Judith Walters - US grants

In Parkinsons disease (PD) patients off medication, atypical oscillatory activity has been found in the beta frequency range. It has been hypothesized that this excessively synchronized activity in the basal ganglia is responsible for the motor impairments seen in these patients and that deep brain stimulation (DBS) of basal ganglia targets is beneficial by disrupting this activity. However, it is also possible that this activity could contribute to non-motor symptoms, such as depression, anxiety and an increased sensitivity to pain which have also been reported in PD. It is hoped that a better understanding of this hypersynchronized state can provide insight into the neuronal circuits undergoing plastic changes after dopamine cell lesion and the relevance of these changes to the symptoms observed in PD. In previous years we have performed simultaneous recordings from the motor cortex and the substantia nigra pars reticulata (SNpr), a basal ganglia output nucleus, in a rat model of PD to explore the potential for this model to provide insight into how this activity emerges in the basal ganglia and whether it is functionally related to the motor symptoms of akinesia and dyskinesia associated with Parkinsons disease. After unilateral dopamine cell lesion, rats show notable motor deficits during treadmill walking in a circular treadmill. Our recording studies have shown that significant increases in local field potential (LFP) spectral power and in SNpr-mCx coherence in the high beta/low gamma 25-40 Hz frequency range emerge in the dopamine-lesioned hemisphere of these rats by day 7 after dopamine lesion which parallel in many ways the increases in oscillatory activity noted in PD patients, indicating that this is a good model for investigating the functional significance of these changes in brain activity. This year we have continued to use this model to better understand the changes in synchronized activity emerging in basal ganglia output after loss of dopamine. In a recently published study, we have used hemiparkinsonian rats performing a treadmill walking task to compare synchronized oscillatory activities in the STN with the motor cortex (MCx) and the medial prefrontal cortex (mPFC), areas involved in motor and cognitive processes, respectively. Data show increases in STN and MCx 29-36 Hz LFP spectral power and coherence after dopamine depletion. These increases are reduced by apomorphine and levodopa treatments. In contrast, recordings from mPFC three weeks after dopamine depletion failed to show peaks in 29-36 Hz LFP power. However, mPFC and STN showed similar and significant peaks in the 45-55 Hz frequency range in LFP power and coherence during the walking task before and 21 days after dopamine depletion. Interestingly, power in this low gamma range was transitory reduced in both mPFC and STN after dopamine depletion but recovered by day 21. In contrast to activity in the 45-55 Hz range, the amplitude of the exaggerated 29-36 Hz rhythm in the STN is modulated by paw movement. Furthermore, as in PD patients, after dopamine treatment, a third band (80-120 Hz) emerged in the dopamine cell lesioned hemisphere. The results suggest that STN integrates activity from both motor and cognitive networks, in a manner that varies with frequency, behavioral state and the integrity of the dopamine system. STN LFP activity can become synchronized with, and presumably modulated by, activity in both limbic and motor cortex networks in a manner that varies with frequency range, behavioral state and the integrity of the dopamine system. We are now following up on 2 relevant observations. First, the low gamma activity observed in the mPFC in normal rats during treadmill walking is dramatically reduced within a day after loss of dopamine. We hypothesize this reflects some direct modulation of local mPFC circuts by dopamine. Interestingly this activity returns to normal levels about 3 weeks after the lesion, showing that the mechanisms governing the cognitively related low gamma activity are relatively plastic. Second, we have also been recording in the anterior cingulate cortex (ACC) together with the STN in the hemiparkinsonian rats. The ACC is another prefrontal area that is well described in the rat and could be relevant to non-motor symptoms in Parkinsons disease. In contrast to the lack of evidence for the high beta activity in the mPFC after dopamine cell lesion, preliminary results show a dramatic increase in high beta activity in the ACC during treadmill walking. The ACC, we have further found is thought to receive input from the ventral medial thalamus, an area receiving from the basal ganglia output, and thus seems part of a larger basal ganglia thalamocortical loop. In a second set of studies, we have completed collecting data for our investigation into changes in spiking and LFP in two areas in the striatum in conjunction with recordings in motor cortex and SNpr in the hemilesioned Parkinsonian rat. Although exaggerated oscillatory activity has been observed in the majority of basal ganglia nuclei in PD, it is still unclear how it emerges and whether it engages the major basal ganglia input nucleus most directly affected by dopamine loss, the striatum. Our results show increases in striatal oscillatory LFP activity after dopamine depletion during treadmill walking in the high beta/low gamma frequency range, and during L-dopa-induced dyskinesia in the high gamma range. However, the oscillatory activity in the striatum was patchy, and striatal projection neurons showed very little phase coupling to the increased oscillatory activity in the 28-36 Hz range in motor cortex and striatum, suggesting a minimum role of the striatum in transmitting this hyper-synchronized activity downstream to the basal ganglia output, the SNpr. On the other hand, we also find evidence that information transfer from the striatum to the globus pallidus pars externa (GPe) may be disrupted. We are in the process of using a stochastic entropy model to study how dopamine depletion alters coding capacity and information flow in the basal ganglia of hemiparkinsonian rats. Analysis of changes in entropy in the spiking of the striatal output is suggestive of changes which may reflect efforts to compensate for loss of dopamine. We are working on a manuscript reporting this data. The absence of oscillatory activity in striatal output calls attention to the role of the STN and GPe circuitry in generating or propagating the excessive activity in the high beta range in the hemiparkinsonian rats. Together with increased phase coupling of cortical, STN and SNpr neurons to the motor cortex LFP activity in the beta frequency range observed after loss of dopamine , these data suggests that this activity propagates via the hyper-direct pathway, that is, from motor cortex to STN, and then to SNpr . A third set of studies have been initiated to explore the role of the STN-GPe circuitry in the initiation and/or propagation of the exaggerated oscillatory acticity emerging in the basal ganglia after loss of dopamine. In these studies we explore the idea that there may be two subsets of neurons in the GPe which become engaged in an antiphase relationship promoting oscillatory activity in conjunction with the STN nucleus. We have begun to use virally-transmitted DREADS to manipulate the activity of these two subsets of neurons to explore the role of these cells in the generation of the high beta rhythms.

In the past year, we have continued to make progress in assessing changes in thalamocortical relationships in the rodent model of Parkinsons disease (PD). Both bradykinetic and dyskinetic states have been shown to be associated with dramatic increases in oscillatory and synchronized activity in the motor cortex. Current results support the idea that these changes in motor cortex LFP activity are driven by alterations in the activity of thalamic input to cortex. However, the role of the high beta 30-35Hz oscillatory activity observed in the VM thalamocortical projection and in the motor cortex during periods of bradykinesia in the parkinsonian animals is still under debate. The time frame of the emergence of the bradykinetic behaviors is not well correlated with the emergence of the high beta activity in the motor cortex. Furthermore, our data does not support the view that the increases in narrow range FTG LFP oscillations in the motor cortex are causally involved in the expression of L-dopa induced dyskinesia, as some have suggested. The emergence of dyskinesia in the rodent model of Parkinsons disease can be disassociated from increases in high gamma range LFP activity in the motor thalamus and motor cortex. We have just submitted a manuscript reporting data from recordings of spike/LFP relationships between basal ganglia output, substantia nigra pars reticulata (SNpr), motor thalamus and motor cortex in hemiparkinsonian rats trained to walk on a circular treadmill. While we are unsure of the actual consequences of the high beta/low LFP oscillations in the motor cortex, the thalamic component of the basal ganglia- thalamocortical loop does seem critical to the emergence of these LFP oscillations after loss of dopamine. Recordings of LFP activity from multiple sites within the motor network show correlated increases in coherence between motor cortex and SNpr, between motor cortex and ventral medial thalamus, and between SNpr and ventral medial thalamus in the 30-35 Hz range after dopamine cell lesion during treadmill walking. Infusion of either the GABAa agonist muscimol or the GABAa antagonist picrotoxin into the ventral medial nucleus to inhibit activity in this nucleus, or block GABAergic input, respectively, induces a reduction of power in both motor cortex and SNpr LFP and reduces coherence between these two sites in the high beta/low gamma range during treadmill walking. This data supports a role for the ventral medial thalamus in induction of high beta/low gamma synchronization of LFP activity in the motor cortex. The data also shows that neuronal activity in the ventral medial thalamus promotes increased coherence within the larger basal ganglia thalamocortical network after loss of dopamine. In a follow up series of studies, we are utilizing electrodes attached to cannula to record locally as we infuse test drugs into the VM nucleus and hope to apply optogenetic techniques to modify spiking activity to more selectively probe the different nodes of this circuit, and the manner in which thalamocortical activity ultimately impacts downstream systems regulating limb movement. We also have a manuscript under review examining the changes in thalamic and thalamocortical activity associated with chronic treatment of L-dopa. The therapeutic effect of treatment of Parkinsons disease patients with the dopamine precursor L-dopa has been well established. However, over time, L-dopa therapy leads to severe motor complications referred as L-dopa-induced dyskinesias (LID). Recently, we have confirmed that there is a strong association between the presence of 80-100 Hz high gamma oscillations in the motor cortex of hemiparkinsonian rats and LID expression. This is especially interesting because high gamma has been observed in human PD patients in recordings through deep brain stimulation electrodes, and the role of this activity in generating dyskinesia is unclear. This activity has been referred to in the clinical literature as finely tuned gamma or FTG . While these results supported relationship between the high gamma oscillatory activity and dyskinesia, other results did not. Unexpectedly, as cortical high gamma power increased, phase locking of cortical pyramidal spiking to high gamma oscillations decreased in the motor cortex. This observation suggests that power in this high gamma range can change dramatically in the motor cortex LFP without an associated changes in spike-LFP phase locking. This raises questions regarding the functional correlation between high gamma and dyskinesia. Indeed, follow up studies are providing further evidence that expression of narrow band finely tuned gamma (FTG) and the dyskinetic behavior can be disassociated. In these studies we first established a role for the motor thalamus in generating the FTG band in the motor cortex, and then explored the consequences of reducing the cortical expression FTG by manipulations of VM thalamic activity. This was further correlated with assessment of the expression of dyskinesia. Evidence that the motor thalamus is driving the FTG in the cortex emerged from these studies of rate and spike-LFP phase-locking in the motor thalamus during LID. Interestingly, changes in firing rate and phase-locking are highly correlated with the changes in power in the high gamma range in the ventromedial thalamus, as opposed to the observations, discussed above, with respect to phase locking in the motor cortex. Firing rate changes in the cortex are not obvious either, in our analysis to date. Gamma activity is generally considered an important modulator of cortical function. However, while systemic administration of a 0.3 mg/kg dose of MK-801, the NMDA GluR antagonist, eliminated both FTG and dyskinesia in the L-dopa-primed dyskinetic rats, a lower 0.15 mg/kg dose of the MK-801 administered i.p. abolished the FTG oscillations in the cortex without affecting dyskinesia. Further studies infusing muscimol into the VM show that suppression of ventromedial thalamic activity by local injection of this GABA receptor agonist completely eliminated aberrant 100 Hz synchronization within the motor cortex, but had nearly no effect on LID. The results suggest that while robust high gamma oscillatory activity in both motor thalamus and motor cortex is evident during LID, this aberrant thalamocortical synchronization does not appear to be requisite for the expression of dyskinesia. Most recently, we are also exploring the role of the cerebellum in contributing to the dramatic changes in thalamocortical activity observed in this model of L-dopa induced dyskinesia. Further studies are underway to investigate the role of the parafascicular thalamus in expressing and propagating the high beta and gamma activity evident in some parts of the basal ganglia thalamocortical circuit after loss of dopamine and treatment with L-dopa. Studies are underway to explore the role of ketamine in generation of high gamma activity in the medial dorsal thalamic nucleus and the prefrontal cortex in collaboration with Dr. Andreas Buonoanno and his graduate student, Katrina Furth in NICHD. Finally, we are initiating some studies in collaboration with some of our colleagues in Bldg 35 to explore the role of the VM thalamus in the changes in pain threshold reported in Parkinsons disease patients. There are several reports in the literature of nociceptive input to the VM thalamus, and other studies showing projections from the VM thalamus to the anterior cingulate cortex. As we have recently observed an increase in high beta activity in the anterior cingulate cortex after loss of dopamine, we are designing studies to determine the effect of mild pain on neuronal on neuronal activity in the VM thalamus and anterior cingulate cortex in the hemiparkinsonian rat.

Our studies have addressed three research areas. 1. Pain mechanisms in PD. Although primarily known as a movement-related disorder, Parkinsons disease (PD) has several non-motor symptoms, such as pain, that have gained increasing attention. High prevalence of pain and increased pain sensitivity have been observed in human PD patients and animal models of PD. Studies have shown the ventromedial thalamic nucleus (VM) to relay nociceptive information from the medullary subnucleus reticular dorsalis to both the BG and a cortical region known to be involved in pain processing, the anterior cingulate cortex (ACC). As recent work from our lab has shown that the ACC and VM, components of the BG-thalamocortical circuit, exhibit pathological beta activity in a parkinsonian rat model, we have hypothesized that the excessive beta oscillations will disrupt pain processing. We have used 6-hydroxydopamine to lesion dopamine neurons in one hemisphere, to induce a unilateral rat model of PD. Electrode bundles are chronically implanted into the subthalamic nucleus (STN), ACC, and VM of the dopamine cell lesioned left hemisphere of rats. To induce a pain response, 1.5% concentration of the toxin formalin was subcutaneously injected into the planar surface of the right hind paw. The formalin test produces a biphasic response consisting of a 3-5 minute interval of pain behavior immediately after injection, a 10-15 minute quiescent interphase during which the rats will not display nociceptive behavior, and finally a 20-40 minute period of inflammatory pain. Behavioral pain response, spiking and LFP activity were recorded for the two hours immediately after formalin injection. Our studies, in collaboration with Dr. Yarimar Carrasquillo, show that the lesioned rats injected with 1.5% formalin have greater alpha (12-19 Hz) LFP power in the VM than sham lesioned rats. Further analysis is ongoing. Additional study into pain mechanisms may lead to the development of better treatment options for PD patients experiencing pain. 2. Cognitive function in PD. Investigation of changes in the ACC cognitive function have provided the basis of a PhD thesis recently successfully submitted by Alex Weiss (NIH Cambridge/Oxford PhD program). These studies are currently being further analyzed and written up for publication and presentation in a poster at the Society for Neuroscience in Nov. PD patients are known to express dopamine-dependent cognitive impairments, implying effects of dopamine loss on PFC function. The electrophysiological correlates of these cognitive symptoms are not well understood. His recent study under the Oxford arm of this project compared electrophysiological data from the BG of PD and non-PD patients during a cognitive task and showed the presence of electrophysiological abnormalities associated with cognitive function in the BG that may be used as biomarkers to help understand the basis of PD cognitive impairment. In light of the ACCs development of parkinsonian exaggerated beta activity, we have investigated the involvement of BG thalamocortical circuits in dopamine-impaired and healthy rats in response to a salient cue that predicts the onset of either tone-to-treadmill-induced walking (an expected event) or tone to-no-treadmill-induced walking (an unexpected event). Rats were trained to expect epochs of treadmill walking after a tone, with subsequent epochs manipulating expectancy through tone-to-walk and tone-to-no-walk epochs. Changes in LFP power, spiking, and coherence data from several areas of the BG thalamocortical circuit involved in decision-making were recorded during epochs surrounding auditory stimuli and treadmill walking in control animals and after dopamine cell lesion. Thus far, results provided a series of observations supporting the view that the BG thalamocortical circuit develops both exaggerated oscillatory activity and spike-to-LFP phase locking during cognition in the hemiparkinsonian rat model of advanced stage PD. Theta frequency in the ACC has been linked with error prediction. Our results show significant increase in theta power the ACC during the first rule reversal in the control rats but not the lesioned rats on day 7 post-lesion. We also observed significant increase of coherence in the theta range between the ACC and the VM from day 7 to day 21 post-lesion during the first rule reversal, as well as significant decrease in the coherence between ACC and the STN. Dopamine lesioned rats exhibited higher ACC alpha power than non-lesioned rats during unexpected outcomes, which may suggest poor memory function or attention in the lesioned animals. In the beta range, LFP power was significantly increased in all areas during the walk epochs and first rule reversal epoch with the exception of the STN during the rule reversal. These results will be further analyzed, and could provide clues to biomarkers for alteration of executive control in hemiparkinsonian rats. As cognitive symptoms in PD patients become an increasingly recognized issue, preclinical animal models for the study of PD cognitive impairment have become increasingly necessary. Despite the limitations of the model, hemiparkinsonian rats display some of the most important cognitive symptoms seen in PD patients, including impairment of executive function, and our results so far have been able to reproduce some of the electrophysiological dysfunctions in the same anatomical sites as observed in human subjects. 3. Role of Parafascicular thalamic nucleus (PF) in PD. The PF nucleus receives inputs from the BG, cortex, and cerebellum and provides feedback to the subthalamic nucleus (STN) and striatum (STR) and thus could be implicated in the pathophysiology of PD. Unlike the VM, however, we have found that exaggerated beta oscillations are not evident in recordings from the PF in the behaving, hemiparkinsonian rat. Instead, our latest results call attention to the potential role of PF output in tonic modulation of STR and STN activity and support the idea that reductions in PF activity may have a therapeutic effect on motor dysfunction in PD. In the present study, one group of hemiparkinsonian rats received infusion into the PF of the inhibitory DREADD virus: AAV2-hSyn-hM4D(Gi)-mCherry (hM4D) and a second group received infusion the CRE-dependent virus (pAAV-hSyn-dF-HA-KORD-IRES-mCitrine) into PF in conjunction with infusion of the retrograde Cre-recombinase virus (AAV pmSyn1-EBFP-Cre) into dorsolateral striatum, allowing for retrograde expression of Cre in the PF neurons projecting to the striatum and subsequent expression of the KORD Cre-dependent virus in the PF nucleus. Histological studies are underway to determine if the KORD virus is selectively expressed in the striatum, as projections from PF neurons could also impact STN activity directly. Recording electrodes were implanted targeting motor cortex (MCx), STR, and SNpr. Histological analyses revealed modest DREADD virus expression in the PF with some spread to surrounding areas. Similar to previous studies, exaggerated high beta power in the MCx and SNpr and coherence between these regions were observed during treadmill walking. At 3-4 weeks post-surgery, clozapine-N-oxide (1-5 mg/kg, ip.) and salvinorin B (1-2 mg/kg, sc.), both of which activate the receptors expressed by the corresponding DREADD virus, substantially improved circular treadmill walking in the clockwise direction without modifying cortical or nigral high beta oscillations. These results suggest that inactivation of PF terminals on PF output neurons reduces motor deficit in a rodent model of PD, similar to the effect of the inhibitory agonist muscimol. Ongoing DREADDs experiments are further exploring the contribution of the PF to high beta oscillations in BG-cortical motor circuits as well as spiking activity in PF, STR, SNpr and MCx.