Reactive oxygen species are involved in ferroportin degradation induced by ceruloplasmin mutant Arg701Trp.

The ceruloplasmin mutant R701W, that causes a dramatic phenotype in the young heterozygous patient carrying this mutation, has been shown to have profound effects also in cell culture models. Here we show that Golgi rearrangement and degradation of the iron exporter ferroportin, that follow transfection of cells with this mutant, are accompanied by the massive production of reactive oxygen species (ROS) in the cell. Scavenging ROS production with different antioxidants, including reduced glutathione and zinc, restores Golgi morphology and rescues ferroportin on the cell membrane. Copyright Â

Introduction

Aceruloplasminemia is a rare autosomal recessive iron overload disease caused by mutations in the gene of the ferroxidase ceruloplasmin (Cp) (Harris et al., 1995). Cp is a multicopper oxidase bearing multiple copper sites, which safely couple the oxidation of substrates to the controlled reduction of oxygen to water. Two isoforms of Cp are found in mammals: secreted Cp is mainly synthesized by hepatocytes and released into the plasma, while a GPI-anchored form (Cp-GPI) produced by alternative splicing is mainly located in the brain, where it resides on the plasma membrane of astrocytes (Patel and David, 1997). Approximately forty mutations of the Cp gene have been so far described. Homozygotes have iron overload mainly in the brain, but also in liver, pancreas and retina. Patients develop retinal degeneration, diabetes mellitus and neurological symptoms, which include ataxia, involuntary movements and dementia (Miyajima, 2003). Onset of clinical manifestations usually occurs in adulthood. The actual pathogenesis of aceruloplasminemia is still unclear, but iron-mediated oxidative stress is thought to contribute to tissue injury and neuronal cell death. Recent data have demonstrated that the ferroxidase activity of Cp is required for proper iron homeostasis: lack of oxidase-active Cp leads to internalization and degradation of ferroportin (Fpn), the only known mammalian iron exporter (De Domenico et al., 2007). Our previous finding has provided a straightforward explanation for brain iron overload in patients with aceruloplasminemia, where lack of a functional Cp would lead to defective export of iron from cells due to degradation of Fpn (Bonaccorsi di Patti et al., 2009). Heterozygous individuals usually have only partial Cp deficiency with normal iron metabolism and no clinical symptoms. However, a very young heterozygous patient carrying an Arg- > Trp mutation at the highly conserved position 701 was shown to have an extremely severe neurologic symptoms despite the presence of the wild type allele (Kuhn et al., 2005). We consistently found that, among several Cp mutants that have been tested for their ability to impair Fpn stability and induce intracellular iron retention, Cp R701W is peculiar, in that it is dominant over wild type Cp, thus acting in heterozygosis. Moreover, both secreted and GPI-anchored isoforms of this mutant induced a significant rearrangement of the Golgi apparatus (GA) and re-localization of the copper ATPase ATP7B (Bonaccorsi di Patti et al., 2009). Oxidative stress and GA fragmentation could represent early phases of cell death. Indeed, GA fragmentation appears to be a general hallmark of neurodegenerative diseases, as it is found in Alzheimer’s disease, amyotrophic lateral sclerosis, Creutzfeldt-Jacob disease, multiple system atrophy, Parkinson’s disease, spinocerebellar ataxia type 2, and Niemann-Pick disease type C (Gonatas et al., 2006). Thus, the morphological status of GA may be a reliable index of degeneration activity. GA fragmentation has been reported to occur in neuronal and non-neuronal cell types after various forms of stress, but the molecular processes are still far from being understood. Since oxidative stress has been widely linked to pathogenesis of neurodegenerative diseases, we explored the possibility of reactive oxygen species (ROS) formation induced by expression of the Cp R701W mutant and its role in GA fragmentation. C6 glioma cells were used as glial cellular model in analogy to our previous studies.

Materials and methods

Materials

Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), gentamicin, Trypsin/EDTA, N-acetyl cysteine (NAC), reduced (GSH) and oxidized (GSSG) glutathione were from Sigma–Aldrich, (Milan, Italy). The siRNA oligonucleotide pool for rat Cp-GPI was from Dharmacon (ON-TARGETplus SMART pool L-089853). Lipofectamine, Plus reagent and Oligofectamine were from Invitrogen.

Constructs

The coding sequence for human Cp-GPI was cloned SacI-XhoI in the pCMVTag4b vector (Stratagene), and Cp mutants were produced as described (Bonaccorsi di Patti et al., 2009). The pFpn-EGFP expression plasmid was a generous gift from J. Kaplan.

Cell culture, transfection and treatments

Rat C6 glioma cells from the ATCC were maintained in DMEM supplemented with 10% FBS and 40 μg/ml gentamicin. Confluent monolayers were subcultured by conventional trypsinization, and cells were seeded in 35-mm tissue culture dishes. For transfections cells were used at a confluency of 50–60% and transfected with pFpn-EGFP and pCMVhCp constructs using lipofectamine enhanced by the Plus reagent. Cells were grown for 18–24 h and then processed for immunofluorescence microscopy. In some experiments, cells were silenced using 100 nM siRNA oligonucleotide pool, matching selected regions of rat Cp-GPI specific for the rat protein, allowing expression of transfected human Cp.

Fluorescence microscopy

Cell staining for immunofluorescence microscopy was performed with either rabbit anti-Cp (1:100, Dako) or rabbit anti-giantin (1:1000, Novus Biologicals), followed by treatment with either Alexafluor 594- or fluorescein isothiocyanate-conjugated goat anti-rabbit IgG (1:750, Invitrogen). Intracellular ROS levels were determined with the redox-sensitive probe 2′,7′-dichlorodihydro-fluorescein diacetate (DCFH-DA, Invitrogen) (Bacsi et al., 2005). C6 cells were loaded with 1 ml of 5 μM DCFH-DA in phosphate buffer saline and maintained at 37 °C in the dark. After 60 min, loading buffer was removed and cells were incubated at 37 °C for additional 30 min in a prewarmed growth medium. NAC, GSH and GSSG were dissolved in water at a stock concentration 0.5 M.

Cells were visualized using an inverted DMI 6000 confocal scanner microscope TCS SP5 (Leica Microsystems CMS GmbH) with a 63× oil immersion objective.

Glutathione measurement

GSH and GSSG analyses were performed according to Sessa et al. (2009). Briefly, cells were treated with 100 μl of 5% trichloroacetic acid. After centrifugation, 50 μl of the supernatant were filtered through a Microcon YM-10 centrifugal filter device (Amicon), and analyzed by reversed phase HPLC on a C18 column (Atlantis 5 μm, 4.6 × 200 mm, Waters) in isocratic conditions (1 ml/min, 20 mM phosphate buffer, pH 3). Eluate was monitored by an electrochemical four-channel colorimetric sensor (ESA model 6210) at 700, 800, 900 and 1000 mV.

Results

Effects of Cp R701W on GA morphology

Ferroxidase-competent Cp is crucial for the maintenance of Fpn on the plasma membrane of astrocytes. We also demonstrated, by using the trans-Golgi network marker TGN-38, that the transient expression of Cp-GPI R701W – but not of any other Cp mutant involved in aceruloplasminemia – was able to induce a significant rearrangement of GA in C6 cells and the formation of vesicles originating from the trans-Golgi network (Bonaccorsi di Patti et al., 2009). Moreover, Cp-GPI R701W was dominant over Cp wild-type, as its effects were observed even in the presence of endogenous Cp. A time course experiment (data not shown) revealed that GA rearrangement is already evident at 9 h post-transfection of C6 cells with Cp-GPI R701W, and that it is maintained up to 24 h. Fpn degradation is also already completed at 9 h. At the earlier time point analyzed (6 h), GA morphology was unaffected.

Effects of Cp R701W on the levels of ROS

GA dissociation and chronic oxidative stress are frequently associated with the pathogenesis of several neurodegenerative disorders. Therefore, we investigated the levels of ROS production in C6 cells expressing Cp-GPI R701W. As representatively shown in Fig. 1, ROS levels were significantly and invariably higher in C6 cells expressing R701W vs. wild-type at 24 h post-transfection. Fluorescence was not linked to transfection, as untransfected cells were undistinguishable from cells transfected with the wild-type construct. Expression levels of recombinant Cp-GPI wild-type or R701W were found to be similar, as already reported (Bonaccorsi di Patti et al., 2009). Undistinguishable results were obtained when cells were silenced for endogenous Cp before transfection with the human mutant. The effect was highly specific for Cp-GPI R701W mutant, as it was not elicited by any other tested Cp mutant involved in aceruloplasminemia. In particular, according to our previous classification (Bonaccorsi di Patti et al., 2009) we tested both functionally competent mutants (Q146E, W264S, G876A) and functionally uncompetent mutants (R701W, G969S, G631R). Moreover, other Cp mutants were also assayed which are not involved in aceruloplasminemia but have been shown to be relevant either because they disrupt a prooxidant copper site (H426A), or because they remove a proteolytic site (K887A) or again because they perturb the flexible external loops of the protein (C181S, C881S). All these mutants turned out to be uneffective in inducing ROS formation (data not shown). To determine whether ROS production induced by Cp-GPI R701W was affected by ROS scavengers, cells were treated with well-known antioxidants, i.e. NAC or GSH, for all the time of transfection up to 24 h. As shown in Fig. 1, the presence of 2 mM NAC or GSH totally abolished ROS production induced by the Cp mutant. As expected, NAC or GSH alone did not alter ROS levels in C6 control cells. At variance with GSH, GSSG was ineffective in reducing ROS levels in Cp R710W transfected cells (data not shown), consistent with the need of reducing power to scavenge ROS. It is known that many zinc-dependent enzymes and transcription factors contribute to maintaining the proper redox balance in the cells. Therefore, we tested the effect of zinc on Cp-GPI R701W-induced ROS production. As also shown in Fig. 1, as low as 5 μM zinc completely prevented ROS production in C6 cells expressing the Cp mutant. To further characterize the free radicals induction by Cp-GPI R701W, the levels of ROS were monitored at different time points (6, 9, 24 h post-transfection). As for the GA fragmentation and Fpn degradation, ROS generation became evident starting at 9 h post-transfection (data not shown). Since cellular GSH plays a fundamental role in the regulation of the cell redox status, we also investigated the levels of intracellular GSH. No significant depletion of GSH content in C6 cells expressing Cp-GPI R701W at any time point post-transfection was observed. GSH cell content was 26.1 ± 3.0, 24.2 ± 2.3 and 25.6 ± 4.5 nmol/mg of protein for cells transfected with empty vector, Cp-GPI wild type and Cp-GPI R701W, respectively.

Fig. 1

ROS levels increase in Cp-GPI R701W-transfected C6 cells and antioxidants can prevent ROS production. C6 cells were transfected with human Cp-GPI wild type or R701W. After 24 h cells were loaded with DCF-DA (5 μM) for 1 h and then washed with PBS three times. Finally, cells were visualized by epifluorescence. The scavenging effect of 2 mM NAC, 2 mM GSH or 5 μM zinc is also shown. The figure is representative of five independent experiments.

Effects of ROS scavengers on Golgi morphology and Fpn degradation

To verify the involvement of ROS in Cp-GPI R701W-induced GA rearrangement and Fpn degradation we performed experiments on transfected C6 cells pre-treated with 2 mM NAC or GSH and the status of Golgi and Fpn was assessed. Both NAC and GSH were able to completely prevent Cp-GPI R710W-induced Fpn degradation, as shown in Fig. 2 for GSH. Golgi morphology also apparently recovered in the presence of the antioxidants, as demonstrated by staining with anti-giantin antibody (Fig. 2). These findings suggest that ROS generation plays a causative role in Fpn degradation induced by Cp-GPI R701W, possibly through rearrangement of the Golgi apparatus.

Fig. 2

Reduced glutathione can prevent the Cp-GPI R701W-induced Golgi fragmentation and Fpn degradation. C6 cells were transfected with Fpn-GFP and human Cp-GPI wild type or R701W, in the latter case cells were cultured in the absence or in the presence of 2 mM GSH. At 24 h posttransfection cells were analyzed by epifluorescence (Fpn-GFP) or immunofluorescence for the Golgi marker giantin. The figure is representative of five independent experiments.

Discussion

The majority of the missense aceruloplasminemia-associated mutants, characterized up to now, lack their ferroxidase activity due to inability to bind copper at the multinuclear active sites. We have previously indicated a novel mechanism causing aceruloplasminemia, that is defined by mutations on Cp which do not abrogate per se the ability of the protein to bind copper but rather impair the copper loading process requiring the Cu-ATPase ATP7B. Impaired copper binding to Cp can be detrimental, and it has been recently highlighted how copper from Cp could elicit pathological effects by promoting production of reactive oxygen species, which in turn play an important role in the aetiology of cardiovascular disease (Shukla et al., 2006).

Oxidative stress plays a role in the pathogenesis of several neurodegenerative diseases. Here we demonstrate, for the first time, that intracellular ROS generation is invariably associated with Fpn degradation induced by the Cp mutant R701W. Failure to incorporate copper in Cp together with aberrant localization of ATP7B and significant fragmentation of GA appear to be critical determinants of the severity of the phenotype observed in the patient carrying the Cp R701W mutation. This mutation is unique, leading to severe neurological symptoms in the young age, despite the heterozygous genotype (Kuhn et al., 2005). Moreover, zinc supplementation is beneficial in this case, with significant reduction of extrapyramidal and cerebellar-mediated movement disorders (Kuhn et al., 2007). Our data show that pretreatment of cells with zinc sulfate greatly reduced ROS production induced by Cp R701W, with effects comparable to those exerted by NAC and GSH. While the antioxidant properties of zinc as well as its effects on negative iron absorption are firmly established (Powell, 2000), the mechanisms of antioxidation are not fully understood, although induction of metallothionein synthesis is considered to be the most relevant aspect.

As far as the molecular aspects are concerned, it is well known that ROS can weaken copper binding to Cp, leading to disruption of iron exchange through impairment of ferroxidase activity. This in turn could lead to increased intracellular iron and thus more ROS formation by a cascade mechanism (Shukla et al., 2006). Indeed, free copper ions catalyze and amplify ROS formation through Fenton-type reactions. Furthermore, reduced copper catalyzes hydroxyl radical formation, which in turn can generate even more superoxide via Haber–Weiss reactions. Intracellular copper can be dislodged by ROS from copper-binding proteins other than Cp, consequently free copper becomes available to enter into reactions with other pro-oxidative factors to promote the formation of ROS. Impairment of the functions of antioxidant proteins relying on copper as an essential cofactor, in particular superoxide dismutases can also occur. These proposals presuppose that the increase in abundance of ROS elicited by the failure to incorporate copper in Cp-GPI R701W by ATP7B together with its aberrant localization, constitute a primary lesion which then has a knock-on effect of disrupting Cp/copper homeostasis. An alternative possible explanation for a ROS-dependent dominant effect of Cp-GPI R701W could be found in the ROS-induced degradation of Cp mRNA. Indeed, it has been recently demonstrated that Cp is downregulated by ROS generation by a mechanism involving the formation of a protein complex with its RNA 3′-untranslated region (Tapryal et al., 2009). However, this hypothesis seems unlikely if we consider, on one hand, that expression levels of both endogenous Cp-GPI wild type and recombinant Cp-GPI R701W appear to be normal; on the other, that Cp-GPI R701W exerts its dominant effect even on transfected recombinant Cp wild type missing the original 3′ end (Bonaccorsi di Patti et al., 2009).

The ability of antioxidants such as NAC and GSH to prevent the effect of R701W mutant on GA fragmentation clearly indicates the involvement of free radical generation in this pathway. Golgi fragmentation can be caused by a variety of mechanisms including accumulation/aggregation of misfolded proteins. Early-stage protein aggregation, often due to oxidative stress, represents a common fundamental molecular mechanism underlying the pathogenesis of neurodegenerative diseases. In this respect, some mutations of superoxide dismutase which are involved in the onset of familial amyotrophic lateral sclerosis are associated to fragmentation of the Golgi apparatus in spinal cord motor neurons (Stieber et al., 2004). While further studies are needed for a deeper understanding of the molecular mechanism accountable for our observations, it is clear that GA dispersal in transfected C6 cells was related to R701W-induced ROS generation and Fpn internalization.

A final point should be made on how Cp-GPI R701W can induce ROS formation. Cp is generally described as an antioxidant because of its ability to inhibit lipids oxidation and to sequester free copper ions. The ferroxidase activity may also contribute to the antioxidant capacity of Cp because conversion of Fe2+ to Fe3+ may reduce the oxidant capacity of iron by inhibiting the Fenton reaction. In contrast, it has been shown that Cp can have pro-oxidant activity, being able to oxidize LDL (Mukhopadhyay et al., 1997). In this respect, it should be taken into account that the molecular defect carried by some Cp mutants, such as R701W, could lead to a potentially toxic gain of function, and in this case the presence rather than the absence of Cp might play an active role in the development of the disease.

In conclusion, our results show for the first time that the form of aceruloplasminemia due to Cp-GPI R701W expression is in principle reversible, as the responsible missense mutation leads to free radicals generation that could be prevented by treatment with antioxidants able to restore Golgi morphology and Fpn integrity. Our findings could have noteworthy consequences on possible therapeutic strategies on this form of aceruloplasminemia.