|Home | About | Journals | Submit | Contact Us | Français|
Mutations in DJ-1 cause familial Parkinson’s disease (PD). The expression pattern of DJ-1 in the brain remains controversial. In the present study, we used DJ-1 deficient mice as negative controls and examined DJ-1 mRNA expression in mouse brains. By sequential double labeling on the same sections, in situ hybridization of DJ-1 mRNA was followed by immunofluorescence detection for cell type markers. We found that DJ-1 mRNA was expressed in the majority of neurons in all brain areas examined. In particular, all dopamine neurons in the ventral midbrain expressed DJ-1 mRNA. Interestingly, the choroid plexus and ependymal cells lining the ventricles were the only non-neuronal regions strongly expressing DJ-1 mRNA. However, DJ-1 mRNA was not detected in astrocytes. DJ-1 mRNA expression in all nigra dopamine neurons but not in astrocytes suggests that its potential neuroprotective role could be cell-autonomous. The fact that DJ-1 expression is not restricted to substantia nigra dopamine neurons suggests that DJ-1 mutations may collaborate with other predisposing factors to cause the relatively selective dopamine neuron degeneration in Parkinson’s disease.
Several missense and deletion mutations of DJ-1 cause early onset familial Parkinson’s disease. Several studies suggested DJ-1 may be involved in oxidative sensoring , ubiquitin proteasome system  and post-transcriptional regulation .
The expression pattern of DJ-1 in the brain remains controversial. It is unclear whether DJ-1 is expressed in astrocytes or neurons and whether DJ-1 is expressed in dopamine neurons. Three published studies on human brains found DJ-1 protein was localized in astrocytes but not in neurons using several antibodies [2, 3, 14]. However, two other groups found DJ-1 protein was present mostly in neurons and less in astrocytes in human brains [4, 17]. In the mouse brains, whether DJ-1 protein is present in astrocytes has also raised controversies [1, 3, 15, 18, 19]. In situ hybridization studies suggested DJ-1 mRNA expression in neuron-like but not astrocyte-like cells judged solely by the morphologies of DJ-1 positive cells [1, 5, 18]. Two study even found that DJ-1 protein was not present in dopamine neurons in rats [10, 19].
In the present study, our goal was to characterize DJ-1 mRNA expression pattern in the mouse brain and determine whether astrocytes and neurons, especially dopamine neurons, express DJ-1 mRNA. We used DJ-1 deficient mice as negative controls to ensure the specificity of in situ hybridization. Moreover, we took advantage of superior cellular resolution of non-radioactive in situ hybridization technique in combination with immunofluorescence staining for cell type specific markers.
All our data were obtained with the inclusion of DJ-1 deficient brain sections to control for staining specificity. The lack of DJ-1 mRNA and protein in our DJ-1 deficient mice brains were confirmed by RT-PCR (Fig. 1A) and western blot. Signet rabbit anti-DJ-1 antibody (Covance, Emeryville) and Santa Cruz rabbit anti-DJ-1 antibody (C-16, Santa Cruz Biotechnology, Santa Cruz), recognized one major 20 kD mouse DJ-1 band in wild type mice, which was absent in DJ-1 deficient mice (Fig. 1B).
Two to three-month old C57BL/6J mice and DJ-1 deficient mice  were used for the present study. The animals were housed in a 12:12 h light/dark cycle with ad libitum food and water. All animal procedures were approved by the Institutional Animal Care and Usage Committee of The University of Chicago. Mice were transcardially perfused with 4% paraformaldehyde. A full-length DJ-1 cDNA (NM_020569) was labeled with nonradioactive DIG RNA labeling kit (Roche, Indianapolis). For nonradioactive in situ hybridization, sections (20μm) were treated with 25μg/ml Proteinase K (Roche, Indianapolis) for 30 minutes, then incubated with labeled probe for 16 hours at 63°C followed by RNase A digestion (50μg/ml, Ambion, Austin) for 30 minutes at 37 °C. DJ-1 mRNA signal was detected with Anti-Digoxigenin-AP (Roche, 1:5000) and visualized by NBT/BCIP. All images were captured under Zeiss Stemi SV6 stereo microscope at constant exposure and magnification and analyzed with NIH ImageJ software to assess relative DJ-1 mRNA expression level. Three thresholds were set and labeling intensities were rated (Table 1).
To identify cellular types expressing DJ-1 mRNA, DJ-1 in situ hybridization was followed by immunofluorescence staining on the same sections. Tyramide signal amplification (TSA) technique was used to amplify the immunofluorescence signal. Briefly, after in situ hybridization, sections were blocked and detected with following antibodies: mouse anti-NeuN (1:500, Millipore, Bedford), mouse anti-TH (1:500, BD Transduction, Lexington) and rabbit anti-GFAP (1:500, DAKO, Carpinteria). Sections were then incubated with a biotinylated secondary antibody followed by peroxidase conjugated avidin-biotin complex (Vector Laboratories, Burlingame). TSA reagent (PerkinElmer, Boston) was used for visualization under Zeiss Axioplan 2 fluorescence microscope.
To quantify the percentage of NeuN or TH positive cells that express DJ-1 mRNA, pictures were taken under 20× objective lens at three randomly picked locations in each brain regions. All NeuN or TH positive cells in the entire field were counted manually. The percentage of DJ-1 expression in NeuN or TH positive cells was calculated by dividing all NeuN+DJ-1+ cells or TH+DJ-1+ cells by all NeuN+ or TH+ cells, respectively.
We found that DJ-1 mRNA was expressed throughout the mouse brain (Fig. 1C) except the white matter. The relative DJ-1 mRNA expression level in brain was summarized in Table 1. As a negative control, DJ-1 mRNA staining was absent in DJ-1 deficient mouse brains (Fig. 1C).
In olfactory bulb, DJ-1 mRNA was expressed in all cell layers with the highest expression in mitral layer and the lowest in granular layer (Fig. 1D). In the dorsal striatum, DJ-1 mRNA expression was lower than that of most other regions. DJ-1 positive cells showed neuronal morphology. Most DJ-1 positive cells were likely medium spiny neurons which comprise > 90% of the total striatal neurons. There were few densely labeled cells scattering evenly throughout dorsal striatum (Fig. 1E). The number of these large cells (20–35 μm) ranged from 20 to 40 cells per striatal section, which could be certain types of interneurons . DJ-1 mRNA expression in the cerebral cortex was higher than that in the striatum. DJ-1 expression was seen in all layers of the cortex with the highest expression in layer 5, which contained large pyramidal neurons innervating subcortical regions (Fig. 1G). The ventral midbrain had strong DJ-1 mRNA expression (Fig. 1J). DJ-1 mRNA expression level was comparable between substantia nigra pars compacta (SNc) and ventral tegmental area (VTA). The substantia nigra pars reticulata (SNr) had few DJ-1 positive cells.
To examine the cell type identity of DJ-1 positive cells, we performed sequential double labeling on the same sections. There was almost complete overlap of DJ-1 positive cells with NeuN positive cells (Fig. 2A). To quantify the percentage of DJ-1 expression in neurons, we counted cells positive for DJ-1, NeuN or both in several brain areas. In the frontal cortex, 96.3% cells (473 out of 491 NeuN positive cells) expressed DJ-1 mRNA. Conversely, few DJ-1 positive cells were negative for NeuN. Since most of these NeuN negative cells had typical neuronal morphology, we attributed these cells to extreme loss of NeuN protein due to proteinase K digestion for in situ hybridization. In addition to the frontal cortex, NeuN and DJ-1 were highly colocalized in the thalamus, dorsal striatum and amygdala, where more than 95% of NeuN positive cells expressed DJ-1 mRNA (Fig. 2 table). DJ-1 in situ hybridization signal was absent in brain sections of DJ-1 deficient mice (Fig. 2B).
GFAP signal apparently showed no colocalization with dark blue staining of DJ-1 mRNA in both gray matter and white matter (Fig. 3A, C). Absence of colocalization of DJ-1 mRNA and GFAP staining was most evident in corpus callosum (Fig. 3A). In corpus callosum, DJ-1 mRNA staining was virtually absent while GFAP positive astrocytes were abundant (Fig. 3A, left panel). Examination under 40× objective lens further confirmed the absence of DJ-1 mRNA in astrocytes (Fig. 3A, right panel). However, DJ-1 mRNA was not exclusively expressed in neurons. Strong expression of DJ-1 mRNA was observed in ependymal cells lining the ventricles and the choroid plexus (Fig. 1I).
Lastly, we tried to address if DJ-1 mRNA was expressed in dopamine neuron in SNc and VTA. Published DJ-1 in situ hybridization studies [1, 5, 18] showed DJ-1 mRNA expression in SNc and VTA but didn’t specify whether dopamine neurons expressed DJ-1 mRNA. On the other hand, two studies found absence of DJ-1 protein in dopamine neurons in rat brains [10, 19]. Therefore, it is important to demonstrate definitively whether DJ-1 is expressed in dopamine neurons. At low magnification, DJ-1 positive cells were abundant throughout VTA and SNc. Similar distribution patterns were observed between DJ-1 positive cells and TH positive cells (Fig. 4A). In VTA and SNc, virtually all TH positive cells (424 TH positive cells were counted) showed colocalization with DJ-1 positive cells (Fig. 4B and table). Some DJ-1 positive cells didn’t display TH immunity, indicated by asterisks in Fig. 4B. These cells had the typical neuronal morphology and scattered throughout the dense TH positive cell populations. Non-dopamine neurons in both VTA and SNc may partially account for these cells . Besides VTA and SNc, we also found nearly all TH positive cells expressed DJ-1 mRNA in the periglomerular region of olfactory bulb, retrorubral area (Fig. 4C, D and table) and hypothalamus (Fig. 4 table). In the locus coeruleus, DJ-1 mRNA was also expressed in nearly all TH positive cells which are presumably norepinephrine neurons (Fig. 4E and table). Therefore, our data clearly established that DJ-1 mRNA was strongly expressed in nearly all dopamine neurons.
This is the first DJ-1 in situ hybridization study using DJ-1 deficient mice as negative controls and combining immunofluorescence staining with cell type specific markers. With the highly specific labeling controlled by deficient mice sections, we unequivocally established that DJ-1 mRNA was ubiquitously expressed in nearly all neurons. We also firmly established that virtually all dopamine neurons expressed DJ-1 mRNA in the VTA and SNc. No DJ-1 mRNA expression was detected in astrocytes in adult mouse brains. However, it can not be ruled out that DJ-1 mRNA could be expressed by astrocytes in aged animals and certain experimental conditions, such as stroke  and neurotoxin treatment . Cells in the choroid plexus and ependymal cells represent the only non-neuronal cells strongly expressing DJ-1 mRNA in our study. It has been reported that DJ-1 protein is detected in human cerebrospinal fluid (CSF). The choroid plexus and ependymal cells might be one of the sources secreting DJ-1 into CSF.
Previous immunohistochemical studies on DJ-1 protein expression with various DJ-1 antibodies generated conflicting results. Two studies from the same group using three different antibodies, including KAM-S100 from Stressgen, consistently found DJ-1 protein was localized in astrocytes but not in neurons [2, 3]. Another group found similar results using KAM-S100 in human brains . However, using the same KAM-S100 antibody, two other groups found DJ-1 protein present mostly in neurons and less in astrocytes, which was further confirmed by several antibodies developed in their own labs [4, 17]. In another study on monkey brains, DJ-1 protein was found in astrocytes of cortex but in neurons of striatum and nigra . Specificity of antibodies used in all these studies could strongly influence the experimental results. It is well recognized that antibody specificity could not be guaranteed even with preabsorption control experiments. That could result from crossactivity for the same epitope of different proteins . In addition, DJ-1 could be oxidized and dimerized. It is possible that different forms of DJ-1 could be recognized preferentially by different antibodies. A third possible explanation for the conflicting results is the length of fixation. It was found that longer fixation decreased DJ-1 staining in neurons but increased DJ-1 staining in astrocytes .
It is possible that DJ-1 protein expression pattern could be very different from that of mRNA expression. There is evidence that DJ-1 can be secreted into cell culture medium, human CSF and serum and secreted DJ-1 may be picked up by cells [8, 10, 12]. It was reported that DJ-1 protein was absent in the soma of dopamine neurons [10, 19] and selectively enriched in striatal processes . This discrepancy is reminiscent of some axonally transported proteins, such as glutamate decarboxylase and dopamine transporter.
Species difference between primates and rodents has been reported. It could result from the different epitope accessibility in astrocytes and neurons between rodents and primates . However, such species difference of DJ-1 expression was not observed by an in situ hybridization study . It is possible that DJ-1 expression could change with aging since most human studies were carried out on postmortem specimen from aged individuals. If this were the case, gradual decrease of DJ-1 protein with age could deprive neurons, especially dopamine neurons, of protection from DJ-1.
In summary, the present study for the first time mapped out DJ-1 mRNA expression pattern and cell type in the mouse brain in great detail and with high specificity. DJ-1’s neuronal expression pattern, especially its expression in dopamine neurons, suggests that its potential neuroprotective role could be cell-autonomous. The discrepancy between DJ-1 mRNA expression pattern and reported DJ-1 protein expression pattern suggests that the majority of DJ-1 protein might be axonally transported to terminal regions. Future studies on DJ-1 protein expression pattern and cell identity in comparison with our mRNA expression data could offer important insights to the normal function of DJ-1 as well as its role in the pathogenesis of PD.
We thank Wanhao Chi for excellent technical assistance. This study was partly supported by a grant of the American Parkinson Disease Association (to L.C.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.