Role of deep brain stimulation targeted to the pedunculopontine nucleus in Parkinson’s disease

Gait disturbance and postural instability (PI) are among the most medication-refractory motor symptoms of Parkinson’s disease (PD). Recent reports, however, suggest that deep brain stimulation (DBS) of the pedunculopontine nucleus may ameliorate these symptoms in PD patients. In this editorial, we summarize the literature and propose future directions and methodological considerations for research utilizing this promising treatment modality.

Many unmet needs still exist for patients with PD, especially those with levodopa-resistant symptoms: gait and balance problems, cognitive dysfunction, depression, psychosis, sleep disorders and autonomic dysfunction. Despite the improvement in therapy for the motor symptoms of PD, motor fluctuations and levodopa-induced dyskinesia (LID) can be challenging to manage, particularly in the advanced stages of PD. DBS targeted to the subthalamic nucleus (STN) [1] or globus pallidus interna (GPi) are important options for the management of patients with levodopa-resistant motor symptoms. Unfortunately, gait and PI often become levodopa resistant in mid- to late-stage disease and may respond only partially to STN-DBS [2]. Emerging animal and human data suggest that DBS of the pedunculopontine nucleus (PPN) may prove beneficial for this group of patients.

Parkinson’s disease & deep-brain stimulation

The benefits of STN-DBS are well established in the PD population and include improvement in activities of daily living as measured by the Unified Parkinson’s disease Rating Scale (UPDRS) Part II and improvements in motor function, as judged by the UPDRS Part III (motor subsection). Depending on the study, ‘off’ time decreases by approximately 50% [3,4]. In a meta-analysis of STN-DBS studies, Kleiner-Fisman and colleagues report the average reduction of LID to be 69.1% [5]. Bilateral STN stimulation appears to be superior to unilateral stimulation, but not twice as effective [6–8]. While some reports indicate that STN-DBS improves LID better than GPi-DBS [9–11], most experts believe GPi-DBS to be slightly more effective via direct antidyskinetic effects [4]. STN-DBS typically allows for a 50% reduction in levodopa equivalents, a phenomenon not as prominent with GPi [12,13]. On the other hand, preliminary data indicate that adverse effects, including cognitive deterioration, are reported less frequently with GPi-DBS [4]. The dramatic shift in the surgical management of movement disorders has led to a proliferation of centers offering stimulation surgery. Unfortunately, the lack of consensus on standardization has raised issues concerning the safety of DBS and the prevention of DBS failures [14–16]. Many basic questions regarding patient selection, optimal target location and mechanisms of action continue to be debated [17]. Multicenter studies designed to address these issues are currently underway [18].

Pedunculopontine nucleus

Many manifestations of PD, including PI, freezing, falls and sleep disturbances, are often resistant to levodopa therapy, and thus cannot be solely explained by degradation of nigrostriatal dopaminergic pathways. Thus it is likely that in addition to the traditional concept of nigro–striatal–pallido–thalamo–cortical neuronal circuits controlling voluntary movements, descending basal ganglia pathways to the brain-stem nuclei also play a role. Recent evidence suggests that the PPN may play a role in a number of parkinsonian deficits. PPN is part of the mesencephalic locomotor region, a functionally defined portion of the brain-stem that controls locomotion in animals [19]. PPN regulates and relays basal ganglia activity to influence sleep, waking, learning, reward and other cognitive functions [20]. Interestingly, postmortem studies of PD and progressive supranuclear palsy reveal that the extent of PPN degeneration correlates with the degree of premortal gait dysfunction [21].

Neuroanatomy & neurochemistry

The PPN is composed of two groups of neurons: one containing acetylcholine and the other containing GABA and glutamate [22–25]. The PPN is connected reciprocally with the limbic system, basal ganglia nuclei and the brainstem reticular formation [22,26]. Specifically, the PPN is connected to many neuronal structures through a variety of pathways: mainly cholinergic dorsal ascending fibers to thalamic nuclei; noncholinergic ventral ascending fibers to the globus pallidus, substantia nigra (SN), STN and ventral tegmental area; brainstem nuclei; bilateral descending projections to the medullary and pontine reticular formation; and bilateral direct spinal projections [25,27]. The PPN is reciprocally connected with the ipsilateral prefrontal motor cortex and the cortico–tegmental tract, which terminates in the PPN. Neurons in the lateral STN send efferent fibers directly to the PPN and tracer studies have identified profuse projections from the PPN to bilateral STN [25,28]. Combined, these neuronal connections lay the foundation for how PPN stimulation may enhance multiple brain targets leading to improved PD therapy [25,29].

Animal studies of PPN

Animal models have provided great insight into the control of locomotion. The caudally directed corticolimbic–ventral striatal–ventral pallidal–PPN–pontomedullary–reticular nuclei–spinal cord pathway appears to be involved in initiation, acceleration–deceleration and termination of locomotion. This pathway is under the control of deep cerebellar and basal ganglia nuclei at the level of the PPN, particularly from the STN, GPi and SN [25–27,29,30]. Spontaneous locomotion appears to be mediated, in part, by ventral pallidal projections to the PPN, as evidenced by carbachol injection into unilateral PPN, resulting in a reduction in spontaneous locomotor activity and normalization of locomotor activities produced by amphetamine injection into the nucleus accubens and procaine injection into the PPN, suppressing the enhanced locomotion from hippocampus stimulation [31–33]. PPN activity is also involved in the rhythmicity of locomotion. Up to 70% of the neurons in the areas ventral or dorsal to the PPN show bursting activity associated with cyclic frequencies of locomotion. Conversely, 77% of neurons in the PPN fire tonically and are active during, before or after locomotor activities [34]. They seem to turn the rhythmic bursting neurons on or off, and fire transiently to initiate or terminate locomotion [34]. Further animal studies suggest that PPN activity may play a role in turning behavior. Amphetamine injection that leads to the damage of approximately 75% of cholinergic neurons within the PPN causes ipsilateral turning behavior, whereas noncholinergic damage results in contralateral turning [35].

Pharmacologic activation or electrical stimulation of the PPN increases motor activity [36,37]. Microinjections of bicuculline, a GABA antagonist, into the PPN was observed to alleviate akinesia in the parkinsonian primate [37]. Macaque monkeys that undergo unilateral PPN lesions with kainic acid are protected from 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-mediated parkinsonism, presumably via suppression of glutamatergic input to nigrostriatal neurons [38]. Jenkinson and colleagues have studied the effects of unilateral PPN-DBS prior to and after inducing MPTP parkinsonism in a macaque monkey [36]. Using constant amplitude and pulse width, three variable conditions were examined: DBS off, 5-Hz and 100-Hz stimulation. A 5-Hz stimulation increased movement counts/h and 100 Hz decreased the same measures. Low-frequency PPN stimulation was as effective as levodopa, but high-frequency DBS was not.

Human studies of PPN

The first report of benefits with bilateral PPN-DBS involved two patients with PD who displayed dominant symptoms of freezing of gait and PI in off and on medication states before surgery [39]. Evaluations occurred preoperatively and postoperatively at day 42 for the first patient and day 16 for the second. All evaluations included off and on medication scores while postoperative evaluations also included off and on stimulation evaluations. Based on the UPDRS and Tinetti gait/balance assessment tool, the best clinical response was observed with stimulation at 20–25 Hz. Total UPDRS improved 53% and the motor subscore improved 57% when averaged between the two patients. High-frequency stimulation actually worsened gait in one patient. This benefit of low- rather than high-frequency stimulation is an unusual aspect of PPN-DBS when compared with STN- or GPi-DBS, and one that will likely prolong battery life.

Mazzone and colleagues described electrophysiologic properties of PPN in two patients with PD during bilateral PPN/STN-DBS implantation [40]. Their report focused on the unique electrophysiologic properties of PPN. While very interesting, this article lacks extensive clinical information. Nevertheless, they found low-frequency stimulation (10 Hz) was associated with a feeling of ‘wellbeing’ that correlated with modest improvements in motor function (bradykinesia and rigidity). Unfortunately, the effects of posture and gait were not reported in this initial investigation.

Stefani and colleagues reported their experience with simultaneous bilateral PPN/STN-DBS placement in six PD patients with severe (UPDRS Part III >70) symptoms including both disabling dyskinesia and PI [41]. STN-DBS alone may be anticipated to help with LID and motor fluctuations, but not significantly with levodopa-resistant gait dysfunction, hence the rationale for proceeding with multiple leads. Patients were evaluated at baseline, and 2 months and 6 months after implantation in the medication off/on states. Postoperative evaluations included both STN and PPN stimulation off and on. Low-frequency (25 Hz) PPN stimulation off medications improved UPDRS Part III by 45% initially, but benefits waned to 32% 3–6 months following the procedure. STN stimulation improved the UPDRS part III by 54% and this improvement was sustained at 6 months. Combination of PPN/STN stimulation did not provide further benefit beyond STN stimulation in the medication off state. In the on medication state, however, dual stimulation of PPN/STN provided further improvement beyond the extent of either single target.

Future directions

We have enough anatomical, neurochemical, pathological, animal and preliminary human evidence to support the hypothesis that PPN-DBS is a promising therapy for levodopa-resistant gait dysfunction and PI, but many questions remain unanswered and more studies are needed to validate the preliminary results. The most recent publication claims that PPN-DBS stimulation alone does not improve motor function significantly in the off medication state to justify its use as ‘a fully alternative target’, thereby implying the need for combination PPN/STN-DBS [41]. On the other hand, the first clinical report argues that PPN-DBS alone can improve bradykinesia and rigidity (but not tremor) along with gait dysfunction [39]. In our opinion, the therapeutic benefit of PPN-DBS in the absence of STN-DBS should be better characterized before the implantation of four simultaneous electrodes (bilateral STN/PPN-DBS) becomes routine. Owing to the clear and significant improvement seen with STN- or GPi-DBS in moderate-to-advanced PD patients, it is difficult to justify performing PPN-DBS primarily for levodopa-responsive symptoms at this time.

Current reports discuss patients implanted with bilateral PPN-DBS. The possibility that unilateral stimulation alone may be sufficient should be explored as these nuclei have more midline projections than STN and GPi. Rather than proceeding immediately with multiple electrodes, it may be safer to first implant one PPN electrode. This would allow several other questions to be answered about PPN-DBS, including whether bradykinesia, rigidity or tremor improve (or if levodopa requirements may be decreased), without the confound of other surgeries and other targets. It is also unclear whether PPN stimulation modulates parkinsonism immediately (as with tremor) or in a delayed fashion (as with dystonia) and whether these effects are sustained. Optimal programming parameters remain to be studied and reported in a systematic manner.

Perhaps the most important consideration as we move toward studying the PPN as a target for DBS will be methodological. Study design, strict inclusion/exclusion criteria, and validated instruments will be crucial to success. We will need to encourage investigators to go beyond the UPDRS and to seek objective measures that capture subtle changes in gait using validated clinical scales [42] or biomechanical technology [2]. When reporting results on PPN-DBS, the methods (including targeting techniques) need to be carefully outlined and the results (such as the final lead location and all side effects [delayed or immediate, device related or not]) must be documented. Given the aforementioned complex interactions between PPN and many other regions of the CNS, cognition and mood should be monitored closely using validated clinical scales. Negative results and study failures are equally, if not more, important to publish and disseminate.

In conclusion, PPN-DBS may improve levodopa-resistant gait dysfunction in PD, an important unmet need at this time. There remain several unanswered questions that must addressed before a larger, definitive study establishing the role PPN-DBS in the treatment of PD is undertaken. Given the negative effect of DBS on other Parkinson-plus syndromes, more experience from human PD patients is probably needed before PPN-DBS can be attempted for other conditions with prominent gait dysfunction, such as progressive supranuclear palsy, multiple system atrophy, vascular parkinsonism or normal pressure hydrocephalus. Finally, since PPN-DBS may modulate PPN glutamatergic activity, this target may exert a disease-modifying effect though there is a lack of evidence at this time to support this conclusion.

Acknowledgements

The authors would like to acknowledge the support of the National Parkinson Foundation Centers of Excellence at Baylor College of Medicine and the University of Florida. The authors would also like to acknowledge the assistance of Chris Haas PhD.