Targeting DBS to the centrolateral thalamic nucleus improves movement in a lesion-based model of acquired cerebellar dystonia in mice

Abstract

Dystonia is the third most common movement disorder and an incapacitating co-morbidity in a variety of neurologic conditions. Dystonia can be caused by genetic, degenerative, idiopathic, and acquired etiologies, which are hypothesized to converge on a "dystonia network" consisting of the basal ganglia, thalamus, cerebellum, and cerebral cortex. In acquired dystonia, focal lesions to subcortical areas in the network - the basal ganglia, thalamus, and cerebellum - lead to a dystonia that can be difficult to manage with canonical treatments, including deep brain stimulation (DBS). While studies in animal models have begun to parse the contribution of individual nodes in the dystonia network, how acquired injury to the cerebellar outflow tracts instigates dystonia; and how network modulation interacts with symptom latency remain as unexplored questions. Here, we present an electrolytic lesioning paradigm that bilaterally targets the cerebellar outflow tracts. We found that lesioning these tracts, at the junction of the superior cerebellar peduncles and the medial and intermediate cerebellar nuclei, resulted in acute, severe dystonia. We observed that dystonia is reduced with one hour of DBS of the centrolateral thalamic nucleus, a first order node in the network downstream of the cerebellar nuclei. In contrast, one hour of stimulation at a second order node in the short latency, disynaptic projection from the cerebellar nuclei, the striatum, did not modulate the dystonia in the short-term. Our study introduces a robust paradigm for inducing acute, severe dystonia, and demonstrates that targeted modulation based on network principles powerfully rescues motor behavior. These data inspire the identification of therapeutic targets for difficult to manage acquired dystonia.

Keywords: Cerebellar Peduncle; Cerebellum; Deep Brain Stimulation; Dystonia; Thalamus.

Figure 1.

The cerebellar network and electrolytic lesion paradigm. A simplified diagram showing example primary afferent projections to the cerebellar cortex, such as the pontine nuclei (PN), which transmit descending cerebral cortical information to the cerebellar cortex via mossy fibers (MF) (A). Inferior olivary neurons (IO) project exclusively to the cerebellar cortex via climbing fibers (CF) (A). Purkinje cells (PC) then assimilate all cerebellar cortical computation and form the sole output neurotransmission of the cerebellar cortex, projecting mainly to the cerebellar nuclei (CN) (A). The CN project via the superior cerebellar peduncle (Red) to the thalamus (Thl) (A), among other regions. For electrolytic lesioning, tungsten electrodes are cut at 3.5mm, and then tested using an oscilloscope to verify capability to deliver a 50 Hz 1600 mA charge (B). The modified electrodes are then stereotactically inserted to a target location in the brain and the 1600 mA, 50 HZ charge is delivered for 60 seconds (C). Nissl stain at site of lesion reveals cystic changes at target site (D, lower panel). Immunohistochemistry with GFAP (red) and neurofilament heavy chain (NFH; white) reveal gliosis and neuron degeneration at the lesion site, respectively (D, upper panel; scale bar = 250 um). Lesion volume in the coronal (E) and axial (F) plane are shown (lobules are as noted, PFL=paraflocculus). Serial sections are shown through the lesion with distance from bregma indicated in the inset. Red circles represent the same lesion across sections. Lesion volume is roughly a sphere with diameter of 0.4 mm. All sections used in the imaging were obtained from animals 2 weeks after lesioning (scale: gross image = 1 mm; insets = 500 um).

Figure 2.

Electrolytic lesioning of the cerebellar outflow tract. Lesions were targeted to the superior cerebellar peduncle (SCP) as it emerges from fastigial (FN) and interposed (IN) nuclei (A, blue represents a section at the approximate coronal plane at which the lesion was made). This site was chosen based on the work of Albazron et al[3]. During lesioning, three classes of movements were noted (B). With sham stimulation, no movements were noted; with lesions that were smaller than expected there was mild forepaw movement and tremor of the upper body; with larger lesions hind paw movement was noted, more prominent movement of the forepaws, and tremor with contraction of trunk was seen (B). Lesion location (C, right) was centered on the site of convergence of the SCP, FN, and IN (C, left; lobules are as noted, PFL=paraflocculus; scale bar = 1mm). Sham lesions reveal no change to the brain parenchyma (D, upper); smaller lesions reveal gliosis (increased cell density indicated within red dotted lines on D middle, and + on inset image) without cystic changes (D, middle); larger lesions reveal more extensive gliosis (increased cell density indicated within red dotted lines on D bottom left, and + on inset image) with cystic changes (asterisks on inset D, lower). Scale bar on coronal sections = 2mm, inset = 500um. All sections were obtained 2 weeks after lesioning.

Figure 3.

Cerebellar outflow tract lesions lead to acute, severe dystonia. Lesioning of the cerebellar outflow tracts induces an acute severe dystonia apparent as the animal recovers from anesthesia. At 30 mins, the animals display a fixed posture with version of the neck, stiff limbs, and splayed digits (A, 30 min); at 1 hour, there is sometimes improvement, and some steps taken (A, 1 hr); at 2 hours the mice are more often able to perform limited ambulation (A, 2 hrs); by 24 hrs, there is still frequent dystonic posturing but limited ambulation is apparent (A, 24hrs). Asterisks indicate limb/digit dystonia, lines indicate cervical/trunkal dystonia, pluses indicate tail dystonia (A). The dystonia rating scale is used to numerically define these phenotypic observations (B) [63]. Compared to sham lesions (n=5), significantly higher scores on the scale were noted in lesion mice (n=8; two-way mixed method ANOVA p<0.0001) (C).

Figure 4.

Lesioning of the cerebellar outflow tract does not alter Purkinje cell firing activity. To diminsh the potential effect of anethesia and mechanical damage from craniotomy, Purkinje cell (PC) recordings were performed at 48 hrs, at which time lesion mice continued to have significantly higher dystonia rating scale scores compared to sham lesion animals (unpaired t-test p<0.0001; A). Example traces from single cell recordings of Purkinje cells in a sham lesion mouse (B, upper) and lesion animal (B, lower). Asterisks indicate complex spikes, which were used to confirm that the recording was from a Purkinje cell, scale = 200 ms (B). Purkinje cell simple spike firing properties from sham (N=2, n=5) and lesion (N=3, n=6) revealed no differences in firing rate, CV, or CV2 (C, upper). Purkinje cell complex spike firing properties from sham (N=3, n=11) and lesion (N=3, n=12) mice also revealed no differences in firing rate, CV, or CV2 (C, lower).

Figure 5.

Targeting nodes in the dystonia network with Deep Brain Stimulation (DBS). The DBS implant consists of 2 twisted bipolar tungsten electrodes bonded together at the appropriate distance to target the desired brain region bilaterally (A). DBS stimulation parameters included biphasic square pulses of 30 mA at 130 Hz (B). After DBS stimulation, mice were sacrificed at 2 weeks post-lesion and Nissl stain was used to verify appropriate targeting. Coordinates were obtained from Paxinos and Franklin[62]. Centrolateral thalamic targeting was 2.06 mm posterior from bregma, +/− 1.55 mm along the medio-lateral plane, and 2.55 mm ventral from the brain surface (C). Dorsal striatum was targeted at bregma, +/− 1.75 mm along the mediolateral plane, and 2.00 mm ventral from the brain surface (D). A simplified scheme of the “dystonia network” with lesion location (pink circle) and DBS targets (yellow octagon) indicated. The cerebellar nodes (cerebellar nuclei (CN), pontine nuclei (PN), and inferior olive (IO) are indicated in orange hues, the thalamo-cortical loops (thalamus (Thl), cortex (Ctx)) are indicated in blue, and the basal ganglia (BG) nodes are indicated in green (E). The basal ganglia are divided into the striatum (upper left loop) and deep gray nuclei (lower right loop, consisting of globus pallidus interna, globus pallidus externa, subthalamic nucleus). The short latency disynaptic projection linking the CN to BG is indicated in mauve.

Figure 6.

Outcomes of Deep Brain Stimualtion (DBS). Prior to centrothalamic DBS, at 24 hrs post-lesion, mice have ongoing severe dystonia with stiff, splayed limbs (arrows) and trunk contraction (A, left). After 24 hrs of centrothalamic DBS, there is improvement in gait and improvement of tone in the trunk and extremities (A, right). Using the dystonia rating scale (DRS), we note no change in movements pre to post DBS with stimulation of centrothalamic nuclei in sham animals (B; left) or in lesion animals with dorsal striatal DBS (B, right). However, all lesion animals with centrothalamic DBS demonstrated improvement on the DRS, leading to significant group effect (ratio paired t-test p<0.0001). To determine whether muscle contraction was affected by DBS in the three conditions, we used nuchal EMG, with 2 silver wire electrodes inserted into each the left and right trapezius (C). EMG yielded raw burst traces as indicated in (D). Example 15 second traces from an individual sham lesion mouse with centrolateral DBS pre and post stimulation (beige), an individual lesion mouse with centrolateral thalamic DBS pre and post stimulation (blue), and an individual lesion mouse with dorsal striatal DBS pre and post stimulation (green) are shown (D). To evaluate how muscle contraction was affected in individual animals, we calculated total burst duration (a proxy for muscle contraction) over the 5 mintues prior to stimulation and the five minutes just after DBS was switched off in individual animals. We did not find a significant difference using a paired t-test in sham lesion with centrolateral thalamic DBS (n=5, paired t-test p=0.1735; E, left) or lesion animals with dorsal striatal DBS (n=7, paired t-test p=0.2842; E, right). With centrolateral thalamic stimulation in lesion animals, we noted a significant decrease in burst duration from the 5 mins pre DBS to post DBS (n=8, paired t-test p=0.0013; E, center). To compare between conditions, percent change in burst duration after 1 hr of stimulation was examined, revealing a significant percent decrease in the lesion mice with centrolateral thalamic DBS compared to the other two conditions (Brown-Forsythe p=0.0309 and Welch’s p=0.0016 ANOVA tests; F).