = 12). These results are indicative of a prolonged somal translocation state due to the reduced transport by the Loa mutant motors.
We find that Loa mutant mice exhibit defects in cortical and hippocampal development. Using live cell imaging techniques, we found direct evidence for a migration delay of bipolar neurons. Our results further indicate that this effect is not associated with defects in dynein interactions with LIS, NudE, or dynactin, and is therefore likely to result directly from the inherent mechanochemical defects caused by the Loa mutation .
The abnormal cortical lamination we observe in the Loa/Loa mouse brain is similar to that seen in LIS1 compound heterozygous mutant mice [5, 9]. LIS1 mice exhibit severe disorganization of pyramidal cells in the CA1, CA2, and CA3 regions, as well as a reduced density of granule cells in the dentate gyrus [9, 39]. We observe a clear delay in the migration of granule neurons in the Loa/Loa mouse to form the dentate gyrus. Although the formation of the dentate gyrus has not been examined in the LIS1 mice, a delay in granule cell migration could provide a potential explanation for the diminished size of the dentate gyrus in these mice.
We detected a decrease in axon extension in Loa/Loa neurons both in vivo and in vitro, as well as a decrease in axon length. Dynein as well as LIS1 are necessary for the reorganization and extension of the growth cone during axonogenesis [7, 35, 36]. These proteins are required to help microtubules extending into the peripheral zone of the growth cone resist retrograde actin flow . A Loa mutant dynein detaches from microtubules about twice as frequently as the wild-type motor protein . Therefore, microtubule penetration into the peripheral zone could be compromised, providing an explanation for altered axon elongation.
Altered expression of LIS, cytoplasmic dynein, and NudE have each been found to affect mitotic index in the developing brain [6, 7, 40]. These effects reflect direct roles for these proteins in mitosis [31, 41, 42]. Furthermore, LIS1 and cytoplasmic dynein RNAi interfere with apical migration of radial glial nuclei to the ventricular surface, which has been found to be essential for mitotic entry [7, 43]. The similarity in mitotic index between Loa/Loa and wild-type mouse brains suggests that any effect on mitotic entry or progression must be relatively minor, though further live analysis of mitosis might still reveal altered mitotic behavior.
The current study indicates that the Loa mutation affects neuronal migration in a similar manner as reduced LIS1 expression [5, 7, 9]. However, the Loa mutation affects dynein processivity , whereas LIS1 contributes to dynein force production . Our results indicate, therefore, that different defects in dynein function can lead to a common developmental outcome. The mechanical basis for this result is uncertain, but is likely a consequence of the coordinate behavior of multiple dynein molecules during high load functions. In in vitro studies, LIS1 was found to convert cytoplasmic dynein to a persistent force-producing state, though the level of force generation by individual dynein molecules was unaffected. However, force generation by multiple motors is summated, and this effect was further enhanced by LIS1 as revealed in in vitro laser trap bead assays and in computational simulations . Thus, LIS1 appears to be ideally suited for a role in high load functions such as nuclear transport [12, 44]. The Loa mutation, in contrast, had a clear effect on motor processivity . How the Loa mutation might affect transport of high load structures is uncertain, but could also involve coordination of multiple dynein molecules. The Loa mutation results not only in decreased single molecule processivity in vitro, but also in decreased run-lengths for lysosomes/late endosomes in vivo, which are thought to be driven by an average of approximately five to seven dyneins [23, 45, 46]. Computational modeling indicates that, although cargo run length increases with motor number, the Loa processivity defect extends to the multimotor condition . Thus, defects in neuronal migration could conceivably be explained by a reduction in run length during nuclear transport. We reason, however, that Loa should also affect aggregate force production by multiple motors, which is a reflection of the number of individual motors simultaneously 'engaged' with microtubules at any given time. This value should be reduced for Loa as a simple consequence of its increased microtubule detachment rate. Thus, the defect in neuronal migration we observe in the Loa/Loa mouse could result from shorter nuclear run lengths, decreased force, or both. The prolonged state of somal distortion we observe during neuronal migration in the Loa/Loa neurons could be indicative, in particular, of a reduction in nuclear translocation forces, but further work will be needed to directly determine the effects of the Loa mutation at the multiple motor level.
The defects we observe in the mutant neurons may explain the perilethality of the Loa/Loa mice. The Loa homozygotes die within 24 hours of birth due to an inability to move their mouths, and therefore feed. A 2-day delay in the migration of facial motor neurons to their destination may be responsible for this facial paralysis. Correspondingly, a delay in axon extension could also contribute to facial paralysis due to a lack of muscle innervation.
Our results reveal a clear effect of the Loa mutation on neuronal migration, in addition to its previously reported effects on retrograde vesicular transport and on motor and sensory neuron viability. Our observations establish the utility of the Loa mouse as a hypomorphic dynein mutant model for the investigation of not only neurodegeneration but also general cytoplasmic dynein function.
Mice were housed in standard cages on a 12:12 light:dark schedule at 25°C. All experimental research was performed according to the guidelines of the Institutional Animal Care and Use Committee (IACUC). All experimental analyses were performed on mutant animals with littermate-matched controls.
Brains were fixed in 4% paraformaldehyde for 24 hours after extraction, then transferred to phosphate-buffered saline (PBS) until sectioning. Brains were sectioned on a Vibrotome (Leica VT 1200S) into 50-μm thick slices. For staining with primary antibodies, slices were blocked at room temperature for 1 hour in PBS with 10% donkey serum and 1% Triton X-100. For antibodies requiring antigen retrieval (such as Foxp2), slices were boiled in 10 mM sodium citrate buffer (pH 6), then cooled back to room temperature before blocking. Primary antibodies were applied overnight at 4°C in PBS with 1% donkey serum and 0.1% Triton-X-100 at the following dilutions: anti-ctip2, 1:500 (rat; Abcam, Cambridge, MA, USA); anti-foxp2, 1:200 (rabbit; Abcam); PH3, 1:500 (rabbit; Millipore, Billerica, MA, USA); Prox1, 1:500 (rabbit; Abcam); Pax6, 1:200 (rabbit; Covance, Princeton, NJ, USA); and TujI, 1:500 (mouse; Covance). Slices were then washed in PBS and incubated with Cy2, Cy3, or Cy5 secondary antibodies (1:1,000; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Sections were mounted onto slides using VECTASHIELD HardSet Mounting Medium with DAPI (VectorLabs, Burlingame, CA, USA), then imaged on an Olympus IX81 inverted microscope. At least three brains per genotype were analyzed for each experiment.
For the BrdU birthdating assay in Figure 2, pregnant mice were injected intraperitoneally with BrdU at 100 μg/g of body weight at either E16 or E18. The BrdU-labeled females were left to give birth and their offspring were sacrificed on P0. BrdU-positive cells in the cerebral cortex were detected by immunostaining with an anti-BrdU antibody (1:200, rat; Abcam). This antibody required pre-treatment with 2N HCl for 15 minutes at 37°C prior to blocking. Staining was performed as described above. To analyze the BrdU-positive neurons in Loa/Loa mutants and their wild-type littermates, each brain section was divided into six equal regions, and the relative density and distribution of the BrdU-stained neurons in each region was plotted.
For the BrdU pulse-chase experiment in Figure 5B, pregnant mice were injected intraperitoneally with BrdU at 100 μg/g of body weight at E16. The mice were then sacrificed 2 hours after the injection to analyze the number of BrdU pulse-labeled cells at the ventricular surface in Loa/Loa and wild-type embryonic brains.
Plasmids were transfected using intraventricular injection followed by in utero electroporation as previously described . Briefly, 1 to 2 μl of GFP cDNA (concentration of 1 to 3 μg/μl) were injected into the ventricle of embryonic brains at E15. A pair of copper alloy oval plates attached to the electroporation generator (Harvard Apparatus, Holliston, MA, USA) transmitted five electric pulses at 40 V for 50 milliseconds at 1 second intervals through the uterine wall. The embryos were then sacrificed 3, 4, or 5 days later and their brains were prepared as described above. The number of GFP-labeled neurons at the ventricular zone, subventricular/intermediate zone, and cortical plate was plotted for Loa/Loa and wild-type littermates to assess the progression of neuronal migration from the ventricular surface.
These experiments were performed using a modified procedure of previously described methods [28, 47]. Briefly, E16 brains were collected and sectioned coronally on a vibratome (Leica), then placed in a 50-mm MatTek dish (MatTek Corporation, Ashland, MA, USA) containing imaging medium (F12/MEM medium with 10% fetal bovine serum and 1 × penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA) with 10 μg/ml of Oregon Green BAPTA-1 488 AM (Invitrogen). Slices were then incubated for 1 hour at 37°C with 5% CO2. After the incubation, the slices were washed in warmed PBS, then covered with a thin layer of Matrigel (BD Biosciences, San Diego, CA, USA) diluted 70 × in imaging medium, then incubated again for 30 minutes. The dish was then imaged in 1 ml of additional imaging media with 10 mM Hepes at 37°C with 5% CO2 on an Olympus IX81 inverted microscope. Images were captured every 10 minutes for 3 to 5 hours using an Hamamatsu ORCA-R2 CCD camera and Metamorph software (Molecular Devices, Sunnyvale, CA, USA. The speed of migration and length of axon were analyzed with Metamorph software. The position of the soma center was measured in each image and displacement of the soma was defined by the difference in location of the center between consecutive frames. Therefore, the somal speed is described here as the displacement per unit time. The length of the axon was defined as the distance from the basal end of the soma to the tip of the process. Live cell imaging experiments and axon length measurements were performed on three brains per genotype.
Whole brain lysates from P0 wild-type and Loa/Loa mouse pups were used for the NudE and CC1 pull-downs. For each experiment, the lysate was first pre-cleared with glutathione agarose 4B (USB Corporation, Cleveland, OH, USA) in order to prevent non-specific binding in the real assay. The beads were then spun down, and the pre-cleared lysate was used for the pull-downs. The pull-down solution contained the brain lysate, phosphate-glutamate buffer (10 mM sodium phosphate, 100 mM sodium glutamate, pH 7.0 with 1 mM MgSO4, 1 mM EDTA, 1 mM DTT, and 0.2% NP-40) up to 350 μl and either beads alone, 1 μg/μl GST-NudE  or 1 μg/μl GST-CC1 [37, 38]. The solutions rotated at 4°C for 2 hours, then the beads were spun down to separate the supernatants and pellets. The supernatants and pellets from each pull-down (beads alone, NudE, or CC1) were collected and analyzed by SDS-PAGE and western blot probing for dynein heavy chain (rabbit, 1:500)  and dynein intermediate chain (mouse, 1:1500; clone 74.1 provided by K Pfister, University of Virginia, Charlottesville, VA, USA). In the case of the NudE pull-down, we also probed for LIS1 (mouse, 1:1500; clone 338, Sigma, St. Louis, MO, USA).
days in vitro
green fluorescent protein
Legs at odd angles
We thank Richard McKenney for providing the GST-NudE and GST-CC1 constructs, as well as for useful discussion. This work was supported by NIH grants GM47434 and HD40182 to RBV and GM008798-09 to KMOM, and the Columbia University Motor Neuron Center.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.