OUP user menu

Postmortem Delay Has Minimal Effect on Brain RNA Integrity

John F. Ervin BA, Erin L. Heinzen PhD, Kenneth D. Cronin BS, David Goldstein PhD, Mari H. Szymanski RN-C, James R. Burke MD, PhD, Kathleen A. Welsh-Bohmer PhD, and, Christine M. Hulette MD
DOI: http://dx.doi.org/10.1097/nen.0b013e31815c196a 1093-1099 First published online: 1 December 2007


The Bryan Alzheimer Disease Research Center obtains postmortem human brain tissue from patients with Alzheimer disease (AD) and cognitively normal control subjects for molecular and genetic research programs. A growing body of research suggests that variations in gene transcript levels may contribute to the onset and progression of disease. Identifying how the regulation of gene expression may affect AD requires the use of high-quality mRNA from banked human brains. The present study was conducted to establish the quality and suitability of available banked brain tissue for future gene expression studies. We chose 32 AD cases with Braak stage IV, V, or VI. These AD cases were matched to 36 normal control cases by age and sex when possible. Multiple regions from each brain were sampled, including frontal cortex, temporal cortex, occipital cortex, and cerebellum. Hippocampus was also available for study from 14 control cases. A comparison of several antemortem and postmortem variables, such as postmortem interval, agonal state, ventricular cerebrospinal fluid pH, and cause of death were analyzed. RNA was isolated from at least 1 area from every brain and most brains yielded intact RNA from all regions tested. Analysis of the clinical variables did not reveal any features that correlated with the ability to recover intact mRNA. We conclude that undegraded mRNA may be isolated from most brain regions many hours postmortem and that neither the pH of ventricular fluid nor postmortem interval is predictive of mRNA integrity.

Key Words
  • Human brain
  • mRNA
  • Postmortem delay


Evidence from genetic research studies suggests that variations in gene transcript levels may contribute to the onset and progression of Alzheimer disease (AD) (1). To comprehensively investigate how transcriptional regulation may contribute to disease we are using microarray and quantitative real-time polymerase chain reaction (PCR) investigations to measure expression levels. A critical step in accurately quantifying transcription levels is to determine the quality of mRNA available from banked human brain samples.

The Kathleen Price Bryan Brain Bank, established in 1985, is a multidisciplinary endeavor that successfully procures human brain tissue from enrolled donors all over the United States. The tissue is retrieved using established protocols and is subsequently used for a variety of neurodegenerative disease research projects (1-3). Approximately 60 specimens are collected each year. Banked specimens generally consist of 1 cerebral hemisphere fixed in formalin and 1 hemisphere frozen and stored at −80°C (4). Consequently, a large number of well-characterized human brains are available for current and future research.

Definition of the variables affecting postmortem integrity of biomolecules is becoming increasingly important as research expands to include a variety of methods. To maximize the use of the tissue available for future molecular and genetic research, we were interested in determining whether agonal state, postmortem interval (PMI), or prolonged storage at −80°C had an adverse effect on RNA integrity.

Materials and Methods

Brain Tissue Collection

Human brain tissue samples were selected from the Kathleen Price Bryan Brain Bank. Donor enrollment procedures have been described previously (4,5) and have been approved by the Duke University Institutional Review Board. At the time of death, autopsy was performed according to institutional guidelines and in accordance with Centers for Disease Control and Prevention universal precautions to prevent the spread of infectious disease (6). Controls and AD donor brains were examined by a neuropathologist (C.M.H.) and diagnosed according to the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) (7) and National Institute on Aging-Reagan Institute criteria (8). Neurofibrillary change was staged according to Braak (9).

Brain tissue samples from cognitively normal control subjects (n = 36) and AD subjects (n = 32) were used. There were 17 males and 19 females in the control group. The age range for the control group was 56 to 97 with a mean of 82 years. The AD group consisted of 8 males and 24 females. The AD group's age range was 79 to 95 with a mean of 83.6 years. CERAD-definite AD subjects were Braak stage IV (n = 5), V (n = 22), or VI (n = 5). The controls were normal (CERAD 1A; n = 20) or had few to moderate neuritic plaques (CERAD 1B; n = 16). In control cases the Braak stages were I (n = 21), II (n = 9), and III (n = 6). Apolipoprotein E (ApoE) genotype was determined from a sample of cerebellum taken at the time of autopsy as described previously (10). The ApoE genotypes and other demographic information for each case are shown in the Table.

View this table:

Routine preparation of banked brain tissue for future research studies was begun immediately after death and included placing bags of wet ice on the subject's head as soon as possible and during transport to the autopsy suite. Placing bags of ice on the face caused some minor cosmetic distortion to the face. Therefore, in January 2000, this procedure was altered and the ice was placed under the subject's head. In addition, the brain tissue was kept cool with wet ice during the dissection process. After dissection the tissue was transported immediately on wet ice to the laboratory for long-term storage in a −80°C freezer. Brain tissue was then quickly frozen between chilled metal plates stored at −80°C, as previously described (2). Brain tissue was collected with a PMI ranging from 40 minutes to 30 hours. The time required for postmortem tissue processing, including tissue dissection and transportation to the laboratory for long term storage ranged from 30 to 60 minutes. The length of time each specimen was kept frozen varied between 8 months and 19 years. Once frozen, each specimen remained frozen until the RNA extraction procedure had begun. Frozen samples from frontal, temporal, hippocampal, occipital and cerebellar brain regions were used for RNA analysis.

RNA Extraction and Assessment of Quality

Total RNA was isolated from approximately 30 mg of tissue ground in liquid nitrogen (mortar and pestle). Total RNA was extracted from brain tissue using the RNeasy Lipid Tissue Purification Kit (Qiagen, Valencia, CA) according the manufacturer's instructions as described previously by Heinzen et al (11). RNA was quantified spectrophotometrically at 260 nm. RNA quality was assessed using standard agarose gel electrophoresis (glyoxal gel protocol; Ambion, Austin, TX). The amount of intact RNA was assessed using densitometric quantitation of the 28S, 18S, and degradation products on the gel. Traditionally, the 28S/18S area ratio has been used to assess the total RNA quality. However, this approach does not quantify the presence of degradation products well. To better quantify the amount of intact RNA we measured the areas of 28S and 18S bands and areas of degradation present between the 28S and 18S bands and degradation products after the 18S ribosomal band. Intact RNA was calculated using the following equation:

Embedded Image

Figure 1 shows representative highest and lowest quality samples obtained in this study and their corresponding percentage of intact RNA values for comparison. RNA from each sample (1 μg) was reverse transcribed into cDNA using a High Capacity cDNA Synthesis Kit (Applied Biosystems, Foster City, CA) according to the product instructions. A TaqMan assay was used in real-time PCR studies to quantitate β-actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA transcript expression (Applied Biosystems). β-Actin and GAPDH mRNA expressions were examined by TaqMan real-time PCR assays in 2 regions (temporal cortex and cerebellum). All real-time PCR assays were purchased through Applied Biosystems and run per the manufacturer's protocol using 10 to 20 μg of total RNA (converted to cDNA) per reaction. Standard curves were run for a range of total RNA (converted to cDNA, 9 standards ranging from 0.01 to 30 ng). The number of cycles necessary to reach a threshold fluorescence reading (CT) was plotted against log RNA amount and a line was fitted to the data points using linear regression. The following equation was used to quantitate the number of transcripts present in an experimental sample:

Embedded Image

Representative RNA agarose gels for samples with highest and lowest quality RNA obtained in this study and their corresponding percentages of intact RNA values for comparison.

In this equation, CT, Sample is the number of cycles to reach the fluorescence threshold for a given sample, and y0 and α are, respectively, the intercept and slope of the line defining the relationship of CT versus log (RNA amount) determined from the standard curve.

Information on the pH of postmortem ventricular fluid was available for 26 subjects. The pH of postmortem ventricular fluid samples was determined after the fluid had been transported to the laboratory and centrifuged to remove any debris. The pH was measured within 30 minutes after autopsy and before the fluid was placed in the freezer for long-term storage. The pH of postmortem ventricular fluid samples ranged from 6.26 to 8.0. Antemortem clinical variables including fever or sepsis, drug use, provision of supplemental oxygen in the agonal period, and temporal events surrounding death such as sudden death or protracted illness had been documented. Complete autopsy data were available for 10 AD subjects and 13 control subjects.

Data Analysis

ImageJ Software (National Institutes of Health) was used for densitometric quantitation. Analysis of variance or linear regression analysis was used to assess associations between RNA integrity and age, gender, ApoE status, PMI, cerebrospinal fluid pH, and brain tissue structure. Student t-tests were used to compare AD samples with control samples. All statistical analyses were performed using SigmaStat (version 2; SPSS, Inc., Chicago, IL).


RNA integrity in stored autopsy brain tissue samples was determined in 36 cognitively normal control subjects with a PMI ranging from 1 hour to 30 hours and 32 AD subjects with a PMI ranging from 40 minutes to 30 hours. The tissue had been stored at −80°C for periods ranging from 8 months to 19 years. We extracted mRNA from multiple regions including frontal cortex, temporal cortex (TC), occipital cortex (OC), and cerebellum (CB). Because of limited tissue availability, all 4 regions were not examined for several cases. Four controls were missing 1 region, 2 CB, 1 OC, and 1 TC; therefore, only 3 regions were examined. One control case was missing 2 regions, TC and OC, so that only 2 regions were examined. Two AD cases were missing 1 region each, 1 TC and 1 OC. In addition to these 4 main regions, hippocampal samples were examined in 14 controls. A total of 152 samples were examined from control cases and a total of 126 samples were examined from the AD cases. Thus, 278 samples were examined from 68 autopsy brains.

Approximately 0.8 μg of RNA was extracted from each milligram of frozen brain tissue. There was no effect of anatomic location on yield. For example, cerebellar samples had the same yield as neocortical and hippocampal samples (data not shown). As a measure of integrity of specific mRNA, β-actin and GAPDH mRNA expressions were examined by TaqMan real-time PCR assay in the TC and CB regions (Fig. 2). There were no differences in β-actin or GAPDH expressions in either TC or CB in relation to PMI (data not shown).


RNA agarose gels from subjects with extreme intervals from 45 minutes to 30 hours postmortem. CB, cerebellum; FC, frontal cortex; HP, hippocampus; OC, occipital cortex; TC, temporal cortex.

Comparisons using available clinical data such as disease status, cause of death, ApoE genotype, PMI, ventricular fluid pH, and length of time in frozen storage were evaluated for correlations with RNA integrity. Data on pH of the postmortem ventricular fluid was available for 24 cases; 9 of these were controls and 15 were AD subjects. There were no apparent correlations found between these variables and the percentage of intact RNA (Fig. 3).


(A) Combined controls and AD subjects comparing PMI and percentage of intact RNA. (B) Combined controls and AD subjects comparing age and percentage of intact RNA. (C) Subjects with postmortem ventricular fluid pH data and percentage of intact RNA. (D) Combined controls and AD subjects comparing storage time in years and percentage of intact RNA. Recovery of RNA from different brain regions from a single individual varied slightly. Data from all available regions were averaged for each subject.

In contrast, there was a slight inverse relationship between age and the pH of postmortem ventricular fluid (Fig. 4). Although age of onset of AD is associated with certain ApoE allelic frequency, no trends associated with the ApoE genotype and pH were identified (data not shown).


Comparison of age and postmortem ventricular fluid pH shows a slight inverse relationship.

Complete autopsy data were available for 23 cases: 10 AD cases and 13 controls. Pneumonia (n = 12), myocardial infarction (n = 3), and cancer (n = 4) were the most common causes of death found at autopsy. Four deaths were due to urinary tract infection, dehydration, and cardiopulmonary arrest. Twelve of the cases with complete autopsy data (6 controls and 6 AD) also had data about the pH of the postmortem ventricular fluid available. Cerebrospinal fluid pH for cases with complete autopsy ranged from 6.26 to 8.0 and did not correlate with agonal state or cause of death.


Many investigators desire brain tissue that has been frozen with a very short PMI for research studies, but this criterion excludes many banked human brains from use. The present study was designed to examine a spectrum of matched tissues for correlation between mRNA integrity and antemortem and/or postmortem variables. Therefore, we designed an experiment that could function both as a quality assurance tool and provide supportive evidence to encourage the utilization of larger numbers of banked specimens for research. We set out to establish whether there are technical variables that influence the usefulness of banked human brain tissue.

Our findings support a growing number of published studies, summarized in reviews (12-14), which demonstrate that acceptable PMI for human autopsy brain tissue destined to be used for a variety of biomolecular studies is much longer than commonly perceived by the basic science community. Our findings, along with those of others (15,16), demonstrate that it is reasonable and practical to use human postmortem tissue, sometimes obtained with a prolonged PMI, for a variety of investigative techniques.

Other studies also suggested that the manner in which the specimens are collected and stored for biomolecule studies may negate some of the concerns about tissue pH and agonal period (17-19). Because tissue acidity has been thought to correlate with impaired biomolecule stability and prolonged agonal state, most investigators recommend postmortem tissue pH that approximates in vivo pH. Thus, the pH of the brain specimens examined here, ranging from 6.26 to 8.0, would be considered quite good by many experts. Nevertheless, some RNA was degraded, indicating that pH is not always a reliable measure of tissue integrity. In contrast, we have shown here that postmortem ventricular fluid pH does not appear to correlate with RNA integrity or agonal events.

Similar to the findings of Harrison et al (18), we also identified a slight inverse trend between age at death and pH of postmortem tissue. Although Harrison et al analyzed postmortem homogenized tissue and we analyzed postmortem ventricular fluid, our data suggest that postmortem tissue pH decreases as age increases.

We also believe that the procedures used routinely at this institution to harvest and store tissue aid in biomolecule preservation. The rate and manner in which the body and head is cooled after death, the cooling of the brain tissue during the dissection process, and rapid freezing could have a great influence on biomolecule integrity and stability. In addition to refrigeration of the body before autopsy, donors had their heads cooled with wet ice during transport to the autopsy suite. The brain specimen was further cooled with wet ice after removal from the cadaver and before coronal slicing and dissection. Coronal slicing routinely begins at the frontal pole and proceeds toward the occipital pole. The cerebellum and brainstem are removed from the cortex and placed in ice while each coronal slice is cut and bagged before they are placed on ice. The occipital lobe is the last region to be bagged and put on ice for transport to the laboratory freezer. Cooling of the frontal cortex and cerebellum in ice first may partially explain the why the frontal and cerebellum have the highest 28S/18S ratio, which suggests higher quality than the temporal cortex, hippocampus, and occipital lobe (Fig. 5).


Average 28S/18S ratio for all samples for each region tested. CB, cerebellum; FC, frontal cortex; Hippo, hippocampus; OC, occipital cortex; TC, temporal cortex.

Most subjects and regions tested yielded a ratio or percentage of intact mRNA that could be used in research. Recovery of RNA from different brain regions from a single individual varied slightly (detailed data not shown). This variation indicates that there was not a global problem with mRNA integrity in that individual subject. Interestingly, others reported similarly that the frontal lobe tissues of the brain were better preserved compared with posterior regions. Schuenke and Gelman (20) have reported that the frontal and temporal lobes were better than the occipital lobe or cerebellum for acceptable microglia cell culture in postmortem brain. These authors speculate that the position of the body at autopsy allows the blood to drain from the anterior regions and pool in the posterior regions, resulting in inadequate postmortem microglia cell recovery and viability. We propose that delayed cooling and/or storage of these regions postmortem may also be a contributing factor.

After dissection and transport to the laboratory, tissue is then allowed to freeze quickly between precooled metal plates as recommended by Vonsattel et al (21). Brains are routinely stored at -80°C for very long periods. We have confirmed preliminary work by this group (2) and others (22) and demonstrated that RNA can be recovered from postmortem human brain. We have further confirmed that intact RNA can be recovered from brains stored at -80°C for 19 years. The selection of this storage temperature may also be an important factor in the preservation of this material. Previous studies (22) have indicated that the time between death and refrigeration of the body may be important. They also suggested that the time between opening of the cranial cavity and completion of tissue dissection may be important. Therefore, it is imperative that all procedures for handling the cadaver after death be performed quickly and efficiently.

Although we have not addressed protein recovery in this study, Ferrer et al (23) have recently demonstrated that brain protein preservation depends upon storage temperature. We conclude that postmortem tissue processing and storage may be as important as antemortem events when postmortem human brain tissue is prepared for research.

In addition, PMI is important but is not the only factor with an impact on biomolecular integrity. We have found that intact mRNA may be obtained from human brain tissue many hours postmortem. Autopsy performed up to 30 hours after death does not have an adverse effect on mRNA integrity. Importantly, RNA integrity in postmortem human brain tissue may depend in part on anatomical region; so that frontal cortex may be more likely to yield intact RNA than occipital lobe. All cases examined here appeared to yield intact RNA suitable for genetic screening and linkage analysis from at least 1 area of the brain. However, preservation of specific biomolecular and neurochemical entities may require enhanced procurement and tissue preservation efforts.


The authors are grateful to the families of the Duke Autopsy Program participants. This work would not be possible without their generosity. We also thank Steve Conlon for expert help with preparation of the figures. This work was presented in abstract form at the XVI International Congress of Neuropathology, San Francisco, CA, Sept. 13, 2006 and published in Brain Pathology 2006;16:S125.


  • This work was supported by National Institute on Aging Grants P50 AG05128 and P30 AG028377 NIA and by GlaxoSmithKline.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
View Abstract