Global transcriptional profiling of the postmortem brain of a patient with G114V genetic Creutzfeldt-Jakob disease
- Chan Tian
- Di Liu
- Chen Chen
- Yin Xu
- Han-Shi Gong
- Cao Chen
- Qi Shi
- Bao-Yun Zhang
- Jun Han
- Xiao-Ping Dong
- Corresponding author:
- Xiao-Ping Dong [ email@example.com ]
Published online on: Thursday, January 10, 2013
- Pages: 676-688
- DOI: 10.3892/ijmm.2013.1239
Prion diseases, also known as transmissible spongiform encephalopathies (TSEs), are a group of fatal neurodegenerative disorders in humans and animals, including Creutzfeldt-Jakob disease (CJD) in humans, bovine spongiform encephalopathies (BSEs) in cattle, and scrapie in sheep and goats. The conversion of prion protein (PrP), coded by the PRNP gene, from its cellular isoform PrPC to its pathogenic isoform PrPSc through a post-translational process is considered the etiology of these diseases. As a result of the conversion, the portion of β-sheets within PrP is increased (from 3 to 45%) whereas that of α-helices decreases (from 42 to 30%), therefore causing PrP to become detergent insoluble and resistant to denaturant (isoform PrPSc) (1). The conversion also causes a series of histopathological changes, including the depositions of PrPSc, spongiform degenerations, neuronal loss and astrogliosis.
According to the pathomechanisms, CJD can be classified into sporadic CJD (sCJD), familial or genetic CJD (fCJD or gCJD) and iatrogenic CJD (iCJD). fCJD accounts for approximately 10–15% of all CJD cases, characterized by the genetic changes of PrP (2). To date, 56 different mutations, including residue substitutions, insertions and deletions, have been reported (3). For instance, proline to leucine mutation at the 102nd position (P102L) causes Gerstmann-Straussler-Scheinker syndrome (GSS) characterized by the impairment of cerebellum (4). Nevertheless, D178N leads to fatal familial insomnia (FFI) (4), mainly harming the thalamus region. The varied clinical manifestations indicate that different mutants may trigger different regulatory pathways.
An efficient approach to uncovering the regulated genes and the signaling pathways caused by PrPSc is using global transcriptional profiling on prion-infected samples. Studies on various prion strains in mice by Hwang et al (5) described a network of differentially expressed genes (DEGs) on functional pathways and discovered that the gliosis fibril acidic protein gene (Gfap) and a set of complement activation associated genes are highly expressed (5). Xiang et al (6) performed their studies using human sCJD brains and described the upregulation of immune and stress-response factors and elements involved in cell death and cell cycle, and the downregulation of genes encoding synaptic proteins (6). However, there is a lack of information on the global transcriptional profiles of gCJD patients, since gCJD and sCJD may utilize different regulatory mechanisms.
In our previous study, we reported a Chinese gCJD patient with a G114V mutation in PrP (7), showing an sCJD-like neuropathological abnormality with large amounts of PrPSc deposit, spongiform degeneration, astrogliosis and neuron loss in the cortex regions (8). To further investigate the molecular mechanisms and to compare those with sCJD, the full transcriptome pattern of the parietal cortex lobe of this G114V gCJD patient was profiled with microarray analysis (Affymetrix Human Genome U133+ 2.0 Chip) with a commercial normal human parietal cortex RNA pool as control. This investigation revealed numerous DEGs and pathways and our results provide information regarding gCJD that is useful for the development of novel diagnostic and therapeutic approaches.
Materials and methods
Brain sample of the G114V fCJD patient
Brain tissue of the parietal cortex obtained from a patient definitely diagnosed with G114V gCJD was enrolled in this study. The patient was a 47-year-old (at onset) Han-Chinese woman, whose clinical and genetic characteristics have previously been described (7). Neuropathological assays of 10 different brain regions revealed typical sCJD-like abnormality and PrPSc deposits (8). A commercial normal human parietal cortex total RNA (cat. #636571; Clontech) pooled from four males/females aged 35–89 years was utilized as a control. Usage of the stored human brain samples in this study was approved by the Ethics Committee of the National Institute for Viral Disease Prevention and Control, China CDC. We obtained written informed consent from all participants in our study.
Total RNA of the parietal cortex of the patient with G114V gCJD was extracted with an RNeasy Mini kit (cat. #74104; Qiagen), according to the manufacturer’s instructions. The quality and quantity of extracted RNAs were verified by 1.2% formaldehyde agarose electrophoresis and ultraviolet spectrophotometry (NanoDrop, ND-1000). The processes of labeling, hybridization and scanning were performed at a platform of CapitalBio Corporation. Briefly, 200 ng of each total RNA preparation was taken for synthesis and amplification of first strand cDNAs, double-stranded cDNAs and biotin-labeled antisense RNAs, using a MessageAmp™ Premier RNA Amplification kit (cat. #AM1792; Ambion) on a PCR apparatus (MJ, PTC-225). After measuring the concentrations of the labeled RNAs by ultraviolet spectrophotometry, 15 μg of each preparation was fragmented and verified using 1.2% formaldehyde denatured agarose electrophoresis.
The biotinylated cRNAs were hybridized to a commercial gene chip, GeneChip® Human Genome U133+ 2.0 (Affymetrix Inc., Santa Clara, CA, USA) containing 47,000 transcripts, at 45°C for 16 h with constant rotation at a speed of 60 rpm. After washing and staining automatically on an Affymetrix fluidics station 450 with a GeneChip Hybridization, Wash and Stain kit (cat. #900720; Affymetrix), the chips were then scanned on Affymetrix scanners 3000 7G.
After scanning the gene chips, the CEL images were processed using the Affymetrix GCOS 1.4 software. The generated documents were analyzed according to the ‘Affymetrix Statistical Algorithms Description Document’. Briefly, the raw data were subjected to processes including masking unusable data, background subtraction, probe values calculation, scaling, single chip analysis, and calculation of P-value. A probe-set was considered as expressed if the corresponding detection P-value was <0.04. Genes were considered to be differentially expressed if their ratio (patient/control) was higher or lower than 2-fold.
The gene functions were formatted for both gene ontologies (GO) and molecular function (9). The GO and molecular functions of the genes with 6-fold differential expression compared to controls were further analyzed by CapitalBio® Molecule Annotation System V3.0 (MAS3.0) (http://bioinfo.capitalbio.com/mas3/). The P-value was calculated according to a probability formula of hypergeometric distribution, reflecting the importance of the selected pathway or GO. The smaller the P-value, the higher its significance. The pathways were ranked in the order of P-value and the 10 most important ones were chosen for further analyses. Meanwhile, a Q-value corresponding to the P-value was calculated to evaluate the false discovery rate (FDR) of significant pathway and GO through screening by using a single P-value as cut-off. The smaller the Q-value, the lower the FDR.
All data are MIAME compliant and the raw data have been deposited in the GEO database (10) with the accession number of that, platform of GPL570, samples of GSM759883 and GSM759884, as well as series of GSE30643.
Quantitative RT-PCR (qRT-PCR)
Prior to qRT-PCR, the RNA extracts were treated with a commercial RQ1 RNase-Free DNase (cat. #M6101; Promega) for 1 h at 37°C according to the manufacturer’s instructions. For cDNA synthesis, 2 μg of treated RNAs were mixed with the reagents in Reverse Transcription System (cat. #A3500; Promega). The real-time PCR was carried out on an ABI Prism 7900 sequencing detector, at the conditions of denaturing at 95°C for 15 sec, annealing at 50°C for 2 min and extension at 62°C for 1 min, 40 cycles in total. β-actin gene was used as an internal control to normalize the expression levels of target mRNAs. The primers for each gene are shown in Table I.
Global transcriptional profiling of the G114V gCJD patient
Following our previous studies on the G114V gCJD patient (7,8), we further investigated all the transcriptional patterns of the brain sample and compared them to those of normal brain RNA pool (control). To better understand the expression level of screened genes, we ranked them with the relative difference ratio using the following strategy: after single chip normalization, each probe set was marked as present (P) or absent (A) according to the comparison to background noise. Probe sets marked with A in both the experimental sample and control were discarded. Thus, we found the differentially expressed genes based upon two criteria, i) present in sample but absent in another sample, or ii) with the gene expression level altered over 2-fold. The genes that met both conditions were considered DEGs. After purging the redundant probe sets reflecting same genes, a total of 8,774 genes were determined to be differentially expressed in G114V gCJD brains. Among them, nearly one-third (2,769) was upregulated and two-thirds (6,005) were downregulated. Due to the lack of sufficient sample/data from other gCJD patients, we were unable to make parallel experiments/comparisons to minimize the gene set of gCJD specific genes. Furthermore, we did not use data from sCJD (6) or prion-infected mice (5), as the different molecular background may have induced unpredictable bias.
According to the annotations of Affy-Chip, 8,494 of the 8,774 genes were either assigned biological functions or were similar to genes with known functions, and 280 genes remain annotated as encoding hypothetical proteins. GO assignment determined these DEGs to be involved in 1,552 biological processes, with 819 molecular functions, and to be located in 368 cell components. In view of the significantly altered biological processes, 87 contained >10 DEGs. The predominantly altered processes covered the major basic cellular functions, including regulation of transcription, signal transduction, development and transport, oxidation reduction and apoptosis (Table II). Regarding the molecular functions, most DEGs were related to molecular (protein, nucleotide, ion) binding/interactions. The cellular component assignment showed that most functional genes were located in the membrane system (plasma membrane, mitochondria, endoplasmic reticulum). Since the sample was collected from the gCJD patient postmortem, the transcriptional profile represents the terminal stage of the infected brain.
Identification of the most DEGs
According to the microarray results, the most upregulated genes included UBE3A and RBBP6, which are involved in the ubiquitin protein degradation system, suggesting that the ubiquitin-dependent catabolic processes are essential in the gCJD patient. ASPH was the second most highly expressed gene that plays an important role in calcium homeostasis. It is well known that destruction of calcium homeostasis is common in neurodegenerative diseases (11). On the contrary, the most downregulated genes included genes involved in iron ion binding (HBB), transcriptional regulation (NR4A2), signal transduction (OR2L13) and cell skeleton formation (NEFL) (Table III).
To evaluate the microarray results, we performed real-time PCR, targeting some specific genes from the brain of the G114V gCJD patient and the control RNA pool from the normal brain. Real-time PCR showed that the transcriptional levels of the downregulated genes HBB, CBLN4, PHLDA2 and NR4A2 from the microarray were markedly lower than those of the normal control, particularly PHLDA2 and NR4A2, which were >40-fold decreased and consistent with the micro-array result. The upregulated genes EIF5B, UBE3A, ZNF396 and ZNF292 in the microarray were also significantly higher than the control in real-time PCR assay, particularly EIF5B, UBE3A and ZNF396 showing a >30-fold increase (Fig. 1). These findings indicate that the results of the microarray are reliable.
Confirmation of microarray results for the selected genes by real-time RT-PCR. Relative expression levels of the genes were indicated by the fold changes in expression level in G114V gCJD vs. control cases. The y-axis indicates fold changes and a negative value indicates downregulation.
Involvement of significant pathways in G114V gCJD
To examine the changes of cell signaling pathways in the brain cortex of the G114V gCJD patient, the genes expressed at least 6-fold higher or lower than that of normal control were subsequently analyzed with the CapitalBio® Molecule Annotation System V3.0 (MAS 3.0) using KEGG pathways (12). In total, 169 different pathways with significant difference in expression (P<0.05) were identified. Briefly, 82 pathways were involved in metabolism, 28 in human diseases, 19 in organismal systems, 16 in environmental information processing, 14 in genetic information processing and 10 in cellular processes, according to the KEGG classification of functional pathways.
The 10 most altered pathways ranked in the order of P-value are summarized in Table IV. Most of the genes in these path-ways were downregulated (Fig. 2). Notably, the first and the third most altered pathways were Alzheimer’s disease (AD) and Parkinson’s disease (PD). Within these two pathways, 31/34 genes and 27/28 genes were significantly downregulated in AD and PD pathways, respectively (Table IV). The down-regulated genes related to these two pathways are involved in mitochondrial dysfunction, ER stress and apoptosis. These data suggest that G114V gCJD may share similar mechanisms to AD and PD.
The 10 most altered pathways and the differentially expressed genes in them. The red color represents the upregulated genes and the green indicates downregulated ones. The gene symbols, RefSeq transcript ID, and gene titles of differentially expressed genes in each pathway are listed on the right. The altered ratios corresponding to various genes involved in pathways are indicated on the left.
Genes related to oxidative phosphorylation and purine metabolism were also significantly altered. Out of 30 differentially altered genes of involved in oxidative phosphorylation, 29 were downregulated, including NADH dehydrogenase, succinate dehydrogenase, cytochrome c reductase, cytochrome c oxidase and F-type ATPase (Fig. 2). Regarding the purine metabolism pathway, all 24 altered genes were downregulated, including adenylosuccinate lyase, RNA polymerase II and some phosphodiesterases (Fig. 2). These observations indicate that there is a severe failure in the mitochondria and a severe dysfunction of cell metabolism at the terminal stage.
Two pathways related to the cytoskeleton were also markedly downregulated; 29/32 altered genes involved in the regulation of actin cytoskeleton were significantly downregulated. All 17 affected genes related to pathogenic Escherichia coli infection were downregulated; among them some gene products were cell structure proteins, i.e. CDC42 and tubulin. Additionally, other possible severely disrupted pathways included those related to gap junctions (17/20 DEGs downregulated), which are involved in direct communication between the cytosolic compartments of adjacent cells and the pathway of axon guidance (24/24 DEGs downregulated), which helps axons extend to their correct targets. The downregulation of these two pathways suggests that there is damage to cell communication and neuronal development in the G114V gCJD brain.
Among the 10 most affected pathways, the mitogen-activated protein kinase (MAPK) signaling pathway was the only one related to environmental signaling processing. The processes seemed to be markedly repressed as well, with 33/35 altered genes being downregulated, including kinases (MAPK1, MAPK1, MAP2K1, MAP2K4, PRKCB, PAK1 and STK4), phosphatases (DUSP4, DUSP5, DUSP6, PPP3CA and PPP3CB) and some regulatory factors (FGF13, FGF14, MEF2C and RASGRF2) (Fig. 2). These genes were distributed in the three sub-pathways, classical MAP kinase pathway, JNK and p38 MAP kinase pathway, and extracellular-regulated kinase 5 (ERK5) pathway. This indicates that the local information processing in the brain of the gCJD patient was severely impaired.
The transcriptional pattern of important prion disease associated genes
Our previous study demonstrated large amounts of PrPSc deposits and severe gliosis in the cortex regions of this G114V gCJD patient, while the transcription levels of PRNP did not vary as much as in the PrPSc deposit among 10 different regions (8). Furthermore, microarray data showed no difference in PRNP transcription between the patient’s brain and normal control, even slightly downregulated in the patient’s brain with several PRNP probes, possibly indicating deposits of PrPSc in brains do not lead to enhancing the PRNP transcription. The transcriptional level of the GFAP gene in the patient was ∼2-fold increased relative to that of the control, highlighting an active gliosis. Nevertheless, a spectrum of neuronal biomarkers was downregulated in the patient’s brain, such as NSE (7.25-fold), tubulin-β III (3.88-fold), MAP2 (8.13-fold), NF-M (38.46-fold), NF-H (12.35-fold) and NF subunit NF-L (27.78-fold), demonstrating severe neuron loss. These data are consistent with the pathological characteristics of G114V gCJD.
In the present study, we analyzed the global expression patterns in the parietal cortex of a G114V gCJD patient with a commercial gene chip containing 47,000 transcripts. This is the highest-capacity approach to gene expression analysis used in human prion disease thus far. After purging the redundant transcripts, we identified 2,769 upregulated and 6,005 down-regulated genes. Further qRT-PCRs for several differentially expressed genes confirmed the results of the microarray. Notably, more downregulated genes in the brain of G114V gCJD are consistent with the results of a previous study on sporadic CJD with a relatively lower throughput microarray (18,000 transcripts), in which 275 genes out of 287 differentially expressed genes were downregulated (6).
In line with the observations of the pathological abnormalities in the G114V gCJD patient (8,13), the transcriptional level of the GFAP gene associated with gliosis is increased and a series of genes associated with neurons are decreased. Although the brain tissues are severely damaged pathologically, the expression levels of prion protein gene PRNP do not differ distinctly compared with those of normal control, which is in accordance not only with the data of PRNP transcription in this patient with qRT-PCR, but also with the previous microarray findings in the sCJD patients (6), mice infected with scrapie or CJD agents (14) and cattle infected with the BSE agent (15). Maintenance of active transcription of the PRNP gene in CNS tissues at the terminal stage of human and animal prion diseases may indicate a special environment that facilitates the replication of prion agents locally by supplying enough PrPC as the substrates for PrPSc replication.
The most differentially expressed genes in G114V gCJD seem to be involved in multiple cell processes, such as regulation of transcription, ion transport, cell adhesion, signal transduction, nervous system development, oxidation reduction, protein transport, RNA splicing and synaptic transmission. In the brains of naturally-occurring or experimental animal and human TSEs, as well as in some prion infected cell lines (5,6,15–17), abnormal alterations in ion transportation, transcription, cell adhesion, signal transduction and synaptic transmission have been repeatedly observed. Numerous differentially expressed genes involved in different cell processes or networks in brain tissues of this G114V gCJD patient reflect an extensive brain dysfunction at the final period of the disease.
Based on the classification of the KEGG database, 169 different pathways were significantly altered in the brain of the patient with G114V gCJD. Most of the differentially expressed genes in the 10 most significantly altered pathways were down-regulated, revealing a deeply suppressed expression status of the relevant functions. Two metabolic pathways, oxidative phosphorylation and purine metabolism, were markedly repressed in our study. In the oxidative phosphorylation pathway, the expression of several key elements, such as NADH dehydrogenase, succinate dehydrogenase, cytochrome c reductase, cytochrome c oxidase and F-type ATPase were decreased. This result is in accordance with previous studies by both microarray (5) and proteomics (18), reflecting a complete failure of mitochondria. Purine metabolism includes the biological synthesis, degradation and salvation of purines, an essential component of nucleotides (19). Abnormality in this pathway has not previously been observed in the prion-infected cells, or human and animal TSEs. Besides, the pathway of cellular proteasome in the brain of this gCJD case is significantly involved, in which various proteasome subunits are downregulated. The dysfunction of the cellular proteasome system is often noticed in several neurodegenerative disorders (18), including prion disease (20). The disability of protein degradation, especially in clearance of misfolded protein, may contribute to the accumulation of PrPSc in brain tissues.
It has been repeatedly observed that the cytoskeleton and microtubule are severely destroyed in the brain of prion disease (21–23). In our study, the pathway of regulation of actin cytoskeleton and the pathway involved in the expressions of cell structure proteins, such as CDC42 and tubulin, were clearly suppressed. Reduction of expression levels of those genes results in rearrangement of the cytoskeleton, disruption of barrier function and an increase in monolayer permeability.
The MAPK cascade is a highly conserved pathway involved in various cellular functions, including cell proliferation, differentiation and migration. Our microarray experiment illustrated the increased expression of p38 MAP kinase and nuclear factor-κB (NF-κB) in the brain of gCJD. The increase of those two factors has been reported in cells treated with the peptide PrP106–126. PrP106–126 has been demonstrated to activate p38 MAP kinase in human microglia accompanied by upregulation of NF-κB (24), and to induce a p38 MAP kinase-dependent apoptosis in SH-SY5Y neuroblastoma cells independently from the amyloid fibril formation (25). The decrease of ERK in our microarray is also in line with the observation that PrP fragment (aa 90–231) activates p38 MAP kinase by inhibiting the activation of extracellular-regulated kinases 1/2 (ERK1/2), followed by the caspase-3-dependent cell apoptosis in SH-SY5Y cells (26).
Genes involved in axon guidance and gap junctions, which are critical for cell communication and cell development, are rarely investigated in prion diseases. Axon guidance is a subfield of neural development concerning the process by which neurons send out axons to reach the correct targets. Its role in prion disease is rarely described, until recently a group performed systematical analyses on the gene changes in the brains of eight mouse adapted-prion strains throughout the progression of the diseases. Axon guidance disturbance was found in the mice with shorter incubation times (5). Gap junctions are involved in direct communication between the cytosolic compartments of adjacent cells. Apart from the changes of MAPK associated genes, some receptors of monoamines and other biogenic amine neurotransmitters, such as β-1 adrenergic receptor (ADRB1), dopamine receptor D2 (DRD2), 5-hydroxytryptamine (serotonin) receptor 2 (HTR2) and mGluR are suppressed, resulting in abnormal regulation of the expressions of the genes downstream and subsequently inducing the dysfunction of calcium signaling pathway and transportation of other biological masses.
The most significantly altered pathways in human diseases in the brain of the G114V gCJD patient are those of AD and PD, strongly indicating that G114V gCJD shares the similar gene expression profiles as these two neurodegenerative diseases. Among these two pathways, cell death induced by changes of oxidative phosphorylation in the mitochondria is a critical factor for neuron loss in AD and PD (18,27). In G114V gCJD, distinct impediment of oxidation phosphorylation is also observed. This includes abnormal phosphorylation, ATP depletion, collapse of mitochondrial membrane potential, increase of reactive oxygen species (ROS). Moreover, reduction of the expression of the relevant genes in this gCJD case highlight the presence of the similar ER stress-induced cell death observed in AD and PD, and Fas-induced cell apoptosis observed in AD, which have also been described in BSE (28).
Aside from numerous genes downregulated in this gCJD case, there are several genes showing upregulation. Among the 12 upregulated genes in sCJD described previously (6), eight genes were increased in G114V gCJD, including RAB13 (RAB13, member RAS oncogene family), inositol 1,4,5-trisphosphate 3-kinase B (ITPKB) and transcriptional coactivator with PDZ-binding motif (TAZ) that are more than 2-fold increased, and GFAP, cysteine and glycine-rich protein 1 (CSRP1), tropomyosin 2 (TPM2), promoting factor 1 (PTN) and RNA binding motif, single stranded interacting protein 3 (RBMS3) that are more than 1.5-fold increased. The proteins of the Rab family regulate specific tethering/docking of incoming vesicles to the correct target organelle (29), and they are involved in various biological processes, protein transport, small GTPase mediated signal transduction, vesicle-mediated transport, modification-dependent protein catabolism, ER to Golgi vesicle-mediated transport, endocytosis and some related regulation processes (12). Several Rab family members have been found to be associated with the PrPSc propagation and accumulation in the prion-infected cells (30), as well as with the clinical manifestations (31). In line with the observation in sCJD, Rab genes are unregulated in the cerebral cortex of G114V gCJD. These phenomena illustrate the similarity of gene expression profiles between G114V gCJD and sCJD. More up and downregulated genes in G114V gCJD rely heavily on the usage of a relatively larger capacity gene chip in this study. In addition, a series of genes that appeared dozens of times increased both in microarray and qRT-PCR will provide useful insights to further explore potential biomarkers for the diagnosis of CJD.
We thank Dr Christopher J. Vavricka from the Institute of Microbiology, Chinese Academy of Sciences, Beijing, China, for kindly checking the manuscript. This study was supported by the Chinese National Natural Science Foundation Grants 30800975, 81101302 and 30800640, the National Basic Research Program of China (973 Program) (2007CB310505), the China Mega-Project for Infectious Disease (2009ZX10004-101 and 2008ZX10004-008), and the SKLID Development Grant (2008SKLID102).
Collins S, McLean CA and Masters CL: Gerstmann-Straussler-Scheinker syndrome, fatal familial insomnia, and kuru: a review of these less common human transmissible spongiform encephalopathies. J Clin Neurosci. 8:387–397. 2001.
National Center for Biotechnology Information, US National Library of Medicine; Bethesda, MD: http://www.ncbi.nlm.nih.gov/geourisimplehttp://www.ncbi.nlm.nih.gov/geo.
KEGG: Kyoto Encyclopedia of Genes and Genomes. http://www.genome.jp/kegg/urisimplehttp://www.genome.jp/kegg/.
Kordek R, Liberski PP, Yanagihara R, Isaacson S and Gajdusek DC: Molecular analysis of prion protein (PrP) and glial fibrillary acidic protein (GFAP) transcripts in experimental Creutzfeldt-Jakob disease in mice. Acta Neurobiol Exp. 57:85–90. 1997.
Tang Y, Xiang W, Hawkins SA, Kretzschmar HA and Windl O: Transcriptional changes in the brains of cattle orally infected with the bovine spongiform encephalopathy agent precede detection of infectivity. J Virol. 83:9464–9473. 2009.
Zabel C, Nguyen HP, Hin SC, Hartl D, Mao L and Klose J: Proteasome and oxidative phoshorylation changes may explain why aging is a risk factor for neurodegenerative disorders. J Proteomics. 73:2230–2238. 2010.
Fabrizi C, Silei V, Menegazzi M, et al: The stimulation of inducible nitric-oxide synthase by the prion protein fragment 106–126 in human microglia is tumor necrosis factor-alpha-dependent and involves p38 mitogen-activated protein kinase. J Biol Chem. 276:25692–25696. 2001.
Corsaro A, Thellung S, Villa V, et al: Prion protein fragment 106–126 induces a p38 MAP kinase-dependent apoptosis in SH-SY5Y neuroblastoma cells independently from the amyloid fibril formation. Ann NY Acad Sci. 1010:610–622. 2003.
Corsaro A, Thellung S, Chiovitti K, et al: Dual modulation of ERK1/2 and p38 MAP kinase activities induced by minocycline reverses the neurotoxic effects of the prion protein fragment 90–231. Neurotox Res. 15:138–154. 2009.
Higgins GC, Beart PM, Shin YS, Chen MJ, Cheung NS and Nagley P: Oxidative stress: emerging mitochondrial and cellular themes and variations in neuronal injury. J Alzheimers Dis. 20(Suppl 2): S453–S473. 2010.