Indian J Med Res 129, February 2009, pp 189-199

Biochemical analysis of protein stability in human brain collected at different post-mortem intervals Ramesh Chandana, R.B. Mythri, Anita Mahadevan*, S.K. Shankar* & M.M. Srinivas Bharath

Departments of Neurochemistry &* Neuropathology, National Institute of Mental Health & Neuro Sciences Bangalore, India

Received November 21, 2007 Background & objectives: Rising prevalence of neurodegenerative disorders with a steady increase in aged-population necessitates studies of the human brain to understand their pathophysiology. As animal models are not available, medical centers have established “brain banks” to provide autopsy brain samples for such research. Frozen tissues must be of optimal quality to permit molecular and protein studies. Post-mortem interval (PMI) is an important factor affecting tissue quality although its effects on brain physiology are unclear. We undertook this study to analyze the biochemical effects of PMI on protein stability in human brains collected at autopsy and stored at the brain bank of a tertiary care neurosciences institute in south India. Methods: Different neuroanatomical areas including frontal cortex (FC), cerebellum (CB), caudate nucleus (CD) and substantia nigra (SN) from autopsy human brains (n=9) with varying PMI (4-18 h) were analyzed for pH, protein insolubility, protein oxidation/ nitration and protein expression of glial fibrillary acidic protein (GFAP), synatophysin and neurofilament (NF). Histological changes at different PMI were also assessed. Results: An increase in tissue pH was noted with increasing PMI. Although there was no significant alteration in solubility of proteins, SN showed increased protein oxidation/nitration events, GFAP and NF expression with increasing PMI. No major abnormalities in cell morphology or tissue integrity were noted. Immunohistochemistry with GFAP and NF did not show any significant increase in signal in FC at high PMI. Interpretation & conclusions: In post-mortem human brains, although there were no gross structural changes at the tissue level with increasing PMI, biochemical events such as oxidative and nitrosative damage of cellular proteins, tissue pH could be considered as markers of tissue quality for biochemical research. Further, SN was found to be most susceptible to PMI related changes. Key words Glial fibrillary acidic protein - histology - human brain - neurofilament - 3-nitrotyrosine - post-mortem interval - protein oxidation - tissue pH - Western blot

has necessitated detailed studies of pathogenesis to evolve therapeutic strategies. Though animal experimentation has allowed major advances in

Increase in geriatric population in the society associated with pathophysiological changes in the nervous system and neurodegenerative disorders 189

190

INDIAN J MED RES, FEBRUARY 2009

understanding neurological disorders, it has become clear that extrapolation to human system is not realistic because of species barrier, diversity in anatomy, physiology, biochemistry and genetics, except for some phenomenological similarities. Therefore, medical centers have initiated “brain banks” to collect and store human nervous tissue and body fluids for research1. Human Brain Tissue Repository for Neurobiological Studies (HBTR-Human Brain Bank) has been established in 1995 in India at the department of Neuropathology, National Institute of Mental Health and Neuro Sciences (NIMHANS), Bangalore, as a national research facility to collect human nervous tissue and tissue fluids and provide to researchers1. Such brain banks subserve the function of evolving a database on standardized samples and also to carry out “question-directed” studies to find clinically translatable answers. The use of brain tissue collected at autopsy for proteomic study of human neurobiology has increased with the Human Proteome Organization-Brain Proteome Project (HUPO-BPP) an international consortium aimed to promote exchanging raw data between groups and co-ordinate proteomic activities using the human brain2,3. But for successful application of proteomic technologies, the biochemical, molecular and structural integrity of the tissue should be well sustained. Premortem factors like metabolic state of the deceased, infections, seizures, hypoxia, and consumption of toxic substances/drugs, terminal agonal state after death, post-mortem delay in extracting the brain, storage environmental temperature form critical modulators. Traditionally, a low post-mortem interval (PMI) has been a feature of high tissue quality and reliability of human brain4-8. Recently, tissue pH and RNA quality have also been introduced as quality markers9. Effect of post-mortem delay is important while studying post-translational modifications in human degenerative diseases. Proteomic studies of brain proteins to evaluate post-mortem changes have been carried out in animal models10 but similar studies in human brain are scarce. To facilitate detailed proteomic studies and to validate the nature of the brain tissue stored at HBTR, the present study was carried out, to evaluate the effect of PMI (the time interval between the time of death to the time the brain is collected at autopsy, sliced and stored at -70 0C in a freezer) on the biochemical events in the brain. Material & Methods All the chemicals and solvents used were of analytical grade. Bulk chemicals and solvents were

obtained from Merck & Co. Inc (Whitehouse Station, NJ, USA). Anti-neurofilament (phosphorylated H and M) mouse monoclonal antibody (clone NE-14), anti-glial fibrillary acidic protein (GFAP) mouse monoclonal antibody (clone GA-5), anti-synaptophysin mouse monoclonal antibody (clone Snp88) were obtained from Biogenex (San Ramon, CA, USA). Nitrocellulose membrane was obtained from Millipore (Billerica, MA, USA). Horse radish peroxidase conjugated secondary antibodies were obtained from Bangalore Genei (Bangalore, Karnataka, India). Antinitrotyrosine antibody (NT), anti-dinitrophenyl (DNP) antibody and protease inhibitor cocktail were procured from Sigma (Eugene, OR, USA). Tissue samples: Tissue samples were collected from subjects between 25-35 yr to exclude the effect of age on biochemical parameters studied. Within 1 h of death, the body was transferred to a refrigerator maintained at 2-40C with a recorder with uninterrupted power supply. Following medico-legal autopsy with inquest from police, with no prejudice to the medico-legal evaluation, brains were collected, with written informed consent from next of kin to use the material for research/teaching purposes. The Institutional Scientific Ethics Committee has approved the study protocol. The victims had no history of neurological disorder prior to death, as indicated in the records. The brains were sliced coronally and kept flat on salt-ice mixture (-15 to -180C) during dissection and then transferred in plastic zip lock bags into a box to be stored at -800C in HBTR. The procedure of dissection took 3045 min and the brain slices were transferred immediately into the deep freezer. Frozen brain samples from nine subjects (age 30 ± 5 yr, non-alcoholics, non-diabetics, not on any medication) with no visible signs of brain damage, from four anatomical areas: frontal cortex (superior frontal gyrus) (FC), cerebellum (CB), caudate nucleus (head) (CD) and substantial nigra (linear darkstrip above crus cerebri, both sides at the level of superior colliculus) (SN) were collected for biochemical study. The mirror image bits fixed in buffered formalin were processed for histological evaluation. Preparation of protein extracts: Approximately 100 mg of frozen tissue from each sample was dissected and immediately transferred to cold 1x phosphate buffered saline (PBS) containing protease inhibitor cocktail (Sigma, Eugene, OR, USA) and manually homogenized on ice (15 strokes). The homogenate was sonicated (20 sec on ice) in a Sonics-vibra cell sonicator (Sonics and Materials Inc, CT, USA). Total protein in the sonicated samples was estimated by

CHANDANA et al: PROTEIN ANALYSIS IN POST-MORTEM HUMAN BRAINS

Bradford method11, aliquoted and stored at -80 0C until further processing. For preparation of triton-soluble extracts, the sonicate from each sample was diluted with 1x PBS containing protease inhibitor and Triton X-100 (Loba Chemie, Mumbai, India) (1% final concentration) up to 200 µl at 5 µg/µl protein concentration and mixed thoroughly for complete solubilization12. The solubilized material was centrifuged at 15,000 g (15 min, 4 0C) and 10 µl of supernatant (Triton-soluble fraction) was loaded onto 12 per cent SDS PAGE. Similarly, the pellet (Triton-insoluble fraction) was solubilized in 10 µl SDS PAGE loading dye, boiled and loaded onto a 12 per cent SDS PAGE followed by staining with coomassie brilliant blue R-250 (Sigma, USA). From corresponding soluble and insoluble fractions, band intensities were quantified by a densitometric scanner (Bio-Rad laboratories, Hercules, CA, USA) and (insoluble fraction)/ (insoluble + soluble fraction) ratio, which is a gross indicator of the level of insoluble proteins, was calculated. SDS PAGE and Western blot: Brain total protein extracts (approximately 50 µg/lane) were run on 12 per cent SDS PAGE at 100 V for about 2.5 h and gels were stained with coomassie brilliant blue R-25013. For Western blot, proteins from SDS gels were electrophoretically transferred to nitrocellulose membranes in a semi-dry apparatus (Sree Maruthi Scientific Works, Bangalore, India) (2 h at 125 mA)13. Non-specific binding was blocked by incubating membranes in 1x PBS containing Tween-20 and 5 per cent skimmed milk powder (Nandini Milk Products, Bangalore, India) (for 1 h at room temperature or overnight at 4 °C). The membranes were incubated with primary antibodies diluted in PBS/Tween-20 containing 5 per cent BSA for 1.5 h at room temperature. Blots were washed with PBS/Tween-20 and then incubated for 1.5 h at room temperature with HRP-conjugated secondary antibodies (in PBS/Tween-20 containing 5% BSA). Membranes were washed with PBS/Tween-20 and the immune reaction was visualized by the colour reaction developed in 1x PBS containing diamino benzidine (DAB) [1mg/ml (w/v)] (Sigma) and 0.1 per cent H2O213. Anti-tubulin (Calbiochem, San Diego, CA, USA) westerns were used as internal control. To detect endogenous protein nitration, equal amounts of protein (100 µg) from different brain samples were spotted in triplicate onto a nitrocellulose membrane. The membrane was washed with PBS/

191

Tween-20 followed by Western blot with polyclonal anti-3-NT antibody. Band intensities in westerns were quantified by a densitometric scanner and the values normalized against the respective anti-tubulin signal. Estimation of protein carbonyls (oxyblot): Oxyblots were carried out based on the protocol published earlier14. Brain protein extract prepared as indicated was centrifuged (14,000 g/ 15 min/ 40C). The supernatant at 4 mg/ml protein concentration was derivatized by dinitrophenyl hydrazine (DNPH) (Sisco Research Laboratories Pvt. Ltd., Mumbai, India) (in 50% sulphuric acid) in a final reaction volume of 20 µl in the presence of 12 per cent SDS for 20 min at room temperature. The reaction was stopped by addition of neutralization buffer (2M Tris in 30% glycerol). Sample (5µl) was spotted in triplicate onto a nitrocellulose membrane. The membrane was washed with PBS/Tween-20 followed by Western blot with polyclonal anti-DNP antibody. pH determination: pH determination in brain tissue with FC as representative for each brain was carried out based on the method described earlier9. Approximately 150 mg of frozen frontal cortex tissue from different subjects was homogenized in 5 ml of saline (pH=7.0) and centrifuged (3 min/8000 ×g/4°C). The pH of the supernatant was measured in duplicate with a pH meter (Control Dynamics Instrumentation, Bangalore, India) that was previously calibrated with known standards. Histological study: The paraffin embedded tissue sections from four different anatomical areas from all the subjects were stained with haematoxylin-eosin (HE), Nissl and Luxol fast blue for myelin. The sections were examined blind to post-mortem delay for myelin pallor, neuronal staining character and anoxic changes if any. The sections from lowest PMI (ID B-183) and longest PMI (ID T-144) were immunostained with antibodies to GFAP (Glial marker, 1:200) and neurofilament (axonal marker, 1:100) to note if there is any difference in staining character. Appropriate positive and negative controls were incorporated during immunostaining. The antigen retrieval was carried out by microwaving at level III, in citrate buffer followed by standard immunoperoxidase technique, with DAB/H2O2 as chromogen. Statistical analyses: Quantitative data wherever mentioned were accumulated from at least three independent experiments. Differences between mean values were analyzed by one-way analysis of variance (ANOVA).

192

INDIAN J MED RES, FEBRUARY 2009

Table I. Details of brain samples with increasing PMI collected from road traffic accident victims-Human Brain Bank (n=9: average age 30 ± 5 yr) Sample ID

B-183 T-149 B-180 B-178 T-146

B-182 T-150 T-148 T-144

Age (yr)/sex Interval between Interval from time Total interval Type of injury Regions studied head injury to of death to freezing from injury to death brain (post mortem freezing brain at delay) –800C 27/F 2 h 45 min 4h 6 h 45 min Poly trauma, no brain injury FC, CB, CD 25/M 12 h 30 min 7 h 30 min 20 h SDH, Lt parietal, Rt. frontal FC, CB, CD, SN contusion 25/M 13 h 7 h 30 min 20 h 30 min Acute SDH-evacuated cerebral FC, CB, CD oedema. CT-brainstem hypodensity 35/M 30 h 15 min 12 h 42 h 15 min Lt. Temporo-parietal EDH-drained FC, CB, CD cerebral oedema 33/M 11 h 14 h 30 min 25 h 30 min Acute SDH-evacuated Lt fronto FC, CB, CD, SN temporal contusion, cerebral oedema 25/M 17 h 30 min 14 h 30 min 32 h High cervical cord injury-sudden FC, CB, CD, SN cardiac arrest, Brain normal 30/M 3h 16 h 30 min 19 h 30 min Medullary tear. Normal brain FC, CB, CD, SN 35/M 8h 17 h 25 h Right fronto- temporal SDH, FC, CB, CD, SN cerebral oedema 25/M 27 h 45 min 18 h 45 h 45 min Cervical cord injury, Brain normal, FC, CB, CD, SN Sudden cardiac arrest

All deaths occurred during January-April 2007, between 18:45 - 05:30 h. Ambient temp. 10-22oC. Body shifted to mortuary freezer within 60 min after death. Mortuary freezer maintained at 2-4 0C with temperature recording. At emergency service for the injury victim oxygen saturation was maintained by intubation and oxygen mask. All received only antioedema measure fluid correction and terminally one vial of adrenaline, atropine and steroid when indicated. None of the victims were alcoholics/diabetic/not on any medication. SDH, subdural haematoma; EDH, extradural haematoma; FC, frontal cortex; CB, cerebellum; CD, caudate nucleus; SN, substantia nigra.

Results Biochemical analysis of FC, CB, CD and SN regions from nine human brain samples collected at autopsy with increasing PMI (4-18 h) (Table I) was carried out. The tissue pH in FC revealed an apparent increase with advanced PMI in collection of tissue (4 h PMI- pH 6.25 to 18 h PMI- 6.77) (Table II). Total proteins from Triton X-100 soluble and insoluble fractions when subjected to SDS-PAGE and quantitation of the profile showed no significant elevation of insoluble proteins with increasing PMI in all the four brain regions (Fig. 1). Slot blot assay with anti-3-NT antibody as a marker of protein nitration and quantitation following normalization with anti-tubulin signal (Figs. 2&3) showed significant increase in nitration levels in SN, mild in CB and no alteration in FC and CD with increasing PMI. On similar lines, oxyblot of total protein carbonyls in protein extracts from human brain samples with increasing PMI showed an elevation in protein carbonyl levels as well in SN whereas there was no significant change in other regions (Fig. 4).

Table II. pH estimation in frontal cortex obtained across different post-mortem interval (PMI) Sample no. B-183 T-149 B-180 B-178 T-146 B-182 T-148 T-144

PMI (h) 4 7.5 7.5 12 14.5 14.5 17 18

pH value + 0.1 (SD) 6.25 6.22 6.42 6.01 6.3 6.56 6.82 6.77

Western blot analysis of selected proteins in the brain such as glial fibrillary acidic protein (GFAP), synaptophysin and neurofilament (NF) at increasing PMI and quantitation (normalized against tubulin) showed no significant change in GFAP signal in FC, CB and CD, while SN revealed a significant increase in GFAP signal (Fig. 5). Quantitation of Western blot signal of synaptophysin, a synaptic vesicle glycoprotein present in virtually all neurons in the brain that participate in synaptic transmission showed no significant change in CB, CD and SN, albeit there was a mild increase in FC (Fig. 6). Neurofilaments (NFs) are intraneuronal and axonal cytoskeletal framework proteins. NF proteins

CHANDANA et al: PROTEIN ANALYSIS IN POST-MORTEM HUMAN BRAINS

Fig. 1. Quantitation of insoluble proteins in human autopsy brain with increasing PMI. Quantitative ratio of insoluble proteins [insoluble fraction/ (soluble+insoluble fractions)] from FC (n=9), CB (n=9), CD (n=9) and SN (n=7) of human autopsy brains (sample details given in Table I) with increasing PMI plotted as arbitrary intensity units (A.U.) vs. PMI.

193

Fig. 2. β-tubulin westerns in different brain regions with increasing PMI. Total protein extract (~50 µg) from FC, CB, CD and SN (sample ID as in Table I) with increasing PMI were run on 12 per cent SDS PAGE followed by anti-β-tubulin Western blot.

Fig. 3. Analysis of 3-nitrotyrosine (3-NT) modification in post-mortem human brain. (A) Total protein extracts (~100 µg) from FC, CB, CD and SN (sample ID as in Table I) with increasing PMI were spotted on nitrocellulose membrane in triplicate followed by anti-3NT Western blot. (B) Plot of 3-NT signal (normalized with β-tubulin signal) vs. PMI.

on SDS PAGE separate into three polypepides based on increasing molecular weight (NF-L, Mol wt: 68-70 kDa; NF-M, Mol. Wt: 145-160 kDa and NF-H, Mol. wt: 200-220 kDa). With increasing maturity of the human brain, NF-L levels recede while the levels of NF-M and NF-H become apparent. In this study on adult brains, Western blots could detect only the NF-M and NF-H species (Fig. 7). In some samples, lower molecular weight bands (~25 and 50 kDa), probably representing degraded peptides of NF were noted but not relating with PMI. Quantitation of the NF signal (normalized against tubulin signal) showed a significant (P