Introduction

Parkinson’s disease (PD) is the second most common neurodegenerative disease worldwide after Alzheimer’s disease, affecting 1–2 % of population older than 60 years (Healy et al. 2008; Tarsy 2012). The classic clinical manifestations of PD include bradykinesia, resting tremor, rigidity and postural instability, which are largely caused by the deficiency of dopamine (DA) in the striatum due to the selective and progressive loss of dopaminergic (DAergic) neurons in the substantia nigra pars compacta (Fahn 2003). PD is believed to be caused by both genetic and environmental risk factors (Warner and Schapira 2003; Verstraeten et al., 2015; Polito et al. 2016). Pathologically, neuroinflammation, mitochondrial dysfunction, protein degradation failure, endoplasmic reticulum stress, and reactive oxygen species overproduction, have been implicated to be involved in PD pathogenesis (Hirsch et al. 2013; Kansara et al. 2013; Shen et al. 2013; Zuo and Motherwell 2013; Tang et al. 2014; Ciechanover and Kwon 2015; Michel et al. 2016). Unfortunately, until now, the precise mechanisms for PD are still largely inconclusive and most PD therapies are symptoms-based rather than preventive or disease-modifying.

Clinically, motor symptoms are the main features of PD onset and progression, but nonmotor symptoms, including depression, fatigue, hyposmia, sleep disorders, automatic dysfunction, cognitive impairment and dementia, also could be evident in some PD patients. Since 1960s, treatment for PD has been focused on the replacement or supplement of DA. Various DAergic drugs or novel mechanism-based therapeutic candidates have been developed and are subjected into clinical usage or clinical trials for PD treatment (Dong et al. 2016a, b). As the most effective pharmacological medications for PD, levodopa and other DAergic drugs benefit almost all PD patients (Poewe 2009). However, long-term use of levodopa is often accompanied by motor complications, including levodopa-induced dyskinesia (LID) and motor fluctuations, which range in severity from mild and non-disabling to incapacitating. Once motor complications emerge, it means that PD patients have entered the advanced stage, in which the dosage of levodopa should be modified and the formulation of levodopa should be changed to combine with DA agonists or other drugs to control the adverse symptoms. In addition, although some motor dysfunctions of PD, such as tremor and dyskinesia, may be alleviated with drug therapy, characteristics such as postural instability and nonmotor symptoms, especially neuropsychiatric disorders, are less responsive to current medications and require alternative approaches (Nirenberg and Fahn 2005; Schapira and Olanow 2005).

All these above-mentioned limitations of pharmacological treatments generate a notion that non-pharmacological or complementary and alternative management may offer both symptomatic relief and disease-modifications that are beneficial for PD patients. Here, we outline recent studies and summarize evidence on the effectiveness of both traditional and modern therapeutic modalities for PD (Results of clinical trials or meta-analysis of alternative therapies for PD are sumarized in Table1). Moreover, the methodological challenges and possible directions of these therapeutic options for future research are also discussed separately.

Traditional complementary and alternative therapy of PD

Complementary and alternative management of PD is normally defined as a diverse group of therapies, practices or products that share in common their exclusion from conventional and well-accepted medicinal or pharmacological therapies of PD. The variety of traditional complementary and alternative management is increasing yearly, mainly including physio-exercise therapy, Tai chi, Qi gong, relaxation training, acupuncture, traditional herbal medicines, and daily beverage and neutraceuticals.

Physio-exercises therapy

Exercise is an integral part of PD management because physical activity has been shown to retard the deterioration of motor functions, to prolong functional independence and to improve the cognitive disabilities (Lugassy and Gracies 2005; Goodwin et al. 2008; Dibble et al. 2009; Tomlinson et al. 2013, 2014; van der Kolk and King 2013). As for the underlying mechanisms, it has been shown that physio-exercise has a protective effect on the dopaminergic function of PD animals by enhancing neurotrophic factors expression, increasing mitochondrial function and stimulating neuroplasticity (Zigmond and Smeyne 2014; Vivar et al. 2016). More precisely, recent studies further reveal that physio-exercise reduces the alteration of the DAergic neurons in the substantia nigra and contributes toward reconstituting the function of the basal ganglia involved in the motor command through increasing brain-derived neurotrophic factor (BDNF) expression and recovering DA and glutamate neurotransmission (Speelman et al. 2011; Wu et al. 2011), restore mitochondrial function, and attenuate neuroinflammation (Lau et al. 2011), thereafter play neuroprotective roles and restore motor deficits (Frazzitta et al. 2013; Zigmond and Smeyne 2014; Sconce et al. 2015; Svensson et al. 2015; Tuon et al. 2015).

Recent clinical trials have further suggested that aerobic exercise including aerobic walking and stretching could ameliorate motor functions such as gait, balance, physical performance, and nonmotor functions such as fatigue, depression and cognition, but not for fall prevention in PD patients (Uc et al. 2014; Canning et al. 2015). Moreover, it has been reported that intensive training modalities could improve muscle strength and mobility (Cruickshank et al. 2015; Uhrbrand et al. 2015), and low-intensity exercise caused a better performance on gait speed than high-intensity (Shulman et al. 2013). Additionally, resistance-based exercises that address deficits in balance and strength have shown positive effects (Scandalis et al. 2001; Hirsch et al. 2003; Dibble et al. 2006). However, the efficiencies of physio-exercises therapy for PD are not compliant to all clinical symptoms and need more long-term and follow-up studies. In addition, the safety and dose–response relationships (i.e., frequency, intensity, and duration) of physio-exercise therapy for PD are still requires monitoring.

Tai chi and Qi gong

Compared with conventional physio-exercises, Tai chi, an ancient traditional Chinese exercise combining with conscious breath control, relaxation, and slow movements, has been proved effective in reducing balance impairment and falls, with additional functional improvement (Li et al. 2012; Tsang 2013). According to the recent meta-analysis, Tai chi showed positive effects in motor function and balance, but not in gait velocity, step length and gait endurance improvements (Yang et al. 2014). Except for improving motor function, Tai chi could also relieve stress and is beneficial for improving quality of life and mood (Esch et al. 2007; Nocera et al. 2013). It is likely that the benefits of Tai chi are a combination of physical training of strength, balance, stretching, and breathing, and additional improvements in mood and stress. Although the exact mechanisms underlying its therapeutic potential is still barely investigated, Tai chi has been proved to be a safe and feasible exercise that improves quality of life, and it could be a good exercise strategy for PD patients with mild to moderate severity.

Qi gong is another traditional Chinese exercise like Tai chi focusing on the movement of internal energy (chi) through the practice of meditation-like practice and focused movements. One randomized controlled trial (RCT) has suggested that Qi gong could improve Unified Parkinson’s Disease Rating Scale (UPDRS)-III score, together with several nonmotor symptoms amelioration (Schmitz-Hübsch et al. 2006). In contrast, another small-scale RCT demonstrated that there was no significant motor benefit after Qi gong practice when compared with aerobic training (Burini et al. 2006). Therefore, it still needs more studies to confirm the exact efficacies and explore the best training pattern for anti-Parkinsonism potential of Qi gong.

Mindfulness meditation, Yoga, and other relaxation training

Mindfulness is an ancient spiritual practice as well as a form of stress-reductive meditation training, in which participants are guided towards greater awareness and acceptance of the current moment (Kabat-Zinn 2009). Fitzpatrick et al. (2010) have found that mindfulness has a qualitatively beneficial effect on the PD patients’ ability to cope. Subsequently, a quantitative controlled study by Pickut et al. (2015) has suggested an improvement in the observing subscale of the Five Facet Mindfulness Questionnaire (FFMQ) and a reduction in motor disability, but no change in depression measured using the Beck Depression Inventory. In line with previous studies, one recent study found that modified Mindfulness-Based Stress Reduction training improve both motor and nonmotor symptoms of PD, as evidenced by a significant increase in the FFMQ-Observing subscale, a reduction in anxiety and depression scores, an improvement in cognition, a decrease in postural instability and gait difficulty symptomatology, and a reduction in symptom distress observed from the Outcome Questionnaire-45 (Dissanayaka et al. 2016). As for the underlying mechanism for the ameliorating effects of mindfulness against PD, Pickut et al. (2013) have found that, versus non-treated control patients, PD patients subjected to 8 weeks of a mindfulness-based training showed significant increases in gray matter density in the cuneus and lingual gyrus of the left occipital lobe, the left thalamus, the left parahippocampus, and bilaterally in the temporoparietal junction. Most of these brain regions are responsible for emotion, anxiety, and cognitive and motor functioning relevant to PD (Pickut et al. 2013). To date, however, formal evaluations of potential benefits have not been reported. Given the low cost and risk, mindfulness is a reasonable complementary practice to PD patients pending the results of further research testing.

Yoga is a traditional mindfulness-based exercise and is another form of exercise which combines multiple physical elements with relaxation and breathing. It has been shown to significantly improve measures of gait, flexibility, muscle force, fatigue, and quality of life in healthy elderly and people with medical disorders including back pain, arthritis, hypertension, anxiety, and depression. One pilot study has shown that, after 6–12 weeks training, Yoga participants report a significant UPDRS scores improvement and positive trends of ameliorations in tremor, depression scores, body weight and forced expiratory volume (Sharma et al. 2015). Another pilot study has demonstrated that after an 8-week Yoga program, some texts such as sit-and-reach text, single-leg balance text improved significantly, and depression was alleviated to some extent (Boulgarides et al. 2014). Except for these small studies, until now, there is still no big-scale RCT about Yoga in PD treatment. It still requires larger size of individuals and further investigation to ascertain the therapeutic efficiency of Yoga for PD patients.

Besides mindfulness and Yoga, various forms of massage have been indicated to be one of the most common forms of relaxation and alternative therapy for PD (Rajendran et al. 2001). Massage is a broad category that includes different forms of soft tissue manipulation, often incorporated with relaxation. Two months of Japanese massage shows a positive effect in various PD symptoms including shoulder stiffness, muscle pain and fatigue (Donoyama and Ohkoshi 2012). Another before–after study has also suggested that after 40-min Anma massage, the movement difficulties of PD patients were generally improved (Donoyama et al. 2014). A small pilot study of whole body therapeutic massage in PD patients has found an improvement in quality of life and self-reported function (Paterson et al. 2005). Alexander technique uses hand contact to assess and manipulate changes in muscle activity by addressing the relationship between thought and the resultant muscle activity. One RCT of 3-month massage and Alexander technique in 93 PD patients by Stallibrass et al. (2002) has found that, compared with no intervention control group, the Alexander technique participants show improved on self-assessment disability scores and depression ratings. Reflexology is another form of relax and stress-reductive therapeutic massage and is reported to induce improvements of nonmotor symptoms of PD (Johns et al. 2010). Another study compared massage therapy with simple muscle relaxation techniques and found that massage therapy resulted in significantly lower urinary stress hormone levels, and also self-reported improvements in sleep quality (Hernandez-Reif et al. 2002). In addition, Craig et al. (2006) have demonstrated that neuromuscular therapy, a technique similar to massage but which relies on direct compression of trigger points, was more effective than relaxation at improving motor UPDRS scores in 36 PD patients over a 4-week intervention period. However, these studies did not see a significant increase of urinary DA in PD patients. Moreover, a small RCT recently found that massage significantly reduced salivary cortisol levels in PD patients directly after the treatment, but had no long-lasting effect on diurnal cortisol levels (Törnhage et al. 2013). Although these forms of massage have been shown to be safe and beneficial, larger clinical trials are needed to more fully document their therapeutic efficacy in PD.

Dance and music therapy

Dance is a multi-dimensional activity offering auditory, visual and sensory stimulation, musical experience, social interaction, memory, motor learning and emotional perception, expression and interaction (Kattenstroth et al. 2010). As an intervention for PD patients, dance could improve both motor and nonmotor symptoms. The recent meta-analysis suggests that short-term dance significantly improves UPDRS scores, balance and gait as compared with no intervention controls (Sharp and Hewitt 2014). Dance, especially Tango, has been reported to alleviate motor function and balance, as compared with conventional exercise (Hackney et al. 2007). Additionally, Tango is also thought to offer cognitive benefits such as improving cuing strategies, as it incorporates aerobic activity and movements that challenge gait and balance with progressive motor skill learning in the presence of external cues provided by the partner and the music. Other forms of dance such as waltz/foxtrot, may show similar benefits to Tango in PD (Hackney and Earhart 2009). However, each dance form has different qualities and researchers have hypothesized that certain qualities will target specific PD symptoms. For example, Tango requires frequent movement initiation and cessation, spontaneous directional changes and movement speeds, which may target movement initiation, turning and bradykinesia. In contrast, ballet challenges strength and flexibility to emphasize posture, body alignment, projection of eye focus and limb extension, as well as whole body coordination (Houston and McGill 2013). Although there is sufficient evidence that dance therapy is enjoyable and effective for improving gait and balance in PD patients, long-term outcome data are still lacking. In addition, the safety of dance programs also needs to be adequately reported to ensure the safe and appropriate implementation of dance interventions.

Music therapy uses music or any of its elements (sound, rhythm, melody, or harmony) to facilitate and promote mobilization and expression in order to meet physical, emotional, mental, social, or cognitive requirements. Previous studies have suggested that music may have an ability to impact both motor (including gait and dexterity) and nonmotor (including cognition, anxiety, apathy and depression) functions of PD patients. Pacchetti et al. (2000) have found that the PD patients in the music therapy group show significant improvement in certain motor and quality-of-life measures. de Bruin et al. (2010) also demonstrated improvements in stride length, gait velocity, and cadence among PD patients walking to an individualized playlist compared to controls. Music has been associated with the release of some specific neurochemicals and hormones in both animal and human studies (Möckel et al. 1994; Knight and Rickard 2001; Chanda and Levitin 2013). Additionally, functional magnetic resonance imaging studies have demonstrated an association between music therapy and increased mesolimbic dopamine release (Menon and Levitin 2005).

Acupuncture

Acupuncture, as a vital part of traditional Chinese medicine, has been applies for over 3000 years. In general, acupuncture needle inserted at the specific acupoint stimulates nerve receptors both directly or indirectly, through mechanical coupling via the connective tissues surrounding the needle, then through the local reflex and the central nervous system, induces endocrine, neuroendocrine, autonomic, and systemic behavioral responses. Clinical trials or practices have demonstrated that acupuncture either manual or electro relieves some motor symptoms in PD patients (Yang et al. 2006; Ren 2008; Cho et al. 2012; Toosizadeh et al. 2015) and markedly improves many nonmotor symptoms such as psychiatric disorders, sleep problems, and gastrointestinal symptoms (Cho et al. 2012; Zeng and Zhao 2016). Additionally, previous studies have demonstrated that acupuncture could improved therapeutic efficacy and reduced dosage of levodopa and also ameliorate drugs-induced side effects or complications (Zou 2006; Kim et al. 2014). Acupuncture studies in PD animal models have reported multiple mechanisms for its neuroprotective effects and therapeutic potential. Acupuncture elicited significant recovery and reduced degeneration of substantia nigra dopaminergic neurons in 6-hydroxydopamine (6-OHDA) lesion rat model (Park et al. 2003). Electroacupuncture may modulate apoptosis and neuroinflammation in the 1-methyl 4-phenyl 1,2,3,6-tetrahydropyridine (MPTP) mouse model, suggesting neuroprotective effects (Jeon et al. 2008), possibly via p53-related signaling (Park et al. 2015). Acupuncture or electroacupuncture has also been reported to modulate the expressions of dopamine transporter (DAT) and DA receptors, increase postsynaptic dopamine neurotransmission, and inhibit microglial activation and inflammatory events (Kang et al. 2007; Kim et al. 2011; Rui et al. 2013). Using functional magnetic resonance imaging technique, another randomized trial revealed that acupuncture brought significant improvement of motor function with putamen and primary motor cortex activation, as compared with placebo control groups (Chae et al. 2009). All these findings suggested that acupuncture may serve as a promising complementary and alternative therapy for PD, although more studies, either comparative effectiveness research or high-quality placebo-controlled clinical studies are still warranted.

Traditional herbal medicines

Traditional herbal medicines have been used for centuries. Recent preclinical studies have suggested that some of herbal medicines, either as single agent or in various combinations, may have neuroprotective effects in PD animal models. Ginseng is a plant substance which has been used in Eastern Asian countries for centuries and is proposed to have anti-inflammatory properties, improve fatigue and cognition. Multiple components of Ginseng, such as Rb1, Rg1, Rd, Re, Notoginsenoside R2, and Pseudoginsenoside-F11 have attracted remarkable interest as promising agents due to their beneficial effects in various PD animal models (González-Burgos et al. 2015). These compounds exert their neuroprotective activity through different mechanisms including inhibition of oxidative stress and neuroinflammation, decreases in apoptosis and nigral iron levels, and regulation of N-methyl-d-aspartate receptor activity (Wang et al. 2009; González-Burgos et al. 2015; Zhou et al. 2016). Ginkgo biloba, as another one widely used herbal medicine with potential antioxidant and neuroprotective properties, has beneficial effects in MPTP animal models (Rojas et al. 2012a, b). The Ginkgo biloba extract EGb 761 and its main constituent flavonoids and ginkgolides increase extracellular dopamine levels in the rat prefrontal cortex (Yoshitake et al. 2010). Clinically, a case report has demonstrated that PD patients showed a dramatic improvement after supplementation with ginkgo and a multivitamin–multimineral supplement (Conrad 2014). Interestingly, herbs containing high concentrations of levodopa, such as Mucuna pruriens and Vicia faba, have been reported to have neuroprotective effects, including recovering endogenous DA production and reducing oxidative stress in the substantia nigra (Manyam et al. 2004; Yadav et al. 2013), and have also been shown to potentially result in less dyskinesia (Rabey et al. 1992; Kasture et al. 2009; Lieu et al. 2010, 2012). Traditional Chinese decoction, San-Huang-Xie-Xin-Tang, composed of Coptidis rhizoma, Scutellariae radix, and Rhei rhizoma, possesses beneficial protection against MPTP model of PD via antioxidative and antiapoptotic effects (Lo et al. 2012). One pilot study reported that dietary extract rikkunshi-to could reduce gastroparesis in terms of shortening gastric emptying time in PD (Doi et al. 2014). Yokukansan is another kind of herbal extract, which is efficient in ameliorating neuropsychiatric symptoms of PD, such as hallucinations, anxiety and apathy, according to a small-scale exploratory trial (Hatano et al. 2014). Through evaluating some neurotransmitters in the brain, Bushen huoxue formulas are considered to enhance the levels of serotonin and norepinephrine, and to improve the depression of PD (Wang et al. 2014). Although previous studies have provided evidence supporting the positive therapeutic potential of herbal medicines for PD, there is still a long way to go for the clinical usage of herbal medicines in PD therapy due to the deficiency of exact bioactive components spectrum data and the lack of sufficient clinical data and safety profile.

Life habits and nutrition supplements

Previous studies have reported an inverse association between cigarette smoking and PD (Morens et al. 1995; Kyrozis et al. 2013). Additionally, numerous case–control studies or cohort studies provide evidence for the protective effect of smoking (Sasco and Paffenbarger 1990; Grandinetti et al. 1994; Gorell et al. 1999; Benedetti et al. 2000; Li et al. 2015). As for the possible mechanisms, previous studies have demonstrated that the anti-PD and neuroprotective effect of cigarette smoking might be due to the reducing enzymatic activity of type B monoamine oxidase (MAO-B) (Fowler et al. 1996), which catabolizes DA and may activate neurotoxicants. Moreover, 2,3,6-trimethyl-1,4-naphthoquinone, one major naphthoquinone component isolated from tobacco, has also been found to exert MAO-B-inhibiting activity and attenuate the DAergic system toxicity in MPTP mouse model (Castagnoli et al. 2001). An animal model further suggests that nicotine, another one major component of cigarette, may act as an antioxidant or prevent excitotoxicity (Halliwell 2006). However, there is no available clinical evidence to indicate that smoking or the use of nicotine and other components of tobacco is beneficial once PD has appeared. Moreover, the healthy concern of tobacco also limits its clinical usage as an alternative therapy for PD.

Similar with smoking, results from epidemiological studies reveal a consistent inverse association between coffee consumption and the risk of PD. A meta-analysis of eight case–control and five cohort studies conducted in four countries between 1968 and 2001 found a 31 % lower risk of PD in coffee drinkers than non-coffee drinkers (Hernán et al. 2002). In addition, this protective potential of coffee against PD shows significant gender preference and hormone sensitivity, as evidenced by the lower PD risk in man and the impacts of postmenopausal hormones on PD risk in women (Ascherio et al. 2003). The protective effect of coffee on PD could be ascribed to caffeine, which is the most important component of coffee and is recently considered as potent Adenosine A2A receptor antagonist. Caffeine and other A2 adenosine receptor antagonists have been shown to stimulate DA release and protect against DAergic neurotoxicity in animal models (Chen et al. 2001; Trevitt et al. 2009; Petzer and Petzer 2015). Although these clinical or preclinical studies have suggested a protective potential of coffee consumption against PD, however, the mechanisms involved are not fully understood and it is premature to recommend increasing coffee consumption to prevent PD, especially in women taking postmenopausal hormones.

In contrast to smoking and coffee, the previously reported relationship between alcohol consumption and PD risk is contradictory. While Palacios et al. (2012) demonstrated that the consumption of beer, wine, or liquor was not associated with PD risk, a more recent meta-analysis of observational studies has revealed an inverse association between alcohol consumption and risk of PD (Zhang et al. 2014). Moreover, multiple lines of evidence have suggested that resveratrol, one natural product present in grape and red wine, could ameliorate neuroinflammation, restor mitochondrial function, reverse oxidative stress over-production, and thereafter provide benefits to various PD animal models (Anandhan et al. 2010; Zhang et al. 2010; Ferretta et al. 2014).

Among the various lifestyle factors, tea consumption has attracted increasing interest, as one of the traditional and most consumed beverages in the world. Previous case–control studies and a cohort study have reported an inverse association between tea drinking and PD risk (Chan et al. 1998; Checkoway et al. 2002; Tan et al. 2003, 2008). Polyphenols components in tea may play important roles in delaying the onset or halting the progression of PD. Green tea and black tea are rich sources of polyphenols, including the most abundant and well-known epigallocatechin-3-gallate (EGCG) and theaflavins. There is now consistent mechanistic data on the neuroprotective and neuroregenerative effects of tea and its major bioactive components (Xu et al. 2006), indicating that they do not just possess anti-oxidant, anti-neuroinflammation or anti-chelating properties (as reviewed by Caruana and Vassallo 2015) but may directly inhibit α-synuclein aggregation and modulate intracellular signaling pathways, both in vitro and in vivo (Bieschke et al. 2010; Camilleri et al. 2013; Chen et al. 2014). However, despite significant data on the potential neuroprotective effects of tea and its bioactive components, clinical studies are still very limited and to date only EGCG has reached phase II trials.

Modern medicine practices that integrate traditional beliefs often focus on nutrition or nutraceuticals, as part of a healthy lifestyle, with the expectation that lowering inflammation and free radical damage may protect against further neuronal death and thus delay or halt disease progression. Evidence from epidemiological studies have been presented that various nutritional supplements, especially vitamins, decrease the risk for PD. For example, vitamin B6 intake has been associated with a significantly decreased risk of PD in smokers (de Lau et al. 2006), vitamin D is also inversely associated with PD (Knekt et al. 2010; Wang et al. 2015). Despite vitamins, other nutraceuticals, such as coenzyme Q10, fish oil and selenium, have shown therapeutic potential for PD. Coenzyme Q10, as the electron acceptor for mitochondrial complexes I and II, can improve complex I activity and reduced levels of coenzyme Q10 have been identified in the mitochondria of PD patients (Shults et al. 1997). However, a phase III trial demonstrates that high-dosage of coenzyme Q10 shows no benefits in early PD (Beal et al. 2014). Similar as coenzyme Q10, although fish oil, glutathione and seletium have been reported to exert antioxidant and anti-neuroinflammation properties, there is still no sufficient evidence to support their clinical usage for PD and the toxicological profiles and side effects of these nutraceuticals should be concerned.

Modern non-pharmacological therapy of PD

Despite those above-mentioned traditional complementary and alternative modalities for PD, recent progress of modern technologies greatly promote the progression of the development of novel strategies for PD therapy. These minimally surgical or non-surgical therapeutic strategies attract attentions and greatly enrich the scope of PD alternative therapy and their clinical benefits for future PD management.

Deep brain stimulation (DBS)

Deep brain stimulation has generally been accepted as an alternative therapy for the amelioration of parkinsonian motor symptoms. Subthalamic nucleus (STN) and globus pallidus internus (GPi), two most hyperactive brain regions during PD progression, are usually used as targets for DBS. In addition, STN-DBS could be preferred to manage patients with advanced PD who have predominantly posture and gait disorders due to its larger improvements in off time (Odekerken et al. 2013). The beneficial effects of DBS on appendicular symptoms such as tremor, rigidity, limb bradykinesia and dyskinesia are well established (Roper et al. 2016), although the exact underlying mechanisms for DBS still remain poorly understood. After long-term observation, both STN- and GPi-targeted DBS therapy showed significant improvement in “on–off” conditions, dyskinesias and motor fluctuations (Weaver et al. 2012; Rizzone et al. 2014). Recently, DBS treatment at 60 Hz frequency shows a promising application potential to improve swallowing, gait freezing, and axial motor signs, almost overall motor signs of PD (Khoo et al. 2014; Xie et al. 2015). Despite the STN and Gpis, DBS therapy with pedunculopontine tegmental nucleus (PPNs) as stimulation targets has recently been found to be an alternative to STN-DBS (Follett et al. 2012; Katz et al. 2015). Moreover, a novel 32-contact electrode of DBS, brings more potential benefits via widened therapeutic window and increased effectiveness (Contarino et al. 2014). As for the mechanisms for therapeutic potential of DBS against PD, preclinical trials, by using either non-human primates or other mammalian species, have shown that there is an improvement of DAergic neurons survival and an increase of BDNF level in the substantia nigra and primary motor cortex after STN-DBS exposure, suggesting the neuroprotective effect of DBS (Wallace et al. 2007; Spieles-Engemann et al. 2011). The local effects of DBS tend to result in the inhibition of neuronal-cell bodies and the excitation of neighboring axons. In addition, astrocytes are stimulated to release calcium, which may lead to a release of glutamate and adenosine, improved microvascular integrity, as well as local increases in cerebral blood flow. Finally, there is evidence that deep brain stimulation can induce local and possibly distal proliferation of neural precursor cells (Okun 2012; Pienaar et al. 2015). From a neurophysiological view, the “disruption hypothesis” seems to be more and more accepted, in which DBS dissociates input and output signals, resulting in the disruption of abnormal information flow through the stimulation site (Chiken and Nambu 2016). Although gross motor improvement is frequently observed following DBS therapy in PD, lack of uniformity in methodology and measured outcomes has clouded the interpretation of DBS benefits across the motor domain. In addition, despite these therapeutic benefits to motor performance of PD, the effects of DBS on nonmotor cognitive and psychiatric symptoms of PD have been controversial. A progressive worsening of neuropsychological performance and movement disorder are even observed (Merola et al. 2011;  Baizabal-Carvallo and Jankovic J 2016), although some scholars consider the impairment of neurocognition might be due to the disease progress and medication reduction, not the DBS itself (Sáez-Zea et al. 2012; Weaver et al. 2012; Albuquerque et al. 2014). Finally, further large sample size trials for brain target site analyses should be performed for clarify the different therapeutic outcomes of various DBS stimulation targets, including STN, GPi and PPNs (Diamond and Jankovic 2005; Baizabal-Carvallo and Jankovic 2014).

Repetitive transcranial magnetic stimulation (rTMS)

Surgical therapies, including DBS, improve advanced symptoms above the best medical therapy, however, only limited (less than 5 %) PD population is suitable candidates for this surgical procedure (Morgante et al. 2007). During the past two decades, rTMS has been closely examined as a possible therapeutic option for PD (Pascual-Leone et al. 1994a, b; Fregni et al. 2005; Elahi et al. 2009). As a noninvasive therapeutic procedure, rTMS delivers repeated magnetic pulses to a specific brain area within a short time through a stimulation coil placed over the scalp without surgery or anesthesia. The repeated magnetic pulses not only alter excitability at local site of stimulation but also influence distant brain regions via the cortico-basal ganglia-thalamo-cortical motor circuit (Strafella et al. 2001; Kim et al. 2008; González-García et al. 2011). Additionally, different types of rTMS may cause different response of cortical excitability. For examples, while high-frequency rTMS induces cortical excitability (Pascual-Leone et al. 1994a, b), low-frequency rTMS causes a depressed cortical activity (Chen et al. 1997). Continuous high-frequency rTMS decreases cortical excitability, whereas intermittent high-frequency rTMS increases cortical excitability (Huang et al. 2005). Because rTMS can produce changes in neural activity and motor signs last well after stimulation, this technique has generated much interest as a potential therapeutic intervention for patients with PD and many rTMS trials have been uptaken to investigate its clinical benefits (Chou et al. 2015; Zanjani et al. 2015). Moreover, several types of rTMS were reported to be effective for drugs-induced dyskinesia and also been tried for non-pharmacological treatment of nonmotor symptoms of PD including depression (Shirota et al. 2016). One most recent meta-analysis further advocates the beneficial effect of rTMS on upper limb function in the short term and on walking performance and motor signs measured by UPDRS III in both the short- and long terms (Chung and Mak 2016). The use of rTMS has a minimal risk of adverse events in the PD patient population and should be encouraged as an alternative treatment for PD so long as it is thought to produce clinically relevant improvements in motor functions. However, the potential risk for seizure as well as a small risk of transient headache and scalp pain should be concerned (Vonloh et al. 2013).

Near-infrared light (NIr) therapy

NIr has been applied in clinical practice mainly for treating tissue contusion for many years. There have also been many in vivo studies of NIr-induced neuroprotection in various PD animal models. Previous preclinical studies have demonstrated that NIr could improve behavior deficits and DAergic neurons survival in parkinsonian mice (Shaw et al. 2010; Johnstone et al. 2014; Moro et al. 2014; Reinhart et al. 2015a, b; El Massri et al. 2016). Remarkably, a recent primate trial has further supported the notion that NIr was neuroprotective but not toxic to brain, which brought a step closer to clinical translation (Darlot et al. 2016). Although substantial evidence supports the therapeutic efficacy of NIr in PD animal models, there have been only limited research reports to date on the efficacy of NIr treatment in PD patients. There is a recent non-controlled and non-randomized clinical report indicating improved speech, cognition, freezing episodes and gait after extracranial NIr therapy in PD patients (Maloney et al. 2010); there are also some clinical reports suggesting improvements in parkinsonian signs in PD patients after intranasal NIr therapy (Zhao et al. 2003). We should note that extracranially delivered NIr may not reach the zones of pathology in the brainstem of PD patients and would only provide symptomatic effects. Hence, an intracranial system might be developed and suggested for PD patients to play neuroprotective and disease-modifying roles to maximize its therapeutic effects.

Table 1 Results of clinical trials or meta-analysis of alternative therapies for PD
Full size table

Neurotrophic factors (NTFs) therapy

Most of the current pharmacological therapies for PD focus on managing symptoms, rather than the molecular pathogenic causes. Therefore, the greatest unmet demand in PD therapy is to discover and develop disease-modifying strategies, which can halt or reverse the ongoing progression of dopaminergic neurons degeneration. One of the promising candidates for disease-modifying therapies is NTFs. NTFs are secreted growth factors that regulate survival, growth, differentiation and maintenance of neurons in both the central nervous system and the peripheral nervous system. Although many animal research suggesting the therapeutic potential of neurotrophic factors for treating neurodegenerative diseases, the clinical trials have not produced conclusive evidence to supporting their utilities as therapies (for reviews see Sullivan and Toulouse 2011; Hegarty et al. 2014; Bartus and Johnson 2016a, b). Following the series of disappointing outcomes from systemic administration of NTFs, including GDNF, CNTF, BDNF and IGF-1, to improve their therapeutic efficiency, NTFs was later tested using intrathecal, Intraventricular or intraparenchymal delivery. However, these types of delivery are associated with several practical problems, such as the poor targeting and rapid biometabolism of NTFs by endogenous enzymes in vivo. Gene therapy can overcome these limitations by incorporating the gene for the therapeutic protein into brain cells, achieving long-term and targeted delivery. Recombinant adeno-associated viral (AAV) or lentiviral vectors (LV) have been used to induce over-expression of NTFs in animal models of PD (for recent review see Kelly et al. 2015) and led to considerable advancements in the ongoing development of clinical neuroprotective therapies for PD. However, in contrast to the successful studies which used GDNF in the 6-OHDA lesion rat models, GDNF delivery by either AAV or LV vectors was not effective in preventing neurodegeneration in the AAV-α-synuclein model (Decressac et al. 2011). Further investigation demonstrated that α-synuclein overexpression downregulates the expression of Nurr1 and its downstream target, the GDNF receptor Ret, in DA neurons of the SN (Decressac et al. 2012). These findings suggest that the development of DAergic NTFs might be Ret-dependent and Ret/NTFs combined gene delivery approach or other Ret-independent NTFs should be explored (Sullivan and O’Keeffe 2016).

Gene therapy

In general, gene therapy methodology is achieved by counteracting or replacing a malfunctioning gene within the cells adversely affected by the condition. For the most part, viruses-based vectors are used to deliver therapeutic molecules including glutamic acid decarboxylase (GAD), aromatic l-amino acid decarboxylase (AADC), neurturin, and neurotrophic factors by now. A phase 1/2 trial with one-year follow-up has shown that ProSavin, a LV vector-based gene therapy to deliver tyrosine hydroxylase, GAD and AADC, significantly improves UPDES-III scores of PD patients without serious side effects (Palfi et al. 2014). Transfer of GAD with AAV2 can modulate γ-aminobutyric acid production with a great improvement of UPDRS scores over 6 months as well (LeWitt et al. 2011). Others like AAV2-hAADC and AAV2-neurturin (CERE-120) have also shown the similar therapeutic benefits and safety profiles (Marks et al. 2010; Mittermeyer et al. 2012; Bartus et al. 2013). Moreover, modified virus-based vectors and non-virus vectors are developed constantly. For examples, tropism-modified Ad5 vectors exert neuron-selective targeting property to enhance gene delivery efficiency (Lewis et al. 2014). Angiopep-conjugated nanoparticles for cellular uptake and gene expression can carry specific genes (Huang et al. 2013). A recently developed non-viral vector has been found to deliver short interfering RNA (siRNA) against the α-synuclein gene specific for neuronal cells, and prevent PD-like symptoms both in vitro and in vivo. Interestingly, this non-viral vector not only helps siRNA duplexes cross the blood–brain barrier in mice, but also stabilize these siRNAs leading to a sustainable knockdown of α-synuclein protein (Javed et al. 2016). Although recent clinical trials of gene therapy have shown remarkable therapeutic benefits and an excellent safety record, however, at this time, no study has demonstrated robust clinical benefits using rigorous double-blind assessments. In addition, as currently practiced, gene therapy for PD only addressed the cardinal motor symptoms, and the possible impacts of gene therapy on other clinical signs of PD should be further investigated.

Fig. 1
figure 1

Complementary and alternative therapies of Parkinson ‘s disease

Full size image

Cell replacement therapy

Cell transplantation is available nowadays and numerous preclinical studies and several clinical trials have shown therapeutic effects of various types of stem cell transplantation, such as motor signs improvement or medication dosage reduction (O’Keeffe et al. 2008; Venkataramana et al. 2010; Gonzalez et al. 2015a, b; Han et al. 2015). Although transplantation of stem cells-derived DAergic neuron can alleviate motor deficiencies of PD, previous studies also found that tumor formation was an unacceptable complication associated with stem cells, especially embryonic stem cells (ESCs) transplantation (Bjorklund et al. 2002; Kim et al. 2002). To minimize the risk of tumor formation, a number of techniques have been developed, including prolonged pre-differentiation of ESCs, selection of differentiated cells for transplantation and genetic engineering to block tumourigenic pathways (Ambasudhan et al. 2014). In addition, Acquarone et al. (2015) pretreated undifferentiated mouse ESCs with mitomycin, then injected into striatum in nude mice. After 15-month follow-up, mitomycin-treated ESCs alleviated motor functions dramatically without unlimited cell proliferation that would be a novel replacement therapy for PD. Besides, reprogrammed neurons, such as combination of new transcriptional therapy may decrease the tumorigenic potential (Yamashita and Abe 2014). Using human unfertilized cell or induced pluripotent stem cells (iPSCs) also offers an unlimited supply for transplantation. The animal experiments confirm its safety and efficiency on motor symptoms (Gonzalez et al. 2015a, b; Han et al. 2015). In a long-term 14-year observation after DAergic neuron transplantation, the majority of transplanted neurons maintain healthy and functional, as shown by the persistent expression of DA transporters and normal mitochondrial morphologies, which proves the rationality and feasibility of cell transplantation as alternative therapy in PD (Hallett et al. 2014). Although many types of stem cells, such as ESCs, iPSCs and NSCs, have been induced into DAergic neurons for PD treatment in various animal models, so far few has been clinically proved functional for patients with PD. Besides the tumorigenic potential, the cell sources, optimal transplantation protocols, including reliable delivery system, transplantation locations and timing, also need to be further explored before clinical use of stem cells-derived DAergic neurons for the transplantation into PD patients.

Environmental enrichment (EE)

Enriched environments are normally defined as a combination of complex inanimate and social stimulation. For rodent studies, EE is generally constituted by bigger cages, with a running wheel and a few toys that are periodically changed to stimulate animal curiosity and exploration. Previous studies have suggested that EE could protect brain damage and promote neuroplasticity in various neurodegenerative animal models including PD. Faherty et al. (2005) have demonstrated that exposure to an EE procedure (a combination of exercise, social interactions and learning) during adulthood totally protects against MPTP-induced Parkinsonism. Furthermore, this neuroprotection of dopamine neurons in the nigrostriatal system might be due to the increased GDNF and decreased DAT and vesicular monoamine transporter (VMAT2) levels (Faherty et al. 2005; Zhu et al. 2005). Similar neuroprotective effects of enriched environment in PD have also been confirmed in 6-OHDA animal model (Steiner et al. 2006; Anastasía et al. 2009). Although all these preclinical findings predict significant implications of this non-pharmacological approach for the prevention and/or treatment of PD, the exact molecular mechanisms are far from being clearly understood. Much more importantly, till now the clinical efficacy of EE therapy in PD patients is rarely reported. Moreover, the containing elements and the training procedure in EE therapy for PD should be further explored before the translation of this non-pharmacological modality from benchside to bedside.

Conclusion

PD patients exhibit various degrees of motor impairment as well as nonmotor symptoms which include depression, apathy, cognitive impairment, sleep disturbances, and autonomic dysfunctions. In the absence of any disease-modifying therapy, the mainstay of PD treatment is pharmacologic therapy aimed at replacement of DAergic function. While this is generally effective for motor symptoms, long-term medication may fail to treat, or even worsen, troublesome nonmotor symptoms and may lead to motor complications. It is not surprising then that nearly 40 % of PD patients in US were using complementary and alternative practices in addition to or instead of conventional treatment options (Rajendran et al. 2001). Centuries of experience using conventional medical modalities (such as herbal medicines and acupuncture) and traditional physical training (includes Tai chi, Qi gong and Yoga) in Eastern countries have been proved to be effective to relieve some of PD symptoms via various mechanisms. In addition, the rapid development of modern science and technology has add new meaning into this traditional remedies and greatly promote the discovery of novel non-pharmacological therapies of PD, including DBS, NIr, gene therapy and cell replacement therapies, as well as some other complementary management (Fig. 1). All these newly developed therapeutic options have been demonstrated to be able to provide potential clinical benefits to patients with PD to relieve either motor or nonmotor deficiencies or drugs-induced side effects. Much more importantly, following the clinical usage of these alternative therapeutic options, we could learn more about the precise pathogenic mechanisms of PD and found more therapeutic targets by which to discover more disease-modifying drugs. In addition, it is proposed that these new therapies may bring promise to give new cures to not only PD, but also other neurodegenerative diseases (Nithianantharajah and Hannan 2006; Perlmutter and Mink 2006; Ljubisavljevic et al. 2013; Soligo et al. 2013).