Influences of dopaminergic lesion on epidermal growth factor-ErbB signals in Parkinson's disease and its model: neurotrophic implication in nigrostriatal neurons


Address correspondence and reprint requests to Hiroyuki Nawa, Department of Molecular Biology, Brain Research Institute, Niigata University, Asahimachi-dori 1–757, Niigata 951–8585, Japan.


Epidermal growth factor (EGF) is a member of a structurally related family containing heparin-binding EGF-like growth factor (HB-EGF) and transforming growth factor alpha (TGFα) that exerts neurotrophic activity on midbrain dopaminergic neurons. To examine neurotrophic abnormality in Parkinson's disease (PD), we measured the protein content of EGF, TGFα, and HB-EGF in post-mortem brains of patients with Parkinson's disease and age-matched control subjects. Protein levels of EGF and tyrosine hydroxylase were decreased in the prefrontal cortex and the striatum of patients. In contrast, HB-EGF and TGFα levels were not significantly altered in either region. The expression of EGF receptors (ErbB1 and ErbB2, but not ErbB3 or ErbB4) was down-regulated significantly in the same forebrain regions. The same phenomenon was mimicked in rats by dopaminergic lesions induced by nigral 6-hydroxydopamine infusion. EGF and ErbB1 levels in the striatum of the PD model were markedly reduced on the lesioned side, compared with the control hemisphere. Subchronic supplement of EGF in the striatum of the PD model locally prevented the dopaminergic neurodegeration as measured by tyrosine hydroxylase immunoreactivity. These findings suggest that the neurotrophic activity of EGF is maintained by afferent signals of midbrain dopaminergic neurons and is impaired in patients with Parkinson's disease.

Abbreviations used

epidermal growth factor


enzyme immunoassay


heparin-binding epidermal growth factor


neuron-specific enolase




Parkinson's disease


post-mortem interval


sodium dodecyl sulfate


substantia nigra


transforming growth factor alpha


tyrosine hydroxylase

Parkinson's disease (PD) is a neurodegenerative disease, caused by progressive degeneration of nigrostriatal dopaminergic neurons. Genetic linkage studies have identified the genes responsible for familial types of PD and hint at the molecular mechanism of this illness (Mayeux 2003; Warner and Schapira 2003 for review). More than 95% cases of PD, however, are sporadic and the etiology is still poorly understood. Dopaminergic neurons in the substantia nigra (SN) appear to receive neurotrophic signals from their main target region of the striatum (Unsicker 1994; Olson 1997; Kholodilov et al. 2004). Surgical lesions of the nigrostriatal pathway result in degeneration of the dopaminergic neurons, but the administration of neurotrophic factors rescues them from cell death or enhances their regeneration (Ventrella 1993; Beck et al. 1995; Tseng et al. 1997; Volpe et al. 1998). What is the molecular nature of the trophic support for the SN dopaminergic neurons? In vitro culture studies indicate that a variety of neurotrophic substances support cell survival and growth of SN dopaminergic neurons. Such neurotrophic factors include brain-derived neurotrophic factor, fibroblast growth factors, IL-1, glial cell-derived neurotrophic factor, epidermal growth factor (EGF), heparin-binding EGF (HB-EGF), and transforming growth factor alpha (TGFα; Casper et al. 1991; Hyman et al. 1991; Alexi and Hefti 1993; Beck et al. 1993; Lin et al. 1993; Fawcett et al. 1995; Ho and Blum 1997; Takei et al. 1998; Hanke et al. 2004). These protein factors enhance the survival of mesencephalic dopaminergic neurons in culture as monitored by tyrosine hydroxylase immunoreactivity or by dopamine uptake. Because some neurotrophic factors can act on glial cells to stimulate the production of other factors, their trophic actions on dopaminergic neurons might be mediated by co-cultured glial cells (Casper et al. 1994; Ho and Blum 1997; Hanke et al. 2004). Whether their neurotrophic actions on SN dopaminergic neurons are direct or indirect, previous studies indicate that neurotrophic signals for SN dopaminergic neurons are abnormal in PD patients as well as in an animal model of PD (Mandel et al. 2000; Mogi et al. 2001). To test the therapeutic application of trophic factors produced with recombinant techniques, the neurotrophic activity of these factors in vivo was evaluated in animal PD models that were produced by dopaminergic neurotoxins, 6-hydroxydopamine (6-OHDA) or 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (Date et al. 1993; Frim et al. 1994; Rosenblad et al. 1998; Yasuhara et al. 2004). To supply the dopaminergic neurons with such neurotrophic activities, a variety of techniques including pump implantation, gene delivery, and cell transplantation have been attempted in the animal PD model as well as in PD patients (Gill et al. 2003).

EGF and structurally related peptides such as TGFα, HB-EGF, amphiregulin, betacellulin, and epiregulin, all interact with EGF receptors of erbB1 gene products and enhance survival and process outgrowth of brain neurons, and influence synaptic plasticity in the CNS (Morrison et al. 1987; Ferrari et al. 1991; Ishiyama et al. 1991; Reynolds et al. 1992). In agreement, ErbB1 receptors are widely present in both the developing and mature CNS (Gomez-Pinilla et al. 1988; Werner et al. 1988; Misumi and Kawano 1998). Molecular studies in knockout mice revealed that gene targeting for TGFα and GDNF, but not brain-derived neurotrophic factor, influences the number of midbrain dopaminergic neurons (Jones et al. 1994; Moore et al. 1996; Blum 1998; Dluzen et al. 1999), suggesting a redundancy of trophic support for this neuronal population (Kholodilov et al. 2004). The findings in TGFα knockout mice indicate the biological importance of TGFα/ErbB1 signals in dopaminergic development (Xian and Zhou 2000). Peripheral EGF administration also up-regulates the expression of tyrosine hydroxylase in postnatal rats and elevates dopamine turnover in the striatum (Futamura et al. 2003). Additional findings in transgenic mice overexpressing TGFα, as well as in ErbB1 knockout mice, have further strengthened the argument that ErbB1 ligands act on monoaminergic neurons and neocortical neurons in vivo (Hilakivi-Clarke and Goldberg 1993; Cirulli and Alleva 1994; Sibilia et al. 1998). In spite of the accumulating evidence that ErbB1 receptor ligands exert trophic activity on dopaminergic neurons, their potential contribution to PD pathology has not been characterized.

In the present study, we quantified proteins of EGF and its related peptides in post-mortem PD brains and a rat PD model. Protein levels of EGF receptor subunits (ErbB1–4) were measured in parallel. Causative contribution of the EGF/ErbB1 change to the dopaminergic pathology was assessed in the PD model. Specific changes in EGF and ErbB1/2 levels are discussed in view of their neurotrophic role in the dopaminergic neurodegeneration of PD.

Materials and methods

Human tissue sampling

Brain tissue was collected from post-mortem samples from nine patients with chronic PD (four men, five women) and from 10 controls (six men, four women) that had no pathologic signs of neurodegeneration in the brain (Tables 1 and 2). All the PD patients had taken dopamine agonists, dopamine precursors, or dopamine uptake inhibitors (Table 1). In each case, the left cerebral hemisphere was fixed in formalin solution for diagnostic confirmation and the right hemisphere was frozen at −80°C. All the pathologic analyses were performed at the author's institute. The post-mortem diagnosis of PD corresponded to more than two of the Hoehn and Yahr (H & Y) scores (Ginanneschi et al. 1988). We used only post-mortem PD samples that displayed typical neurodegeneration in the SN. Tissue samples were taken from the prefrontal cortex (approximately Brodman's area, BA9) and the putamen; the areas were identified in frozen coronal slices according to a published atlas of the human brain (Nieuwenhuys et al. 1988). The families of the control and PD subjects provided written informed consent to allow us to use the brain tissues of these patients as material for pathologic investigations.

Table 1.  Autopsy and clinical dataa
No.SexAge at death (y)Cause of deathH & Y scaleDuration of illnessDrugsFreezer time (m)PMI (h)
  1. M, male; F, female; DA, dopamine agonists; DP, dopamine precursors; DR, dopamine releasers; AC, anticholinergic agents; DIC, disseminated intravascular coagulation.

Parkinson's disease
P-1F74Multiple organ failureIV6DP, DR453
P-2F60Pulmonary embolismII2.5AC, DP751.5
P-3F68PneumoniaIII9AC, DP874
P-4M84AsphyxiaIII7.3DP, DA611.5
P-5M65SymptosisIII8AC, DP2213
P-6F85PneumoniaV16DP, DA143.7
P-7M70Pulmonary candidiasisV7DP, DA263.5
P-8M79Heart failureIII10DP463
P-9F85SymptosisIV8AC, DP + DA + DR963
C-1M82Gastric cancer   446
C-2F64Sepsis   452
C-3M75Intestinal bleeding   493
C-4F83Natural causes   505.7
C-5M62Renal failure   539
C-6M71Pneumonia   693.7
C-7M68DIC   783
C-8F83Multiple organ failure   814.7
C-9F49Idiopathic portal hypertension   813
C-10M72Hemorrhagic pancreatitis   953
Table 2.  Profiles of patients and normal control subjects for the brain samples
  1. Data indicate the mean ± SD of age, PMI and FT. M, male; F, female; PMI, post-mortem interval; FT, freezer time.

n10 (M = 6, F = 4)9 (M = 4, F = 5)
Age (years)70.9 ± 10.274.4 ± 8.8
PMI (h)4.3 ± 2.02.9 ± 0.8
FT (months)64.5 ± 17.574.6 ± 57.7

Dopaminergic lesions

Male Wistar rats (8 weeks old, initial weight 220–270 g) were housed under a 12-h light/dark cycle with free access to food and water. Intra-nigral infusion of 6-OHDA was performed under aseptic conditions and pentobarbital (45 mg/kg) anesthesia. A cannula (24 gauge) was inserted unilaterally into the SN (4.8 mm posterior and 2.2 mm right lateral measured from the bregma, 8.7 mm below the skull) and 2 µL of 6-OHDA (4 mg/mL) or saline was administered over 2 min. After 1 week, neurotoxic depletion of the dopamine neurons was confirmed by immunoblotting for tyrosine hydroxylase (see below). The animals were killed by inhalation of carbon dioxide gas, then the brains were harvested rapidly from decapitated rats, and each brain region was dissected on ice. All of the animal experiments were performed in accordance with the Animal Use and Care Committee guidelines of Niigata University.

Subchronic EGF administration

Male rats (8 weeks old) were subjected to pump implantation together with 6-OHDA- or saline-injection to the SN. A cannula (30 gauge; Terumo, Tokyo, Japan) was implanted simultaneously in the striatum (0.5 mm anterior and 3.0 mm right lateral measured from the bregma, 5.5 mm below the skull), glued to the skull and connected to an Azlet osmotic minipump (model 2002; Azla Corp., Palo Alto, CA, USA) by medical grade vinyl tubing. Pumps were filled with human recombinant EGF (0.3 mg/mL; Higeta Syoyu, Chiba, Japan) or 0.9% saline and implanted subcutaneously in the nape of the neck. The scalp incision was closed with surgical staples and treated with a topical antiseptic. In parallel with the acute injection of 6-OHDA or saline in the ipsilateral side of the SN, rats were infused continuously at a rate of 0.5 µL/h with either saline or EGF for 7 days.

Enzyme immunoassay

For determination of growth factor concentrations, brain tissues were homogenized in 10 volumes of homogenization buffer (phosphate-buffered saline) containing protease inhibitors (aprotinin, 200 kallikrein units/mL), 0.1 mm phenylmethylsulfonyl fluoride, 0.1 mm benzethonium chloride, 1 mm benzamidine, and 1 mm ethylenediaminetetraacetic acid (all from Sigma Chemical Co., St Louis, MO, USA). Brain homogenates were centrifuged at 14 000 g for 30 min at 4°C, and supernatants containing soluble proteins were stored at −80°C until use. The protein concentrations in the samples were determined using a Micro BCA kit (Pierce, Rockford, IL, USA) with bovine serum albumin as a standard.

Growth factor levels in the soluble fraction were measured by sandwich enzyme immunoassay (EIA), as described previously (Nawa et al. 1995; Futamura et al. 2002). Briefly, tissue extracts (100 µL of 0.5–1.5 mg/mL protein; in duplicate) or growth factor standards (0.1–30 pg; in triplicate) were loaded into wells of EIA titer plates that had been coated (30–80 ng/well) with anti-human EGF (Ab3, Oncogene, Cambridge, MA, USA), anti-rat EGF (Sigma), anti-HB-EGF (R & D Systems, Minneapolis, MN, USA), or anti-TGFα (Ab4, Oncogene) antibodies as a primary antibody. Biotinylated anti-HB-EGF (R & D Systems) and anti-TGFα (R & D Systems) antibodies (secondary antibody, 3–14 ng/mL) were added to the wells and allowed to bind to HB-EGF and TGFα, respectively. The biotinylated secondary antibodies were detected using streptavidin β-galactosidase (1 : 10000, Sigma). Alternatively, the secondary antibody for EGF (R & D Systems for human or Santa Cruz Biotechnology, Santa Cruz, CA, USA, for rat) was applied and detected using antimouse IgG β-galactosidase (1 : 1000, Sigma). The β-galactosidase activity retained in each well was measured by incubation with 200 μm 4-methylumbelliferyl-d-galactosidase (MUG, Sigma). The amount of resulting fluorescent product was monitored using a Fluorolite-1000 fluorometer (Dynatech Laboratories, Alexandria, VA, USA) with excitation at 364 nm and emission at 448 nm. An average of two measurements per sample was standardized by the protein concentration. The EIA specificity for each ligand was confirmed by assaying other ErbB1 ligands; the cross-reactivity of all EIAs was less than 1/1000 (data not shown).

Immunoblot analysis

Protein levels of ErbB family members, tyrosine hydroxylase, and neuron-specific enolase (NSE) were determined by immunoblot analysis using methods similar to those described previously (Iwakura et al. 2001). Protein samples for immunoblotting were prepared from the same tissue extracts as for the two-site EIAs by denaturing them with 2% sodium dodecyl sulfate (SDS), separated by SDS–PAGE (polyacrylamide gel electrophoresis), and transferred to a nitrocellulose membrane (Schleicher and Schull, Dassel, Germany) by electrophoresis. The membrane was probed with anti-ErbB1 receptor antibody (1 : 1000, Santa Cruz Biotechnology), antiphospho-ErbB1 receptor antibody (1 : 500, Santa Cruz Biotechnology), anti-ErbB2 antibody (1 : 1000, Santa Cruz Biotechnology), anti-ErbB-3 antibody (1 : 1000, Santa Cruz Biotechnology), anti-ErbB-4 antibody (1 : 1000, Santa Cruz Biotechnology), antityrosine hydroxylase (TH) monoclonal antibody (1: 2000, Chemicon, Temecula, CA, USA) or anti-neuron-specific enolase (NSE) antibody (1: 2000, Chemicon). After extensive washing, the immunoreactivity on the membrane was detected with antirabbit/mouse immunoglobulin conjugated to horseradish peroxidase, followed by a chemiluminescence reaction (ECL kit, Amersham Biosciences, Piscataway, NJ, USA). Densitometric quantitation of band intensities was performed.

Statistical analysis

Given the non-Gaussian distribution of the data values, comparisons between experimental/PD group and control were made by a non-parametric analysis of Mann–Whitney U-test. The Spearman rank correlation was used for correlation of relative protein levels. A probability level of less than 0.05 was considered to be statistically significant.


Post-mortem samples of the brain regions were obtained randomly from nine chronic PD patients and from 10 controls who had no sign of neurodegeneration in the brain (Table 1). There was no significant difference in the sample indices of post-mortem interval, freezer time, or age of subjects between the two groups (Table 2). We determined the content of the three neurotrophic factors in the EGF family (EGF, TGFα, and HB-EGF) in two brain regions, prefrontal cortex and putamen, which are major targets of the midbrain dopaminergic neurons. The two-site EIA revealed marked alteration in EGF levels in the PD group: EGF protein levels were significantly decreased to 64% in the frontal cortex (U = 8.5, p = 0.002) and 60% in the putamen (U = 14, p = 0.008), compared with controls (Fig. 1). There was no difference in either region in the other members of the EGF family (HB-EGF and TGFα) between groups. Correlation analysis between EGF and several sampling indices suggested that post-mortem interval (PMI), storage time, and age were not co-factors in any of the brain regions of either PD patients (Spearman's test: p = 0.860–0.887 for PMI, p = 0.540–0.786 for storage time, p = 0.209–0.583 for age) or control subjects (p = 0.080–0.322 for PMI, p = 0.083–0.096 for storage time, p = 0.151–0.887 for age).

Figure 1.

Protein levels for members of the EGF family in the brains of patients with Parkinson's disease and control subjects. Scatter plots displaying the variability and differences in soluble protein levels of EGF, TGFα, and HB-EGF in the striatum (a) and the prefrontal cortex (b) of PD patients (n = 9) and control subjects (n = 10). Significant differences from controls are indicated as follows: **p < 0.01, Mann–Whitney U-test. bsl00066, patients with Parkinson's disease (PD); ○, controls. Horizontal lines represent mean values. Note that protein in the cytoplasmic fractions was subjected to EIA.

There is considerable feedback regulation between the ligand stimulation and expression of ErbB1 receptors. EGF stimulation down-regulates the expression of the ErbB1 receptors in the brain, thereby further desensitizing EGF signaling (Canesi et al. 2000). EGF binds to the single ErbB1 subunit and triggers its homomeric or heteromeric dimerization with another EGF receptor subunit such as ErbB2 (Neu), ErbB3, or ErbB4 (Ullrich and Schlessinger 1990; Earp et al. 1995; Walker 1998). To evaluate the influence of PD on all of the ErbB family proteins, the post-mortem samples were subjected to immunoblotting for ErbB1–4 receptors (Figs 2 and 3). The expression of ErbB1 protein was detected as a 170-kDa band, which is consistent with the apparent molecular size of non-truncated EGF receptors (Ullrich et al. 1984). ErbB1 immunoreactivity was significantly lower in the striatum (Fig. 2) and prefrontal cortex (Fig. 3) of PD patients compared with control subjects (51% decrease, U = 0, p < 0.001 for the striatum and 30% decrease, U = 9, p = 0.009 for the prefrontal cortex). There was a parallel decrease in ErbB2 immunoreactivity in the striatum and prefrontal cortex of the PD patients (41% decrease, U = 8, p = 0.003 for the striatum and 30% decrease, U = 13, p = 0.027 for the prefrontal cortex). Immunoreactivity for the other EGF receptors, ErbB3 and ErbB4, was not different between groups. A striatal decrease in tyrosine hydroxylase levels confirmed the authenticity of the PD samples (Fig. 2).

Figure 2.

Expression of ErbB receptors in the striatum of patients with Parkinson's disease. Immunoblot analysis in the brain homogenates from the striatum in the same PD patients and controls. The protein was separated by SDS–PAGE on an 7.5% gel, transferred to a nitrocellulose membrane, and incubated with anti-ErbB1, anti-ErbB2, anti-ErbB3, and anti-ErbB4 antibodies, as well as with anti-NSE antibody as an internal control. Immunoreactivity for tyrosine hydroxylase was also measured to confirm the dopaminergic neurodegeneration in patients: Tyrosine hydroxylase (TH) levels in the striatum were reduced to 7.3 ± 1.6% of the mean control level while NSE levels were 71 ± 18% of the control levels. Immunoblotting for ErbB1–4 receptors and densitometric measurement data are shown as means ± SEM *Significantly different from controls (*p < 0.05, **p < 0.01 and ***p < 0.001, Mann–Whitney U-test). The immunoreactive bands are re-arranged from left to right in order of their intensity.

Figure 3.

Expression of ErbB receptors in the prefrontal cortex of patients with Parkinson's disease. Brain homogenates from the prefrontal cortex of the same patients with PD and controls were similarly subjected to immunoblot analysis (see Fig. 2 legend). Immunoblotting for ErbB1–4 receptors and densitometric measurement data are shown as mean ± SEM. *Significantly different from controls (*p < 0.05 and **p < 0.01, Mann–Whitney U-test). The immunoreactive bands are re-arranged from left to right in order of their intensity.

Neurotrophic interaction between a target and innervating neurons is often regulated mutually by neurotransmitters as well as via neurotrophic factors (Nawa and Takei 2001; Poo 2001 for review): Neurons transport neurotransmitters anterogradely toward the target tissue. In turn, the target tissue synthesizes and releases neurotrophic molecules that are transported retrogradely toward the presynaptic neurons. To examine the influence of dopaminergic lesions on the production of these ErbB1 receptor ligands in the target brain regions, protein levels of EGF, TGFα, and HB-EGF were determined in an animal model of PD. The PD model was prepared by administering a neurotoxin, 6-OHDA, into SN in one hemisphere. One week after an injection of 6-OHDA, levels of tyrosine hydroxylase, a rate-limiting enzyme for dopamine synthesis, decreased to less than 8.5%, confirming the completion of the chemical lesion of the nigrastriatal dopaminergic neurons (Fig. 4). In parallel, there was a marked reduction in EGF levels on the lesioned side of the striatum compared with its control hemisphere (U = 0, p < 0.001). In contrast, there was no difference in protein levels of other members of the EGF family, HB-EGF and TGFα, between hemispheres. The present animal study suggests that dopaminergic deafferentation play a primal role in decreasing the local EGF production in the striatum.

Figure 4.

 Selective reduction in EGF levels in the 6-OHDA-lesioned hemisphere of rats. Effects of dopaminergic lesion on the EGF family peptides were estimated in rats treated with 6-OHDA. 6-OHDA was administered into the SN in the right hemisphere. Soluble protein levels of EGF, TGFα, and HB-EGF in the striatum were measured in the right hemisphere (□; n = 10) and in the left hemisphere as controls (▪; n = 10) and are shown as mean ± SEM. Protein expression of tyrosine hydroxylase was monitored and significantly decreased in the right hemisphere (to 8.5 ± 2.3%), compared to the left hemisphere. Significantly different from controls (**p < 0.01, Mann–Whitney U-test).

To estimate the expression of the ErbB receptor family in the PD animal model, immunoreactivity for ErbB receptors was also determined in the striatum as well as in the SN of the 6-OHDA-treated rats (Fig. 5). Tyrosine hydroxylase levels were markedly reduced in the striatum of the 6-OHDA-treated hemisphere (6.9 ± 2.3%; Fig. 5b). ErbB1 immunoreactivity in the striatum was reduced in the lesioned hemisphere than in the control hemisphere (U = 0, p = 0.009; Fig. 5a). There was a similar decrease in ErbB1 levels in the SN (U = 3, p = 0.049; Fig. 5c). In parallel, phosphorylation levels of the ErbB1 receptors were also decreased in the striatum and SN compared with their control hemispheres (68 ± 5.6% in the striatum, U = 2, p = 0.028; 79 ± 4.2% in the SN, U = 3, p = 0.047), suggesting a decrease in trophic signals for dopamine neurons. Dopaminergic lesion also decreased ErbB2 levels in the SN (U = 3, p = 0.047; Fig. 5d). The parallel decrease in ErbB1 and ErbB2 after dopaminergic lesion is likely to indicate biological importance of EGF/ErbB signaling in the nigrostriatal dopaminergic neurons. In contrast, there was no significant alteration in ErbB3 and ErbB4 receptors, however (U = 12, p = 0.34 for ErbB3, and U = 16, p = 0.75 for ErbB4).

Figure 5.

A specific decrease in EGF receptor (ErbB1 and ErbB2) levels in 6-OHDA-treated rats. Protein levels of ErbB1 receptors (a) and tyrosine hydroxylase (b) were measured in both hemispheres of the striatum (R+, right; L−, left) by immunoblot analysis in 6-OHDA-treated or vehicle-treated rats (n = 5 each). ErbB1 (c) and ErbB2 (d) levels in the SN were also monitored in the same sets of rats (n = 5 each). The dopaminergic lesion was performed unilaterally in the right hemisphere of the brain as described above. A typical immunoblot (n = 2 each) is shown for display. Significantly different from controls (*p < 0.05, **p < 0.01 ***p < 0.001, Mann–Whitney U-test). The same samples were also subjected to immunoblotting for ErbB3 and ErbB4 receptors. Data for ErbB3 and ErbB4 are presented in the text.

Does the EGF decrease following 6-OHDA treatment contribute to the progression of dopaminergic neurodegeneration in the above model? Although ErbB1 receptor ligands support cell survival and growth of cultured SN dopaminergic neurons (Casper et al. 1991, 1994; Ho and Blum 1997; Hanke et al. 2004), the chronic effects of these ligands on dopaminergic neurons have not been fully characterized in vivo. To address this question, we attempted to rescue dopaminergic neurons from 6-OHDA-triggered neurodegeneration by supplying EGF to the striatum. In parallel with the 6-OHDA- or saline-injection to the SN, we administered EGF (3.6 µg/d) or saline to the striatum using a mini-osmotic pump (Fig. 6). Although 6-OHDA treatment alone down-regulated ErbB1 phosphorylation, intrastriatal EGF administration reversibly increased the phosphorylation to control levels (Fig. 6a). Immunoblotting for tyrosine hydroxylase indicated that the EGF treatment almost fully inhibited formation of the dopaminergic lesion surrounding the cannula route in the striatum, but saline administration did not (Fig. 6c). In contrast, there was no apparent EGF effect on neuronal degeneration in the striatum as monitored by immunoblot for the neuronal marker, NSE (Fig. 6b). In the 6-OHDA rats, there was an overall reduction in tyrosine hydroxylase levels in the SN region in which the cell bodies of the dopaminergic neurons are located (Fig. 6d). These results suggest that the striatal EGF decrease in the PD model is involved in the initiation or progression of neurotoxin-induced dopaminergic lesions.

Figure 6.

Neurotrophic effects of striatal EGF administration in 6-OHDA-treated rats. EGF (150 ng/h; egf) or saline was administered into the ipsilateral hemisphere of the striatum of SN-saline- or SN-6-OHDA-treated rats for 7 days from an osmotic minipump. Striatal tissue sounding the cannula (5 mm diameter, right hemisphere) and the saline- or 6-OHDA-injected SN tissue (right hemisphere) were obtained from rats and subjected to immunoblotting. Protein levels of phosphorylated ErbB1 receptors (a) and NSE (b) were measured in 6-OHDA/egf-treated, 6-OHDA/saline-treated or Saline/saline-treated rats (n = 5 each). Tyrosine hydroxylase levels in the striatal (c) and SN tissues (d) were also monitored in the same sets of rats (n = 5 each). A typical immunoblot (n = 2 each) is shown for display. Protein levels in the striatum and SN of unoperated rats were set as controls (100%) (data not shown). Marks with lower letters (egf, saline) indicate striatal pump infusion and those with capital letter(s) (6-OHDA, Saline) represent an injection to the SN. Significantly different from controls (**p < 0.01, Mann–Whitney U-test).


Using the post-mortem brain samples that matched the typical PD criteria (H & Y scale ≥ 2), we demonstrated that severe PD decreases specifically a neurotrophic factor, EGF, in the striatum and prefrontal cortex. As there are lesions of the nigrostriatal fibers and mesolimbic dopaminergic fibers in severe PD (Jellinger 1991), we assumed that dopaminergic depletion was responsible for the EGF reduction observed in both striatum and prefrontal cortex. The results also suggest that the EGF reduction reversibly promotes the initiation or progression of the dopaminergic lesion. To test these hypotheses, we performed two sets of animal experiments: a PD model produced by cytotoxic lesion of the SN dopaminergic neurons and striatal EGF administration to the PD model. Dopaminergic lesions induced by 6-OHDA mimicked the specific EGF reduction in the striatum and produced a decrease in ErbB1 phosphorylation. These results demonstrated that dopaminergic deafferentation is responsible for the reduction in EGF and its signaling in the target brain regions. This explanation agrees with our preliminary data: Cultured striatal neurons produce EGF and liberate it in response to dopaminergic stimulation (Iwakura et al. 2002). Secondly, we evaluated the pathologic implication of the EGF decrease in the striatum of PD patients and its model. We tested whether local supplement of the neurotrophic factor, EGF, inhibits dopaminergic degeneration. EGF administration rescued the surrounding tyrosine hydroxylase-immunoreactivity in the striatum, whereas it failed to prevent the overall neurodegeneration of SN dopaminergic cells. The present results are compatible with previous studies of other PD models: EGF exerts neurotrophic effects on SN dopaminergic neurons in animals with nigrostriatal lesions induced by surgery or treatment with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine treatment (Hadjiconstantinou et al. 1991; Ventrella 1993). With the given strong trophic activity of EGF, it is possible that the neurotrophic decline of EGF contributes to the progression of the dopaminergic degeneration in PD patients as well. Once partial degeneration/deafferentation of the dopaminergic nerves occurs in the early stages of PD, the local EGF production should be down-regulated and the trophic starvation further promotes the dopaminergic degeneration/deafferentation.

Previous pathologic examinations indicated that several alterations in molecular markers sometimes represent the influences of chronic medication (Sokolov 1998). Thus, the present post-mortem study might have detected the neurochemical changes that were produced by anti-PD drugs such as dopamine precursors and dopamine agonists. The fact that there was the same trend for EGF content as well as for its receptor expression in the animal PD model, however, does not support this possibility. We also needed to consider the technical factors in the use of human post-mortem samples. There was no significant difference in PMI, freezer time, or age between PD samples and control subjects, however. Multiple regression analysis revealed that PMI, freezer time, and age of the sample collection did not correlate with growth factor levels in any brain region in either PD samples or control subjects. Accordingly, it is less likely that our findings reflect artificial changes in the post-mortem sampling or during storage.

The present results indicate that SN dopaminergic afferents positively regulate the production and reception of this neurotrophic factor in the striatum. The loss of ErbB1 receptors in the striatum as well as in the SN was apparent only in the lesioned hemisphere treated with the neurotoxin, 6-OHDA. The same decrease in ErbB1 levels was also observed in the striatum of PD patients. In situ hybridization studies of ErbB receptor mRNAs indicates that ErbB1 mRNA is distributed in many brain regions including neurons and glial cells, but particularly enriched in nigrostriatal dopaminergic neurons (Gomez-Pinilla et al. 1988; Werner et al. 1988; Seroogy et al. 1994; Misumi and Kawano 1998). Thus, the dopaminergic lesion only resulted in partial reduction in ErbB1 levels in the striatum as well as in the SN. These observations suggest that a neurotrophic interaction between SN dopaminergic neurons and its target of the striatum presumably, at least, involves the EGF/ErbB1 activity. The specificity of the dopaminergic lesion for EGF and ErbB1 subunits suggests that the loss of ErbB1 is one of the molecular pathologic features of dopaminergic lesions or denervation. These observations strengthen the argument that neurotrophic EGF/ErbB1 activity is impaired in the brains of PD patients.

Previous in vivo studies also suggest that there is a tight biological link between endogenous neurotrophic activity via ErbB1 signaling and dopaminergic development/function in the brain: Null-mutant mice for TGFα exhibit a reduction in midbrain dopaminergic neurons in early postnatal stages (Blum 1998). EGF administration up-regulates dopamine synthesis and its turnover in the striatum in developing rats (Futamura et al. 2003). Together with the present findings, all these observations suggest the possibility that endogenous EGF and dopamine positively interact with each other to regulate the development and function of nigrastriatal dopaminergic neurons through the target–axon interactions.

In addition to the ErbB1 ligands examined in the present study, amphiregulin, betacellulin, and epiregulin also bind to the ErbB1 receptor and activate its intracellular cascades. In this context, we cannot rule out the possibility that other ErbB1 ligands, which we did not examine in the present study, also contribute to the neurotrophic abnormality of SN dopaminergic neurons in PD patients. A similar post-mortem study of schizophrenia suggests that there is an interaction between EGF/ErbB1 activity and dopaminergic function. Chronic schizophrenic patients that display the neurobehavioral defects related to the dopaminergic system have higher levels of ErbB1 receptors in the prefrontal cortex (Futamura et al. 2002). The neurochemical abnormality of schizophrenia is also specific for the combination of EGF and ErbB1, not for HB-EGF, TGFα and ErbB3–4. Of note, the direction of ErbB1 alteration in schizophrenia patients is opposite to that in PD. Therefore, an abnormal neurotrophic increase of EGF/ErbB1 signals is implicated in the hyper-dopaminergic pathology that occurs in some of schizophrenic patients. Future studies should examine how the alterations in EGF content and ErbB1 levels are involved in the emergence or progression of dopaminergic degeneration in PD patients and how it associates with pathophysiology of this illness.


This work was supported by a grant-in-aid for Creative Scientific Research, the Targeted Research Grant for Brain Sciences (MECSST) and a Promotion Grant for Young Investigators. YI is a research fellow of Japan Society for the Promotion of Science.