Persistent motor dysfunction despite homeostatic rescue of cerebellar morphogenesis in the Car8 waddles mutant mouse
- 195 Downloads
Purkinje cells play a central role in establishing the cerebellar circuit. Accordingly, disrupting Purkinje cell development impairs cerebellar morphogenesis and motor function. In the Car8wdl mouse model of hereditary ataxia, severe motor deficits arise despite the cerebellum overcoming initial defects in size and morphology.
To resolve how this compensation occurs, we asked how the loss of carbonic anhydrase 8 (CAR8), a regulator of IP3R1 Ca2+ signaling in Purkinje cells, alters cerebellar development in Car8wdl mice. Using a combination of histological, physiological, and behavioral analyses, we determined the extent to which the loss of CAR8 affects cerebellar anatomy, neuronal firing, and motor coordination during development.
Our results reveal that granule cell proliferation is reduced in early postnatal mutants, although by the third postnatal week there is enhanced and prolonged proliferation, plus an upregulation of Sox2 expression in the inner EGL. Modified circuit patterning of Purkinje cells and Bergmann glia accompany these granule cell adjustments. We also find that although anatomy eventually normalizes, the abnormal activity of neurons and muscles persists.
Our data show that losing CAR8 only transiently restricts cerebellar growth, but permanently damages its function. These data support two current hypotheses about cerebellar development and disease: (1) Sox2 expression may be upregulated at sites of injury and contribute to the rescue of cerebellar structure and (2) transient delays to developmental processes may precede permanent motor dysfunction. Furthermore, we characterize waddles mutant mouse morphology and behavior during development and propose a Sox2-positive, cell-mediated role for rescue in a mouse model of human motor diseases.
KeywordsPurkinje cell Granule cell Proliferation Stem cells Lobule Ataxia Tremor Dystonia
- (SS) CV
(Simple spike) Coefficient of Variation
- (SS) CV2
(Simple spike) Coefficient of Variation of adjacent intervals
Carbonic anhydrase VIII
External granule layer
Hematoxylin & eosin
Inner granule layer
1,4,5-triphosphate receptor type 1
Normal donkey serum
Normal goat serum
Purkinje cell layer
Standard error of the mean
Tibilias anterior muscle
The cerebellum controls motor coordination, motor learning, posture, and balance. Therefore, damage to its circuits causes a number of motor disorders, such as ataxia, dystonia, and tremor [1, 2, 3, 4, 5, 6, 7]. A range of Purkinje cell alterations may be at the center of all cerebellar disease phenotypes, irrespective of whether the causative insult occurs during development or in adulthood . In adults, Purkinje cells typically contribute to movement defects because they degenerate, whereas during development they impact motor function by failing to properly assemble the circuitry [5, 9, 10, 11, 12, 13]. This study tests how motor dysfunction arises even when the assembly of cerebellar structure is rescued, and no obvious degeneration occurs.
The functional architecture of the cerebellum is established through multiple morphogenetic, patterning, and remodeling processes . These processes include lobule formation, zonal patterning of zebra-like striped compartments, and afferent fiber synaptogenesis. Each of these processes operates within a very dynamic environment, whereby the cerebellum rapidly grows in size and the folds deepen to form sub-divisions that generate the recognizable foliation pattern in adults . Cerebellar growth and foliation are dependent on several molecular cues including sonic hedgehog, engrailed1/2, and Fgf signaling [16, 17, 18]. Regardless of the initial genetic defect, granule cell proliferation is almost always affected if cerebellar size and growth are impacted. And because of bidirectional signaling between developing granule cells and Purkinje cells , when one cell type is affected the other is as well. Disrupting Purkinje cell morphogenesis can also cause dendrite anomalies  and miswiring of afferents [21, 22, 23], which both lead to circuit firing problems [24, 25]. From an experimental perspective, it has been challenging to parse out how developmental defects impact the onset of motor dysfunction in animal models of movement disease. Major reasons for this challenge are that the affected Purkinje cells can degenerate, undergo cell death, or even look normal but function poorly due to a compromised circuit. Such variations call into question how developmental mechanisms intersect with structure, functional circuitry, and behavior in ataxia and related disorders [26, 27, 28, 29].
Car8wdl mice are ideal for testing how morphogenesis and wiring impact motor dysfunction . In the brain, CAR8 protein is expressed predominantly in Purkinje cells. Its expression is initiated during embryogenesis and maintained into adulthood [31, 32]. CAR8 belongs to a family of zinc metalloenzymes that catalyze the reversible hydration of CO2 , although CAR8 lacks the catalytic domain that would make it an active carbonic anhydrase . It does, however, bind to inositol 1,4,5-triphosphate receptor type 1 (IP3R1), where it has the proposed effect of decreasing the affinity of IP3 for its receptor . Car8wdl mutant mice have ataxia, tremor, and appendicular dystonia, with cerebellar microcircuit abnormalities [30, 35] supposedly occurring without gross anatomical defects . In humans, mutations in the orthologous gene, CA8, also cause ataxia and a predisposition for quadrupedal locomotion [37, 38]. To better understand how the motor deficits arise and then persist without any accompanying severe morphological defects, we used a combination of genetics, gene expression analysis, in vivo electrophysiology, and mouse behavior to follow cerebellar morphogenesis and function in postnatal developing Car8wdl mice. We uncover transient defects in cerebellar size, Purkinje cell morphology, and granule cell proliferation during development in Car8wdl mice. Although most of the structural deficits are corrected by weaning, neural circuit function remains impaired and behavior deficits persist in adulthood.
Car8wdl mutant mice (Stock 004625), C57BLKS/J control background strain, and NpyGFP mice (B6.Cg-Tg(Npy-MAPT/Sapphire)1Rck/J, Stock 008321) were purchased from The Jackson Laboratory (Bar Harbor, ME) and then maintained in our animal colony at Baylor College of Medicine. We bred the control and mutant mice using timed pregnancies, and we designated noon on the day a vaginal plug was detected as embryonic day (E) 0.5 and the day of birth as postnatal day (P) 0. We used a standard PCR genotyping protocol to differentiate the mutants from the controls using the same primer sequences as previously described [30, 36]. Mice of both sexes were studied. They had food and water ad libitum. All animal studies were carried out under an approved IACUC animal protocol according to the institutional guidelines at BCM.
Perfusion, basic histology, and tissue staining procedures
After being anesthetized under 2,2,2-tribromoethanol (Avertin), mice (ages P5, P10, P15, P17, P20, P180, and P360 adults) were transcardially perfused, first with 0.1 M PBS (pH 7.2) then with 4% paraformaldehyde (PFA). Dissected brains from the perfused mice were post-fixed in 4% PFA for at least 24 h then transferred onto 2% agar for whole mount imaging, transferred sequentially through a series of sucrose solutions (18 and 30%) for cryoprotection, or embedded in paraffin. Whole mount images were taken using Zeiss AxioZoom VI6 to compare cerebellar size and lobulation pattern across 5 ages (P5, P10, P15, P20, Adult) in Car8wdl and C57BLKS/J mice. Cryoprotected and frozen tissues were cut sagittally on a cryostat into 40 μm (or 80 μm, as for Golgi-Cox staining) sections and stored at 4 °C, free-floating in PBS. Midline sections of the cerebellum were either stained following Golgi-Cox and immunohistochemistry staining protocols published previously [39, 40, 41, 42] or following a modified protocol for manual hematoxylin and eosin (H&E) staining of frozen tissue . Tissues embedded in paraffin first were dehydrated by overnight incubations in 70% ethanol, 95% ethanol, 100% ethanol, and chloroform. Dehydrated tissue was then transferred to plastic cassettes, where they soaked in 2 changes of paraffin (Paraffin #1 for ~ 18 h, Paraffin #2 for ~ 2 h) in a 65 °C oven. Hot paraffin covered the brains and using forceps, brains were positioned on their sides for sagittal sectioning. Tissue was left overnight at room temperature until the paraffin solidified. Paraffin-embedded tissues were cut on a microtome into 10 μm sections, mounted onto slides from a warm water bath, and transferred to a 65 °C oven overnight for ~ 16–18 h.
We performed the Golgi-Cox staining  with a modified procedure using the FD Rapid GolgiStain Kit (FD Neurotechnologies, Inc) on 80 μm-thick tissue sections that were processed as described by the manufacturer using a free-floating approach. Stained tissue sections were imaged using a Zeiss AxioImage.M2 microscope (additional details below).
Tissues stained via immunohistochemistry were first incubated in 10% Normal Donkey Serum (NDS, Sigma) or 10% Normal Goat Serum (NGS) blocking solution (0.1% Tween 20 and 0.1 M PBS) for 1–2 h then were incubated with primary antibodies in new aliquots of 10% NDS or 10% NGS blocking solution overnight (~ 16–18 h). A Calbindin D-28 K monoclonal mouse antibody (Swant #300) was used to visualize Purkinje cell soma, axons, and dendrites at a 1:10,000 dilution. A CAVIII polyclonal rabbit antibody (Santa Cruz sc67330) differentiated Purkinje cells with carbonic anhydrase 8 (CAR8) from other cells that lacked this metalloenzyme at a 1:500 dilution. To visualize proliferating cells, a polyclonal mouse Ki67 antibody (BD Pharmingen #55069 Lot 33,112,568) was used at a 1:400 dilution with NeuroTrace 435/455 Blue Fluorescent Nissl Stain (ThermoFisher Scientific #N21479) or DAPI (Vectashield Antifade Mounting Medium with DAPI #H-1200) to distinguish the External Granular Layer (EGL). To visualize mitotically active cells, a polyclonal mouse PH3 (Ser10) antibody (Cell Signaling 9701S Lot 16) was used at a 1:200 dilution with NeuroTrace 435/455 Blue Fluorescent Nissl Stain (ThermoFisher Scientific #N21479) or DAPI (Vectashield Antifade Mounting Medium with DAPI #H-1200). Sox2 (Invitrogen PA1–094 Lot RH235771 and Millipore Sigma Cat. No. AB5603) polyclonal antibodies were used at a 1:50 and 1:200 dilution, respectively, to determine if cerebellar stem cells (or granule cell precursors expressing Sox2) were present in the P5 and P10 EGL. There were no obvious differences in Sox2 staining between the two antibodies on the cerebellar tissue tested. To evaluate the zonal patterning of Bergmann glia and Purkinje cells during Car8wdl development, a rabbit polyclonal anti-HSP25 (StressGen #SPA-801) and a chicken anti-GFP (Abcam Cat. No. ab13970) were used at a 1:500 and a 1:2000 dilution, respectively. Tissue was rinsed 3–4 times for 5 min each with 0.1 M PBS before incubating the tissue with fluorescent Alexa 488-, 555-, or 647- immunoglobins (Invitrogen Molecular Probes Inc., Eugene, OR, USA #A-21202, #A-31572, and #A-31573) diluted 1:1500 in new aliquots of 10% NDS or 10% NGS blocking solution for 2 h. Following additional 4 X 5-min rinses, the tissues were mounted onto electrostatically coated glass slides, coverslipped with FLUORO-GEL (with Tris Buffer) and then imaged using a Zeiss AxioImage.M2 microscope (additional details below).
Tissues stained with H&E followed one of two procedures. If tissues were sectioned on the cryostat, they were first mounted onto slides and allowed to dry for approximately 5 min. Once there appeared to be no more moisture on the adhered tissues, the slides were placed in hematoxylin for approximately 5 min, transferred to lithium until the sections turned a deep blue, then placed in eosin for about 1 min. Sections were next subjected to a series of ethanol washes for dehydration (70, 95, 100, 100, 100%, xylene, xylene). Tissues were dipped 20 times in both 70 and 95% ethanol, 10 times in the first change of 100% ethanol, then 5 times in the last 2 changes of 100% ethanol. Slides were transferred to the first change of xylene for 3–5 min, and then moved to the second change of xylene for 3–5 min. Slides were then cover-slipped with Cytoseal and imaged using a Zeiss AxioZoom VI6 microscope after at least 24 h in the fume hood. Paraffin-embedded tissues were left on slides overnight in a 65 °C oven. To remove the paraffin, slides containing the sections were relayed from 2 changes of xylene, to 2 changes of 100% ethanol, to 95% ethanol, and lastly to 85% ethanol. In each of the solutions, slides were immersed for 2 min. After the 85% ethanol wash, slides were placed in hematoxylin for approximately 3 min, transferred to lithium until the sections turned a deep blue, then placed in eosin for about 30 s. Like the cryostat-sectioned tissues, after eosin staining, paraffin-sectioned tissues were processed through the same dehydration series as indicated above. Paraffin-embedded tissue sections were also cover-slipped with Cytoseal and imaged using a Zeiss AxioZoom VI6 microscope approximately 24 h after staining.
In vivo EdU labeling
The protocol followed was adapted from Mead and Lefebvre (2014) . In vivo EdU labeling was performed using the Click-iT EdU 488 Imaging Kit (Invitrogen Catalog #C10337). P5 and P10 mutant and control pups were injected either subcutaneously (P5) or intraperitoneally (P10) with 100 μL of 5-ethynyl-2′-deoxyuridine (EdU) per 10 g. After 4 h, the injected pups were anesthetized with isoflurane, and then decapitated to remove the brain. The brain was immersed in 4% paraformaldehyde (PFA) for 48 h at 4 °C and then transferred to a demineralization solution consisting of 1% PFA and 0.5 M Tris-EDTA (pH 8) at 4 °C. The demineralization solution was replaced with fresh aliquots every other day for 7 days. After 7 days, the brains were cryoprotected, embedded, and sectioned following the methods described above (see Perfusion, basic histology, and tissue staining procedures). At least 3 sections from 4–5 mutants and 5–6 controls at P5 and P10 were mounted onto electrostatically coated glass slides and dried on the bench for 1 h at room temperature (22–25 °C). The mounted tissue was washed 3 times in 1X PBS for 2 min each before being incubated in Click-iT reaction cocktail for 30 min. The Click-iT reaction cocktail consisted of a 1X Click-iT reaction buffer, copper sulfate (CuSO4), Alexa Fluor 488 azide, and 1X reaction buffer additive. Following incubation, the tissue was washed twice for 5 min each in 1X PBS. Slides were coverslipped with FLUORO-GEL (with Tris Buffer) and then imaged using a Zeiss AxioImage.M2 microscope (additional details below).
Image acquisition and quantification
Photomicrographs of some of the NpyGFP reporter stained tissue sections were captured using a Leica DFC 360 FX camera mounted on a Leica DM5500 microscope. Images of the rest of the tissue sections were acquired and analyzed using Leica Application Suite software. For the H&E and fluorescently stained tissue sections, at least 2 sections from at least 3 animals were analyzed per developmental age studied (P5, P10, P15, and P20). Sagittal tissue sections containing all 10 lobules were selected within ±0.36 mm from the midline , as determined by the lack of cerebellar nuclei seen in the white matter. The medial-lateral coordinates of the tissue sections selected were matched across genotype for each of the observed ages. The H&E stained tissues that were used to quantify cerebellar size were imaged on the Zeiss AxioZoom VI6. The areas of the sagittal cross-sections were calculated using ImageJ software. The fluorescently stained tissues that were assessed for CAR8, Ki67, calbindin, PH3, Sox2, and EdU expression were imaged on the Zeiss AxioImage.M2 using the Apotome and/or Z-stack features. ImageJ was used for automated and manual counting of Ki67-positive, PH3-positive, EdU-positive cells in the EGL of lobules V and VI. Lobules V and VI were selected for analysis due to their roles in locomotion [47, 48, 49, 50, 51]. The long, straight regions of the cerebellar cortex in these lobules (particularly the depths within the primary fissure) are also ideal for anatomical analysis since this architecture allows the acquisition of consistent measurements. To estimate the density of PH3-positive cells in P5 and P10 animals, cells were therefore counted from lobules V and VI. ImageJ was also employed to measure ML thickness, EGL area, Purkinje cell dendritic tree length, width, area, and to manually count the number of primary (i.e. from the soma) and secondary (i.e. from the primary branches) dendrites on Purkinje cells in lobules V and VI. The ML thickness was defined as the distance from the Purkinje cell layer (PCL) to the EGL. Therefore, the distance between the Purkinje cell soma and the concentrated DAPI that demarcates the EGL were measured from 3 locations (left, center, right) per ML per section. The EGL was defined by DAPI staining and its boundaries were delineated from a lack of Calbindin staining and the surface of the lobule. The area of 2 EGLs was measured per tissue section. The inner EGL refers to the ~ 27,000 μm2 (P5) or the ~ 14,000 μm2 (P10) of EGL immediately adjacent to the ML (i.e. ~ 50% of the EGL as characterized by an absence of calbindin and a sparser labeling of DAPI). Here, the area of inner EGL was calculated by roughly dividing the total control EGL area at P5 and P10 (Fig. 5) by 2. For the Purkinje cell morphology measurements, 15 non-neighboring cells were chosen from different regions of each section, which were taken from 3 different mutants and controls at P5 and P10. At P15, 8 Purkinje cells were chosen from each section belonging to 3 different mutants and controls for quantification. The number of primary and secondary branches were checked and confirmed in ImageJ through adjusting the threshold to better visualize individual Purkinje cells. The dendrite lengths, widths, areas, and branch numbers of each Purkinje cell (8 or 15) within a section were averaged. Measurements were then averaged across sections to obtain one representative measurement per parameter per animal. Student’s t-tests (p < 0.05) or a one-way ANOVA (p < 0.05) with a Tukey’s multiple comparisons post-hoc test was performed to compare the mutant and control groups at the ages studied, for ML thickness, EGL area, and the different measures of granule cell proliferation and Purkinje cell morphology. All of the histology data was plotted according to the number of animals per genotype, per age. Since each point plotted represents an animal, each point also refers to the average calculated from the pool of tissue sections.
We used the CatWalk apparatus (Noldus Information Technology; Leesburg, VA) to visualize and measure the altered gait of Car8wdl mice. P30 mice of both sexes were placed into the corridor of the apparatus and the placement of the paws and the body motion were recorded with a high-speed camera during locomotion (n = 3 of each genotype). The illuminated footprints are easily observed as the mice walk in a relatively straight line, which can be seen on the short movies provided (Additional file 1: Movie S1 and Additional file 2: Movie S2). The print area, reported as cm2, is a parameter that described the total floor area contacted by the paw during the stance phase. An unpaired, two-tailed Student’s t-test (p < 0.05) was performed to calculate the statistical significance between the footprints of the mutant and control animals. The data points plotted correspond to the average print length, width, and area per animal.
We used the Ametek Chatillon (DFE/EDFE series) meter to determine the grip strength (gf/g) of Car8wdl mutants (n = 8) compared to control (n = 8) mice. The forepaws of female and male adult mice (≥P60) were placed on a grated extender and then the mouse slowly pulled away from the apparatus. Measurements for grip strength were taken 3 times from each animal and then the values averaged. An unpaired, two-tailed Student’s t-test (p < 0.05) was performed to calculate the statistical significance between the grip strength of the mutant and control animals. The data points plotted correspond to the average grip strength per animal.
Electromyography (EMG) was used to assess motor coordination in postnatal developing Car8wdl mice and control littermates. P19–20 (n = 6 of each genotype) mice were deeply anesthetized with isoflurane so that 2 silver wires could be implanted into the tibialis anterior (TA) and gastrocnemius (GC) muscles of the left hindlimb. A ground wire was implanted into the neck and a head mount connecting all the wires was secured to the head with Metabond and dental cement. Mice were provided with a pre- and post-operative analgesic and given at least 2 days to recover before the recording sessions began. The mice were attached to a pre-amplifier via the EMG head mount and allowed to move freely in a glass container during the recording. Muscle activity was recorded for at least 15 min and the data was collected using Spike2 software. After selecting a 30-s interval of locomotion, as defined by the alternating firing of hindlimb muscles, TA and GC channels were rectified and a cross-correlation waveform analysis was performed in Spike2. Burst analysis was also performed on the same 30-s interval in Spike2. The probability of synchronous muscle contractions, the number of bursts over time, the number of spikes per burst, and the mean burst length in the TA muscle were averaged among all mice and compared between-groups using the Student’s t-test (p < 0.05). The data points plotted correspond to the average burst number, burst frequency, and burst length per TA in each mouse.
In vivo electrophysiology in behaving mice
For head-fixed, awake recordings, P60-P90 mice were implanted with custom made headplates and a craniotomy was made above the cerebellum [52, 53]. After 72 h of recovery, mice were trained for 30 min/day in a head-fixed apparatus for 3 days before recording. Cells were recorded and categorized based on standard stereotaxic coordinates measured from Bregma . Purkinje cells were identified by the presence of complex spikes. Single unit recordings were attained with 5–8 MΩ tungsten electrodes (Thomas Recording, Germany) and then the signals digitized into Spike2 (CED, England), where single units were verified with principal component analysis. Analysis of continuous traces of > 300 s was performed with Spike2, MS Excel, and MATLAB (Mathworks, Natick, MA, USA). We examined simple and complex spikes over a pre-defined period of recording. Firing pattern variability, or regularity, is defined as a measure of the consistency of time intervals between neuronal spikes (inter-spike time interval or ISI = seconds). To quantify the average variability in Purkinje cell firing pattern, the coefficient of variation of the ISI (CV) was calculated as the ratio of the standard deviation of ISIs to the mean ISI of a given Purkinje cell. We measured local regularity with the coefficient of variation of adjacent ISIs (CV2). CV2 measures firing pattern variability within a short period of two interspike intervals (CV2 = 2|ISIn + 1–ISIn|/(ISIn + 1 + ISIn)) . CV and CV2 were computed and reported as mean ± standard error of the mean (SEM). Statistical analyses were performed with unpaired, two-tailed Student’s t-tests. Statistical significance is indicated in the graphs for p < 0.01 and p < 0.0001 with ** and ****, respectively. Each data point plotted represents the firing activity of a single Purkinje cell.
Transient anatomical alterations precede motor dysfunction
Granule cell proliferation is impaired in Car8 wdl mice
Dynamic changes in the EGL and ML in postnatal developing Car8 wdl mice
Purkinje cell dendrite morphology recovers in Car8 wdl mice
Normal mitotic progression of developing granule cells
Sox2 is upregulated in the EGL of Car8 wdl cerebella
Reorganized zonal patterning after loss of Car8
The anatomical defects we found in Car8wdl mice transiently affect either Purkinje cell or granule cell development. Although these abnormalities disappear by the third postnatal week, it is possible that permanent defects manifest to affect behavior. For instance, several developmental processes—Purkinje cell dendritogenesis, granule cell proliferation, and granule cell migration—are interdependent. When the timing or execution of one process is altered, for example Purkinje cell dendrite outgrowth and branching, so is the timing and execution of subsequent events such as granule cell proliferation and migration, which are critical for establishing the mature cerebellar circuit [69, 70, 71].
Here, we started by examining the tissue at the level of individual stripes and found that loss of Car8 causes a dorsal-ventral misplacement of Purkinje cell somata within the molecular layer at P17 (Fig. 9c, yellow arrowheads). This offers another possible explanation for the somewhat paradoxical finding that although Purkinje cell dendrites are reduced in multiple parameters (Fig. 6), molecular layer thickness is increased (Fig. 5). The data suggest the possibility that abnormal stacking of Purkinje cells in the ML accounts, at least in part, for the pronounced molecular layer phenotype that we observed in the Car8wdl mutants.
In many cases, alterations to zonal patterning can predict disease because of underlying impairments to the cerebellar circuit . These alterations involve Purkinje cells, mossy fibers, climbing fibers, Golgi cells, unipolar brush cells, and Bergmann glia since all of these cell types and afferents use the patterning of molecular compartments to guide their own development [73, 74]. Of interest to us, was to investigate whether loss of Car8 modifies the zonal positioning of Bergmann glia given that loss of Car8 impacts Purkinje cell and granule cell development. In addition, loss of Car8 has been shown to delay zone formation and alter afferent positioning . Bergmann glia direct Purkinje cell dendritic outgrowth into the ML as well as granule cell migration into the inner granule cell layer (IGL) [75, 76]. Therefore, Bergmann glia are normally located next to Purkinje cells in the Purkinje cell layer and have long processes for scaffolding Purkinje cell dendrites and granule cell somata (Fig. 9b). Alterations to Bergmann glia morphology, or their positioning within the cerebellar cortex, can contribute to the pathogenesis of ataxia . We predicted that the delay in HSP25 zone formation  might cause an equivalent delay in Bergmann glia zonal patterning. Indeed, we found that Bergmann glia zones are defective in Car8wdl. When comparing the distribution of HSP25-labeled Purkinje cells and GFP-labeled Bergmann glia in control and Car8wdl posterior cerebella at P17, we found that there is more defined labeling in mutants, especially in lobule VIII (Fig. 9c-d). Normally, HSP25 is downregulated in lobule VIII, giving way to sharp zones only in lobules VI/VII and IX/X . Despite the specific errors in Purkinje cell patterning and dorsal-ventral positioning of the somata within the tri-lamina cerebellar cortex, loss of Car8 function does not compromise the overall compartmentalization of Purkinje cells and Bergmann glia into zones but does affect the establishment of normally robust and sharp expression of zonal marker boundaries (that is, a topographic map still forms in the Car8wdl mutants, albeit defective in its fine architecture).
Motor dysfunction develops despite rescue of cerebellar structure
Cerebellar diseases are devastating. They can cause the failure to execute a range of movements that are critical not only for fine motor skills, but also for everyday actions such as locomotion. Several of these diseases, for example ataxia, can initiate during development. Developmental cerebellar disorders involve improper circuit assembly with an ultimate consequence of disrupting function. However, in the Car8wdl mouse model, circuit anatomy and patterning are initially disorganized, but surprisingly by P20, the basic architecture of the cerebellum is normal [30, 80]. Here, we report that the early postnatal defects in Car8wdl mice include abnormalities in cerebellar size, Purkinje cell morphology, patterning, and granule cell proliferation. Most structural defects were transient, whereas Purkinje cell misfiring, muscle co-contractions, and motor dysfunctions persisted through adulthood.
Several morphogenetic events must occur simultaneously to enlarge the cerebellum during postnatal development. This includes Purkinje cell morphology becoming more elaborate, Bergmann glia directing Purkinje cell dendritic outgrowth into the molecular layer, and granule cells proliferating in the EGL [75, 76, 81, 82]. It is the proliferation of granule cells in the EGL that largely contributes to the rapid growth of the cerebellum [83, 84]. Since these processes are interdependent, when Purkinje cells degenerate, or Bergmann glia morphology is abnormal, granule cell proliferation and cerebellar size are affected because of compromised Purkinje cell-granule cell communication and scaffolding [82, 85, 86, 87]. Although perturbations affecting Purkinje cell, granule cell, or Bergmann glia development often have drastic effects on cerebellar size, studies have shown that both the ML and EGL are, at least in certain contexts, resilient. For instance, the size of the ML fluctuates during development due to synaptogenesis and behavioral changes [62, 88]. Through the trans-differentiation of astroglia to granule cells and the differentiation of Nestin-positive or Sox2-positive stem cells into granule cells, the EGL has the potential to recover through molecular mechanisms that reconstitute the population of cells [63, 64].
The rescue of the EGL that we found could be explained by the dynamics of GCP proliferation. Normally, GCP proliferation spans from ~E14 to ~P15 in mice . During this time, GCPs have characteristic divisions. The process starts with GCPs dividing symmetrically, and then as the GCPs near differentiation, their number of divisions peak . We show that decreased cerebellar size in Car8wdl mice is compensated for by an atypical presence of robust granule cell proliferation at P15 (Fig. 4b,g-h). This is in contrast to control mice because normally by P15 the EGL is almost completely depleted of GCPs . We also found a control-sized EGL in P5 and P10 mutants (Fig. 5b-f). These data suggest that granule cell proliferation is delayed in Car8wdl mice and that a large precursor pool is likely present at P5 and available to reconstitute the EGL. This process may contribute to the homeostatic rescue of cerebellar morphogenesis in the Car8wdl model. Analysis of granule cell development during mitosis and interphase indeed supports this hypothesis. We used PH3 expression, Sox2 staining, and EdU in vivo labeling to determine how loss of Car8 transiently delays granule cell proliferation. Our findings reveal that granule cell proliferation and cerebellar size may be recovered by upregulating Sox2 expression either in granule cell precursors or through recruiting Sox2-positive stem cells to the EGL. Sox2-positive stem cells have been shown to differentiate into cerebellar granule cells in vivo and in vitro [63, 64, 90]. A Sox2-positive population of cells occupies the PCL early in postnatal development [63, 64]. This pool of multipotent progenitors decreases with age and is greatly diminished by P8 . When the EGL is injured, this Sox2-postive pool is capable of repopulating granule cells by migrating from the PCL to the EGL and changing their glia and interneuron fates . This migration and subsequent change in fate extends cell proliferation in the EGL into later developmental stages . Therefore, one way in which Sox2-positive cells may be recruited to compensate for granule cell loss is through the modulation of signaling pathways associated with sonic hedgehog function. For example, in sonic hedgehog-dependent medulloblastoma, there are more Sox2-positive cells in the EGL, followed by more granule cells . When sonic hedgehog signaling is dampened, so is the amount of ectopic Sox2-positive cells and granule cells . These findings support previous data that aberrant signaling between Purkinje cells and granule cells alters morphogenesis. It is interesting to speculate that altered signaling in Car8wdl mice may trigger a compensatory response from Sox2-positive cells to rescue communication between Purkinje cells and granule cells.
In addition to revitalizing the EGL during postnatal morphogenesis, Purkinje cell architecture achieves normal dimensions, and the innervation of the Purkinje cells by climbing fibers reaches typical patterns , but what is perhaps most interesting about the Car8wdl model is that these defects are all repaired within the normal window of development. That is, cerebellar size and structure are rescued during the “critical period,” which in the cerebellum occurs until ~P17 . But, Purkinje cell firing seems to lack the potential for repair in this model. There are several possibilities for how structure and function are differentially impacted in Car8wdl mice. Most notably, loss of Car8 delays granule cell proliferation (Fig. 4), and we postulate that this could lead to a delay in granule cell migration and parallel fiber formation. This delay can ultimately manifest as reduced excitatory input and ultrastructural defects in Purkinje cells, as supported by microarray studies that show a correlation between a loss of Car8 and decreased expression of genes involved in synapse formation and maintenance [35, 93]. Previous manipulations of the EGL further support the hypothesis that transient morphological defects in developing Car8wdl mice contribute to the severe phenotypes that result. For example, transiently injuring the EGL from P1-P4 to temporarily decrease granule cell proliferation, transiently prolonging the cell cycle of GCPs at P3 to delay granule cell migration, and delaying granule cell differentiation cause permanent wiring defects as well as persistent motor and vestibular deficits [69, 71, 94]. The findings that loss of Car8 alters postnatal patterning of Purkinje cell stripes (Fig. 9) and that Car8 modulates Purkinje cell morphology , further suggest the possibility that defective cerebellar firing could also arise due to intrinsic molecular defects in the mutant Purkinje cells.
Here, we show that loss of a Purkinje cell gene causes a temporary increase in ML size at P20, which could involve several circuit deformities that impact the EGL. As granule cells migrate inwards after proliferation to create the inner granule layer (IGL), they leave behind their axons, which form synapses with Purkinje cell dendrites and interneurons in an ascending order. This “stacking” process contributes to the expansion of the ML [88, 96]. In Car8wdl mice, the temporary increase in ML size at P20 may be attributed to impaired synaptogenesis or perhaps to even increased ambulation. In mice, locomotion develops during the first two weeks after birth . After P15, locomotion becomes more complex, with features such as the step cycle and joint biomechanics maturing to support the need for increases in the speed of locomotion and improvements in coordination . In rodents, the relative amounts and type of activity as well as other behaviors such as exploration impact Purkinje cell dendritic development and synapse size [98, 99]. Therefore, increased movement from P15 to P20 in Car8wdl mice, and any functional abnormalities therein, may also contribute to fluctuations in ML size.
Despite the restoration of cerebellar morphology and the main features of the circuit diagram, the function of the cerebellum is less amendable to automatic repair. Analysis of Purkinje cell firing in awake behaving mice confirms our previous report in anesthetized mice showing that Car8wdl Purkinje cells fire with an irregular pattern , a measure of neuronal function that is currently thought to translate directly into behavioral deficits such as ataxia and dystonia because Purkinje cell excitation and spiking correlate with the stimulation of agonist/antagonist muscles during locomotor activity [6, 7, 25, 30, 59, 60, 61, 100, 101, 102, 103]. The neuronal mechanism for how the irregular firing of Purkinje cells influences behavior likely involves their direct targets, the cerebellar nuclei. When Purkinje cells fire erratically they induce similar irregularities in the firing of cerebellar nuclei output neurons [6, 7, 79, 101]. What is intriguing about the Car8wdl mice is that they exhibit ataxia (Fig. 10, Additional file 1: Movie S1 and Additional file 1: Movie S2) and tremor , which both presumably have origins in defective cerebellar output. Jiao et al. (2005) performed a qualitative neurological analysis and concluded that they also have appendicular dystonia . In addition, Jiao et al. (2005) found that the Car8wdl mice have abnormal paw placements . We performed a quantitative analysis of muscle function using EMG, footprinting, and CatWalk to confirm that Car8wdl mice have co-contractions of agonist and antagonist muscles in the hindlimb, abnormal stepping, and abnormal paw contacts with the floor (Figs. 10 and 11; Movie S1 and Additional file 1: Movie S2). These are all features consistent with dystonia-like behavior in rodents and using the combination of EMG, footprinting, and CatWalk allowed for a comprehensive view of movement dynamics [104, 105]. Interestingly, cerebellar function has previously been shown to impact how the hindlimb paws contact the ground. Robertson and McCollum (1991) examined the different receptive fields in the cerebellum innervated by climbing fibers . The receptive fields, and by extension the climbing fibers, that are activated affect hindlimb paw placement . Therefore, in Car8wdl mice and in other rodent models of dystonia where climbing fiber to Purkinje cell communication is impaired [2, 7], it might be expected that erratic Purkinje cell firing accompanies abnormal stepping. Altogether, our data suggest that it is possible that the altered pattern of Purkinje cell activity in Car8wdl reflects combined firing changes that culminate into motor system-wide defects that promote phenotypes such as ataxia, tremor, and dystonia.
Purkinje cells play a central role during cerebellar development and they are critical for motor and non-motor functions. It is therefore not surprising that Purkinje cell defects cause diverse behavioral deficits. However, in some cases, compensatory mechanisms during development can mask the cellular and circuit deficits that are typical of diseases such as ataxia. Here, we use the Car8wdl mouse model to demonstrate that even if postnatal cerebellar morphogenesis is rescued, neuronal misfiring, muscle dysfunction, and overt behavioral abnormalities can persist. Therefore, our data highlight the importance of timing during development. While the brain is equipped with many safeguards and compensatory mechanisms, short delays during critical periods may be enough to cripple subsequent circuit function and animal behavior.
We thank Colleen Brady for help with tissue staining and Samri Gebre for the initial breeding of our Car8wdl colony.
This work was supported by funds from Baylor College of Medicine (BCM) and Texas Children’s Hospital, BCM IDDRC Grant U54HD083092 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (The IDDRC Neuropathology and Neurobehavior Sub-Cores were used for anatomy and behavior), The Hamill Foundation, and by National Institutes of Neurological Disorders and Stroke (NINDS) R01NS089664 and R01NS100874 to RVS.
Availability of data and materials
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.
LNM, JJW, TL, AMB, and KJO designed and performed experiments. LNM and JJW analyzed data. LNM and RVS wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
All animal studies were carried out under an approved Institutional Animal Care and Use Committee (IACUC) animal protocol according to the institutional guidelines at Baylor College of Medicine (BCM).
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 12.Perkins EM, Clarkson YL, Sabatier N, Longhurst DM, Millward CP, Jack J, et al. Loss of beta-III spectrin leads to Purkinje cell dysfunction recapitulating the behavior and neuropathology of spinocerebellar ataxia type 5 in humans. J Neurosci. 2010;30:4857–67.PubMedPubMedCentralCrossRefGoogle Scholar
- 46.Paxinos G, Franklin K. The mouse brain in stereotaxic coordinates: Elsevier Academic Press; Amsterdam. 2001.Google Scholar
- 51.Eccles JC, Provini L, Strata P, Taborikova H. Topographical investigations on the climbing fiber inputs from forelimb and hindlimb afferents to the cerebellar anterior lobe. Exp Brain Res. 1968;6(3):195–215.Google Scholar
- 62.Altman J, Bayer S. Development of the cerebellar system in relation to its evolution, structure and function. Boca Rat: CRC Press; 1997.Google Scholar
- 89.Espinosa JS, Luo L. Timing neurogenesis and differentiation: insights from quantitative clonal analyses of cerebellar granule cells. J Neurosci. 2008;28:22301–12.Google Scholar
- 105.Tuschi K, Clayton PT, Gospe SM, Mills PB. Dystonia/parkinsonism, Hypermanganesemia, polycythemia, and chronic liver disease. Seattle: University of Washington; 2017.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.