Quantitative clinical proteomic study of autopsied human infarcted brain specimens to elucidate the deregulated pathways in ischemic stroke pathology

Highlights

• Identification of human ischemic proteome (1520 proteins, FDR = 0.1%) by iTRAQ

• Global failure of the cellular energy metabolism in the ischemic infarcts

• Concomitant reduction of all participating proteins of the malate-aspartate shuttle

• Iron-mediated oxidative imbalance as seen by the elevation of ferritin

• Presence of reactive gliosis along with an increased anti-inflammatory response

Abstract

Ischemic stroke, still lacking an effective neuroprotective therapy is the third leading cause of global mortality and morbidity. Here, we have applied an 8-plex iTRAQ-based 2D-LC–MS/MS strategy to study the commonly regulated infarct proteome from three different brain regions (putamen, thalamus and the parietal lobe) of female Japanese patients. Infarcts were compared with age-, post-mortem interval- and location-matched control specimens.

The iTRAQ experiment confidently identified 1520 proteins with 0.1% false discovery rate. Bioinformatics data mining and immunochemical validation of pivotal perturbed proteins revealed a global failure of the cellular energy metabolism in the infarcted tissues as seen by the parallel down-regulation of proteins related to glycolysis, pyruvate dehydrogenase complex, TCA cycle and oxidative phosphorylation. The concomitant down-regulation of all participating proteins (SLC25A11, SLC25A12, GOT2 and MDH2) of malate-aspartate shuttle might be responsible for the metabolic in-coordination between the cytosol and mitochondria resulting in the failure of energy metabolism. The levels of proteins related to reactive gliosis (VIM, GFAP) and anti-inflammatory response (ANXA1, ANXA2) showed an increasing trend. The elevation of ferritin (FTL, FTH1) may indicate an iron-mediated oxidative imbalance aggravating the mitochondrial failure and neurotoxicity. The deregulated proteins could be useful as potential therapeutic targets or biomarkers for ischemic stroke.

Biological significance

Clinical proteomics of stroke has been lagging behind other areas of clinical proteomics like Alzheimer's disease or schizophrenia. Our study is the first quantitative clinical proteomics study where iTRAQ-2D-LC–MS/MS has been utilized in the area of ischemic stroke to obtain a comparative profile of human ischemic infarcts and age-, sex-, location- and post-mortem interval-matched control brain specimens. Different pathological attributes of ischemic stroke well-known through basic and pre-clinical research such as failure of cellular energy metabolism, reactive gliosis, activation of anti-inflammatory response and aberrant iron metabolism have been observed at the bedside. Our dataset could act as a reference for similar studies done in the future using ischemic brain samples from various brain banks across the world. A meta-analysis of these studies could help to map the pathological proteome specific to ischemic stroke that will guide the scientific community to better evaluate the pros and cons of the pre-clinical models for efficacy and mechanistic studies.

Infarct being the core of injury should have the most intense regulation for several key proteins involved in the pathophysiology of ischemic stroke. Hence, a part of the up-regulated proteome could leak into the general circulation that may offer candidates of interest as potential biomarkers. In support of our proposed hypothesis, we report ferritin in the current study as one of the most elevated proteins in the infarct, which has been documented as a biomarker in the context of ischemic stroke by an independent study. Overall, our approach has the potential to identify probable therapeutic targets and biomarkers in the area of ischemic stroke.

Introduction

Ischemic stroke, still lacking an effective neuroprotective therapy, continues to be a major socioeconomic burden throughout the world. The successive clinical failures in new drug development along with a bleak epidemiological landscape have led to the emergence of several novel concepts. Increased emphasis has been given on the neurovascular unit and the interaction between its different components instead of the neuron alone. The participation of peripheral organs through bidirectional communications with the brain following an ischemic stroke has been highlighted [1]. Consequently, a combination therapy instead of a mono-therapy or the incorporation of a multi-target drug has been suggested [2]. Therefore, a global comprehensive study of the altered system(s) is the prerequisite to understand and tackle this multifactorial disorder.

Quantitative neuroproteomics has emerged as an advanced technique in the post-genomic era for unbiased probing of perturbed pathways in complex biological systems which is not possible by traditional reductionist approaches. Recently an iTRAQ-based neuroproteomic strategy has been applied to improve our mechanistic understanding of axon injury and Alzheimer's disease (AD) [3], [4]. In the area of ischemic stroke, we pioneered the employment of the iTRAQ-2D-LC–MS/MS-based quantitative proteomic strategy to study an in vitro neuronal model of ischemic penumbra [5] and an in vivo rodent model of transient focal stroke [6] to delineate the molecular complexities and to propose potential therapeutic targets for further studies. However, the direct translation of a hypothesis generated through pre-clinical models into clinical application is limited by the complexity and fundamental difference of the biological systems as exemplified by successive clinical failures in translating the targets. The reasons may be related to the extreme heterogeneity of human stroke, the absence of long-term environmental influence or co-morbidity or risk factors in the pre-clinical models or flaws in the clinical trial design [7]. Moreover, the majority of the animal studies used rodents that are phylogenetically separated from humans millions of years ago [8]. Hence, a wealth of bedside-back-to-bench data using post-mortem brain specimens is needed to exclude model-specific artifacts and to identify stroke-specific ‘target space’ encompassing different pathological events and brain cell types. However, studying the molecular pathophysiology of ischemic stroke using clinical samples is confronted with various difficulties and challenges that may account for pre-analytical or ex vivo stresses [9]. Additionally, for a pathological condition like stroke, considerable neuropathological and diagnostic expertise is needed as only specific locations have to be targeted for isolation. Unsurprisingly, although several methodological studies using different sub-structures of human brain have been reported for profiling either whole proteome [10], [11], [12] or sub-proteome [13], only a few studies addressed pathological questions related to ischemic stroke [14], [15]. The first one used a targeted approach by combining laser microdissection with protein array to focus on different matrix metalloproteinases [14]. The other one is the only reported profiling study to examine different areas of ischemic brain where 2DE with an off-line MALDI MS/MS was applied [15]. Using a complementary online shotgun proteomic approach, such as the multi-dimensional protein identification technology [16] with isobaric labeling of peptides (i.e. iTRAQ) will allow simultaneous quantification of proteins in up to 8 samples by avoiding intra-experimental variation.

Given that the iTRAQ-2D-LC–MS/MS strategy has already been successfully applied on the pre-clinical models of ischemic stroke in our laboratory [5], [6] here we extend to quantitatively study three well-documented and pathologically-characterized infarcted tissues of human brain by an 8-plex iTRAQ experiment. Ethnicity-, age-, sex-, location- and post-mortem interval (PMI)-matched areas of the non-infarcted brains were used as control for an appropriate comparison. The location-matched perturbed proteome obtained from the three pairs of infarct–control specimens were validated by comparing the expression of each protein with an independent control and infarcted sample included in the same iTRAQ experiment during the data analysis. Mining the data set revealed a mitochondrial dysfunction that may have been caused by the uncoupling of cytosolic and mitochondrial metabolism along with an iron-mediated oxidative imbalance.