- Top of page
- Materials and methods
- Supporting Information
Transition metals have been suggested to play a pivotal role in the pathogenesis of Parkinson's disease. X-ray microscopy combined with a cryogenic setup is a powerful method for elemental imaging in low concentrations and high resolution in intact cells, eliminating the need for fixation and sectioning of the specimen. Here, we performed an elemental distribution analysis in cultured primary midbrain neurons with a step size in the order of 300 nm and ~ 0.1 ppm sensitivity under cryo conditions by using X-ray fluorescence microscopy. We report the elemental mappings on the subcellular level in primary mouse dopaminergic (DAergic) and non-DAergic neurons after treatment with transition metals. Application of Fe2+ resulted in largely extracellular accumulation of iron without preference for the neuronal transmitter subtype. A quantification of different Fe oxidation states was performed using X-ray absorption near edge structure analysis. After treatment with Mn2+, a cytoplasmic/paranuclear localization of Mn was observed preferentially in DAergic neurons, while no prominent signal was detectable after Mn3+ treatment. Immunocytochemical analysis correlated the preferential Mn uptake to increased expression of voltage-gated calcium channels in DAergic neurons. We discuss the implications of this differential elemental distribution for the selective vulnerability of DAergic neurons and Parkinson's disease pathogenesis.
Parkinson's disease (PD) is the most frequent neurodegenerative movement disorder with a prevalence of about 1% in a population older than 70 years and about 3% in individuals older than 80 years (Strickland and Bertoni 2004). Pathological hallmarks include, but are not confined to, the preferential loss of dopaminergic (DAergic) neurons within the substantia nigra pars compacta and the presence of intracytoplasmic inclusions containing α-synuclein and ubiquitin, so-called Lewy bodies (Braak et al. 2003).
Next to the generation of free radicals in the course of dopamine metabolism, the increased presence of iron in the midbrain of PD patients has been assumed as a contributing pathogenic factor (Berg and Hochstrasser 2006). Iron participates in metabolic processes by undergoing oxidation–reduction reactions, a common property among transition metals, which allows this metal to undergo interconversion between the divalent cationic or ferrous (Fe2+), and trivalent cationic or ferric (Fe3+) states. These electron exchange processes can also lead to significant oxidative damage via free radical production within the brain when excess iron is present. In DAergic neurons during the dopamine catabolism, hydrogen peroxide is produced that in presence of iron favors Fenton reactions and oxidative stress [reviewed in (Papanikolaou and Pantopoulos 2005)].
Other transition metals such as Mn or Zn have been suggested to contribute to the pathogenesis of PD. For example, exposure of workers to high concentrations of manganese is an established risk factor for the development of a Parkinsonian syndrome, which clinically greatly mimics idiopathic Parkinson's disease [reviewed in (Guilarte 2010)]. Similar symptoms have been observed in psychostimulant drug abusers after repetitive intravenous injection of manganese-containing substances (Sikk 2011). Excessive levels of brain manganese have been linked to the loss of dopamine in the striatum, death of non-DAergic neurons in the globus pallidus, and damage of other neuronal pathways such as glutamatergic and GABAergic projections, all of which contribute to altered behavior, motor dysfunction, and cognition deficit (Erikson 2002). Manganese elevates intracellular H2O2 and related peroxides (HaMai et al. 2001) and reduces tyrosine hydroxylase activity and intracellular antioxidant levels (GSH, thiols, catalase) in DAergic neurons (Migheli 1999).
Iron and manganese are known to pass across the blood–brain barrier [reviewed in (Yokel 2009; Moos 2007)]. Because of their chemical similarity, both metals share and compete for transport proteins in organisms ranging from bacteria to mammals. As such, during conditions of low iron, abnormal manganese accumulation occurs. Conversely, when manganese concentrations are altered, the homeostasis and deposition of iron and other transition metals are disrupted. However, the knowledge on the subcellular allocation of trace elements upon exposure to manganese or iron remains incomplete. Indirect methods, like radioactive 54Mn tracing after cellular extraction, are able to quantify manganese in cell extracts, but provide only limited spatial information (Kalia et al. 2008).
The principle of X-ray fluorescence (XRF) is based on the irradiation of sample atoms with X-rays with energy sufficient to eject an electron from one of the atom's inner shells. The relaxation process is accompanied by the emission of a fluorescence photon. The energy of the emitted X-rays is characteristic of the excited element enabling the identification of the composition of the atoms constitutive of the sample. Beside element composition, X-ray spectromicroscopy performed using an X-ray nanoprobe beam is a unique method to determine the oxidation state of elements in cells and tissues (Qin 2011). X-rays can penetrate thick cells and tissues, eliminating the need of invasive preparation and sectioning of the specimen. Several studies have thus recently demonstrated the use of X-rays for trace element mapping of different cell types (Ortega 2007; Bacquart 2007; Ducić 2011).
This study was designed to investigate the subcellular distribution of manganese and iron in primary midbrain neurons and to compare this distribution in the DAergic and non-DAergic subpopulations by X-ray fluorescence microscopy. Previous studies have been limited to cell lines, post-mortem tissue, or did not take into account changes in the oxidative state during sample processing (Bacquart 2007; Szczerbowska-Boruchowska 2007). In this study, we visualized the transition metal distribution upon exposure in primary midbrain neurons and used the cryo-preserving (so-called cryo-embedding) technique to preserve oxidative states of the metal ions, which, to the best of our knowledge, is the first description exploiting primary midbrain neurons culture model with X-ray imaging techniques.
- Top of page
- Materials and methods
- Supporting Information
Transition metals, such as manganese and iron, have been suggested to be important pathogenetic factors for a number of neurodegenerative diseases, including the most frequent neurodegenerative movement disorder, PD. In particular, metal ions have not only been implicated in mediating oxidative stress, e.g., by catalyzing the Fenton reaction, but have recently also been shown to increase aggregation of α-synuclein (Kostka 2008; Uversky et al. 2001). Although PD is now increasingly recognized as a system disorder with a spreading pathology (Braak et al. 2003), which finally affects most regions of the brain, the special vulnerability of the dopaminergic nigrostriatal projections is particularly intriguing. In this study, we aimed to correlate the localization of the transition metals iron and manganese with the transmitter phenotype of midbrain neurons and evaluate in more detail the subcellular localization of these transition metals after treatment with different oxidative states of Fe and Mn.
We used primary MDN cultures from transgenic mice expressing GFP under the TH promoter (Matsushita 2002), which allowed us to clearly identify DAergic neurons in this mixed primary neuron culture and to correlate metal distribution to the transmitter phenotype (Figure S1 and Fig. 1). The metal ion concentrations employed in this set of experiments were chosen to achieve a slight, but quantifiable effect on the survival of DAergic neurons (Figure S2), which, of course, only partially are able to mimic the chronic exposure in PD patients.
To detect transition metals, we combined the above-mentioned culture technique with synchrotron radiation XRF nanoprobe analysis, which is a multielemental analytical technique enabling the simultaneous imaging of chemical elements at trace concentrations (Carmona et al. 2008; Ortega et al. 2009). In general, in comparison with X-ray electron microanalysis, synchrotron X-ray beams possess a higher spatial resolution (100 nm vs. 1000 nm with electron probe microanalysis), higher sensitivity (0.1 μg/g vs. 100 μg/g), and the ability to operate spatially resolved compounds analysis using XAS (Bacquart 2007). Using the X-ray microscope at ID21 gave us the unique opportunity to examine Fe and Mn in neural cells with both, high spatial and high spectral resolution.
Several previous reports demonstrated measurements of transition metals, which, however, were mostly performed in freeze-dried or chemically fixed samples under conditions that permit oxido–redox processes (Ide-Ektessabi and Rabionet 2005; Szczerbowska-Boruchowska 2007). Bacquart and coworkers proved better repeatability and sensitivity with no oxidation state modification as well as minimal beam damage when cells were analyzed in a frozen-hydrated state, as compared with freeze-dried cells (Bacquart 2007). Our goal was to show the localization of metals and their relative concentrations without changing their oxidative state. This was possible only by keeping the neuronal cultures deep frozen under liquid nitrogen temperature. Rapid cryogenic cooling and subsequent measurement under cryo conditions is at the moment the best option to achieve structure and chemical preservation in the sample and protection from radiation damage (Paunesku 2006).
To make a quantitative conclusion about metal distribution in different neuron subtypes, we present here data from a large number of imaged neurons, which is in contrast to many studies presenting only single-cell data. To the best of our knowledge, this is the first report directly measuring the distribution of the transition metals iron and manganese in primary midbrain neurons with a clear differentiation between DAergic and non-DAergic neurons in culture by using XRF methods combined with cryo conditions.
In our primary culture model, colocalization of iron and neurons was observed almost exclusively in Fe2+-treated cultures, whereas after Fe3+ application iron was found mostly in the extracellular space (Fig. 3). Moderate survival impairment was, however, observed after application of iron in both oxidation states (Figure S2). Our data thus support the hypothesis that the toxic effects of the redox-active Fe2+ are at least partially because of internalization and possibly cytoplasmic production of free radicals. Using XANES analysis after Fe2+ treatment, we found Fe to accumulate within the cells mostly in oxidized Fe3+ or ferritin-bound form (Fig. 5). Fe2+ is known to act as a catalyst for redox reactions and it is oxidized to Fe3+, e.g., via Fenton reactions, thereby fostering the production of intracellular free radicals [reviewed in (Jomova 2010)]. Our results of the XANES analysis support the notion that this oxidation indeed occurs. Iron uptake in vivo occurs mainly via ferritin, but non-transferrin-bound iron, e.g., as citrate has also been shown to be taken up by neurons in vitro (Moos 2007). On the other hand, we found that Fe is also present in the ferritin-bound form, the main iron store within the cell. The heavy chain of ferritin possesses ferroxidase activity, that involves the conversion of iron from the ferrous (Fe2+) to ferric (Fe3+) forms (Harrison and Arosio 1996). This limits the deleterious reaction that occurs between ferrous iron and hydrogen peroxide known as the Fenton reaction that produces the highly damaging hydroxyl radical.
Conversely, the toxic effects of Fe3+ are most likely because of its action at the cell membrane as treatment with Fe3+ did not result in a detectable uptake. Interestingly, using XANES analysis, we could detect a substantial amount of Fe2+ after treatment with Fe3+, which suggests the presence of extracellular reducing agents, e.g., in the cell culture medium.
The divalent metal transporter (DMT)-1 also mediates iron uptake in DAergic neurons, and rats deficient for DMT-1 are protected against iron-mediated neurodegeneration (Salazar 2008), while over-expression leads to increased dopaminergic cell death that is further aggravated by mutant α-synuclein (Chew 2011). In PC12 cells, Fe2+ was also shown to enter the cell via VGCC in a competitive manner with calcium (Gaasch 2007). VGCC in midbrain DAergic neurons could thus mediate iron entry into the cell independent of the DMT-1, and our analysis suggests that DAergic neurons indeed show a higher VGCC expression in comparison with non-DAergic neurons (Fig. 6). A fraction of the neurons examined in our study were DAergic, however, we did not observe a preferential colocalization with iron in this neuronal subtype. For the treatment time of 3 h, which was used in our studies, the uptake of iron did not differ significantly in DAergic versus non-DAergic cells. In humans, iron is preferentially found in DAergic neurons containing neuromelanin (NM), and NM has been suggested to act as an iron store (Fasano et al. 2006). NM, which accumulates in the aging human nigral DAergic neurons (Zecca 2003), however, is not present in these mouse primary cultures because of their embryonic age. This could be one explanation why we did not observe an increased colocalization of iron and DAergic neurons in this short-term-treatment model, which only insufficiently reproduces age-related alterations of the DAergic system.
Our observations in manganese-treated cultures suggest on one hand that manganese is taken up mostly in its Mn2+ oxidation state, whereas Mn3+ uptake is one magnitude lower, and that DAergic neurons show a more prominent uptake than non-DAergic neurons (Fig. 4). Moreover, manganese seems to accumulate in the apical region of the neuronal soma, possibly the axon hillock, as well as in neurites (Fig. 4). Carmona et al. showed that in manganese-treated PC12 cells, manganese was localized predominantly in the Golgi apparatus after exposition to 100 μM MnCl2, but it became more cytoplasmic when cells were treated with a higher concentration. The authors claim that the Golgi apparatus may act as a manganese storage, which becomes overloaded at toxic concentrations (Carmona et al. 2008). Especially low concentrations of Mn were measured after Mn3+ treatment, which was used in a 10-fold lower concentration to achieve the similar effect on survival as with the other treatments (Figure S2).
Unlike for iron, manganese uptake in the brain very likely does not involve the DMT-1 (Crossgrove and Yokel 2004), but manganese may be taken up by other transporters or via calcium channels such as VGCC (Crossgrove and Yokel 2005; Mertz et al. 1992). Our data suggested that the DAergic subpopulation might be more prone to manganese toxicity as a result of increased manganese uptake after Mn2+ incubation. The immunohistochemical analysis of MDN at 2 and 5 days in culture revealed that there is a significant difference in VGCC expression between DAergic and non-DAergic neurons, such that DAergic neurons show a stronger VGCC expression (Fig. 6). While this may not be the only factor, stronger VGCC expression could be at least partially responsible for increased manganese uptake in DAergic neurons and thus explain their selective vulnerability toward this transition metal.
Imaging studies in non-human primates suggest that manganese exposure markedly reduces striatal dopamine release and results in a degeneration and gliosis in the globus pallidus, the nucleus subthalamicus, and the pars reticulata of the substantia nigra, whereas the pars compacta was largely unaffected (Eriksson 1987; Olanow 1996). Together, these findings argue in favor of a pre-synaptic damage of dopaminergic neurons induced by manganese resulting in altered dopamine release [reviewed in (Guilarte 2010)]. Interestingly, a recent study suggested that LRRK2, one of the most commonly mutated proteins in familial PD, may act as a manganese sensor, establishing a link between genetic and environmental factors in the pathogenesis of PD (Covy and Giasson 2010). In addition, manganese can increase the rate of α-synuclein fibrillization in vitro (Uversky et al. 2001). Recently, dopamine secreted by the neurons and not intracellular dopamine was shown to be directly involved in the generation of toxic reactive oxygen species after treatment with Mn. The extracellularly active enzyme dual-oxidase and the dopamine reuptake transporter were proposed to be the major mediators for increased oxidative stress in DAergic neurons exposed to manganese in C. elegans (Benedetto 2010). Thus, the increased vulnerability of DAergic neurons may be because of their specific neurotransmitter metabolism combined with preferential manganese uptake.
In summary, our data underscore the importance of single-cell analysis and correlative imaging including novel X-ray microscopy techniques for the understanding of pathophysiological processes in models of neurodegeneration. Our study sheds light to the selective trace metal distribution in primary DAergic neurons suggesting that this subpopulation is more prone to accumulate Mn intracellularly than other neurons, which could not be observed for Fe. Uptake of divalent ions was preferred and the distribution of Mn and Fe was different regarding the cellular compartments. This differential localization and distribution could contribute to selective vulnerability of DAergic neurons in Parkinson's disease and suggests trace metals as molecular targets in future therapeutic options of this neurodegenerative disease.