|Home | About | Journals | Submit | Contact Us | Français|
Cerebral ischemia-induced accumulation of unfolded proteins in vulnerable neurons triggers endoplasmic reticulum (ER) stress. Arginine-rich, mutated in early stage tumors (ARMET) is an ER stress-inducible protein and upregulated in the early stage of cerebral ischemia. The purposes of this study were to investigate the characteristics and implications of ARMET expression induced by focal cerebral ischemia. Focal cerebral ischemia in rats was induced by right middle cerebral artery occlusion with a suture; ischemic lesions were assessed by magnetic resonance imaging and histology; neuronal apoptosis was determined by TUNEL staining; the expressions of proteins were measured by immunohistochemistry, immunofluorescent labeling, and Western blotting. ARMET was found to be extensively upregulated in ischemic regions in a time-dependent manner. The expression of ARMET was neuronal in all examined structures in response to the ischemic insult. We also found that ARMET expression is earlier and more sensitive to ischemic stimulation than C/EBP homologous protein (CHOP). ER stress agent tunicamycin induced ARMET and CHOP expressions in the primary cultured neurons. Treatment with recombinant human ARMET promoted neuron proliferation and prevented from neuron apoptosis induced by tunicamycin. These results suggest that cerebral ischemia-induced ARMET expression may be protective to the neurons.
The assembly and folding of secreted proteins in the endoplasmic reticulum (ER) is regulated by a complicated mechanism that maintains equilibrium between folded and unfolded proteins. However, it is inevitable to produce ‘inferior goods' (unfolded/misfolded proteins) in the ER during this processing. Accumulation of these unfolded/misfolded proteins in the ER lumen triggers ER stress. Eukaryotes use the unfolded protein response to overcome the critical status induced by ER stress (Zhang and Kaufman, 2006). ER stress has been shown to be involved in brain ischemia (Kumar et al, 2001; Paschen et al, 2003; Morimoto et al, 2007; Oida et al, 2008a,2008b), as well as in some neurodegenerative diseases such as Parkinson's disease, Alzheimer's disease, and prion disease (Katayama et al, 2001; Imai et al, 2001; Nishitoh et al, 2002; Hitomi et al, 2004; DeGracia and Montie, 2004; Lindholm et al, 2006; Reijonen et al, 2008), as folding of proteins in the ER lumen requires molecular oxygen (Tu and Weissman, 2002). Protein aggregation in the ER after focal cerebral ischemia/reperfusion (I/R) has been observed using electron microscopy (Leroux et al, 2001). Hypoxia increased the level of activated XBP1 protein, a major transcriptional regulator of the unfolded protein response (Romero-Ramirez et al, 2004). The recent studies showed that a selective inducer of BiP (BIX) induces BiP to prevent neuronal death by ER stress (Kudo et al, 2008; Oida et al, 2008a,2008b).
Arginine-rich, mutated in early stage tumors (ARMET) or arginine-rich protein was originally identified as a human gene highly mutated in a variety of cancers. ARMET was also described as mesencephalic-astrocyte-derived neurotrophic factor (MANF), a survival promoting factor specific for midbrain dopaminergic neurons in vitro (Petrova et al, 2003). Its paralogue human conserved dopamine neurotrophic factor recently identified was proved to protect and rescue midbrain dopaminergic neurons in vivo in a rat 6-OHDA model of Parkinson's disease (Lindholm et al, 2007). ARMET and conserved dopamine neurotrophic factor form a novel protein family of evolutionarily conserved secreted factors with eight cysteine residues of similar spacing, predicting a unique protein folding. GeneChip arrays showed the ARMET gene was induced by ER stress and hypoxia in mouse embryonic fibroblast (MEF) cells (Lee et al, 2003; Romero-Ramirez et al, 2004). The recent studies show ARMET to be a secreted protein, and ER stress upregulates its expression and secretion (Apostolou et al, 2008; Mizobuchi et al, 2007). It was also found that ARMET is widely expressed in mammalian tissues (Mizobuchi et al, 2007; Lindholm et al, 2008) and differently regulated by epileptic and ischemic insults in rodent brain and heart (Lindholm et al, 2008; Apostolou et al, 2008; Tadimalla et al, 2008). These results suggest that ARMET may have important functions both under normal and pathological conditions.
Owing to the ability of ischemia/hypoxia to trigger potentially ER stress, we are interested at the profile of ARMET induction and potential significance in ischemic brain injury. Our preliminary study found that the protein level of ARMET was upregulated by 2h ischemia followed by 2h reperfusion [(I/R) 2/2h], using middle cerebral artery occlusion (MCAO) (Apostolou et al, 2008). Much shorter ischemia (10mins) reported recently also gave rise a transient increase in transcriptional level of ARMET in hippocampus (Lindholm et al, 2008). However, the prolonged I/R on ARMET protein expression as well as the functional consequences of ARMET induction remain to be elucidated. In this study, we showed the characteristics of ARMET expression and the time course after MCAO, and compared its sensitivity to ischemia with C/EBP homologous protein (CHOP). The protective effects of recombinant human ARMET on cultured primary neurons were also investigated.
Seventy male Sprague–Dawley rats (grade SPF, weighing 200 to 240g) were obtained from Shanghai Laboratory Animal Co. Ltd (Shanghai, China). The pregnant Sprague–Dawley rats (grade SPF) were from Anhui Experimental Animal Center (Hefei, China). The rats were kept under standard lighting conditions (12h light/dark cycle). The procedure for animal surgery was performed in accordance with the Guideline of Animal Care and Use Committee of Anhui Medical University.
Focal cerebral ischemia was induced by using MCAO with a suture. The rats were anesthetized with 10% chloral hydrate (3mL/kg body weight, intraperitoneally). The right middle cerebral artery (MCA) was occluded with a nylon filament (0.235mm in diameter) with a rounded tip, which was introduced from the common carotid artery lumen into the internal carotid artery to block the origin of the right MCA. The filament was inserted at least 18mm from the carotid bifurcation and was withdrawn at designed time points (2 or 4h after occlusion) to allow reperfusion for different time periods (2, 4, 12, 24, 36 or 48h). For each time point, 8 to 10 animals were used. In sham operation, a right neck incision was made to expose the arteries, but the nylon thread was not inserted into internal carotid artery.
Rats were anesthetized and placed in a general holder and positioned inside the magnet. The animal's head was held in place inside the 3inch surface coil. All MR studies were performed using a 1.5T superconducting magnet (General Electric Medical Systems, Waukesha, WI, USA). SE T1-weighted (TR/TE=340/12ms) and FSE T2-weighted images (TR/TE=4,000/110ms) were obtained from a 2.0mm thick axial and coronal section without gap using a 40mm × 40mm field of view, and reconstructed using a 256 × 192 image matrix.
Rats were killed under deep anesthesia. The brains were removed carefully and dissected into coronal sections. The fresh brain slices were immersed in a 2% solution of 2, 3, 5-triphenyltetrazolium chloride in normal saline at 37°C for 30mins and fixed by 4% buffered formaldehyde solution for 0.5h at 4°C. The normal area of brain was stained red whereas infarct tissue remained unstained (white). After triphenyltetrazolium chloride staining, digital photographs were taken.
Expression of ARMET and preparation of rabbit anti-ARMET polyclonal antibody have been described earlier. Monoclonal antibody was produced using standard methods. Briefly, 6-week-old female BALB/C mice were immunized three times intraperitoneally with the purified ARMET (200μg per injection) emulsified in incomplete Freund adjuvant at 10-day intervals. Antigen-specific-antibody-secreting hybridomas were identified by ELISA assay. Monoclonal antibodies were immunotyped using isotype kit (Sigma, Saint Louis, MO, USA) to define the heavy and light chains. Specificity of monoclonal antibody was evaluated by determining the fusion protein FLAG-tagged ARMET expressed in eukaryotic cells using Western blot and double-labeled immunofluorescent staining.
The brain tissues were fixed with 4% paraformaldehyde. Serial coronal sections (3μm) were hydrated. After blocking by 5% goat serum, the sections were incubated with primary antibodies overnight at 4°C. Incubation with secondary antibodies (DAKO, Glostrup, Denmark) at 37°C for 1h, the slides were washed and incubated in 3′-3′-diaminobenzidine-tetrahydrochloride for 5mins. Then the slides were stained with hematoxylin, dehydrated in ethanol, and mounted in neutral gum. The identical regions in different samples were selected at low magnification. For each sample, the number of ARMET-positive cells was counted in five randomly selected fields under the high-power field ( × 400 magnification).
Brain sections were hydrated and rinsed in phosphate buffered saline (PBS). After antigen retrieval, sections were permeabilized/blocked in PBS containing 0.5% Triton X-100 and 5% goat serum. The sections were incubated with primary antibody overnight at 4°C. For dual fluorescent staining, the sections were incubated with Alexa Fluor 488-conjugated or 568-conjugated IgG (Invitrogen, Carlsbad, CA, USA) and observed under fluorescent microscopy (Olympus, Tokyo, Japan). Immunocytofluorescent staining was performed as described earlier. The nuclei of cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) (5μg/mL).
The whole tissue of ispilaterally cerebral cortex was homogenized with 10 volumes of lysis buffer (50mmol/L Tris–HCl pH 7.2, 150mmol/L NaCl, 1mmol/L EDTA, 5mmol/L sodium pyrophosphate, 0.5% Triton X-100, 0.5% Nonidet P-40) containing protease inhibitor cocktail. The lysate was centrifuged and processed for SDS–PAGE. The proteins were transferred to polyvinylidene fluoride (PVDF) membranes and blocked in 5% nonfat milk. Primary and secondary antibodies were diluted in 5% milk and incubated for 1h, respectively. Blots were developed using the enhanced chemiluminescence kit (Amersham Biosciences, NJ, USA).
Pregnant Sprague–Dawley rats were killed at embryonic days 16 to 18 under anesthesia. The cortexes and hippocampi were moved into Ca2+- and Mg2+-free Hank's solution. Cells were dissociated using a fire-polished Pasteur pipette. The cells were resuspended in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum and 10% horse serum and seeded onto poly--lysine-coated 24-well plates and incubated at 37°C and 5% CO2 in a humidified atmosphere. After several days of culture, the cells were exposed to cytosine arabinoside for 24h. The medium was changed into low serum (5%) when the cells were treated.
The apoptotic cells were detected using the TUNEL kit (Promega, Madison, Wisconsin, USA) following the manufacturer's instructions. Under the high-power field ( × 400 magnification) of fluorescent microscope, the number of apoptotic cells characterized by the stained nuclei was counted in five randomly selected fields. The mean number of apoptotic cells per field was calculated for further statistical analysis.
Quantitative data were presented as the means±s.e.m. Statistical comparisons were analyzed using a one-way ANOVA followed by Dunnett's test. P<0.05 was considered statistically significant.
Our earlier study has shown that ARMET is an unfolded protein response-upregulated protein and inhibits ER stress-induced cell death (Apostolou et al, 2008). Focal cerebral ischemia was shown to be an in vivo ER stress model in that ischemia induces accumulation of immature proteins in the ER (DeGracia and Montie, 2004). Additionally, ARMET expression was induced in cardiac myocytes after myocardial infarction (Tadimalla et al, 2008). We wondered whether brain ischemic injury induces ARMET expression. To test this hypothesis, the focal cerebral ischemic model was successfully induced by right MCAO in adult male rats. Magnetic resonance imaging was used to estimate the lesions in vivo. Fast spin-echo T2-weighted images are shown in Figure 1A. These axial brain sections were obtained at the different levels, including caudate–putamen complex. Images were collected at different time points, including 24 and 48h after reperfusion indicated in Figure 1A. Diffuse hyperintense signal in lesions of ispilateral brains that were subjected to ischemic injury was noted. The representative images were shown in middle column (I/R, 4/24h) and right column (I/R, 4/48h) of Figure 1A.
Triphenyltetrazolium chloride staining showed that MCAO rats developed prominently focal infarcts at 4/24h (I/R) and 4/48h (I/R) (Figure 1B). The core of infarction was observed as a white area and the penumbra was pink. The histologic assessment showed mild edema in cerebral cortex supplied by the MCA started from 2/2h (I/R) (Figure 4A, indicated by circle). The comparison of focal infarcts by HE staining at 4/24h (I/R) and 4/48h (I/R) was presented in Figure 1C. The number of NeuN-positive neurons was obviously reduced in ischemic cortex at 4/24h (I/R) and 4/48h (I/R) (Figure 1D). These results suggest that the focal ischemic model was induced successfully by MCAO.
C/EBP homologous protein is a stress-inducible nuclear protein (Zinszner et al, 1998), and was usually used as a marker of ER stress (Ohoka et al, 2005). To clarify whether focal ischemia induced by MCAO leads to ER stress, CHOP was determined by using immunofluorescent staining and immunoblotting with anti-CHOP (Santa Cruz, CA, USA) at the designed time points. It was found that CHOP-positive cells were largely dispersed in MCA-innervated cerebral cortex (Figure 2B) at the time of 2/24h (I/R). Meanwhile, small amount of CHOP was found in ispilateral hippocampus of MCAO at that time (Figure 2D). Western blotting showed that expression of CHOP in ischemic cortexes increased beginning from the time point of 4/4h (I/R), compared with the shams (Figures 2E–2F). More interesting, the ischemia-induced expression of CHOP was found to be localized in the different subcellular compartments, including cytosol (Figures 2G–2I), nuclei (Figures 2J–2L), and perinuclear distribution (Figures 2M–2O). Focal cerebral ischemia-induced expression of CHOP showed here further confirms the results reported before that focal cerebral ischemia induces ER stress.
We reported earlier the expression of Armet in ischemic cerebral cortex (Apostolou et al, 2008). To investigate the characteristics of ischemia-induced ARMET expression, we observed the time course of ARMET expression in different brain regions by means of immunohistochemical assay and Western blotting. It was found that ARMET was detectable in many regions of normal rat brain, including cortex, hippocampus, subcommissural organ, and substantia nigra, but the staining is weak (not shown). Compared with the shams (Figures 3A1–3D1, Figures S2 and S3), ARMET was induced in MCAO rats at the early stage of ischemia, starting from 2/2h (I/R) and retaining in the period of observation (up to 4/48h). The typical ARMET-positive cells were found in the regions supplied by MCA at 2/2h (I/R) (Figures 3A2–3D2, Figures S1 and S4), whereas HE staining did not show marked infarct of the brain tissue at that time. Along with the prolongation of reperfusion, not only the immunostaining with anti-ARMET antibody became darker in cytoplasm, but also the number of ARMET-positive cells was increased in the ischemic regions (Figure 3G). Western blotting shown in Figure 3H also conformed that the levels of ARMET in ispilaterally cerebral cortexes of ischemic rats were increased at 2/2h (I/R) and 4/4h (I/R). These results indicate that ARMET was induced by focal ischemia in a time-dependent manner. Furthermore, in the core of infarct tissue, ARMET expression was relatively weak, even absent. The intense ARMET immunostaining was often found in peri-infarct tissue (Figure 3I).
ARMET is a secreting protein. It is synthesized in the ER and transferred to the Golgi for secreting. To observe the subcellular localization of ARMET induced by the stimulation of cerebral ischemia, DAPI was used to counterstain the nuclei. It was found that ARMET had no overlap with DAPI (Figures 3E1–3E3), suggesting that ARMET localizes in cytosol and focal ischemia did not induce its relocalization. To investigate the subpopulations of ARMET-positive cells, we first performed the doubly immunofluorescent staining with antibodies against ARMET and NeuN, a neuron-specific nuclear protein associated with withdrawal from the cell cycle and terminal differentiation (Mullen et al, 1992). We showed that ARMET was expressed in neurons (Figures 3F1–3F3). Meanwhile, we found that some NeuN-negative cells presented ARMET-positive immunoreactions. Further identification with glial markers showed that cerebral ischemia also induced ARMET expression in various glial cell types, including astrocytes, microglial cells, and oligodendrocytes (data not shown).
We also found that ARMET expression was susceptible to focal cerebral ischemia. For example, the early ischemia-induced dissociation of meninx from cortex at 2/2h (I/R) stimulated ARMET expression (Figure 4A2, region III) and tissue edema (indicated by circle), which was not observed in the sham controls (Figure 4A1). This result suggests that ARMET is sensitive to ischemic simulation. Additionally, double-labeled immunofluorescence showed that there were many ARMET-positive cells in ischemic regions at 2/2h (I/R) (Figure 4C2), whereas few CHOP-positive cells was found at that time (Figure 4C1). CHOP was remarkably induced at 4/48h (I/R) (Figure 4D1). These results suggest that ARMET was induced earlier than CHOP.
To further confirm that ARMET expression can be induced by ischemia, we first use SH-SY5Y cells. It was found that the expression of ARMET (Figure 5B2), but not CHOP (Figure 5B3), was moderately increased after deprivation of serum for 24h. However, both CHOP and ARMET were remarkably increased after exposure of cells to ER stress inducer tunicamycin (2.5μg/mL) for 4h (Figures 5C1–5C4). Furthermore, we isolated and cultured the neurons from embryonic day 18 rats and identified with neuron marker NeuN. We found that NeuN was dispersed in cytoplasm during neuron differentiation and it concentrates to nuclei after the neurons get mature (Figure 5D3). The number of neurons was reduced and the nuclei became small and condensed (Figure 5F1) after treatment with tunicamycin for 4h. ARMET and CHOP strongly expressed in the primary cultured neurons after treatment with tunicamycin (Figure 5F2–5F3), suggesting that ER stress induces ARMET expression in primary cultured neurons.
To investigate the physiological significance of ARMET upregulation during ER stress in neurons either in vivo or in vitro, we expressed recombinant human ARMET in bacteria and purified it. After treatment with ARMET (0.5μg/mL) for 2 weeks, the primary cultured neurons formed proliferative clones (Figure 6B, indicated by arrows). As a control, the recombinant human full-length tau protein inhibited the neuronal proliferation (Figure 6C). To test whether ARMET prevents ER stress-induced apoptosis, TUNEL staining was used. The neurons were pretreated with ARMET (0.5μg/mL) for 2 weeks, and then tunicamycin (2.5μg/mL) was added to medium for 4h. We found that there were fewer TUNEL-positive cells in ARMET-treated neurons (Figure 6E4), compared with those treated with bovine serum albumin (Figure 6D4) or tau protein (Figure 6F4). These results indicate that ARMET is protective to neurons during ER stress.
The findings that ARMET/MANF mRNA levels were transiently increased after 10mins of global forebrain ischemia in the adult rat were reported recently by Lindholm et al (2008). Simultaneously, we found that ARMET protein expression in the ispilateral cerebral cortexes of rats was upregulated by unilateral MCAO for 2h followed by 2h reperfusion (Apostolou et al, 2008). Here, we showed in detail the characteristics of the ischemia-induced ARMET expression in protein level, and the functional consequences in this kind of changes of ARMET expression were further analyzed. In this study, ARMET was induced at the early stage of ischemia, starting from 2/2h (I/R) and retaining in the period of observation (I/R, 4/48h). The typically upregulated ARMET expression was extensively found in the ispilateral regions, including cerebral cortex, hippocampus, hypothalamus, substantia nigra, subcommissural organ, amygdaloid nucleus, and so on, at 2/2h (I/R). However, this was in contrast to the observation by Lindholm et al in that no changes were detected in cortical areas except a transient increase of MANF mRNA expression at 2h after 10mins of global forebrain ischemia in the retrosplenial cortex (Lindholm et al, 2008). This difference may be associated with the methods to induce ischemia, the time of I/R, and the indexes detected (protein or mRNA).
In the earlier study, ARMET had been described as MANF (Petrova et al, 2003). However, we found here that only a small number of astrocytes were ARMET positive, whereas the upregulated expression of ARMET was mostly in neurons. This is in accordance with the finding that the expression was neuronal in all examined structures in response to the epileptic and ischemic insults (Lindholm et al, 2008). These findings suggest that ARMET may be a target of the unfolded protein response in vulnerable neurons during cerebral I/R injury and the inducible expression of ARMET has more important role on neuronal protection.
This hypothesis was further supported by our findings that recombinant human ARMET promoted cell proliferation and prevented from ER stress-induced apoptosis in cultured primary neurons. In accordance with that, MANF/ARMET was observed to promote the survival of dopaminergic neurons (Petrova et al, 2003). In addition, in Drosophila the neurotrophic factor DmMANF, a homologous to mammalian MANF and conserved dopamine neurotrophic factor, was associated with the function of dopaminergic axons (Palgi et al, 2009). The latest study showed pretreatment with recombinant human ARMET/MANF significantly reduced the volume of infarction at 2 days after MCAO (Airavaara et al, 2009). These results indicate the inducible expression of ARMET is protective to neurons in cerebral ischemia. Structure analysis suggests the neurotrophic activity of ARMET/MANF may reside in the N-terminal domain where there is a saposin-like lipid-binding domain, which may bind lipids or membranes (Parkash et al, 2009).
This study also reconfirmed that focal cerebral ischemia induces ER stress and upregulates CHOP expression, which is consistent with the earlier studies (Morimoto et al, 2007; Oida et al, 2008a,b). ARMET was significantly induced at translational level at the early stage of ischemia starting from 2/2h (I/R) when CHOP mRNA was not increased (Tajiri et al, 2004), and the injury was not detectable by imaging and histologic analysis at that time. This suggests that the secreting protein ARMET might become a potential serum biomarker for the early ischemic diagnosis.
It was reported that brain ischemia induces neuron death through ER stress (Tajiri et al, 2004; Oida et al, 2008a,b). In an attempt to survive under ER stress, cells priming self-protective mechanisms, which include induction of molecular chaperones in ER, attenuation of translation, and enhancement of ER-associated degradation. In ischemic brain injury, ER stress triggers molecular and cellular repair mechanisms that contribute to recovery in the adult brain (Cramer and Chopp, 2000). BiP/Grp78 was considered as a cytoprotective protein in stressed cells (Rao et al, 2002). BIX, a selective inducer of BiP/Grp78, was reported to reduce the area of infarction (Kudo et al, 2008) and protect against cell death (Oida et al, 2008a,b) induced by focal cerebral ischemia after pretreatment. ARMET induction in ischemic brain may be a protective reaction of cells to the ischemic injury. Meanwhile, ARMET expression is earlier than CHOP in focal cerebral ischemia; this somehow indirectly reflects the different impacts of ARMET and CHOP on cells. Another evidence for the protective effect of ARMET comes from the results that ARMET inhibited ER stress-induced apoptosis in cultured primary neurons, suggesting that this secreting protein might be effective in the treatment of cerebral disorders associated with ER stress, such as cerebral ischemia.
CHOP is expressed at very low levels under physiological conditions, but is strongly induced at the transcription level under ER stress (Ron and Habener, 1992). Overexpression of CHOP leads to growth arrest and apoptosis, and mediates cerebral ischemia-induced neuronal death (Tajiri et al, 2004; Oida et al, 2008a,b). As a nuclear transcriptional factor, CHOP was detectable in cellular nuclei of the ischemic tissues (Lindholm et al, 2008). CHOP was also found to present perinuclear distribution in cerebellar granule neurons (Wang et al, 2007). In this study, CHOP immunoreactivity was observed predominantly in perinuclear distribution. Only a small amount of CHOP was localized in cellular nuclei. More interestingly, CHOP immunoreactivity was found to display in cytosol. The reason why CHOP displays in different ways in brain ischemia remains unclear. We suspect that it may depend on the intensity of ischemic stimuli and the state of cells. CHOP mRNA was increased at 6h and decreased at 24 to 48h in the ischemic striatum and hippocampus (Tajiri et al, 2004), suggesting that induction of CHOP mRNA is an early event in cerebral ischemia. However, under persistent or prolonged ER stress, ATF4-induced expression of CHOP initiates apoptosis.
Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow and Metabolism website (http://www.nature.com/jcbfm)
Conflict of interest
The authors declare no conflict of interest.