Abstract

Well-characterized and preserved human brain tissue that is prepared and stored in brain banks is an essential resource for research in neurological diseases. This study examined the quality of human brain postmortem tissue from multiple laboratories within the BrainNet Europe brain bank network to identify all possible confounding variables and determine how they may affect RNA quality. Antemortem and postmortem information was retrospectively collected for a large cohort of samples. Total RNA was isolated from anatomically defined brain regions using a standardized procedure; RNA quality was assessed using an Agilent 2100 Bioanalyzer. No significant difference in RNA quality was observed in 6 different brain regions. RNA quality deteriorated with increasing numbers of antemortem events such as hospitalization, coma, respiratory illness, and the use of artificial ventilation; accumulation of such events was associated with elevated hypoxia-inducible factor 1α mRNA expression. Brain pH was found to be a good indicator of RNA quality. There was no correlation of postmortem delay with cerebrospinal fluid pH or RNA quality overall, but some individual RNAs decreased in quality with antemortem events and with postmortem delay. RNA quality did not affect total RNA yield. Determining the factors that are best predictors of RNA quality can help brain banks with selection criteria for storing high-quality brain tissue for research.

INTRODUCTION

Providing high-quality brain tissue for research into the molecular pathogenesis and development of treatments for neurodegenerative and neuropsychiatric disorders is the major goal of modern brain banking. If this is to be achieved, methods for optimal collection, preservation, storage, and characterization of such tissues need to be standardized (1). There are many steps involved in obtaining useful brain tissue for research purposes (2), and brain banks need to make constant efforts to achieve optimal preservation (3–7).

A critical factor for conducting molecular and cellular research on human postmortem brain tissue is the quality of extracted RNA and DNA. Achieving optimal tissue quality is particularly relevant to the increased use of RNA and DNA microarrays that have become powerful tools for biomedical and clinical research (8, 9). For animal studies, variable factors from antemortem conditions and death to tissue storage can be under strict control, whereas that is not the case for human tissue. Human tissue can only be collected in a manner that has many inherent confounding variables and over which brain banks have little control. These variables can influence the yield and quantity of RNA extracted from the tissue. Thus, determining the factors that are the best predictors of RNA quality is of great importance to brain banks, and determination of RNA integrity for each case entering a bank can considerably help with selection criteria for the provision of tissue of acceptable quality for specific research projects.

Several studies have investigated the effects of postmortem delay ([PMD] usually defined as the interval between the time of death and the time the tissue has been placed in the storage freezer) on RNA quality using a variety of measures of RNA integrity (10–16). The most common measure of RNA quality currently in use is quantitative digital analysis with Agilent Bioanalyzer 2100-generated electropherograms that calculate an RNA integrity number (RIN), which is based on the ratio of the 28S and 18S peaks (incorporating measurement of the width and height of the peaks and several intermediary regions of the electropherogram). Most studies have found a weak correlation between PMD and RNA quality (2, 10–12, 17). A study of total RNA quality determined by RIN values in more than 450 cases with PMD values up to 200 hours found only a weak significant negative correlation between PMD and RIN (2).

Another major group of variables that affect RNA is the length and severity of agonal events preceding death such as coma or hypoxia. Previous reports have shown that agonal factors significantly affect RNA integrity and have a profound impact on gene expression microarray studies (13). Various scales for the quantification of agonal states have been developed, including a simple 2-item scale (rapid vs slow death) (14), a 4-item scale (violent and fast death, fast death of natural causes, intermediate death, and slow death) (15), a rapidity of death measurement scale (5 items) (16), and a 25-item questionnaire containing information about events 48 hours before death and completed immediately after death by treating staff members (18). In general, slow death leads to lower tissue pH measurements compared with rapid or violent death. Consequently, brain tissue pH has become a good indicator of agonal status and of RNA quality and is considered to be a better indicator of RNA quality than PMD (2, 19–21). Moreover, brain tissue pH remains stable across brain regions and is unaffected during freezer storage (16). On the other hand, cerebrospinal fluid (CSF) pH was found not to correlate with RNA quality (22, 23).

BrainNet Europe II is a consortium of 18 brain banks across Europe the main objectives of which are to optimize and standardize brain banking methodology across member states and to disseminate uniform standards of practice (4). The aims of the current study were to examine human brain postmortem tissue quality within the BrainNet Europe network, to track all possible confounding variables and determine whether those factors had an impact on RNA quality, and to determine the effect of RNA quality on mRNA levels in banked human brain tissue. This is an important step toward producing guidelines for a unified European approach to brain tissue banking. The main findings showed that accumulation of deleterious antemortem (AM) events such as hospitalization, coma, artificial respiration or respiratory illness before death was associated with decreases in RNA quality and increasing evidence of hypoxia. Brain pH was also found to be a good indicator of RNA quality. Interestingly, no effects on RNA quality from all the postmortem variables that are under the control of the tissue banks were identified.

MATERIALS AND METHODS

Tissue Samples

RNA quality from 193 brain tissue samples originating from 9 different tissue banks within the BrainNet Europe Brain Bank Consortium (http://www.brainnet-europe.org/) was assessed. Fully informed consent and ethical approval were obtained for the collection and study of postmortem tissue following local and guidelines recently published by the consortium (4). The tissue samples were snap-frozen as small blocks or tissue punches on dry ice, liquid N2, dry ice-chilled isopentane, or on liquid nitrogen-chilled isopentane.

The cohort included 101 men and 92 women, the mean age at death was 63.9 ± 16.4 years, and the PMD ranged from 0.5 to 101 hours (mean, 17.9 ± 16.4 hours). A frequency distribution with fitted normal distribution curve showed that most of the cases had a PMD of shorter than 24 hours (Fig. 1). The cohort included 12 cases of Alzheimer disease (AD), 11 cases of amyotrophic lateral sclerosis, 66 cases of multiple sclerosis, 24 cases of Parkinson disease (PD) including 3 with dementia and 10 with concomitant AD pathology, 1 case of Guillain-Barré syndrome, 3 cases of multiple system atrophy, 2 cases of stroke (1 with multiple cerebral infarcts, 1 with left medial cerebral artery infarct), 2 cases of progressive supranuclear palsy, and 72 cases with no neurological condition. Subcohorts were used for various aspects of this study.

FIGURE 1

Frequency histogram of postmortem delay (PMD) with fitted normal distribution curve. Most cases studied (n = 161) had a PMD of shorter than 24 hours; 21 cases had a PMD between 1 and 2 days; 11 cases had a PMD longer than 2 days.

FIGURE 1

Frequency histogram of postmortem delay (PMD) with fitted normal distribution curve. Most cases studied (n = 161) had a PMD of shorter than 24 hours; 21 cases had a PMD between 1 and 2 days; 11 cases had a PMD longer than 2 days.

For the anatomical regional study of RNA quality, tissue samples from the superior frontal gyrus (gray and white matter), corpus callosum, thalamus, and cerebellum (gray and white matter) were obtained from 10 cases. The cohort included 5 men and 5 women, the mean age at death was 71.7 ± 10.4 years, and the PMD ranged from 3 to 25 hours (mean, 18 hours). Brain region selection was based on tissue accessibility and on prior studies. To determine the best region for routine screening, the mean and SDs were assessed for each region to determine the region with the highest and least variability in RIN values. A correlation analysis was conducted using the Pearson product-moment correlation test among the regions to determine whether RNA variability was similar from region to region.

To study the effects of AM variables on RNA quality, superior frontal gyrus tissue samples were obtained from 64 brains. These came from 5 different tissue banks: Amsterdam (n = 4); Barcelona (n = 4); Budapest (n = 41); Edinburgh (n = 12); and London (n = 3). The cohort included 40 men and 24 women, the mean age at death was 60.8 ± 18.4 years; and the PMD range was from 30 minutes to 71 hours (mean, 15 hours).

To investigate effects of RIN values and AM and postmortem variables on the mRNA levels of 7 common housekeeping genes, 43 samples from the AM study were used. The range of RIN values for this cohort was between 4.6 and 9.2 (mean, 6.7 ± 1.0); PMDs ranged from 0.5 to 71 hours (mean, 15 hours).

Antemortem and Postmortem Variables

Information on common quantifiable AM and postmortem variables that could influence RNA quality outcome was obtained from each brain bank using a questionnaire (Table 1). For the AM variables, the information was retrospectively gathered by the respective tissue bank manager from hospital notes, nursing homes, or the next of kin. Postmortem variables were recorded by tissue bank staff members. An SG2 pH meter with an InLab Solids electrode (Mettler Toledo, Leicester, UK) was used to measure tissue pH. Tissue pH measurements are done by inserting the electrode into the white matter of the medial frontal gyrus. We initially took several measurements from various regions and did not detect any significant regional variability in pH (data not shown).

TABLE 1

Antemortem and Postmortem Variables Investigated

Antemortem 1 For how long was the patient hospitalized before death? (terminal illness) 
 2 Was the patient in a coma before death? For how long? 
 3 Did the patient have a respiratory illness just before death? 
 4 Was the patient artificially ventilated? For how long? 
Postmortem 5 Which month did the death occur? 
 6 How long was the body left at room temperature after death? 
 7 How long was the body in cold storage after death? 
 8 Length of time from death to autopsy 
 9 Length of time from autopsy to tissue freezing 
 10 How was the tissue transported, that is, cold storage vs ambient temperature? 
 11 How long did it take to dissect the brain? 
 12 CSF pH at time of autopsy 
 13 Brain pH at time of dissection 
 14 Postmortem delay 
 15 Freezing methods 
 16 Freezer interval—duration of storage before RNA extraction from tissue 
Antemortem 1 For how long was the patient hospitalized before death? (terminal illness) 
 2 Was the patient in a coma before death? For how long? 
 3 Did the patient have a respiratory illness just before death? 
 4 Was the patient artificially ventilated? For how long? 
Postmortem 5 Which month did the death occur? 
 6 How long was the body left at room temperature after death? 
 7 How long was the body in cold storage after death? 
 8 Length of time from death to autopsy 
 9 Length of time from autopsy to tissue freezing 
 10 How was the tissue transported, that is, cold storage vs ambient temperature? 
 11 How long did it take to dissect the brain? 
 12 CSF pH at time of autopsy 
 13 Brain pH at time of dissection 
 14 Postmortem delay 
 15 Freezing methods 
 16 Freezer interval—duration of storage before RNA extraction from tissue 

Total RNA Extraction and Purification

Total RNA was extracted from dissected snap-frozen tissue (<100 mg) using the RNeasy tissue lipid mini kit (Qiagen Ltd, Crawley, UK) according to the manufacturer's instructions, and was kept at −80°C until further use. RNA concentration and purity were assessed by spectrophotometry (NanoDrop ND1000; NanoDrop Technologies, Wilmington, DE). RNA integrity was further assessed using an Agilent 2100 Bioanalyzer and lab-on-a-chip platform technology (Agilent Technologies UK Ltd, West Lothian, UK). Sample concentration, the 28S/18S ribosomal ratio, and the RIN values were automatically calculated with the system software. This system assigns samples an RIN ranging from 1 (lowest quality) to 10 (best quality) (24). This standardized measure of RNA integrity does not identify mRNA or quality of in situ mRNA (25).

Quantification of mRNA Expression by RT-qPCR

The 2-step real-time reverse transcriptase quantitative polymerase chain reaction (RT-qPCR) was performed using the QuantiTect reverse transcription kit, the QuantiTect SYBR Green kit, and with QuantiTect primer assays (Qiagen Ltd). For cDNA synthesis, 1 µg of total RNA from each sample was reverse transcribed according to the manufacturer's instructions using the QuantiTect reverse transcription kit with integrated removal of genomic DNA contamination. Real-time PCR experiments were performed using the Mx3000P real-time PCR system with software version 4.01 (Stratagene, La Jolla, CA). For each sample, 25-µL reactions were set up in duplicate, with each reaction containing 12.5 (µL of QuantiTect SYBR Green, 2.5 (µL of QuantiTect primer assay (a mix of forward and reverse primers), 9 (µL of RNAse-free water, and 1 (µL template cDNA (complementary DNA). The QuantiTect primer assays used are listed in Table 2. The sequences of the Qiagen primers are proprietary information of Qiagen. The thermal cycling conditions for samples and internal controls included an initial activation step at 95°C for 15 minutes, followed by 40 cycles of denaturation (94°C, 15 seconds), annealing (55°C, 30 seconds), amplification (72°C, 30 seconds), and a final melting curve analysis with a ramp from 55°C to 95°C. Fluorescence data collection was performed during the annealing step. Samples with no reverse transcriptase reaction to test for contaminating DNA and a negative control containing no RNA template were introduced in each run. To control for variation in RNA levels caused by manipulation, the same cDNA stock was used for all the experiments, and all RT-qPCR assays for a particular gene were undertaken at the same time in a 96-well plate under identical conditions. For all RT-qPCR assays, the expression levels of target genes were normalized to the levels of a reference gene (beclin-1, RRN18S, or β-actin) and calibrated using a standard curve method for quantitation. The calibrator was generated by creating an RNA pool from all the samples or baseline group (control group). Levels of the calibrator represent the baseline with a value of 1 from which all RNA expression were calculated. The standard curve was used to determine relative quantity expression values for each target gene after RT-qPCR analysis of each test specimen. Relative expression values for each target gene are expressed as a ratio of target gene expression level to the reference gene expression level in the same specimen. Finally, mRNA level checks were conducted to ensure that observed variations in expression levels were caused by biologic sample changes and did not correlate with the amount of template RNA synthesized.

TABLE 2

Primers Used for RT-PCR Quantitation of mRNA*

Gene Symbol Official Gene Name Entrez Gene ID Accession Catalogue Number 
ACTB β-actin 60 NM_001101 QT01680476 
B2M β-2-Microglobulin 567 NM_004048 QT00088935 
BECN1 Beclin 1, autophagy related 8678 NM_003766 QT00004221 
GAPDH Glyceraldehyde-3 -phosphate dehydrogenase 2597 NM_002046 QT01192646 
HIP1A Hypoxia-inducible factor 1, α subunit 3091 NM_001530 QT00083664 
PPIA Peptidylprolyl isomerase A (aka cyclophilin A) 5478 NM_021130 QT01669542 
RRN18S 18S ribosomal RNA  X03205 QT00199367 
TUBB β-Tubulin 203068 NM_178014 QT00089775 
Gene Symbol Official Gene Name Entrez Gene ID Accession Catalogue Number 
ACTB β-actin 60 NM_001101 QT01680476 
B2M β-2-Microglobulin 567 NM_004048 QT00088935 
BECN1 Beclin 1, autophagy related 8678 NM_003766 QT00004221 
GAPDH Glyceraldehyde-3 -phosphate dehydrogenase 2597 NM_002046 QT01192646 
HIP1A Hypoxia-inducible factor 1, α subunit 3091 NM_001530 QT00083664 
PPIA Peptidylprolyl isomerase A (aka cyclophilin A) 5478 NM_021130 QT01669542 
RRN18S 18S ribosomal RNA  X03205 QT00199367 
TUBB β-Tubulin 203068 NM_178014 QT00089775 
*

Details of the primers are at the Qiagen GeneGlobe Search Centre (www.qiagen.com/GeneGlobe).

Statistical Analysis

The Pearson product-moment correlation test (2-tailed) and linear regression were applied to determine the relationship between variables. A multiple regression analysis was conducted to assess contribution and effect of all variables. An individual t-test (independent t-test; 2-tailed) was used to determine the group difference and an analysis of variance with post hoc Bonferroni multiple comparison test (2-tailed) for multiple group comparison. The following software packages were used GraphPad Prism 5.01 (GraphPad Software, La Jolla, CA), SPSS 15.0 (SPSS Inc., Chicago, IL), and Microsoft Office Excel 2007 (Microsoft UK Headquarters, Reading, UK). The data are presented as mean ± SE unless otherwise stated. Differences were considered statistically significant if p < 0.05.

RESULTS

RNA Quality Assessment

The RIN values collected from the 193 cases ranged from 2.9 to 9.2 (mean, 6.8 ± 1.0). The frequency distribution histogram showed that most cases had RIN values close to the mean, and only a small percentage (14.5%) had an RIN value less than 6 (Fig. 2).

FIGURE 2

Frequency histogram of RNA integrity number (RIN) with fitted normal distribution curve. The RIN values ranged from 2.9 to 9.2; most cases (66%) had an RIN value greater than 6.5, the point at which RNA is generally considered to be of acceptable quality for molecular research studies.

FIGURE 2

Frequency histogram of RNA integrity number (RIN) with fitted normal distribution curve. The RIN values ranged from 2.9 to 9.2; most cases (66%) had an RIN value greater than 6.5, the point at which RNA is generally considered to be of acceptable quality for molecular research studies.

Regional Study

There was no significant difference in RNA quality observed between the 6 brain regions based on the mean from 10 cases (Fig. 3). Nevertheless, some cases showed more variability in RNA quality than others (Table 3), and some differences in RNA quality among cases based on mean RIN values across regions were also detected. Control and PD cases showed less variability in RIN values than multiple sclerosis cases. This suggests that sampling from any region would offer satisfactory overall RNA quality evaluation of the entire brain based on the mean RIN value for routine brain RNA quality assessment.

FIGURE 3

RNA quality across 6 brain regions. A 1 -way within-subject analysis of variance with post hoc Bonferroni multiple comparison test was conducted to assess RNA quality (RNA integrity number) across 6 brain regions. No significant difference in RNA quality was detected. Means and SEM are plotted. GM, gray matter; SFG, superior frontal gyrus; WM, white matter.

FIGURE 3

RNA quality across 6 brain regions. A 1 -way within-subject analysis of variance with post hoc Bonferroni multiple comparison test was conducted to assess RNA quality (RNA integrity number) across 6 brain regions. No significant difference in RNA quality was detected. Means and SEM are plotted. GM, gray matter; SFG, superior frontal gyrus; WM, white matter.

TABLE 3

RNA Integrity Numbers of Samples From Different Brain Anatomical Regions

Brain Areas (n = 6) 
Case (n = 10) SFG GM SFG WM Corpus Callosum Thalamus Cerebellum GM Cerebellum WM Mean SD 
8.00 7.90 7.30 8.00 7.90 7.50 7.77 0.2687 
7.00 6.90 7.60 7.40 7.20 7.00 7.18 0.2478 
6.20 6.60 7.30 6.30 6.70 5.70 6.47 0.4922 
7.20 6.90 6.30 6.40 6.20 6.10 6.52 0.3976 
7.70 7.30 7.60 7.30 7.60 8.00 7.58 0.2409 
6.40 6.60 7.20 7.00 6.50 6.80 6.75 0.2814 
7.00 7.20 6.90 7.00 7.20 8.00 7.22 0.3670 
5.80 6.00 7.20 7.40 7.00 7.80 6.87 0.7272 
6.80 6.80 5.40 6.20 5.10 7.40 6.28 0.8133 
10 7.80 6.90 5.90 7.30 6.60 6.10 6.77 0.6574 
Mean 6.990 6.910 6.870 7.030 6.800 7.040   
SD 0.7187 0.4999 0.7514 0.5755 0.7888 0.8422   
Brain Areas (n = 6) 
Case (n = 10) SFG GM SFG WM Corpus Callosum Thalamus Cerebellum GM Cerebellum WM Mean SD 
8.00 7.90 7.30 8.00 7.90 7.50 7.77 0.2687 
7.00 6.90 7.60 7.40 7.20 7.00 7.18 0.2478 
6.20 6.60 7.30 6.30 6.70 5.70 6.47 0.4922 
7.20 6.90 6.30 6.40 6.20 6.10 6.52 0.3976 
7.70 7.30 7.60 7.30 7.60 8.00 7.58 0.2409 
6.40 6.60 7.20 7.00 6.50 6.80 6.75 0.2814 
7.00 7.20 6.90 7.00 7.20 8.00 7.22 0.3670 
5.80 6.00 7.20 7.40 7.00 7.80 6.87 0.7272 
6.80 6.80 5.40 6.20 5.10 7.40 6.28 0.8133 
10 7.80 6.90 5.90 7.30 6.60 6.10 6.77 0.6574 
Mean 6.990 6.910 6.870 7.030 6.800 7.040   
SD 0.7187 0.4999 0.7514 0.5755 0.7888 0.8422   

GM, gray matter; SFG, superior frontal gyrus; WM, white matter.

AM Study

The 4 AM variables measured were duration of hospitalization, duration of coma, occurrence of respiratory illness, and duration of artificial ventilation. A modest significant difference in RNA quality was found between cases involving hospitalization before death and those without (t = 2.872; p < 0.01; 2-tailed), suggesting that RNA quality was better with no hospitalization before death (Fig. 4). No significant difference or effect on RNA quality was detected for the other individual variables (Figure, Supplemental Digital Content 1, ).

FIGURE 4

Effect of hospitalization before death on RNA quality. There was a significant difference in RNA integrity number (RIN) between cases that were hospitalized before death (Yes) and those that were not (No) (t = 2.872; p = 0.006; 2-tailed). Mean RIN value was slightly higher when there was no hospitalization before death (*p < 0.05).

FIGURE 4

Effect of hospitalization before death on RNA quality. There was a significant difference in RNA integrity number (RIN) between cases that were hospitalized before death (Yes) and those that were not (No) (t = 2.872; p = 0.006; 2-tailed). Mean RIN value was slightly higher when there was no hospitalization before death (*p < 0.05).

Further analysis was conducted to determine the cumulative effects of the AM variables. The AM agonal events (duration of hospitalization, coma duration, occurrence of respiratory illness, and duration of artificial ventilation) were summed for each case (range, 0–4). There was a significant negative correlation between AM agonal events and RIN (r = −0.335; p < 0.01; Fig. 5A). The number of AM agonal events was used to divide the cases into 3 groups; 0 AM events, 1 AM events, and 2+ AM events. There were only 2 cases with 3 AM events and no case with 4. One-way analysis of variance showed that the number of AM agonal events had a significant effect on RNA quality (F2,62 = 5.36; p < 0.01; Fig. 5B). Mean RIN value was 7.1 ± 0.13 when no AM agonal events were recorded (n = 36) and decreased to 6.45 ± 0.23 (n = 17; p < 0.01) for 1 event and to 6.34 ± 0.28 (n = 11; p < 0.01) for 2 or more events. Finally, using a 6.5 RIN value as a cutoff point above which RNA is considered as acceptable for gene expression profiling or RT-qPCR, more than 70% of the cases had acceptable RNA quality when no events were recorded, but this dropped to 36.36% when 2 or more events were recorded (Fig. 5C).

FIGURE 5

Effect of antemortem (AM) agonal events on RNA quality. (A) There was a significant negative correlation between AM agonal events and RNA integrity number (RIN) (r= -0.335; p = 0.007). (B) A 1 -way between-subject analysis of variance with post hoc Bonferroni multiple comparison test was conducted on the 3 groups to assess the impact of AM agonal events on RIN. There was a significant gradual decrease of RIN with increasing number of AM agonal events recorded (F2,61 = 5.36, p = 0.007). (C) When no AM events were recorded, more than 70% had an RIN value greater than 6.5 (**p < 0.01; ***p < 0.001).

FIGURE 5

Effect of antemortem (AM) agonal events on RNA quality. (A) There was a significant negative correlation between AM agonal events and RNA integrity number (RIN) (r= -0.335; p = 0.007). (B) A 1 -way between-subject analysis of variance with post hoc Bonferroni multiple comparison test was conducted on the 3 groups to assess the impact of AM agonal events on RIN. There was a significant gradual decrease of RIN with increasing number of AM agonal events recorded (F2,61 = 5.36, p = 0.007). (C) When no AM events were recorded, more than 70% had an RIN value greater than 6.5 (**p < 0.01; ***p < 0.001).

The response to hypoxia is known to be mediated principally through the activation of hypoxia-inducible factor la (HIF-1α). Hypoxia-inducible factor 1α expression levels are very low in normal oxygen conditions. Hypoxia increases HIF-la levels by stabilizing expression levels of HIF-1α, which is usually rapidly degraded via the proteasome (26). Hypoxia-inducible factor la mRNA levels were investigated in a subset of samples from the previously mentioned cohort. There was a significant increase (100%) in HIF-1α mRNA levels in cases with 1 or more AM agonal events compared with cases with no such events (p < 0.05; Fig. 6A). Furthermore, using β-actin as a reference gene for normalization, there was a significant negative correlation between HIF-1α mRNA levels and RNA integrity number (r = −0.477; p < 0.05; Fig. 6B).

FIGURE 6

Hypoxia-inducible factor 1α (HIF-1α) mRNA expression. (A) There was a significant increase in HIF-1α mRNA levels in cases with 1 or more antemortem (AM) events (t = 2.468; p < 0.05). (B) Hypoxia-inducible factor 1α mRNA levels correlated with RNA quality. Increased levels of HIF-1α coincided with lower RNA integrity number (*p < 0.05).

FIGURE 6

Hypoxia-inducible factor 1α (HIF-1α) mRNA expression. (A) There was a significant increase in HIF-1α mRNA levels in cases with 1 or more antemortem (AM) events (t = 2.468; p < 0.05). (B) Hypoxia-inducible factor 1α mRNA levels correlated with RNA quality. Increased levels of HIF-1α coincided with lower RNA integrity number (*p < 0.05).

Postmortem Study

All postmortem variables measured and investigated are listed in Table 4. These include both demographic information and as many of the steps from the time of death to the freezer storage interval for which quantitative data could be obtained. RNA quality was slightly higher in men compared with women (p < 0.05; Fig. 7A). There was a significant negative relationship between age at death and RNA quality (r = −0.252; p < 0.001; n = 193; Fig. 7B). Postmortem RNA quality also varied slightly across the different disease groups (F5,186 = 6.53; p < 0.0001; Fig. 7C). Furthermore, postmortem brain pH was significantly correlated with RTN values (r = 0.598; p < 0.05; n = 16), suggesting that lower brain pH is linked to lower RNA quality (Fig. 7D). There was no significant effect of PMD, disease duration, freezer storage duration, CSF pH, freezing methods, or the individual tissue bank on RNA quality. Furthermore, within the PD group (RIN, 4.5–7.9), the presence of a-synuclein-positive inclusions in the superior frontal gyrus did not have an impact on RNA quality. Finally and most importantly, no effect on RNA quality was found for all the variables measured from time at death to freezer storage, such as month of death, time at room temperature after death, duration of cold storage of the body, duration from death to autopsy, duration from autopsy to freezer, tissue transport condition, and dissection duration (Figure, Supplemental Digital Content 1, ).

TABLE 4

Effect of Postmortem Variables on RNA Quality as Determined by RNA Integrity Number

Variables Test p (2-Tailed) No. Samples 
Postmortem delay, hours r = 0.037 0.605 193 
Sex (men, n = 101; women, t = 2.313* 0.022 193 
n = 93)    
Age at death, years r= −0.252** <0.001 193 
Tissue bank (n = 9) F8,184= 1.322 0.235 193 
Freezing method (n = 4) F3,189 = 0.1373 0.938 193 
Diagnostic groups (n = 6) F5,187 = 6.53*** >0.0001 184 
Disease duration r = 0.092 0.368 97 
Freezer duration, months r = 0.0839 0.260 182 
Month of death (n = 9)† F8,53 = 1.319 0.254 64 
Duration at room temperature r = −0.202 0.112 63 
after death, hours    
Cold storage duration of the r = 0.157 0.238 58 
body, hours    
Duration from death to r = 0.136 0.286 63 
autopsy, hours    
Duration from autopsy to r= -0.017 0.891 64 
freezer, hours    
Tissue transport condition t= 1.034 0.305 60 
(CS vs AT)    
Dissection duration, hours r = 0.040 0.751 64 
CSF pH r = 0.101 0.336 91 
Brain pH r = 0.598* 0.014 16 
Variables Test p (2-Tailed) No. Samples 
Postmortem delay, hours r = 0.037 0.605 193 
Sex (men, n = 101; women, t = 2.313* 0.022 193 
n = 93)    
Age at death, years r= −0.252** <0.001 193 
Tissue bank (n = 9) F8,184= 1.322 0.235 193 
Freezing method (n = 4) F3,189 = 0.1373 0.938 193 
Diagnostic groups (n = 6) F5,187 = 6.53*** >0.0001 184 
Disease duration r = 0.092 0.368 97 
Freezer duration, months r = 0.0839 0.260 182 
Month of death (n = 9)† F8,53 = 1.319 0.254 64 
Duration at room temperature r = −0.202 0.112 63 
after death, hours    
Cold storage duration of the r = 0.157 0.238 58 
body, hours    
Duration from death to r = 0.136 0.286 63 
autopsy, hours    
Duration from autopsy to r= -0.017 0.891 64 
freezer, hours    
Tissue transport condition t= 1.034 0.305 60 
(CS vs AT)    
Dissection duration, hours r = 0.040 0.751 64 
CSF pH r = 0.101 0.336 91 
Brain pH r = 0.598* 0.014 16 

Test value, p values and n for postmortem variables are indicated. Significant variables are in bold type.

***

Correlation significant at the 0.0001 level.

**

Correlation significant at the 0.001 level.

*

Correlation significant at the 0.05 level.

No deaths occurred in August, and only 1 death each occurred in July and October. The analysis of variance test did not include these months; therefore, n = 9.

AT, autopsy to freezing duration; CS, cold storage duration.

FIGURE 7

Effect of gender, age at death, disease group, and brain pH on RNA quality. (A) Men showed slightly higher RNA integrity numbers (RIN) than women (6.96 ± 0.09 vs 6.64 ±0.1; t= 2.31 3; p = 0.02). (B) There was a modestly significant effect of age at death on RNA quality; cases with age at death greater than 70 years had slightly lower RIN values (r = -0.252; p = 0.0004). (C) There was a significant effect of disease group on RIN (F5,186 = 6.53; p < 0.0001). (D) RNA quality was significantly lower in more acidic brains (r= 0.598; p = 0.0144). AD, Alzheimer disease; ALS, amyotrophic lateral sclerosis; MS, multiple sclerosis; PD, Parkinson disease.

FIGURE 7

Effect of gender, age at death, disease group, and brain pH on RNA quality. (A) Men showed slightly higher RNA integrity numbers (RIN) than women (6.96 ± 0.09 vs 6.64 ±0.1; t= 2.31 3; p = 0.02). (B) There was a modestly significant effect of age at death on RNA quality; cases with age at death greater than 70 years had slightly lower RIN values (r = -0.252; p = 0.0004). (C) There was a significant effect of disease group on RIN (F5,186 = 6.53; p < 0.0001). (D) RNA quality was significantly lower in more acidic brains (r= 0.598; p = 0.0144). AD, Alzheimer disease; ALS, amyotrophic lateral sclerosis; MS, multiple sclerosis; PD, Parkinson disease.

Effects of AM and Postmortem Variables on Total and Individual mRNA Levels

There was no significant effect of PMD on total RNA yield (micrograms per milligram) (Figure, Supplemental Digital Content 1, ). To determine whether there was an effect of AM or postmortem variables on individual mRNA levels, the effect of RNA quality on mRNA levels from 7 reference genes commonly used for validating candidate genes isolated from microarray studies was investigated. A highly significant positive relationship between RIN and mRNA levels was observed for all common reference genes except for 18S ribosomal RNA (rRNA), showing that lower RNA quality correlated with reduced mRNA levels (Figs. 8A–G). There was also a significant negative correlation between the expression levels of 2 reference genes and AM events (β-actin, r = −0.312, p < 0.05, n = 43; cyclophilin A, r = −0.417, p < 0.001, n = 43) (Figs. 8H, I, respectively). Only a trend was seen for beclin-1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and no significant correlation was found for β2-microglobulin, β-tubulin, or 18S rRNA, suggesting that only some mRNAs are directly affected by increased levels of hypoxia, as indicated by the number of AM agonal events. Finally, only 18S rRNA was affected by PMD, that is, lower mRNA levels were correlated with longer PMDs (r = −0.48, p < 0.001, n = 43; Fig. 8J).

FIGURE 8

Expression levels of 7 common reference genes. Except for 18S ribosomal RNA (rRNA) (F), all other reference genes significantly correlated with RNA integrity number (RIN), suggesting that mRNA levels are RIN dependent (A–E, G). β-actin and cyclophilin A expression levels decreased with cumulative antemortem agonal events (H, I, respectively); 18S rRNA levels were affected by postmortem delay (J). The most stable gene based on coefficient of variance was 18S rRNA, followed by beclin-1 (K). GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

FIGURE 8

Expression levels of 7 common reference genes. Except for 18S ribosomal RNA (rRNA) (F), all other reference genes significantly correlated with RNA integrity number (RIN), suggesting that mRNA levels are RIN dependent (A–E, G). β-actin and cyclophilin A expression levels decreased with cumulative antemortem agonal events (H, I, respectively); 18S rRNA levels were affected by postmortem delay (J). The most stable gene based on coefficient of variance was 18S rRNA, followed by beclin-1 (K). GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

When the levels of the constitutively expressed 7 reference genes were ranked from 1 (the most stable) to 7 (the least stable) using coefficient of variance, 18S rRNA was the most stable, followed by beclin-1, and then in increasing ranking order β-actin (third), β2-microglobulin (fourth), GAPDH (fifth), cyclophilin A (sixth), and β-tubulin (seventh) (Fig. 8K). The Microsoft Excel Visual Basic application geNorm is another commonly used method to identify the most stable reference genes for normalization (27), and with this method, it was determined that beclin-1, GAPDH, and β-actin were the most stable reference genes, followed by cyclophilin A, β2-microglobulin, 18S rRNA, and β-tubulin. Thus, both approaches indicate that beclin-1 is and that β-tubulin is not a stable reference gene in human brain tissue. The methods do disagree, however, with respect to 18S rRNA. Normalization using an internal reference gene is known to reduce or diminish intrakinetic and interkinetic RT-PCR (sample-to-sample and run-to-run) variations (28, 29). Expression results of GAPDH and β2-microglobulin were normalized using 18S rRNA or beclin-1 or β-actin, the most stable reference genes. mRNA expression levels of β2-microglobulin and GAPDH remained RIN dependent using the most stable reference gene (18S rRNA; Figs. 9A, C) but disappeared with beclin-1 or β-actin (Figs. 9B, C).

FIGURE 9

β2-Microglobulin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression levels were normalized with 18S ribosomal RNA (rRNA) and beclin-1 and β-actin. Expression levels of both housekeeping genes remained RNA integrity number (RIN) dependent with 18S rRNA (A, C, respectively); this effect disappeared with beclin-1 and β-actin (B, D).

FIGURE 9

β2-Microglobulin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression levels were normalized with 18S ribosomal RNA (rRNA) and beclin-1 and β-actin. Expression levels of both housekeeping genes remained RNA integrity number (RIN) dependent with 18S rRNA (A, C, respectively); this effect disappeared with beclin-1 and β-actin (B, D).

DISCUSSION

The main objective of this investigation was to determine measurable factors that have the potential to predict or impact the quality of RNA extracted from human brain tissue obtained from multiple laboratories within the BrainNet Europe brain bank network. These predictors can assist brain banks in providing tissue of acceptable quality for molecular studies of neurodegenerative and neuropsychiatric diseases.

Our major findings suggest that accumulation of deleterious AM agonal events such as hospitalization, coma, artificial respiration, or respiratory illness before death is associated with decreases in RNA quality and increasing evidence of hypoxia. Therefore, we suggest that these 4 AM variables represent good predictors of mRNA quality. RNA quality is generally better when no AM events have been recorded and when hypoxia is minimal. Brain pH was also found to be a good indicator of RNA quality. Surprisingly, none of the postmortem variables that are under tissue bank control had a significant effect on RNA quality. Furthermore,

RNA quality did not affect total RNA yield. Most reference gene mRNA levels investigated, however, were found to be RIN dependent, but normalization eliminated that effect except for 18S rRNA; based on RIN assessment, not all mRNAs in the human brain seem to deteriorate in a similar manner. Therefore, we recommend using RT-qPCR testing for the most appropriate internal reference genes in each experimental setting. In addition, because the absolute value of correction might differ with different RIN values, comparative studies should be carried out with samples of similar RIN values to minimize effects of RNA quality variations.

Because a prolonged agonal state may influence RNA quality, the effect of agonal state on brain tissue quality has been of some concern for molecular neuroscience research. The success of microarray experiments is entirely dependent on the quality of RNA (30–33), although meaningful results can still be generated from moderately degraded RNA (34). Tissue pH and near-death physiological stress have been shown to influence expression pattern (13, 20, 35). In general, slower modes of death have been associated with lower brain tissue pH (11, 14–16, 20) and decreased RNA integrity (2, 12). In addition, tissue from subjects with a neurological disorder had a lower pH compared with controls who died suddenly (36).

Collecting information on the mode of death and agonal state events before death in a practical and efficient way is very important for brain banks. We determined that information about any hospitalization, coma, artificial ventilation, and/or respiratory illness would suffice to summarize the agonal state of an individual before death. It is likely that there are other AM events that could affect tissue quality, but for which information is only rarely recorded. The simple questionnaire that we used was not time consuming, and general practitioners and hospital or nursing home staff were willing to find answers to the questions when they were contacted. Our results clearly indicate that the best-quality RNA for biomolecular studies is more likely to be obtained when none of these events have occurred, but eliminating all cases where at least one of these AM events had occurred would have eliminated all of the cases with a neurological condition, leaving only the control subjects. The study cohort had a large PMD range, but PMD did not have a significant impact on RNA quality and, therefore, was unlikely to have been a confounding variable for this study of AM variables. Our results suggest that acceptable RNA quality can be obtained from cases with no or 1 agonal event before death. Thus, although our results provide further evidence that AM agonal occurrences can be confounding variables for RNA quality, records of these events can assist in the selection of acceptable brain tissue.

Brain tissue pH can be measured from homogenized tissue or directly using a portable pH meter reader with a sensor. We found that the latter method provides a good indication of RNA quality, that is, lower brain tissue pH was associated with lower RIN values. By contrast, CSF pH did not correlate with RNA quality, as in previous studies (2, 27). Tissue pH has been found to be lower after prolonged death (20). It is, therefore, becoming common practice to use brain pH as a marker of hypoxia indicative of an agonal state, and we recommend that measurement of brain pH by brain banks be adopted as standard practice.

Because only small tissue samples are required for RNA extraction for ascertaining quality, we assessed RNA quality in multiple brain regions and found no significant variation in RNA quality among 6 brain areas in 10 cases. Some cases did, however, show more variability across regions than others. Thus, RNA preservation may vary to some extent in some brains, and the study of a single region may not necessarily be representative of the whole brain in some cases. Sampling regions affected by specific pathological processes may yield different results in different regions in some brains. This further supports the need for analyzing each RNA sample using a Bioanalyzer before any further study of the RNA (37). Finally, although no statistical significant difference was found for mean RIN values and intersample variability, we suggest that the superior frontal gyrus (in which no difference in RNA quality was detected between white and gray matter) and the cerebellar white matter be sampled for routine testing.

Our finding that none of the postmortem factors studied affected RNA quality is reassuring and may be explained by the fact that the BrainNet Europe member laboratories have already optimized various stages of tissue retrieval, including methods for collection, transport, dissection and freezing methods, and storage. Although no obvious difference in freezing method on RNA quality was observed, freezing small blocks of tissue with liquid nitrogen-chilled isopentane has become the freezing method of choice for optimum RNA and protein preservation. Postmortem delay has always been considered a major problem for studies of human tissues, but this may be unfounded for human brain RNA. The lack of effect of PMD on RNA quality has also been found in previous studies (22, 23, 38–41). In contrast, a number of studies have reported a significant negative correlation between PMD and RTN values, particularly studies that used a greater range of PMD values than in the current study (2, 10–12, 17). All studies concluded, however, that PMD represents only a weak indicator of RNA quality. It should be noted that the studies that showed a significant correlation between PMD and RNA quality used RIN values provided by the Bioanalyzer rather than by standard agarose gel electrophoresis, suggesting that the Bioanalyzer is appropriate for future RNA quality studies. In contrast to their lack of effect on RNA quality, postmortem variables such as PMD (42, 43) or storage temperature (44) have been shown to have an impact on brain protein preservation. The impact of postmortem variables on protein integrity has been recently reviewed (37).

No effects of AM and postmortem variables were found on total RNA yield, in line with other investigations (23). The knowledge that RNA integrity in general is not affected by postmortem variables is clearly not an indication that the expression levels of individual mRNA species can be confidently studied in postmortem brains. Our results show that mRNA levels of some common reference genes are affected by agonal states, and that most, but not all, expression levels were RTN dependent. This effect can be eliminated with appropriate normalization, however. In general, expression levels of most reference genes showed a strong linear correlation between each other (data not shown), as reported by others (40). Nevertheless, to carry out appropriate normalization in real-time PCR, suitable reference genes are essential for generating biologically relevant results (45). There are concerns about some of the more traditional reference genes when investigating mRNA expression levels in the human brain (46). For example, GAPDH is mostly known for its role in glycolysis, but recent reports have demonstrated increased GAPDH activity associated with apoptosis and neurodegeneration (47). Reports have confirmed abnormal expression of GAPDH in several neurodegenerative diseases such as AD, PD, and Huntington disease (48). Increased levels of beclin have been associated with neurons at the site of traumatic brain injury (49), and lower levels were observed in AD brains (50). Similarly, β-actin mRNA levels were differentially expressed in AD brains compared with controls (51). These studies suggest that the identification of a universal human reference gene might prove more challenging than initially thought and that novel organ-specific reference genes may be more appropriate. Thus, because we demonstrated that not all brain mRNA seem to deteriorate in the same manner based on RTN assessment, verification of expression variability is warranted before an internal control is applied in each experimental setting.

In summary, none of the variables from the time of death to adequate storage facilities had a major effect on RNA quality, suggesting that optimal tissue handling is already in place across the BrainNet Europe II network of brain banks. The only factors that impact RNA quality were recorded AM events occurring before death. These factors are not under the control of the tissue banks, but data collected on these events will prove useful to tissue banks in the assessment of RNA quality. Our 4-item questionnaire showed clearly that an increasing number of AM events significantly correlated with a decrease in RNA quality and increased levels of hypoxia, and that brain tissue pH was a good indicator of RNA quality. Finally, we have shown that not all mRNA species degraded at the same rate in the human brain, raising caution regarding the use of internal reference genes for normalization.

REFERENCES

1
Kretzschmar
H
.
Brain banking: Opportunities, challenges and meaning for the future
.
Nat Rev Neurosci
 
2009
;
10
:
70
78
2
Webster
MJ
.
Tissue preparation and banking
.
Prog Brain Res
 
2006
;
158
:
3
14
3
Vonsattel
JP
Amaya M del
P
Cortes
EP
et al
Twenty-first century brain banking: Practical prerequisites and lessons from the past: The experience of New York Brain Bank, Taub Institute, Columbia University
.
Cell Tissue Bank
 
2008
;
9
:
247
58
4
Bell
JE
Alafuzoff
I
Al-Sarraj
S
et al
Management of a twenty-first century brain bank: Experience in the BrainNet Europe consortium
.
Acta Neuropathol
 
2008
;
115
:
497
507
5
Ravid
R
.
Standard operating procedures, ethical and legal regulations in BTB (Brain/Tissue/Bio) banking: What is still missing?
Cell Tissue Bank
 
2008
;
9
:
151
67
6
Haroutunian
V
Pickett
J
.
Autism brain tissue banking
.
Brain Pathol
 
2007
;
17
:
412
21
7
Grinberg
LT
Ferretti
RE
Farfel
JM
et al
Brain bank of the Brazilian aging brain study group—a milestone reached and more than 1,600 collected brains
.
Cell Tissue Bank
 
2007
;
8
:
151
62
8
Trevino
V
Falciani
F
Barrera-Saldana
HA
.
DNA microarrays: A powerful genomic tool for biomedical and clinical research
.
Mol Med
 
2007
;
13
:
527
41
9
Hoheisel
JD
.
Microarray technology: Beyond transcript profiling and genotype analysis
.
Nat Rev Genet
 
2006
;
7
:
200
10
10
Bauer
M
Gramlich
I
Polzin
S
et al
Quantification of mRNA degradation as possible indicator of postmortem interval—a pilot study
.
Leg Med (Tokyo)
 
2003
;
5
:
220
27
11
Harrison
PJ
Heath
PR
Eastwood
SL
et al
The relative importance of premortem acidosis and postmortem interval for human brain gene expression studies: Selective mRNA vulnerability and comparison with their encoded proteins
.
Neurosci Lett
 
1995
;
200
:
151
54
12
Barton
AJ
Pearson
RC
Najlerahim
A
et al
Pre- and postmortem influences on brain RNA
.
J Neurochem
 
1993
;
61
:
1
11
13
Tomita
H
Vawter
MP
Walsh
DM
et al
Effect of agonal and postmortem factors on gene expression profile: Quality control in microarray analyses of postmortem human brain
.
Biol Psychiatry
 
2004
;
55
:
346
52
14
Chevyreva
I
Faull
RL
Green
CR
et al
Assessing RNA quality in postmortem human brain tissue
.
Exp Mol Pathol
 
2008
;
84
:
71
77
15
Hardy
JA
Wester
P
Winblad
B
et al
The patients dying after long terminal phase have acidotic brains: Implications for biochemical measurements on autopsy tissue
.
J Neural Transm
 
1985
;
61
:
253
64
16
Johnston
NL
Cervenak
J
Shore
AD
et al
Multivariate analysis of RNA levels from postmortem human brains as measured by three different methods of RT-PCR
.
Stanley Neuropathology Consortium. J Neurosci Methods
 
1997
;
77
:
83
92
17
Lipska
BK
Deep-Soboslay
A
Weickert
CS
et al
Critical factors in gene expression in postmortem human brain: Focus on studies in schizophrenia
.
Biol Psychiatry
 
2006
;
60
:
650
58
18
Papapetropoulos
S
Shehadeh
L
McCorquodale
D
.
Optimizing human post-mortem brain tissue gene expression profiling in Parkinson's disease and other neurodegenerative disorders: From target “fishing” to translational breakthroughs
.
J Neurosci Res
 
2007
;
85
:
3013
24
19
Bahn
S
Augood
SJ
Ryan
M
et al
Gene expression profiling in the post-mortem human brain—no cause for dismay
.
J Chem Neuroanat
 
2001
;
22
:
79
94
20
Mexal
S
Berger
R
Adams
CE
et al
Brain pH has a significant impact on human postmortem hippocampal gene expression profiles
.
Brain Res
 
2006
;
1106
:
1
11
21
Stan
AD
Ghose
S
Gao
XM
et al
Human postmortem tissue: What quality markers matter?
Brain Res
 
2006
;
1123
:
1
11
22
Cummings
TJ
Strum
JC
Yoon
LW
et al
Recovery and expression of messenger RNA from postmortem human brain tissue
.
Mod Pathol
 
2001
;
14
:
1157
61
23
Ervin
JF
Heinzen
EL
Cronin
KD
et al
Postmortem delay has minimal effect on brain RNA integrity
.
J Neuropathol Exp Neurol
 
2007
;
66
:
1093
99
24
Schroeder
A
Mueller
O
Stacker
S
et al
The RIN: an RNA integrity number for assigning integrity values to RNA measurements
.
BMC Mol Biol
 
2006
;
7
;
3
25
Fleige
S
Pfaffl
MW
.
RNA integrity and the effect on the real-time qRT-PCR performance
.
Mol Aspects Med
 
2006
;
27
:
126
39
26
Dery
MA
Michaud
MD
Richard
DE
.
Hypoxia-inducible factor 1: Regulation by hypoxic and non-hypoxic activators
.
Int J Biochem Cell Biol
 
2005
;
37
:
535
40
27
Vandesompele
J
De Preter
K
Pattyn
F
et al
Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes
.
Genome Biol
 
2002
;
3
:[RESEARCH0034]
28
Fleige
S
Waif
V
Huch
S
et al
Comparison of relative mRNA quantification models and the impact of RNA integrity in quantitative real-time RT-PCR
.
Biotechnol Lett
 
2006
;
28
:
1601
13
29
Wittwer
CT
Herrmann
MG
Moss
AA
et al
Continuous fluorescence monitoring of rapid cycle DNA amplification
.
Biotechniques
 
1997
;
22
:
130
31
, [34–38]
30
Copois
V
Bibeau
F
Bascoul-Mollevi
C
et al
Impact of RNA degradation on gene expression profiles: Assessment of different methods to reliably determine RNA quality
.
J Biotechnol
 
2007
;
127
:
549
59
31
Soverchia
L
Ubaldi
M
Leonardi-Essmann
F
et al
Microarrays—the challenge of preparing brain tissue samples
.
Addict Biol
 
2005
;
10
:
5
13
32
Buesa
C
Maes
T
Subirada
F
et al
DNA chip technology in brain banks: Confronting a degrading world
.
J Neuropathol Exp Neurol
 
2004
;
63
:
1003
14
33
Hynd
MR
Lewohl
JM
Scott
HL
et al
Biochemical and molecular studies using human autopsy brain tissue
.
J Neurochem
 
2003
;
85
:
543
62
34
Schoor
O
Weinschenk
T
Hennenlotter
J
et al
Moderate degradation does not preclude microarray analysis of small amounts of RNA
.
Biotechniques
 
2003
;
35
:
1192
96
, [8–201]
35
Li
JZ
Vawter
MP
Walsh
DM
et al
Systematic changes in gene expression in postmortem human brains associated with tissue pH and terminal medical conditions
.
Hum Mol Genet
 
2004
;
13
:
609
16
36
Yates
CM
Butterworth
J
Tennant
MC
et al
Enzyme activities in relation to pH and lactate in postmortem brain in Alzheimer-type and other dementias
.
J Neurochem
 
1990
;
55
:
1624
30
37
Ferrer
I
Martinez
A
Boluda
S
et al
Brain banks: Benefits, limitations and cautions concerning the use of post-mortem brain tissue for molecular studies
.
Cell Tissue Bank
 
2008
;
9
:
181
94
38
Weis
S
Llenos
IC
Dulay
JR
et al
Quality control for microarray analysis of human brain samples: The impact of postmortem factors, RNA characteristics, and histopathology
.
J Neurosci Methods
 
2007
;
165
:
198
209
39
Miller
CL
Diglisic
S
Leister
F
et al
Evaluating RNA status for RT-PCR in extracts of postmortem human brain tissue
.
Biotechniques
 
2004
;
36
:
628
33
40
Preece
P
Cairns
NJ
.
Quantifying mRNA in postmortem human brain: Influence of gender, age at death, postmortem interval, brain pH, agonal state and inter-lobe mRNA variance
.
Brain Res Mol Brain Res
 
2003
;
118
:
60
71
41
Yasojima
K
McGeer
EG
McGeer
PL
.
High stability of mRNAs postmortem and protocols for their assessment by RT-PCR
.
Brain Res Brain Res Protoc
 
2001
;
8
:
212
18
42
Li
X
Greenwood
AF
Powers
R
et al
Effects of postmortem interval, age, and Alzheimer's disease on G-proteins in human brain
.
Neurobiol Aging
 
1996
;
17
:
115
22
43
Liu
X
Brun
A
.
Synaptophysin immunoreactivity is stable 36 h postmortem
.
Dementia
 
1995
;
6
:
211
17
44
Ferrer
I
Santpere
G
Arzberger
T
et al
Brain protein preservation largely depends on the postmortem storage temperature: Implications for study of proteins in human neurologic diseases and management of brain banks: A BrainNet Europe Study
.
J Neuropathol Exp Neurol
 
2007
;
66
:
35
46
45
Barrachina
M
Castano
E
Ferrer
I
.
TaqMan PCR assay in the control of RNA normalization in human post-mortem brain tissue
.
Neurochem Int
 
2006
;
49
:
276
84
46
Huggett
J
Dheda
K
Bustin
S
et al
Real-time RT-PCR normalisation; strategies and considerations
.
Genes Immun
 
2005
;
6
:
279
84
47
Sirover
MA
.
New insights into an old protein: The functional diversity of mammalian glyceraldehyde-3-phosphate dehydrogenase
.
Biochim Biophys Acta
 
1999
;
1432
:
159
84
48
Chuang
DM
Hough
C
Senatorov
VV
.
Glyceraldehyde-3-phosphate dehydrogenase, apoptosis, and neurodegenerative diseases
.
Annu Rev Pharmacol Toxicol
 
2005
;
45
:
269
90
49
Cao
Y
Klionsky
DJ
.
Physiological functions of Atg6/Beclin 1: A unique autophagy-related protein
.
Cell Res
 
2007
;
17
:
839
49
50
Lee
JA
Gao
FB
.
Regulation of Abeta pathology by beclin 1: A protective role for autophagy?
J Clin Invest
 
2008
;
118
:
2015
18
51
Ruan
W
Lai
M
.
Actin, a reliable marker of internal control?
Clin Chim Acta
 
2007
;
385
:
1
5

Supporting Information

Author notes

Pascal F. Durrenberger and Shama Fernando contributed equally to the study.
This study was supported by the European Commission under the Sixth Framework Programme (Brain Net Europe II, LSHM-CT-2004-503039). The UK Multiple Sclerosis and UK Parkinson's Disease Tissue Banks were supported by the Multiple Sclerosis Society of Great Britain and Northern Ireland and the Parkinson's Disease Society, respectively.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article in the journal's Web site (http://jnen.oxfordjournals.org/).