An Elongator mouse model of ALS spotlights TDP-43 in the motor neuron nucleolus

Dysfunction of the Elongator complex is associated with amyotrophic lateral sclerosis (ALS). To investigate the potential mechanisms via which Elongator contributes to ALS, we generated conditional knockout mouse models in which either Elp1 or Elp3 is selectively ablated in cholinergic neurons including alpha motor neurons of the spinal cord. These mice exhibit numerous ALS phenotypes including a progressive loss of motor strength, motor neuron degeneration, and denervation of the neuromuscular junction. To interrogate the molecular mechanisms that contribute to motor neuron cell death in this model, we examined multiple disease pathways including the expression of TDP-43, a protein whose mislocalization and aggregation are associated with both familial and sporadic forms of ALS. Surprisingly, we see robust expression of TDP-43 in the nucleolus of motor neurons from control mice and clearing of TDP-43 from nucleoli in the conditional knockout. Further analysis of the nucleolar marker fibrillarin indicates that motor neurons in Elp conditional knockout mice exhibit nucleolar shrinkage, an indicator of nucleolar stress. Although the presence of TDP-43 in the nucleus is well characterized, a function for TDP-43 in the nucleolus, the nuclear subdomain where ribosome biogenesis takes place, has not been previously described. Thus, this study directly links dysfunction of the Elongator complex with TDP-43 clearing and nucleolar disruption, two hallmark cellular pathologies of ALS.  

Methods
Mice

All experiments with animals were performed according to the National Institutes of Health Guide for Care and Use of Laboratory Animals and were approved by the Montana State University Institutional Animal Care and Use Committee.  Additionally, all experiments involving live mice were conducted according to ARRIVE guidelines 2.0 where applicable.   Sample sizes were based on pilot studies and sample sizes were calculated using a statistical power of 80%.  Ikbkaptm1a(KOMP)Wtsi− ‘knockout first’ mice containing a frt-flanked LacZ Elp1 (previously known as Ikbkap) reporter that disrupts Elp1 expression before the LoxP flanked 4th exon were obtained from the International Knockout Mouse Consortium.  This strain was generated on a C57Bl/6j background and has been previously described 26. Elp1 CKO mice (Chat-Cre; Elp1LoxP/LoxP) were generated by crosses between Chat-Cre; Elp1+/LoxP X Elp1LoxP/LoxP; Chat-GFP/GFP mice.  A humane endpoint for Elp1 CKO males was defined by a mass < 55% of their control counterparts combined with a hindlimb clasping score of three and constant fasciculations when moving about the home cage.  A humane endpoint for Elp1 CKO females was defined by a weight < 60% of their control counterparts combined with a hindlimb clasping score of three and constant fasciculations when moving about the home cage.  The following strains were purchased from the Jackson Laboratory: ROSAmT-mG, stock no. 007576; Chat-Cre, stock no. 006410; Chat-GFP, stock no. 007902.  Elp3Lox mice have been described previously 36.   

Floxed Elp3 mice were generated on a 129 SvJ background.  Elp3 CKO mice were generated through the same genetic strategy as Elp1 CKO mice (Chat-Cre; Elp3+/LoxP X Elp3LoxP/LoxP) except that the Chat-GFP allele was not incorporated into the Elp3LoxP/LoxP strain.  To visualize cholinergic neurons in Elp3 CKO mice and controls, an anti-ChAT antibody was used (see below).

Hindlimb clasping scoring

Hindlimb clasping was scored using a previously published scoring system30.  Briefly, the mouse is suspended by the tail and hindlimb position is observed for 10 seconds.  If hindlimbs are consistently splayed outward, away from the abdomen, the mouse is assigned a score of 0. If one hindlimb is retracted toward the abdomen for more than 50% of the time suspended, it receives a score of 1. If both hindlimbs are partially retracted toward the abdomen for more than 50% of the time suspended, it receives a score of 2. If both hindlimbs are entirely retracted and touching the abdomen for more than 50% of the time suspended, it receives a score of 3.

PaGE testing

Paw grip endurance (PaGE) testing as a measure of motor function was performed as described in Weydt et al. (2003)49, and was modified slightly.  Briefly, each mouse was placed on a wire lid from a conventional rodent housing cage; the lid was gently shaken to induce gripping and turned upside down (180°). The latency until the mouse released both hind limbs was measured in seconds. Each mouse was tested five times with an arbitrary maximμmof 60 s, and the longest latency to fall or release both hind limbs was recorded. 

Immunofluorescence (IF)

All washing, blocking, secondary antibody, and post-fixation steps were performed at room temperature. All other steps were performed at 4°C unless stated otherwise.  Following fixation, tissues were rinsed in PBS, cryoprotected through a series of sucrose solutions in PBS (15%, 30%), incubated for 2 h in a 1:1 mixture of 30% sucrose and optimal cutting temperature (OCT) compound (Tissue-Tek,Torrance, CA), followed by 2 h in OCT.  Tissues were then embedded in OCT and frozen in a dry ice ethanol bath.  For immunostaining, slides were bathed in TBS (tris-buffered saline) for 10 min, followed by NGS block (10% normal goat serum, 1% glycine, 0.4% Triton X-100 in 30mM Tris, 150mM NaCl) for 1 h, and overnight incubation in primary antibody (in NGS block). Slides were then rinsed in NGS block, incubated in Alexa Fluor secondary antibody (1:2,000 in NGS block) for 1 h, rinsed in 3:1 TBS:NGS block, and mounted in Prolong Antifade Diamond (Invitrogen, La Jolla, CA). Control and experimental embryos were cryosectioned on the same day and sections were incubated in primary antibody on the same day that they were sectioned. For antigen retrieval, slides were incubated in citrate buffer at 95°C for 10 min following sectioning.

Antibodies

Primary antibodies included the following:  GFP (Abcam, ab13970, 1:1000), ELP1 (Abnova PAB12857, 1:800, knockout validated26, 50.  ELP3 (Proteintech, 17016-1-AP, 1:200, knockout validated here using Chat-Cre; Elp3LoxP/LoxP mice), NeuN (Proteintech, 26975-1-AP, 1:4000, see the manufacturer’s website for validation details), TDP-43 RRM2 (Proteintech, 10782-2-AP, 2 ug/ml, knockout validated).  *Note: although the manufacturer’s website describes this antibody as recognizing the TDP-43 N-terminus, Tsuji et al., 201238, showed the antibody recognizes amino acids 203-209 of human TDP-43 which corresponds to the second RNA recognition motif. TDP-43 C-terminus (Proteintech, 12892-1-AP, 2 ug/ml, knockout validated), Fibrillarin (Novus Biologicals, NB300-269, 1:500, knockdown validated), ChAT (Millipore Sigma AB144P, 1:100, see the manufacturer’s website for validation details).  Secondary antibodies used were Alexa Fluor goat anti-rabbit 488, goat anti-mouse 568, donkey anti-chick 488, donkey anti-goat 568 (Invitrogen, 1:2,000). 

Immunohistochemistry

LacZ staining was performed as previously described51.  Briefly, slides with tissue sections (20 um) were incubated for three hours in 1 ml of X-gal solution (1 mg/ml X-gal, 5mM potassiμmferricyanide, 5mM potassiμmforrocyanide, 2mM MgCl2, 0.25% Triton X-100 in PBS) at 30 degrees C.  Tissue sections were then fixed in 1 ml of 4% paraformaldehyde for 10 min, followed by three washes in TBS.  To combine with IF, tissue sections were then incubated in NGS block followed by the application of primary antibody as above.

Alpha motor neuron quantification

The Chat-GFP allele was included in all Elp1 CKO experiments where alpha motor neurons were quantified or measured.  Only cells in the ventral horn (Fig. S6) that were GFP-positive and NeuN-positive were selected for further analysis34.  For adult animals (6 weeks), spinal cords were removed from both control (Chat-Cre; Elp1+/LoxP) and CKO (Chat-Cre; Elp1LoxP/LoxP) animals via hydraulic extrusion52.  The lumbar enlargement was isolated and fixed for 2.5 h at 4°C in 4% paraformaldehyde in PBS.   Sixty 16 μm sections were trimmed to reach the L5/L4 level.  Every other 16 μm section was then collected, filling one slide (approximately 48 sections).  IF was performed as described above using anti-GFP and anti-NeuN antibodies, and the field of view (40X) containing the highest number of GFP-positive neurons in the ventral horn (lamina IX, Fig. S6) was photographed.  For control mice, sections to be photographed were selected randomly using a random number generator.  For the CKO, given the reduced number of large motor neurons, all sections were photographed, and the field of view (40X) was positioned to include the largest possible number of large motor neurons.  The number of alpha motor neurons (GFP-positive and NeuN-positive) per ventral horn were counted and their areas measured using ImageJ.  Averages were then calculated for the 10 sections with the largest number of alpha motor neurons (GFP-positive, NeuN-positive, and an area > 440 μm2) present per section.  The same approach was used to quantify the number of GFP-positive, NeuN-positive neurons with an area less than 440 μm2.  For E18.5, embryos were decapitated and fixed in 4% paraformaldehyde in PBS for 2.25 hours.  Sections were cut at 16 μm and every other section was collected from the mid to upper lumbar axial level, filling two slides.  Every other section was photographed, and the number of alpha motor neurons in a single hemisphere (the hemisphere with the highest number of alpha motor neurons) counted per section and an average calculated for the 12 sections with the largest number of alpha motor neurons present per section.

Neuromuscular junction analysis

The Chat-GFP allele was included in both control (Chat-Cre; Elp1+/LoxP) and CKO (Chat-Cre; Elp1LoxP/LoxP) animals to visualize innervation of the NMJ.  Anterior tibialis muscles were dissected from female controls and CKOs and fixed for 20 min in 4% paraformaldehyde in PBS at 4°C.  The muscles were then worked up for cryosectioning as described above for spinal cords.  Sections were cut at a thickness of 25 μm and every other section collected, filling two to three slides for a total of 60 sections.  IF was performed as described above using an anti-GFP antibody to visualize innervation of the NMJ.  Alpha bungarotoxin 555 (Molecular Probes, #B160; 1 μg/ml) was included at the secondary antibody step.  Sections to be analyzed were selected randomly using a random number generator.  NMJs were also randomly selected by generating a number between 1 and 10 and counting to that NMJ starting from the top right corner of the section.  50 NMJs were analyzed per animal and tallied as being either fully innervated (GFP pattern matches the pattern of alpha bungarotoxin), or not fully innervated (GFP either partially or totally missing).  Anterior tibialis muscles from three control and three experimental animals were analyzed per time point.

TDP-43 and Fibrillarin corrected fluorescence and area analysis

Nu area in Fig. 4j corresponds to the entire nucleus (including the nucleolus).  Fluorescence levels per unit area were calculated as follows: No = (No corrected total fluorescence (CTF) / No area); Nu = (Nu CTF – No CTF) / (Nu area – No area);  Cytoplasm = (total cell CTF – Nu CTF) / (total cell area – Nu area).  For TDP-43 and fibrillarin fluorescence measurements, spinal cords from control (Chat-Cre; Elp1+/LoxP) and CKO (Chat-Cre; Elp1LoxP/LoxP) were collected, processed for cryosectioning, and sectioned as described above (alpha motor neuron quantification).  IF was performed as described above using either anti-TDP-43 and anti-GFP primary antibodies, or anti-fibrillarin and anti-GFP antibodies.  Antigen retrieval was included for fibrillarin.  For controls, a random number generator was used to randomly select sections.  The field of view in the spinal cord ventral horn (lamina IX, Fig. S6) (40X) with the highest density of GFP-positive neurons was photographed.  For CKOs, given the reduced number of surviving motor neurons, all GFP+ cells in the ventral horn were photographed. Identical exposure, gain, and offset settings were used for both control and experimental images. Next, for both controls and CKOs, ImageJ was used to measure the area of photographed, GFP+ cells.  In cells with a soma area > 440 μm2 and that contained a whole intact nucleus via DAPI staining, TDP-43 or fibrillarin areas and fluorescence levels were measured using Image J processing and analysis software (https://imagej.nih.gov/ij/).  A minimum of 20 cells were analyzed per spinal cord for controls and between 8 and 20 for the CKOs given the reduced number of alpha motor neurons with soma areas > 440 μm2 in the CKO spinal cords.  A minimum of three spinal cords were analyzed per genotype and the exact number of spinal cords for each analysis is shown on the individual bar graphs in Fig. 4.

Microscopy

Images were captured using a Nikon TE200 inverted microscope equipped with a QImaging QICAM 12-bit Mono Fast 1394 Cooled camera and SPOT software. Identical exposure times, gain, and offset settings were used to capture control and experimental images.