Smoking and Parkinson’s disease

Tobacco smoking is a lifestyle factor that substantially increases morbidity and mortality. Yet, it enigmatically appears to reduce the risk of developing Parkinson’s disease (PD), reducing the pooled adjusted odds ratio of smokers vs. never-smokers to develop PD to 0.2–0.51. Smoking has been shown to significantly reduce the age of onset in a large cohort of American patients with idiopathic PD2 and seems to have a similar effect in some monogenic forms of PD3,4.

The exceptionally detrimental consequences of smoking on health would indicate that selective mortality in smokers could explain this association since idiopathic PD usually occurs late in life. However, it turned out that potential delaying effects of smoking, selective mortality or other forms of reverse causation do not fully explain this association1,5,6. The same is true for the hypotheses that (a) differences in dopaminergic circuits may account for a preference for smoking, while at the same time protecting from PD, (b) alterations in the dopaminergic system during disease progression may result in reduced smoking habits or (c) low-risk-taking personality traits may be a confounding factor explaining the effects of smoking on PD5,7. Nevertheless, reverse causation can still not be completely excluded and neuroprotective effects of smoking remain to be conclusively demonstrated. PD symptoms such as hyposmia and/or changes in the reward circuits may influence smoking behavior (see hypothesis b), potentially facilitating cessation of smoking prior to the diagnosis of PD due to decreased enjoyment. Accordingly, it is possible that early-stage PD patients stop smoking more frequently than the general population, as recently discussed8.

The potential causal relationship between smoking and PD has recently been extensively summarized1. Nicotine has been considered the primary candidate mediating neuroprotection in smokers. However, this assumption has been refuted at least for continuous (as opposed to intermittent, like in smoking) nicotine uptake regarding disease progression in manifest PD in the NIC-PD phase II trial (NCT01560754). Several other compounds taken up during tobacco smoking like carbon monoxide (CO), monoamine oxidase-B inhibitors or cytochrome p450 enzymes have been proposed to reduce the risk to develop PD but their roles in PD require substantially more research1.

Carbon monoxide: a potential modulator of smoking effects

Rose and colleagues have now made a compelling case for CO as a potentially causal factor of tobacco smoking in the prevention of PD risk in animal models9. By increasing hemoglobin-bound CO-levels in the blood of PD model rats (AAV-A53T) to 4.5–7.8% (mostly below 1% in control rats), they significantly attenuated the loss of dopaminergic (tyrosine-hydroxylase and NeuN positive) neurons and alpha-synuclein pathology in the substantia nigra observed in this model9. This also resulted in increased dopamine levels in the striatum. The authors confirmed their findings in a commonly used toxin model of PD. In C57BL/6 mice, dopaminergic cell loss following a single, intraperitoneal injection of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) was significantly reduced in mice with elevated hemoglobin-bound CO levels (15.6–23.2%)9.

Moreover, Rose and colleagues found that increased CO-levels in rats resulted in the elevation of nuclear levels of the hypoxic response regulating transcription factor hypoxia-inducible factor 1alpha (HIF1a). Consequently, they also observed increased levels of heme-oxygenase 1 (HO-1), an enzyme responsible for decomposition of heme. HO-1 is upregulated in response to oxidative stress and many neurotoxins and can be both protective and detrimental in PD models10. Following hypoxia or ischemia, the dynamic regulation of HO-1 is thought to underlie beneficial adaptations, protecting from future hypoxic injury11. While increased HO-1 activity is associated with antioxidative and anti-inflammatory effects, several of its metabolites exert opposite effects and may promote neurodegenerative processes10. The expression of HO-1 is known to be controlled by HIFs10 but also by other transcription factors, including nuclear factor erythroid 2-related factor 2 (Nrf2), nuclear factor ‘kappa-light-chain-enhancer’ of activated B-cells (NfKB) or activator protein 1 (AP-1), all known to be activated by hypoxic stress12.

Intriguingly, smoking is associated with higher HO-1 levels in cerebrospinal fluid of humans9 and increased HO-1 levels have been reported in postmortem substantia nigra samples, serum, plasma and saliva of PD patients.

How CO might improve oxygen transport in Parkinson’s disease

CO binds to hemoglobin with high affinity, displacing oxygen from hemoglobin and thereby impairing oxygen transport in the blood. While not toxic at low levels, high levels of hemoglobin-bound CO (about 40% and higher in rats13) are highly neurotoxic. Impaired oxygen supply is emerging as a notable feature of PD14,15,16. Whereas chronic deficiencies in oxygen supply or disordered breathing during sleep are thought to be involved in the pathogenesis of PD, innate cellular and systemic responses to transient hypoxic stress can potentially be used to increase the resilience to hypoxia17. The involved beneficial hypoxia responses can induce protective metabolic alterations, improve cellular oxidative stress management and trigger hematological (e.g., increased erythropoiesis and hemoglobin mass) and vascular (angioneogenesis and vascular remodeling) adaptations17. They are strongly mediated by transcription factors like HIFs and Nrf2. These effects may counteract chronic tissue hypoxia and can be induced by mild intermittent hypoxia protocols (hypoxia conditioning)17 and/or “functional hypoxia”, elicited by motor-cognitive training in ambient hypoxia18. Whether intermittent hypoxia is an effective tool to treat manifest PD is investigated in an ongoing clinical trial19.

In sports, inhalation of low-dose CO has recently garnered interest as a potential means to increase hemoglobin mass and exercise performance, although the extent of such effects is debated and ethically questionable20. Still, it highlights the possibility that the blood’s oxygen transport capacity can be modulated by CO-inhalation. CO-induced impaired oxygen transport is thought to mediate similar hematological and vascular adaptations like “altitude training”.

Low dose CO therefore may induce similar physiological responses and adaptations like the exposure to moderate altitude, as suggested by the overlaps in signal pathway activation (e.g., hypoxia-induced transcription factors) and molecular outcomes (e.g., increased expression of HO-1 and its antioxidant effects). These adaptations are thought to improve the capacity of cells and tissues to manage disturbances in oxygen homeostasis. They may therefore protect from harmful consequences of chronic hypoxic conditions in tissues or subsequent severe acute hypoxia, such as during ischemic events. Further supporting overlapping neuroprotective mechanisms underlying low CO doses and mild hypoxia may be similar effects on the promotion of mitochondrial quality control, which may be induced by hypoxia21 and by CO, at least in astrocytes22.

In contrast to these beneficial outcomes, intriguingly both too high levels of CO23 or too severe hypoxia14 can have the exact opposite effects, namely promoting the development of PD. This indicates that the described phenomenon can be classified as a hormesis effect; while low doses (of CO or hypoxia) induce protective adaptations, too high doses are toxic.

The possibility of harmful hyperoxia in Parkinson’s disease

Oxygen levels are strictly controlled in healthy tissues since both too low (hypoxia) and too high (hyperoxia) oxygen availability pose threads to cells12. Severe hypoxia is primarily associated with energetic crisis because oxidative phosphorylation in mitochondria depends on oxygen. Hyperoxia promotes oxidative stress and inflammation. Therefore, cells and the whole organism have to respond quickly to changes in oxygen availability. These responses may be impaired in Parkinson’s disease14,24,25 and mitochondrial dysfunction, a molecular hallmark of PD, may be importantly involved. Complex I of the mitochondrial respiratory system is considered a primary vulnerability in PD brain26. Mitochondria have central roles in oxygen sensing12 and the functioning of complex I is crucial for hypoxia responses27. Mitochondrial complex I dysfunction has also recently been suggested to lead to locally increased oxygen concentrations: in animal models of diseases with complex I dysfunction such as Leigh disease28 or Friedreich’s ataxia29, continuous exposure to ambient hypoxia reduced brain hyperoxia and ameliorated symptoms and disease progression. Since deficits in mitochondrial complex I are characteristic of PD, hyperoxia may occur as well and could possibly be reduced by the administration of CO, by smoking or exposure to moderate hypoxia. Moreover, hyperoxia results in vasoconstriction in the brain, which in turn is associated with oxidative stress30, another pathological hallmark of PD. Importantly, both ambient hypoxia12 and CO31 are well known to counteract vasoconstriction in the brain, in part via HO-1.

Notably, smoking is associated with a later age of onset both in idiopathic PD2 and at least in some familial forms3,4. This indicates that the mechanisms by which it is protective—possibly by modulating oxygen transport—may be an important feature in most forms of PD, even if caused by different mutations and in heterogenous disease manifestations. Since mitochondrial dysfunction is common in idiopathic PD and several monogenic causes of PD are directly linked to mitochondrial deficits32, impaired oxygen supply may indeed be an underappreciated factor and pathological link in different types of PD. That smoking is protective not only in idiopathic PD but also in some familial forms has for example been shown for the LRRK2 (leucine-rich repeat kinase 2) mutation p.Gly2019Ser3 and SNCA (encoding alpha-synuclein) polymorphisms4. However, the association remains poorly defined for most genetic types of PD.

Conclusions

Deficits in cellular and systemic responses to altered oxygen availability are increasingly acknowledged features of PD. Various approaches can restore these responses and possibly benefit people with PD. Smoking—at least in part via increasing CO-levels and thereby repeatedly challenging the oxygen transport system—may reduce the risk for PD by “training” the hypoxia response systems and leading to increased cellular resilience and maintained adaptive capacities on the systemic levels (e.g. hypoxic ventilatory response). However, the adverse effects of tobacco smoking greatly outweigh potentially protective components of smoking or smoking behavior in PD. Smoking is associated with increased risks for many non-communicable diseases, including vascular and respiratory diseases or neoplasms, and therefore clearly is not suitable as a preventive strategy1,33,34,35. Accordingly, although PD smokers may die from fewer neurological causes, their risk to die from smoking-related cancers has been shown to be significantly increased34. Should the beneficial effects of increased CO-levels in PD be confirmed, low-dose CO inhalations may be one interesting novel strategy to target one crucial aspect of PD. If short, repeated exposures to hypoxia have similar consequences like CO, hypoxia conditioning, strategic altitude exposures or adapted breathing exercises will be important alternative approaches.