Abstract
Exposures to particulate matter and gaseous air pollutants have been associated with respiratory tract inflammation, disruption of the nasal respiratory and olfactory barriers, systemic inflammation, production of mediators of inflammation capable of reaching the brain and systemic circulation of particulate matter. Mexico City (MC) residents are exposed to significant amounts of ozone, particulate matter and associated lipopolysaccharides. MC dogs exhibit brain inflammation and an acceleration of Alzheimer’s-like pathology, suggesting that the brain is adversely affected by air pollutants. MC children, adolescents and adults have a significant upregulation of
Keywords
Introduction
The chronic health effects associated with chronic exposures to significant concentrations of air pollutants are an important issue for the 20 million people, including 9 million children, living in Metropolitan MC, as well as for millions more around the world. MC’s geographical setting and climate along with the relatively low mobility of its residents presents an opportunity to identify chronic adverse health effects associated with sustained exposures to urban air pollution.
The purpose of this review is to summarize our work on the health effects of pollutant exposures in clinically healthy pediatric residents of MC, and to correlate the pediatric pathology with the pathology we have observed in MC adults and dogs.
Metropolitan Mexico City Air Quality Trends
An automated air quality monitoring system (>30 stations) measures the ambient concentrations of ozone (O3), PM10 and PM2.5 (particulate matter <10 μm and <2.5 μm in diameter), sulphur dioxide (SO2), nitrogen dioxide (NO2), and carbon monoxide (CO) on an hourly basis in MC. The critical air pollutants in MC are O3, PM10 and PM2.5 with concentrations above their respective daily and annual U.S. air quality standards (Blake and Rowland, 1995; Bravo and Torres, 2002). Ozone is by far the most important air pollutant in terms of frequency of occurrence of high levels, persistence, and spatial distribution (Bravo and Torres, 2002). PM2.5 has significant concentration differences between different areas in the city, reflecting the distribution of the fixed and mobile PM sources. Other pollutants that have been identified in MCs ambient atmosphere include: formaldehyde, peroxyacetylnitrate, and lipopolysacharides (LPS), an important component of PM (Bonner et al., 1998; Osornio-Vargas et al., 2003).
Systemic Effects of Air Pollutant Exposure in Clinically Healthy Children
Apparently healthy children living in MC have systemic inflammation as evidenced by increased serum concentrations of proinflammatory mediators including IL-6, IL-1β, and PGE2 (Calderón-Garcidueñas et al., 2003b). Systemic IL-10 concentrations are also upregulated in MC children (Calderón-Garcidueñas et al., 2003b). IL-10 limits and ultimately terminates inflammatory responses by suppressing proinflammatory cytokine production and human T cell proliferation (West et al., 1994; Ziegler-Heitbrock, 1995; West and Heagy, 2002). Innate and adaptive immune responses in MC children may also be affected by chronic exposure to PM-associated LPS, which is suggested by significant alterations in the systemic expression of several LPS-binding proteins in MC children, including lactoferrin and heat shock protein 60 and the increased expression of CD14 on peripheral blood monocytes (Calderón-Garcidueñas et al., 2005b).
There is also decreased peripheral blood monocyte expression of the human leukocyte DR antigen (HLA-DR), a major histocompatibility complex class II (MHC-II) molecule (L. Calderon-Garciduenas, unpublished observations). LPS-induced downregulation of monocyte MHC-II expression and impaired antigen presentation are features of endotoxin tolerance (Wolk et al., 2003), a transient immunodeficiency that could be elicited in MC children by sustained exposure to PM-associated LPS. LPS-induced suppression of MHC-II expression is independent of IL-10 (Wolk et al., 2003), which also suppresses MHC-II surface expression (Koppelman et al., 1997; Fumeaux and Pugin, 2002). The strong suppression of HLA-DR expression by peripheral blood monocytes in MC children could be the result of a combination of elevated IL-10 and endotoxin tolerance.
The immunosuppressive effects of IL-10 and LPS in MC children are accompanied by a significant decrease in natural killer (NK) cells (Calderón-Garcidueñas et al., 2005b), a further indication of immunosuppression. All of these effects may be regarded as beneficial in the context of respiratory tract inflammation, but chronic diminished innate and adaptive immune responses may be detrimental to the host in the setting of an infectious process or in tumor surveillance (Nolan, 1975; Muehlstedt et al., 2002; West and Heagy, 2002; Michel, 2003).
The Nasal Barrier and the Lung
The nasal cavity warms, and humidifies inspired air and protects the lower respiratory tract from certain air pollutants by filtering the inspired air (Proctor, 1995). This filtering activity, places the nasal epithelium (including the olfactory epithelium) at risk of damage and disease. The olfactory pathway is a direct connection between the environment and the olfactory bulb (OB), the first synaptic relay of the olfactory system in the brain. Olfactory sensory neuron axons exit the nasal olfactory epithelium, coalesce into the olfactory nerve, and terminate on the dendrites of tufted and mitral cells and local circuit neurons in the OB (Lopez-Mascaraque and de Castro, 2002). Axons of the mitral and tufted neurons project to the anterior olfactory nucleus, the pyriform and entorhinal cortices and cortical nuclei of the amygdala (Lopez-Mascaraque and de Castro, 2002).
The entorhinal cortex and the hippocampus contribute importantly to learning and memory (Stockhorst and Pietrowsky, 2004). The olfactory pathway provides a route by which metals and other toxicants that come into contact with the olfactory epithelium can enter the central nervous system without the interference of the blood-brain-barrier (BBB) (Henriksson et al., 1997; Thiel and Audus, 2001). Thus, the nose is a prime portal of entry of pollutants into the brain, especially regions of the brain that are essential for learning and memory.
With age, MC children and adults have a progressive substitution of the normal mucociliary nasal epithelium with a squamous metaplastic epithelium with increased cell proliferation and inflammatory infiltrates in the submucosa (Figure 1A; (Calderón-Garcidueñas et al., 1992, 2001c, 2001d). Ultrafine PM deposition in nasal epithelial cells (Figure 1B), intercellular spaces, and in the transudate between epithelial cells is observed in healthy MC children (Calderón-Garcidueñas et al., 2001d). Furthermore, 22% of 112 MC children who underwent a complete ear nose and throat exam showed a grossly abnormal nasal mucosa whose microscopic counterpart is a squamous metaplastic/dysplastic epithelium (Calderón-Garcidueñas et al., 2003b). These observations have 3 major implications.
First, the nose may contribute to systemic inflammation in MC residents, because the inflammatory response of the nasal epithelium is a significant contributor to the systemic levels of cytokines such as IL-6 and IL-8 (Kenney et al., 1994). Second, uptake of pollutants through the nose may be enhanced in MC residents, because the ability to filter out inhaled gases and PM (mucociliary clearance) is reduced in the absence of an intact mucociliary nasal epithelium. Third, poor mucociliary transport rate prolongs PM retention times resulting in an increased risk of carcinogenic effects (Calderón-Garcidueñas et al., 2001b).
The pathology of chronic air pollutant exposure has been pursued further in MC dogs. Dogs living in MC also exhibit chronic respiratory tract inflammation and breakdown of both the respiratory and olfactory epithelial barriers with accumulation of PM in Bowman’s glands (Figure 1C) and significant metallothionein I and II immunoreactivity in the olfactory mucosa (Figure 1D) (Calderón-Garcidueñas et al., 2001b, 2002, 2003a). The expression of metallothioneins suggests significant exposure to metals, presumably metals associated with PM produced by fossil fuel combustion. These observations in dogs suggest that the nasal respiratory and the olfactory barriers are probably impaired in humans chronically exposed to air pollutants as well.
The major lung findings in MC dogs and humans are bronchiolar pathology (hyperplastic epithelium, smooth muscle cell hyperplasia, peribronchiolar fibrosis) and chronic mononuclear inflammatory infiltrates that surround bronchioles, pulmonary veins and arteries. In dogs, free ultrafine PM (UfPM) is seen in the cytoplasm of macrophage-like cells in the lumen of lung capillaries (Figure 1E) and translocation of UfPM could be seen from alveolar cells to endothelial cells, to intraluminal macrophage-like cells (Figure 1F) and to interstitial cells (Figure 1G) (Calderón-Garcidueñas et al., 2001b). In MC adolescents and young adults, polymorphonuclear neutrophils are often found attached to swollen lung capillary endothelial cells (Figure 1H), a plausible explanation for the marked decrease in neutrophils in venous blood observed in clinically healthy MC children.
Cardiovascular Effects of Air Pollutants
Exposure of MC children to PM2.5 is associated with increased serum concentrations of endothelin-1 (ET-1) (Calderón-Garcidueñas et al., 2005c), a potent vasoconstrictor that regulates pulmonary arterial pressure (reviewed in Galie et al., 2004). This finding is in accordance with reports of elevated circulating ET-1 levels in rodents exposed to PM2.5 (Thomson et al., 2004). In keeping with the elevated ET-1 levels, MC children have elevated pulmonary arterial pressures as well (L. Calderón-Garcidueñas, unpublished observations).
Elevated circulating levels of ET-1 is a feature of endothelial dysfunction, a syndrome characterized by a shift of the actions of the endothelium toward reduced vasodilatation, a proinflammatory state, and prothrombotic activities (Endemann and Schiffrin, 2004). ET-1 stimulates integrin-dependent adhesion of neutrophils to endothelial cells (Lopez et al., 1993; Zouki et al., 1999). Thus elevated ET-1 levels may mediate the margination of neutrophils in lung capillaries (Figure 1H) and the marked decrease in the absolute numbers of peripheral blood neutrophils in MC children. Neutrophils are also often observed in close contact with endothelial cells in brain capillaries of MC adolescents (L. Calderon-Garcideunas, unpublished observations) suggesting that circulating ET-1 may act systemically. Adhesion of neutrophils to lung, heart and brain endothelium has been documented in MC dogs as well (Calderón-Garcidueñas et al., 2001b, 2001d, 2003a). ET-1 also mediates cardiac mast cell degranulation, matrix metalloproteinase activation and myocardial remodeling in rats (Murray et al., 2004). Thus the increased plasma ET-1 levels in MC children may promote myocardial remodeling.
The cardiac pathology of chronic exposure to urban air pollution has been explored in some detail in MC dogs (Calderón-Garcidueñas et al., 2001a). Healthy young MC dogs exhibit myocardial alterations including clusters of mononuclear cells, myocardial fibers with loss of their myofibrillary structures (Figure 2A), degranulated mast cells (Figure 2B), and neutrophils attached to endothelial cells occluding ventricular capillaries (Calderón-Garcidueñas et al., 2001a). UfPM is seen in caveolae of activated endothelial cells (Figure 2C) and in luminal red blood cells in the ventricles of young dogs. The histamine content of human heart mast cells is comparable to that of the lung and skin mast cell pools. Circulating antigens and stimuli (e.g., ET-1) could activate cardiac mast cells to release vasoactive mediators and cause the release of significant amounts of histamine that may give rise to arrhythmias (Genovese and Spadaro, 1997). The major arrhythmogenic effects of histamine are increased sinus rate and ventricular automaticity and decreased atrioventricular conduction (Genovese and Spadaro, 1997).
Alterations in heart rate variability (HRV) in adult populations have been associated with exposure to PM air pollution (Schwartz et al., 2005; Chuang et al., 2005; Vallejo et al., 2006; Wheeler et al., 2006). PM2.5 produced by accelerating traffic seems to play a role in the increased heart rate variability and the increase in the frequency of premature supraventricular beats observed in healthy young patrol officers (Riediker et al., 2004).
Heart rhythm and neurovegetative cardiac status in MC children has been investigated using time-domain analysis of ambulatory electrocardiogram recordings (Calderón-Garcidueñas et al., 2005a). As expected, there was a strong negative age-dependence of the average heart rate and a positive association of age with HRV. When age was a covariable, there was a significant negative association between typical indices of HRV and cumulative PM2.5exposure two days prior to the electrocardiogram recording (Calderón-Garcidueñas et al., 2005a). There was also a significant positive correlation between the frequency of ectopic supraventricular beats and measures of HRV taken between 3 and 4 AM when vagal input to the heart predominates. Taken together, the findings suggested that that exposure to PM2.5 decreased cardiac vagal tone in healthy children.
Central Nervous System Effects
Chronic exposure to urban air pollutants has been associated with evidence of brain inflammation and the generation of reactive oxygen species in healthy MC dogs (Calderón-Garcidueñas et al., 2003a), as indicated by enhanced expression of cyclooxygenase-2 (COX2) and inducible nitric oxide synthase (iNOS) in brain endothelial cells (Figure 2D) and oxidative DNA damage (Calderón-Garcidueñas et al., 2002). The association of exposure to urban air pollution and brain inflammation was confirmed and extended in a cohort of Mexican adults using brain tissues obtained by autopsy of sudden, accidental death victims. COX2 mRNA expression was elevated in frontal cortex and hippocampus (Calderón-Garcidueñas et al., 2004). A subsequent autopsy study focussed upon a younger cohort that included children, adolescents and young adults. We observed upregulated expression of IL-1β and COX2 mRNA expression in the olfactory bulb and frontal cortex of the younger cohort as well, indicating that the association of exposure to urban air pollution with brain inflammation does not depend upon exposures lasting several decades.
Air pollutants could produce adverse effects in the brain through a number of mechanisms. Inflammatory mediators, produced in the respiratory tract as a consequence of chronic pollutant-induced epithelial and endothelial injury, and released into the circulation could activate brain endothelium and cross the blood brain barrier (BBB) (Rivest, 2001). Respiratory tract endothelial and epithelial injury elicits the production and release into the circulation of IL-6, IL-1β, tumor necrosis factor-α (TNF-α) and GM-CSF (Laskin et al., 1998, van Eeden et al., 2001, van Eeden et al., 2005). Brain blood vessels express receptors for TNF-α, IL-1β, and IL-6 (Ericsson et al., 1995; Nadeau and Rivest, 1999). TNF-α and IL-1β can evoke expression of inflammatory mediator genes, such as cyclooxygenase-2 (COX2) (Rivest, 2001) and inducible nitric oxide synthase (iNOS) within brain capillary endothelium (Borgerding and Murphy, 1995; Shafer and Murphy, 1997). Systemic cytokines could also affect the central nervous system (CNS) via sensory nerves such as the vagus. This is a very likely consequence of exposure to air pollutants, because IL-1β is recognized by chemosensory receptors located in vagal paraganglia in the vagus nerve at several levels including cervical, thoracic, and abdominal regions (Elmquist et al., 1997a).
Circulating cytokines can gain access to the brain by being transported across the BBB (Pan and Kastin, 2001; Rivest, 2001; Nguyen et al., 2002). Once across the BBB, they can evoke additional inflammatory mediator expression by vascular-associated microglia (Griffin and Mrak, 2002) and increase the permeability of the BBB (Blamire et al., 2000). The production of nitric oxide can lead to the opening of the BBB (Thiel and Audus, 2001). Thus pollutant-induced systemic inflammation could evoke a brain inflammatory response and open the BBB, compromising its ability to exclude circulating mediators and neurotoxicants.
In fact, vascular lesions are seen in healthy MC adolescents and young adults without risk factors for BBB pathology (L. Calderon-Garciduenas, unpublished observations). Significant leaking of red blood cells and an increment in the number of perivascular macrophages and microglia that express CD163—a scavenger receptor that mediates disposal of haemoglobin-heptaglobin complexes (Kim et al., 2006)—are particularly common in the deep frontal and temporal white matter (Figure 2E).
Alternatively, the carbon core of fine and ultrafine PM and PM-associated chemicals, including LPS, combustion-derived metals, such as vanadium and nickel, and polyaromatic hydrocarbons, might evoke brain inflammation by acting directly on the brain. Controlled exposures of rats to ultrafine particles or metals suggest that the PM or PM components accumulate in the olfactory bulb (Dorman et al., 2002; Oberdorster et al., 2004), and the trigeminal pathway can transport neurotoxins to the brain (Lewis et al., 2005). Thus, fine and ultrafine PM may reach the brain through olfactory receptor neurons and the trigeminal nerves.
Breakdown of the nasal (respiratory and olfactory) barriers in pollution-exposed subjects may contribute to brain inflammation by increasing the access of fine and ultrafine PM (including LPS) to the brain through the olfactory and trigeminal pathways. Ultrafine PM may also be transported to the brain through the systemic circulation (Takenaka et al., 2001) or through phagocytes present in the systemic circulation (Calderón-Garcidueñas et al., 2003a; Oberdorster et al., 2004).
Ultrafine PM can cross red blood cell membranes by non-phagocytic mechanisms (Geiser et al., 2005) suggesting that its spread may not be impeded by the BBB. Indeed, we have evidence of the transport of vanadium and nickel into the olfactory bulbs of MC dogs (Calderón-Garcidueñas et al., 2003a), and of significant levels of vanadium in brain tissues of human MC residents. Having reached the brain, PM components could activate proinflammatory, stress-response or apoptotic signal transduction cascades in the brain by a variety of mechanisms (reviewed in Nel et al., 2006).
Because of its bioactivity, PM-associated LPS may elicit brain inflammation by mechanisms distinct from those employed by other PM components. For example, LPS could act directly on the brain by activating meningeal macrophages and perivascular microglia at the BBB interface, giving rise to the production of cytokines that can activate neighbouring cells, spreading a proinflammatory signal and giving rise to a cascade of inflammatory mediators (Elmquist et al., 1997a). Alternatively, LPS could affect the brain indirectly by activating the vagus nerve in the lung, because LPS is recognized by chemosensory receptors located in vagal paraganglia of the vagus nerve (Elmquist et al., 1997b).
In MC adults and dogs, brain inflammation is accompanied by oxidative DNA damage, neuronal and astrocytic accumulation of the 42 amino acid form of β-amyloid peptide (Aβ42) and the deposition of Aβ42 in diffuse and mature plaques in frontal cortex (Calderón-Garcidueñas et al., 2003a, 2004). Aβ42 is a hydrophobic fragment of β-amyloid, a ubiquitous integral plasma membrane protein having 3 alternatively spliced isoforms that are expressed in a tissue-specific fashion. The accumulation of Aβ42 and its deposition into diffuse and mature neuritic plaques in the brain are hallmarks of AD. A variety of evidence supports the amyloid (or Aβ) hypothesis of AD pathogenesis, which asserts that Aβ42 accumulation is the central pathogenic event in AD (Hardy and Selkoe, 2002, reviewed in Cummings, 2004).
In addition to Aβ42 accumulation AD pathogenesis is characterized by brain inflammation as indicated by activated microglia and reactive astrocytes expressing inflammatory mediators and by increased ROS generation, including nitric oxide. Increased expression of IL-1β in the brain, as we have observed in MC adolescents and adults, is associated with a spectrum of neuroinflammatory processes related to chronic neurodegenerative diseases (Stylianou and Saklatvala, 1998; Ferrari et al., 2004), including AD. Endothelial iNOS expression, as has been seen in MC dogs, likely contributes to the maintenance, self-perpetuation and progression of neurode-generative processes as well (Grammas et al., 1997). Thus the evidence of increased IL-1β, COX2 and iNOS expression and Aβ42 accumulation in MC humans and dogs is consistent with an AD pathogenic process.
The increased circulating ET-1 levels may also play a role in a neurodegenerative process. ET-1 participates in the initiation of gliosis (MacCumber et al., 1990; Yoshimoto et al., 1990), a feature of neurodegeneration (Monnerie et al., 2005; O’Callaghan and Sriram, 2005). Elevated plasma ET-1 levels may penetrate the brain either by transport or leakage across a permeabilized BBB. Focal injections of ET-1 generate focal ischemic white matter lesions in rats (Hughes et al., 2003), and traumatic brain lesions are accompanied by sustained increases in ET-1 levels, which in turn affect brain microcirculation and neural cell function (Steiner et al., 2004).
An additional correspondence between the pathology in MC human residents and AD pathogenesis is the early involvement of the olfactory bulb. There is an accumulation of Aβ42 in endothelial cells, neurons and astrocytes in the olfactory bulb of MC residents (Calderón-Garcidueñas et al., 2004) as early as the second decade (Figure 2F). In neurodegenerative diseases such as AD and Parkinson’s disease olfaction is impaired in the earlier stages and neurodegeneration of the olfactory bulb in AD precedes neurodegeneration of the entorhinal cortex and hippocampus (Braak et al., 1998; Liberini et al., 2000; Kovacs et al., 2001; Tissingh et al., 2001; Hawkes, 2003; Attems et al., 2005).
Diffuse Aβ42 plaques have been observed in the frontal cortex of very young MC dogs (Figure 2G). Similarly Aβ42 accumulation and diffuse and mature Aβ42 plaques have been observed in frontal cortex of MC adolescents and young adults (Figures 2H and 2I respectively). In most cases, the accumulation of Aβ42 occurred in the absence the E4 allele of Apolipoprotein E and a family history of dementia, two well-known risk factors for AD (Selkoe, 2001a, 2001b, 2002). This finding is an indication that exposures to urban air pollutants over several decades are not necessary for the development of the brain pathology described here.
Based on our findings, we suggest that chronic exposure to urban air pollutants including PM-associated LPS may be a risk factor for the development of AD and that neurodegenerative processes may begin in childhood and adolescent years in the context of exposure to urban air pollutants. This issue is of particular relevance to the nine million children residing in MC, because their major clinical and pathological findings include the breakdown of their respiratory epithelial barriers, sustained nasal and bronchiolar inflammation, systemic inflammation, altered systemic expression of LPS-binding proteins, impaired innate and adaptive immune responses and increased plasma concentrations of ET-1. As we have shown, these findings may put them at enhanced risk of developing AD.
Our findings provide a basis for much more extensive epidemiological, forensic, and toxicological studies aimed at identifying the underlying mechanisms of neural damage, and strengthening of the association between chronic exposure to air pollutants, and the risk of developing AD. About half of all U.S. residents live in areas where levels of air pollution are unhealthy and others suffer occupational and indoor exposures to air pollutants. AD is an irreversible, fatal brain disorder that places a burden on the health care system and on caregivers as well. The identification and mitigation of environmental factors that might influence AD pathogenesis is one approach to limiting the future impact of AD.
Footnotes
Acknowledgments
Grant support was provided by 1KO1 NS046410-01A1, 1R21-ES013293-01A1, and US EPA CR829522, National Science Foundation 0346458, and the Montana Board of Research and Commercialization Technology Grant No 04-06, #5 P20 RR015583 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH)