Parkinson’s Disease Is Predominantly an Environmental Disease

Improved diagnosis?

A common response to the rise is that awareness of PD has increased and that our ability to diagnose it has improved over time. While these assertions may be true, they would only explain the relative rise of PD if they were more true of PD than other neurological disorders, such as multiple sclerosis and stroke. However, the opposite is likely the case. Advances in neuroimaging have greatly enhanced our ability to diagnose individuals with these two latter conditions, often at an earlier stage of disease [3, 4]. Yet, neither has grown nearly as rapidly as PD [3]. Moreover, PD is often missed and thus, the true burden is underestimated. Door-to-door studies suggest that anywhere from 12% (Rotterdam, The Netherlands) to 48% (Beijing) to 100% (rural Bolivia) of individuals with the disease are undiagnosed [4].

Aging?

A second explanation commonly offered is that the rise of PD is due to aging. Indeed, the incidence of PD (Fig. 1A) begins to rise in individuals roughly in their 50 s and then doubles or triples every decade thereafter until at least the ninth decade of life. However, the rates of PD have risen in both absolute terms and, when adjusted for age [5]. By contrast, the global burden of Alzheimer’s disease has more than doubled in absolute terms from 1990 to 2016 but, adjusted for age, has changed little [3]. So aging alone cannot explain the rise in PD.

OPEN IN VIEWER

Fig. 1

A) Incidence of Parkinson’s disease by age and sex in the United Kingdom, 2017 [16]. B) Incidence of lung cancer by age and sex in the United Kingdom, 2016–2018 [17].

More importantly, there is hardly any evidence that aging per se is a cause of PD. Aging, though, comes with a time-related decline in physiological functions [6], and these changes could contribute to PD. For example, aging is associated with reduced protein clearance [7] and lysosomal impairment [8], tasks that may be relevant to PD and other neurodegenerative conditions.

Longevity, the length of life, is critical to PD [9]. PD, like heart disease and many cancers, is an age-related disorder. A comparison to lung cancer is illuminating. The increased incidence of PD with age is almost identical to that of lung cancer (Fig. 1B) [10]. Few would argue that lung cancer is due to aging. Indeed, prior to the advent of cigarettes in the early 20^th century, lung cancer was so rare that when confronted with a case, doctors considered it “a once-in-a-lifetime oddity” [10]. However, twenty-five years after the introduction of cigarettes, the rates of lung cancer began to rise, in an age-related manner [10]. Longevity allows for both sufficient exposure to toxicants (e.g., enough pack-years to cause a malignancy) and time to pass for the disease to become manifest. Indeed, a long life may be a prerequisite to a diagnosis of lung cancer and PD as the pathological processes underlying both disorders take years, if not decades, to unfold [11].

PD and lung cancer share another similarity (Fig. 1). In most studies, both are more common in men [12]. Sociology, rather than biology, may be the reason. Historically, men have smoked more often than women; however, as the gender differences in smoking have diminished [13], the incidence of lung cancer in women has risen. In the case of PD, men in many countries are more likely to be engaged in work that exposes them to environmental toxicants, such as pesticides. However, in a few countries, such as Japan, PD is more common in women [14], and there, the majority of farmers historically have been women [15].

Genetics?

The third explanation is that genetic mutations are the principal cause of PD. Genetic factors do contribute to the etiology of PD, as Professors Shen-Yang Lim and Christine Klein detail well in a companion paper in this journal [18]. However, purely genetic causes of PD are rare (Table 1) [19], and its heritability (estimated at 0.34 from twin studies) is low [20]. We have known since Sir William Gowers’s case series of PD in 1903 that only about 15% of individuals with the condition have a family history of the disease [21]. A century later, Tanner and colleagues examined PD in twins and found similar rates in monozygotic and dizygotic twins leading to the conclusion that “genetic factors do not play a major role in causing typical PD” [22]. Compared to other conditions where heritability was also studied in twins, PD is near the bottom in terms of genetic contribution [20].

Purely genetic causes of PD account for about 2% of individuals with the disease. Even “common” genetic causes, such as LRRK2 mutations, are present in only 2–3% of individuals and have incomplete penetrance (25 to 43%) [23]. Other factors, including environmental ones, must be necessary for the disease to develop [24]. Indeed, rather than causing the disease, the principal role of genetics is likely to help determine (along with other environmental factors [25]) who among the exposed is likely to develop the disease. The latter contention is supported by experimental work showing that almost all the principal genetic causes and risk factors tied to PD have known interactions with environmental toxicants (Table 1). In other words, the underlying genetic predisposition likely enhances the risk of nigrostriatal damage following exposure to toxic chemicals.

In addition, variability in the genes encoding for proteins that metabolize some notable toxicants, especially pesticides [26], may be critical in explaining who develops PD and who does not. Indeed, this gene-environment interaction might well explain why some of the older mutations only became relevant when humans began producing certain toxicants. For example, genetic variants in GBA (penetrance 10–20% [27]), encoding the lysosomal enzyme glucocerebrosidase, are common risk factors for PD today. However, the founder effects for these genetic variants are likely thousands of years old [28] with little contribution to PD prior to the 19^th century.

Finally, genetic differences also cannot readily explain the near perfect correlation between use of pesticides and PD prevalence [29], the strong association between agricultural activities and its incidence [30], the 10-fold variation in prevalence within countries [31], or the five-fold difference in age-adjusted prevalence between countries [5]. Nor do they account for the finding that the rates of PD are generally highest in industrialized nations like Canada and the U.S., lowest in less industrialized regions, and rising most rapidly in industrializing nations like India and China [5]. In short, while additional (poly- or epi-) genetic causes of the disease remain to be identified, a century of studies suggest that the overall heritability of PD is low [20].

Pesticides

Pesticides are the environmental toxicants with the clearest relationship to PD. Numerous epidemiological and animal studies argue for a causal relationship between certain pesticides and PD. An important clue came in the 1980s when seven young adults developed end-stage parkinsonism sub-acutely after intravenous injection of the designer drug 1-methyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP) [34]. Damage to the nigrostriatal dopaminergic neurons occurs after MPTP is converted to its metabolite 1-methyl-4-phenylpyridinium (MPP+), which kills dopaminergic neurons through inhibition of mitochondrial complex 1. MPP+ shows a striking structural resemblance to paraquat, one of the world’s most widely used weedkillers which also impairs mitochondrial function [35].

Later epidemiological work further suggested a link between PD and prior exposure to pesticides. For example, Barbeau and colleagues found that the correlation between the prevalence of PD and the use of pesticides in nine areas of Quebec, Canada, was 0.967 [29]. Since then, similar studies have been reported in the U.S. [36], Israel [37], and France [30]. In 2011, Tanner and colleagues found an association between PD and use of the pesticides rotenone (odds ratio 2.5) and paraquat (odds ratio 2.5), both of which cause mitochondrial dysfunction [38].

Importantly, the risk of PD is not confined to those working professionally with pesticides. Those living near their application are also at risk. For example, in France the increased risk of developing PD is not limited to farmers [39, 40] but extends to those living near farmland or vineyards [30]. Pesticide exposure may also lead to faster disease progression after the diagnosis has been established [41, 42].

Numerous laboratory studies support these epidemiological findings. For example, rotenone inhibits complex I of the respiratory chain and causes nigrostriatal dopaminergic degeneration in rats, along with visible signs of bradykinesia and rigidity [43]. Similar findings have been demonstrated with paraquat [44]. Today, several animal models of PD are based on exposure to pesticides and other neurotoxicants [45].

Several pesticides that have been linked to PD (e.g., paraquat, rotenone, heptachlor) are now banned in many countries, but major concerns remain. Some of these toxic pesticides continue to be used in many parts of the world including, as Professor Shen-Yang Lim and colleagues have reported, in the Western Pacific [46]. Paraquat is still marketed in Argentina, Australia, Japan, India and the U.S. [47], where its use has more than doubled from 2013 to 2018 [48]. A visual inspection of the use of paraquat in the U.S. from 1992 [48] and the incidence of PD is concerning (Fig. 2A, B).

OPEN IN VIEWER

Fig. 2

A) Incidence of Parkinson’s disease in the U.S., 2012 [49]. B) Estimated annual use of paraquat in the U.S., 1992 [50]. Map of release of trichloroethylene (orange) and perchloroethylene (in blue) in the U.S, 1987. Map courtesy of Meghan Pawlik, University Rochester, based on data from the Toxic Release Inventory from the U.S. Environmental Protection Agency. D) Map of particulate matter smaller than 2.5 microns in the U.S., 2006 [51].

In addition, serious deficiencies in current regulatory policies prevent adequate assessment of currently used pesticides when it comes the risk of PD and other neurodegenerative disorders. Our concerns are detailed elsewhere [52], but here we highlight three major drawbacks. First, current screens for neurotoxicity depend on the manifestation of parkinsonism in exposed animals but frequently ignore relevant nigrostriatal neuronal loss which often precedes clinically discernible signs. This limitation had been long acknowledged by Parkinson’s experts and is now recognized by the European Food Safety Authority [53].

Second, many safety studies have been performed by the industry itself, which may not have been fully transparent about the health risks of their products. A recent investigative report by the British newspaper The Guardian reviewed legal documents that paraquat’s manufacturer submitted in response to lawsuits from individuals with PD. According to the reporters, the manufacturer had conducted research showing that rats and mice exposed to very high doses of the pesticide displayed a stiff gait or tremors. The studies were conducted in the 1960s [47]. So toxic is the chemical that regulators in 1974 expressed concern about workers “who might inadvertently lick small quantities of paraquat residue off lips, or inhale paraquat mist” [47]. The company was also concerned in 2009 that the scientific community would “conclude from the laboratory and epidemiology data that paraquat exposure is a causal factor (its emphasis) in Parkinson’s Disease or parkinsonism” [47]. Similarly, the manufacturers of glyphosate, another weed killer, withheld a relevant study from regulators which provided evidence of developmental neurotoxicity in young rats [54]. Despite the evidence linking certain pesticides to PD and other diseases (e.g., amyotrophic lateral sclerosis [55]), their use persists. To limit exposure to these toxicants requires greater awareness of industry practices employed to manufacture doubt and accountability for its actions [56–59].

Third, farmers, rural residents, and ultimately almost all of us are exposed to mixtures of multiple pesticides. Such mixtures are either used by farmers or arise from environmental contamination. For example, following use on crops, glyphosate is carried across long distances via the air, and in Germany, is found in measurable levels across the entire country, including in the most protected natural areas [60]. The large international SPRINT stud (Sustainable Plant Protection Transition) identified high concentrations of numerous pesticides, particularly in dust in farmhouses [61]. Multiple pesticides were also found in blood and stool samples of both farmers and people living near farmlands. Due to cumulative or possibly even synergistic effects, such cocktails may contribute to a much higher risk of developing PD. A recent study in California suggests that co-exposure to multiple pesticides is associated with much greater damage to dopaminergic neurons than exposure to any single compound [62], yet current regulatory actions only consider the neurotoxicity of isolated compounds.

The result of these shortcomings is that few of the commonly used pesticides have been adequately tested for their risk of PD. Improved neurotoxicity studies that focus specifically on the risk of PD and evaluate different combinations are urgently needed. And such experiments should be performed by independent researchers and not only by the industry itself.

Trichloroethylene

Pesticides may account for PD in rural areas but are unlikely to be a satisfactory explanation for the high rates of the disease in urban areas [38]. Other environmental toxicants are thus likely to be responsible as well. One may be the ubiquitous chemical, trichloroethylene or TCE [63]. However, despite its widespread use and broad contamination, few are aware that they have ever been exposed [64]. The result is that its influence on PD is likely (vastly) underappreciated.

First created in 1864, use of TCE began in the 1920s, and it has since found a myriad of military, industrial and commercial applications including degreasing metals, decaffeinating coffee, and dry cleaning clothes [63]. At one point, one out of every 12 workers in the U.K. and 10 million Americans are estimated to have worked with the volatile solvent [65]. In Italy, a population study found TCE in the urine of 75% of individuals [66]. TCE is a known carcinogen, and its toxic effects have been known since at least 1932 [67]. Like many pesticides, TCE is a mitochondrial toxicant that inhibits complex I of the respiratory chain [63]. In addition, recent research has found that like certain LRRK2 mutations, TCE induces LRRK2 kinase activity in the brains of rats and causes nigrostriatal dopaminergic degeneration [68].

Since 1969, case reports have linked TCE exposure to parkinsonism or PD [65, 69–71]. In 2008, Gash and colleagues found that among 30 factory workers, three developed PD after using TCE for many years to degrease and clean metal gauges. These three individuals were stationed closest to an open vat of TCE, and 14 of 27 individuals who worked further from the source “displayed many features of parkinsonism, including significant motor slowing” [72]. Four years later, Goldman and colleagues [73] found that the risk of PD among individuals who had hobby or work exposure to the solvent was 500% higher than their twins who did not. The lag time between exposure and PD diagnosis was 10 to 40 years [73]. Gash [72] and others [69, 74–76] have found that the lipophilic compound causes selective loss of dopaminergic neurons in animal studies.

Exposure through work and consumer applications (including in carpet cleaners, paint removers, and typewriter correction fluid [77]) may just be the beginning. Environmental exposure to the chemical through outdoor air, contaminated ground water, and polluted indoor air may be responsible for the bulk of exposures, most of which is invisible to individuals [63]. One of the most glaring examples of TCE (and a closely related chemical, perchloroethylene [63], with likely similar toxicity [73]) contamination is the U.S. Marine base Camp Lejeune, where the drinking water was contaminated with the two chemicals for thirty years at up to 3000 times those permitted by safety standards [4, 78]. A recent study by Goldman and colleagues found that individuals who served at Camp Lejeune had a 70% increased risk of PD compared to those who served at a much less polluted base [79].

However, TCE contamination does not end with water. Beginning in the 1970s, researchers began to realize that the volatile TCE could evaporate from contaminated water and enter people’s workplaces, schools, and homes, usually undetected, through a process called vapor intrusion [63]. Today, a school in Shanghai, China [80], and multi-million dollar homes in Newport Beach, California [81], sit on top of contaminated sites, and their occupants may be subject to the adverse health effect of inhaling this toxicant. The number of TCE-contaminated sites globally are too numerous to count; the state of Michigan alone is estimated to have thousands [82]. Figure 2C shows a map of U.S. release of TCE and other chlorinated solvents into the air from 1987 (the first year for which data are available [83]) and the incidence of PD. The U.S. Environmental Protection Agency just recently proposed a ban [84] on the known carcinogen, an action some European countries have already taken. However, global use, especially in China, is increasing [85].

Air pollution

Concerns about the health effects of outdoor air pollution date to at least the time of Hippocrates, around 400 BCE, who linked multiple ailments to poor air quality [86]. The Roman philosopher Seneca noted that escaping from the polluted air of Rome around 64 CE resolved the “sluggishness in [his] brain” [86]. However, industrial air pollution first took hold in London. There, in 1661, air pollution from burning coal to power steam engines captured the attention of the British writer John Evelyn when he wrote that the glorious city wrapped “her stately head in clouds of smoke and sulphur, so full of stink and darkness” [86]. When Dr. James Parkinson wrote An Essay on the Shaking Palsy in 1817, the people he described would have long been exposed to the toxic effects of the infamous London fog [87].

However, the potential link between air pollution and PD is only now beginning to become clear [88]. The source of air pollution varies by time (open fires were the likely concern of Hippocrates) and place (coal burning in London, car exhaust in Los Angeles, wildfires in Canada). In 2008, Calderón-Garcidueñas and colleagues sought to evaluate the effects of air pollution in Mexico City, which had the world’s dirtiest air in the 1990s. To do so, she examined the brains of children and young adults (ages 11 months to 40 years) who died suddenly [89]. She found Lewy neurites in the olfactory bulb of toddlers, which were present in 68% of children in their second decade of life [90] and alpha-synuclein immunoreactivity in their substantia nigra [91]. Since that time, several studies have linked traffic-related air pollution to an increased risk of PD in Denmark [92], Taiwan [93], and South Korea [94]. Given the high content of particulate matter from wildfires, their effects on PD requires further investigation [95–97].

Tied to these pathological findings may be deficits in the sense of smell [98], which is a common feature of PD that typically antedates the onset of the motor syndrome by many years. It is possible that this hyposmia is a remnant of the toxic effects of air pollutants that damage the olfactory system early in the disease process, while on their way to cause further harm to the brain. The aforementioned observations on the individuals with MPTP-induced parkinsonism provide further fascinating evidence to support this idea. Despite having severe and irreversible clinical parkinsonism, the smell in six MPTP patients was normal, noting that the entry route for the neurotoxin had been an intravenous one [99]. However, the chemist who had manufactured this designer drug developed parkinsonism several years later even though he had never used MPTP intravenously (J.W. Langston, personal communication). His parkinsonism was presumably related to exposure to lower MPTP doses via inhalation. Unfortunately, his sense of smell was never formally tested.

The reason for the relationship between air pollution and PD is uncertain, but it may be particulate matter. Suspended in the smog visible in large urban areas are tiny particles of dirt and soot. The smallest of these (measured in microns or even nanometers) can bypass the body’s normal protective mechanisms and enter noses, airways, and lungs [100, 101]. Hitchhiking on these particles are heavy metals from cars (e.g., lead from gasoline, iron from brakes, platinum from catalytic converters) and industry (e.g., arsenic, manganese, mercury) that are toxic to neurons [102]. Recent research has found that fine particulate matter (smaller than 2.5 microns) in transgenic mice causes mitochondrial dysfunction, triggers the fibrillization of alpha-synuclein, and leads to the formation of alpha-synuclein fibrils [103]. Figures 2A and 2D depict the U.S. incidence of PD [49] and a map of particulate air pollution. The relationship remains to be quantified, and numerous factors, including the appropriate time lag between exposure and incidence and effects of migration, need to be addressed. Nonetheless, on the surface, the maps and similar ones [104] are concerning about a possible relationship, which requires further investigation.