Pathophysiology of gait disorders induced by bilateral globus pallidus interna stimulation in dystonia

Sir, We read with great interest the article by Mahlknecht et al. (2018), who analyse the mechanisms of parkinsonian features induced by bilateral pallidal stimulation in dystonia, including hypokinetic gait disorders with freezing of gait. These gait disorders, which were not present preoperatively, may be relieved by the changes in the stimulation parameters (Tisch et al., 2007; Zauber et al., 2009; Schrader et al., 2011; Wolf et al., 2016). By combining detailed analysis of pallidal stimulation (dystonia improvement, bradykinesia, axial signs) with anatomical mapping of volume of tissue activated by the therapeutic contact in the internal pallidum (GPi), Mahlknecht et al. (2018) pointed out pathophysiological mechanisms related to the current spread to adjacent structures, such as pyramidal tracts or external pallidum (GPe). As bradykinesia and axial motor symptoms were observed concurrent to dystonia improvement, as stimulation overlapped the ‘sweet spot’ for dystonia reduction, they suggested that these parkinsonian symptoms may be relates to deleterious modification of activity of different functional zones within the GPi and to altered outflow of pallido-thalamo-cortical pathways. Besides these observations, structural or functional alteration of the pedunculopontine nucleus (PPN) and its connections may be another mechanism of postoperative gait alteration in dystonia. Projections from the GPi to the PPN (Benarroch et al., 2013) could be functionally altered by pallidal stimulation. As PPN cholinergic deficiency has been recently described in cervical dystonia (Mente et al., 2018), mild structural abnormalities may disrupt PPN subtle tuning of gait (Caggiano et al., 2018) and contribute to parkinsonian gait features in dystonia. In Parkinson’s disease, we observed a relation between the PPN and frontal eye field functional connectivity, altered eye movements (antisaccade latency) and impairment of postural adjustments before gait initiation (Ewenczyk et al., 2017). Starting from the hypothesis that saccade and antisaccade latencies, a simple and robust parameter, may be an indirect marker of the mechanisms underlying stimulation induced gait impairments in dystonia, we designed a study on eye movements in dystonic patients who developed various subtypes of gait disorders besides beneficial effect GPi stimulation (see Supplementary material for methods). To unravel the potential implication of PPN, we focused on antisaccade latency in the altered gait (i.e. ‘bad’ gait) and in the best gait condition (after optimization of stimulation), in relation with the stimulation parameters and the location of the stimulation contact. As dopaminergic denervation may also alter eye movement metrics (Hood et al., 2007), we explored the integrity of the dopaminergic pathways with DaTSCAN. Our study identified three patients with three different patterns of gait disorders out of the 22 patients with isolated generalized dystonia treated with bilateral deep brain stimulation (DBS) in the postero-lateral part of the GPi, from our seminal controlled study (Vidailhet et al., 2005), each patient corresponding to different types of ocular performances during prosaccade and antisaccade tasks (see Supplementary material for detailed results). Patient 1 presented with irregular gait (with asymmetry of leg movements) and freezing of gait (FoG). The eye movement study showed unilateral increase of prosaccade latencies. This points to the disruption of the parieto-collicular pathway, which is crucially involved in the generation of reflexive prosaccades and a border diffusion of the stimulation towards the posterior limb of internal capsule, integral part of parieto-collicular pathway (Gaymard et al., 2003). The postoperative CT scan (superimposed on the preoperative MRI) showed a borderline position of the therapeutic contact close to posterior limb of internal capsule for both electrodes, predominantly more on the left electrode (Fig. 1A). The gait irregularities with asymmetry of limb involvement may have contributed to FoG generation (Plotnik et al., 2008). Increase of stimulation parameters partially restored symmetry of leg movements with preserved beneficial effect on trunk and limb dystonic symptoms. Potential underlying mechanisms of gait disorders in dystonia postoperative axial MRI. DBS-induced: spreading of current to adjacent structures. Postoperative axial MRI slices with localization of intial therapeutic contact with development of gait disorders on follow-up. (A) Contact is located on the border of GPi and posterior limb of internal capsule (IC) (Patient 1). (B) Contact close to external part of GPe (Patient 3). Patient 2, in the ‘bad’ gait stimulation condition, developed severe FoG when initiating gait, with complete inability to make a step and with recurrent falls. We observed unilateral increase of antisaccade latency to the right side, which could be a surrogate marker of altered functional connectivity between frontal eye field and PPN that correlates with antisaccade latency (Ewenczyk et al., 2017), suggesting a possible alteration of the PPN network. In addition, dopaminergic denervation (postoperative DaTSCAN, Fig. 2A) may also contribute to the parkinsonian-like symptoms. A 6-month treatment with l-DOPA 750 mg/day was not beneficial. In contrast, stimulation change (unilateral pulse width increase), with broader GPi volume activation, restored the ability to walk with rare episodes of FoG, and the effect on dystonic symptoms was stable. Potential underlying mechanisms of gait disorders in dystonia. Postoperative DaTSCAN: dopaminergic denervation. Patients 2 (A), and 3 (B) shows signs of presynaptic dopaminergic deficit. (C) Normal DaTSCAN (Patient 1). Patient 3 also presented an increase of antisaccade latency in the ‘bad’ gait stimulation condition, further supporting the hypothesis of functional alteration of PNN connectivity in dystonia. Like Patient 2, he had dopaminergic denervation (DaTSCAN, Fig. 2B), suggesting a ‘double hit’ (cholinergic and dopaminergic alteration) mechanism for gait disorders, although we cannot rule out underlying asymptomatic parkinsonism; however, l-DOPA treatment had no effect. In contrast, adaptation of stimulation parameters markedly improved the slow cautious gait suggesting a beneficial effect of GPi stimulation on akinesia. Further, there were abnormally fast responses in the antisaccade task, with higher error rate, both results suggesting a spread of current to GPe. The GPe contains two types of neurons (Yoshida et al., 2016), the increase-type neurons that might play a role in facilitating antisaccades, and the decrease-type neurons, which could mediate signals for reflexive saccade suppression. In our case the antisaccades were faster with potential overstimulation of decrease-type neurons, which mediated more impulsive release of antisaccades. The therapeutic contact position was close to the GPe/GPi border (Fig. 1B) and current spread to the adjacent GPe, in the ‘bad’ gait condition, could account for the GPe dysfunction and related abnormalities in antisaccade tasks, both being reversible with changes in the stimulation parameters. Beyond the conclusions from Mahlknecht et al. (2018), which favour stimulation-induced local functional alterations of neural elements within the pallidum as the predominant mechanisms of stimulation-induced parkinsonism, we offer another perspective on how impaired neural integration in the various networks implicated in dystonia affects motor processing (dystonic symptoms and gait) and saccade initiation time. Overall, studying eye movements using simple and robust saccade parameters combined with clinical observation and neuroimaging, may reveal subtle perturbations of the pedunculopontine network, as part of intrinsic PPN abnormalities described in dystonia, or deleterious activity of the GPi/PPN connections, with disruptive conditions of pallidal stimulation (with DBS-induced parkinsonism). We also identified asymptomatic dopaminergic denervation that may not be sufficient to produce gait disorders per se, as deleterious stimulation with the GPi may be necessary to allow the emergence of symptoms (that could be reversible with stimulation adaptation). Nevertheless, the concept of different combinations of neural and networks dysfunctions networks provides a conceptual model that may foster further studies on the pathophysiology of differential effect on dystonic symptoms, and fine tuning of gait and motor control. The authors confirm that the data supporting the findings of this study are available within the article or its supplementary material. Raw data were generated at Salpetriere University Hospital, and ICM, Paris France. We are in debt to the Clinical Investigation Centre (Dr Lucette Lacomblez, MD, PhD). This work was supported by Investissement d Avenir ANR-10 AIHU-06. M.V. has no conflict of interest, received travel grants from Movement disorders Society, European Academy of Neurology. E.R. served on scientific advisory boards for Orkyn, Aguettant, Merz-Pharma; received honoraria for speeches from Orkyn, Aguettant, Merz-Pharma, Medday-Pharma, Everpharma, International Parkinson and Movement disorders Society; received research support from Merz-Pharma, Orkyn, Aguettant, Elivie, Ipsen, Everpharma, Fondation Desmarest, AMADYS, Fonds de Dotation Brou de Laurière, Agence Nationale de la Recherche; received travel grant from Vitalair, Aguettant, Merz-Pharma, Ipsen, Merck, Orkyn, Elivie, Adelia Medical, Dystonia Medical Research Foundation, International Parkinson and Movement disorders Society, European Academy of Neurology, International Association of Parkinsonism and Related Disorders. Z.K., S.R-P., P.P., B.G., D.G., M-O.H., J.Y., and S.T. report no competing interests.