Developmental pesticide exposures and the Parkinson's disease phenotype

Abstract

Whereas Parkinson's disease is a neurodegenerative disorder that typically onsets after 60 years of age, the possibility that it could result from insults sustained during development has been proposed. Experimental evidence based on the combined paraquat + maneb model of the Parkinson's disease (PD) phenotype summarized here provides support for such an assertion. Postnatal exposures of mice to these pesticides led not only to a permanent and selective loss of dopaminergic neurons in the substantia nigra pars compacta but also enhanced the impact of these pesticides administered during adulthood relative to developmental only or adult only treatment. Exposure to maneb alone during gestation resulted in a dramatic response to paraquat in adulthood, including notable reductions in levels of dopamine and metabolites and a loss of nigral dopamine (DA) neurons, despite the fact that paraquat does not share structural similarity to or mechanisms of action with maneb. Collectively, these studies provide developmental environmental models of the PD phenotype. In addition, they demonstrate the fact that silent neurotoxicity produced by developmental insults can be unmasked by challenges later during life as well as the potential for cumulative neurotoxicity over the life span. Birth Defects Research (Part A), 2005. © 2005 Wiley-Liss, Inc.

INTRODUCTION

Parkinson's Disease (PD), a neurodegenerative disorder that manifests with tremor, bradykinesia, rigidity, and akinesia, among other symptoms, results from the loss of dopamine (DA) neurons of the nigrostriatal system of the brain. Consequent to this is a loss of striatal dopamine and its control over motor function. Epidemiological studies have indicated a role for environmental pesticide exposures in the etiology of PD (Rajput et al.,1987; Semchuk et al.,1992; Liou et al.,1997; Gorell et al.,1998; Tanner and Ben-Shlomo,1999). In concert with such evidence, we recently developed a model of PD in young adult C57BL/6 mice based on combined exposure to two pesticides, the herbicide/desiccant paraquat (PQ) and the ethylene bisdithiocarbamate fungicide maneb (MB) (Thiruchelvam et al.,2000a,2000b). PQ bears remarkable structural similarity to 1-methyl-4-phenyl-1,2,5,6-tetrahydropyridine (MPTP), the most widely used experimental model of PD (Langston et al.,1984). Both PQ and MB have been shown to adversely impact brain dopamine systems (Miller et al.,1991; Liou et al.,1996; Walters et al.,1999; Manning-Bog et al.,2002), although attempts to mimic PD via the use of PQ in experimental models have not been particularly successful. Human exposure to environmental contaminants, of course, is to mixtures rather than to individual pesticides. The rationale for the combined exposure to PQ+MB in our studies was the premise that concurrent insults to the dopamine system at different target sites, as would be expected from two pesticides with different modes of action, would preclude operations of homeostatic defense mechanisms and thereby increase vulnerability of the system.

In this model, six-week-old mice were exposed two times per week for six weeks to either saline, 10 mg/kg PQ, 30 mg/kg MB, or the combination of PQ+MB (Thiruchelvam et al.,2000b). Measurements carried out five to seven days after treatment revealed PQ+MB-induced reductions in striatal levels of DA and its metabolite DOPAC, and decreased levels of tyrosine hydroxylase (TH; catecholaminergic marker) and DA transporter immunoreactivity and TH protein levels only in response to PQ+MB. These effects were observed in the nigrostriatal DA system, and not in the adjacent mesocorticolimbic DA system. Correspondingly, decreases in TH immunoreactivity and cell counts were found only in response to PQ+MB, and in the substantia nigra, site of DA cell bodies of the nigrostriatal DA system, but not in the ventral tegmental area, the region containing DA cell bodies of the mesocorticolimbic DA system.

Subsequent characterization of the PQ+MB model has confirmed that it produces a selective nigrostriatal dopaminergic neurotoxicity, including dopaminergic cell loss, that appears to be both permanent and progressive (Thiruchelvam et al.,2000b,2002). The PD phenotype associated with PQ+MB, moreover, is substantially enhanced by aging (Thiruchelvam et al.,2003), with age-related increases in vulnerability in mice 6 weeks of age, 5 months of age, or 18 months of age at the beginning of pesticide treatment. Consistent with epidemiological findings (Van Den Eeden et al.,2003), males appear more vulnerable to the dopaminergic neurotoxicity produced by PQ+MB than do females (e.g., see Barlow et al.,2004). Alterations in genetic background can also enhance the PD phenotype. Specifically, mice expressing, in catecholaminergic cells, both mutations of human alpha-synuclein that are associated with the early onset familial form of Parkinson's disease show markedly enhanced neuronal cell loss and associated DA depletion than do either mice expressing normal human alpha-synuclein or nontransgenic littermates (Thiruchelvam et al.,2004).

Although typically having its onset after the age of 60, the possibility that PD could arise from insults sustained during development that only manifest later in life has been suggested (Calne and Langston,1983). To date, evidence to support such an assertion has been limited. There are several possible configurations by which developmental insults could serve as an etiologic basis for adult onset PD, some of which are depicted in Figure 1. These schema are based on the premise that a normal loss of dopamine neurons occurs over the life span (A). Environmental insults sustained in adulthood superimposed on this normal aging-related loss could further reduce the number of DA neurons to an even lower level, such that the loss eventually reaches a level associated with PD signs and symptoms, i.e., approximately an 80% loss (A′). An insult sustained developmentally (B) could accelerate the normal age-related decline in DA number, thus resulting in an earlier age at which the PD symptomatic stage is reached, and such an effect could then be further enhanced by an associated insult during adulthood (B′). Another model is one in which an insult occurring during development reduces the initial number of DA neurons (C) and thereby accelerates the time to symptoms, an effect that could be further augmented by adult insults superimposed on the developmental loss (C′).

Figure 1. Schematic representation of three different potential scenarios leading to the Parkinson's disease phenotype. Scenario A depicts normal aging-related loss of DA neurons. An adult insult superimposed on this process (A′) produces a further reduction in numbers of DA neurons and thereby results in a loss sufficient to enter into the symptomatic PD range. In scenario B, a developmental insult accelerates the normal aging-related loss, leading to a PD phenotype at an earlier age. When an environmental insult is superimposed on this process (B′), it further accelerates the onset of a PD phenotype. A third potential scenario (C) presumes a developmental insult that effectively reduces the numbers of DA neurons at birth from which normal aging-related loss then occurs. Combined with an adult insult (C′), this scenario would have the potential for an even earlier onset PD phenotype. (Modified from Thiruchelvam et al.,2002).

To determine whether pesticide exposures sustained developmentally would also lead to a PD phenotype, consistent with a hypothesized developmental etiology, the PQ+MB model initially implemented in young adult mice was extended to exposures occurring early in development. Studies carried out to date to assess this possibility have been highly supportive of this hypothesis. Indeed, several significant findings have emerged from these studies. First, in concert with the proposal that a developmental insult could lead to adverse effects later in life, postnatal exposures to PQ+MB (0.3 mg/kg PQ + 1.0 mg/kg MB intraperitoneally (i.p.) from days 5 to 19 of age) were found to produce a permanent nigrostriatal DA neurotoxicity in mice that was also progressive. Additionally, mice exposed developmentally showed enhanced vulnerability to the effects of PQ+MB administered again later in life (10 mg/kg PQ + 30 mg/kg MB i.p. twice a week for six weeks) as compared to postnatal-only exposure or to adult only exposure. Furthermore, whereas postnatal exposures to PQ or MB alone produce more modest or even no effects relative to PQ+MB, they nevertheless appear to be associated with silent toxicity that is only unmasked by subsequent pesticide challenges (Thiruchelvam et al.,2002). Finally, recent experiments now indicate vulnerability results from gestational exposures as well, and that pesticide insults can result in cumulative neurotoxicity across the life span.

Figure 2 depicts loss of dopamine neurons in the substantia nigra pars compacta in response to postnatal, adult, or postnatal + adult treatments. As it shows, adult only exposure (Adult) resulted in the most modest TH⁺ neuron loss, with PQ alone and PQ+MB resulting in small but significant losses in TH⁺ neurons. Postnatal-only exposure (Developmental; days 5–19) to PQ and MB alone resulted in small but significant losses in TH⁺ neurons, an effect that was significantly enhanced with combined PQ+MB, in which losses on the order of 36% relative to saline control were found. When postnatal exposure was followed by adult rechallenge with the same pesticides (Developmental + Adult), however, it can be seen that far more pronounced losses were seen, both for each of the pesticides administered alone, as well as for PQ+MB, in which the reduction relative to control now averaged approximately 67%.

Figure 2. Total numbers of TH⁺ neurons in the substantia nigra pars compacta in groups exposed: 1) from days 5 to 19 of age (Developmental) to saline, 0.3 mg/kg PQ, 1.0 mg/kg MB, or the combination and challenged with saline as adults; 2) at 6.5 months of age (Adult) to saline, 10 mg/kg PQ, 30 mg/kg MB, or the combination, or 3) developmentally to saline, 0.3 mg/kg PQ, 1.0 mg/kg MB, or the combination and challenged with saline, 10 mg/kg PQ, 30 mg/kg MB, or the combination in concordance with developmental exposures as adults (Development + Adult). TH⁺ neurons were measured using unbiased stereology two weeks after the last adult treatment (n = 4 for each treatment group) (Thiruchelvam et al.,2002). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]

In an attempt to begin to understand associated mechanisms of the nigrostriatal DA neurotoxicity associated with this pesticide model, gene array analysis of striatal sections from brain was carried out using four mice from each of the different groups exposed to PQ+MB and normalized to saline controls. Of interest, the outcome of this analysis reveals a similar pattern of vulnerability in relation to developmental periods of exposure, as illustrated using functional classification for changes in DA and glutamate systems in Figure 3.

Figure 3. Changes in gene expression in striatum for genes related to dopamine (left) and glutamate (right) systems in adults following adult only exposure to 10 mg/kg PQ + 30 mg/kg MB (A), postnatal exposure to 0.3 mg/kg PQ + 1.0 mg/kg MB and challenge with saline as adults (PN), postnatal exposure to 0.3 mg/kg PQ + 1.0 mg/kg MB followed by adult rechallenge with 10 mg/kg PQ and 30 mg/kg MB (PN+A) as normalized to saline controls.

In this example, the magnitude of the upregulation of changes in D2-receptor like family genes can be seen to be similar to the patterns of TH⁺ neuron loss, with the most pronounced increases in D2-like receptor family genes occurring in mice postnatally exposed to PQ+MB and then rechallenged as adults with PQ+MB (PN+A). Postnatal-only exposure was likewise associated with upregulation of genes associated with this family of receptors, but to a slightly less extent than that seen with postnatal exposure followed by adult rechallenge. The smallest loss of TH⁺ neurons produced by PQ+MB occurred in response to adult-only exposures, and, correspondingly, the increase in expression of D2-like receptor family genes was also of smallest magnitude with adult-only PQ+MB exposure (A) relative to the developmental only, and the developmental followed by adult treatments. Similar patterns of vulnerability were seen in genes related to glutamatergic system function, including those related to NMDA1, NDMA2, and NMDA3 receptors, AMPA1, AMPA2, and AMPA4 receptors, and Kainate 2 and 5 receptors. Thus, the patterns of vulnerability as reflected in TH⁺ neuron loss showed correspondences to the patterns of altered gene expression, at least as examined currently at this level of analysis.

Further efforts have focused on questions of the impact of gestational exposure and whether cumulative neurotoxicity could occur. Indeed, a recent study by our group demonstrated that gestational exposure (days 10 through 17) of mice to MB (1.0 mg/kg via the dam) resulted in marked vulnerability to challenges with PQ (10 mg/kg/day for eight days) in adulthood for reasons that are as yet unclear, since there is no structural or simple mechanistic relationship between PQ and MB (Barlow et al.,2004), only a commonality of target (the nigrostriatal DA system). This was reflected as a reduction in TH⁺ neuron loss (approximately 30%), without any changes in TH⁻ neurons, indicating the specificity of the effect. Furthermore, in concert with effects noted with postnatal exposures, females were largely spared from these effects. Clearly, the sequence of exposures is critical, since prenatal exposures to PQ followed by adult challenges with MB fail to produce any TH⁺ neuron loss (Barlow et al.,2004).

Collectively these findings indicate a broad period of vulnerability to pesticides during development. Furthermore, experiments with both postnatal and gestational exposures indicate that resulting neurotoxicity can be cumulative over the life span, including the unmasking of toxicity that was silent prior to subsequent challenges. Such neurotoxicity, moreover, may occur with nonrelated pesticides, indicative of unpredictable patterns, findings that pose particular challenges for human health risk assessment. Coupled with specific hypotheses and time-course data, gene array analyses may provide insights into particular pathways and sites of vulnerability of the developing nervous system to pesticide exposures.