Region-Specific Neurotrophin Imbalances in Alzheimer Disease: Decreased Levels of Brain-Derived Neurotrophic Factor and Increased Levels of Nerve Growth Factor in Hippocampus and Cortical Areas

Background  Nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin 3 (NT-3), and neurotrophin 4/5 (NT-4/5) are members of the neurotrophin gene family that support the survival of specific neuronal populations, including those that are affected by neurodegeneration in Alzheimer disease (AD).

Objective  To determine whether neurotrophin protein levels are altered in the AD-affected brain compared with control brains.

Methods  We quantitated protein levels of NGF, BDNF, NT-3, and NT-4/5, and calculated neurotrophin/NT-3 ratios in AD-affected postmortem hippocampus, frontal and parietal cortex, and cerebellum, and compared them with age-matched control tissue (patients with AD/controls: hippocampus, 9/9 cases; frontal cortex, 19/9; parietal cortex, 8/5; and cerebellum, 5/7, respectively). We applied highly sensitive and specific enzyme-linked immunosorbent assays in rapid-autopsy–derived brain tissue (mean±SD postmortem interval, 2.57±1.75 h, n=71) to minimize postmortem proteolytic activity.

Results  Levels of BDNF were significantly reduced in hippocampus and parietal cortex (P<.001, and P<.01) as well as BDNF/NT-3 ratios in frontal and parietal cortices (P<.05, and P<.01) in the group with AD compared with the control group. Levels of NGF and NGF/NT-3 ratio were significantly elevated in the group with AD compared with the control group in the hippocampus and frontal cortex (P<.001). Levels of NT-4/5 and the NT-4/NT-3 ratio were slightly reduced in hippocampus and cerebellum in the group with AD compared with the control group (P<.05). In contrast, the levels of NT-3 were unchanged in all brain regions investigated.

Conclusion  Decreased levels of BDNF may constitute a lack of trophic support and, thus, may contribute to the degeneration of specific neuronal populations in the AD-affected brain, including the basal forebrain cholinergic system.

NERVE GROWTH factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin 3 (NT-3), and neurotrophin 4/5 (NT-4/5) are members of the neurotrophin gene family that support the survival, differentiation, maintenance, and repair of vertebrate neurons.¹ Nerve growth factor supports the cholinergic neurons of the basal forebrain system.² Brain-derived neurotrophic factor supports cholinergic, dopaminergic, 5-hydroxytryptamine, and neuropeptides containing neurons.^3-6 Neurotrophin 3 was shown to prevent the death of adult central noradrenergic neurons in vivo.⁷ Neuronal loss in Alzheimer disease (AD) includes the cholinergic neurons of the basal forebrain system, as well as neurons of the noradrenergic and serotonergic system.⁸ Previous reports demonstrated increased levels of NGF in cortical and subcortical brain areas including the frontal and parietal cortices and the hippocampus.^9-11 These increases were attributed to both reduced uptake and retrograde transport of NGF to NGF-sensitive cell bodies, since expression and protein levels of the NGF high-affinity receptor trkA were reduced in target regions of basal forebrain cholinergic neurons, such as the cortical association areas.^12-15 In contrast, BDNF messenger RNA levels were reported to be decreased in hippocampal neurons¹⁶ and BDNF protein levels were reported to be decreased in the entorhinal cortex of patients with AD.¹⁷ These results pointed to an opposite involvement of NGF and BDNF, which are both associated with trophic support of cholinergic neurons of the basal forebrain system, in the neuropathology of AD. To further explore the differential involvement of neurotrophins in AD, we quantitated protein levels of NGF, BDNF, NT-3, and NT-4/5 in well-characterized, rapid-autopsy–derived AD postmortem hippocampus, frontal and parietal cortices, and cerebellum, and compared them with age-matched control tissue. In addition, to control for potential effects of alterations in tissue wet weight on neurotrophin concentrations, we calculated neurotrophin/NT-3 ratios.

The clinical diagnosis of AD was made according to the National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria,¹⁸ and was histopathologically confirmed according to the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) criteria.¹⁹ Tissue samples from age-matched patients without neurological disorders were used as controls . Brain tissues from patients with AD and from controls were frozen in dry ice and stored at -85°C. Characteristics of the brain tissue samples investigated are given in the Table 1. Different brain tissue volumes were available for the analysis of the neutrophin proteins. This fact explains the different sample sizes for the individual measurements. All available brain tissue samples were used for the analyses.

Tissue samples were dissected from the frozen brain sections, weighed, thawed on ice, and supplied with ice-cold homogenization buffer (44 mmol/L of Tris/hydrochloride [pH 7.4], 300 mmol/L of sodium chloride, 2 mmol/L of EDTA, 0.1 mmol/L of phenylmethylsulfonyl fluoride [PMSF], 0.1 mmol/L of benzethonium chloride, 0.1 mmol/L of benzamidine hydrochloride, 1 µg/mL of aprotinin, 0.010% Triton X-100, 0.010% sodium desoxycholate, 0.010% Nonidet P-40, 0.03% bovine serum albumin, 0.025% sodium azide), as described previously.^20,21 Following homogenization with a pestle, the solutions were sheared 5 times by a sterile ice-cold needle (1.2×40 mm). An additional sonication step occurred for 10 seconds within 2-mL Eppendorf tubes (Eppendorf Inc, Hamburg, Germany) placed in an ice bath. After centrifugation (14,000 rpm for 25 minutes at 4°C) the S1 supernatants were collected and stored at −80°C. The pellet-containing tubes were taken and each pellet was resuspended in homogenization buffer with a pH 7.0 via a 5-second sonication step. The probes were subjected to centrifugation (13,000 rpm for 30 minutes at 4°C) and the S2 supernatant was mixed with the former supernatant fraction S1. After an additional sedimentation step (14,000 rpm for 7 minutes at 4°C), the supernatants were directly placed on the prepared enzyme-linked immunosorbent assay (ELISA) plates coated with the first neurotrophin-specific antibody. Brain tissue samples of about 50 mg wet weight were homogenized in 1 mL of homogenization buffer. Thus, about 6-mg weight wet was analyzed in each of the 120-µL brain supernatants used for the ELISAs. Nonspecific binding was blocked with block and sample 5× buffer (Promega, Madison, Wis).

After a probe/standard-incubation period of 20 hours at 4°C, ELISA plates were washed 5 times and incubated with the second antineurotrophin antibody. Before color reaction occurred, wells were incubated with a third enzyme-linked antibody followed by addition of a chromogenic substrate.

Brain tissue levels of NGF were measured by an ELISA as described recently.²² Black 96-well microplates (Nunc Inc, Wiesbaden, Germany) were coated with monoclonal anti-β (2.5S, 7S) NGF antibodies (Ab) (clone 27/21; Boehringer Mannheim Corp, Indianapolis, Ind) diluted in carbonate buffer pH 9.2 overnight at 4°C. The supernatants and standard solutions (120 µL) were added and incubated for 20 hours at 4°C. Plates were washed and incubated with anti-β (2.5S, 7S) NGF-β-galactosidase conjugate for 2½ hours at room temperature. Following an additional washing step, the fluorogenic substrate 4-methylumbelliferyl-β-D-galactopyranoside was added and plates were incubated at 4°C overnight. The reaction was stopped after 1 hour at room temperature and the fluorescent product was measured in the microtiter wells using a fluorometer (Labsystems Fluoroskan Ascent FL; Labsystems Inc, Helsinki, Finland) (excitation wavelength, 355 nm; emission, 460 nm). The detection limit was 1.5 pg/mL; the cross-reactivity with other neurotrophins at 10 ng/mL was less than 2%.

Brain-derived neurotrophic factor, NT-3, and NT-4/5 levels were determined using commercially available ELISA systems (Promega, Madison, Wis) and were performed according to the manufacturer's protocols, but modified as described recently.²³ The supernatants (120 µL) were added to 96-well immmunoplates (Nunc) at 4°C overnight. Anti-BDNF monoclonal antibodies (mAb), anti–human NT-3 polyclonal antibodies (pAb), and anti–human NT-4 pAb, respectively, were used as capture Ab. Anti-BDNF pAb, anti–NT-3 mAb, and anti–NT-4 mAb, respectively, were used as reporter Ab. After incubation with a species-specific Ab (anti-rat IgG) conjugated to horseradish peroxidase as a tertiary reactant and washing, the solution was incubated with the chromogenic substrate 3,5,3′,5′-tetramethylbenzidine. Absorbance was measured at 450 nm using a microplate reader (model MR 7000; Dynatech Labsystems Inc). Brain-derived neurotrophic factor ELISA: range, 7.8 to 500 pg/mL; cross-reaction with other neurotrophins at 100 ng/mL less than 3%; detection limit, 15.6 pg/mL. Neurotrophin 3 ELISA: range, 4.7 to 300 pg/mL; cross-reaction with other neurotrophins at 10 ng/mL less than 3%; detection limit, 10.0 pg/mL. Neurotrophic 4/5 ELISA: range, 4.7 to 300 pg/mL; cross-reaction with other neurotrophins at 100 ng/mL less than 3%; detection limit, 9.4 pg/mL. All brain tissue samples were assayed in duplicate determinations.

Neurotrophin levels given in picograms per milligrams largely depend on tissue wet weight as denominator. Since tissue weight may be susceptible to agonal state, edema, and postmortem variations, we additionally performed ratio analysis of NGF, BDNF, and NT-4/5 to NT-3, because the latter neurotrophin did not show changes in brain distribution in this study.

Statistical analyses of data were performed using the Mann-Whitney test for group comparisons. Correlation analyses were performed by multiple regression using brain levels of neurotrophins, as well apolipoprotein E (APOE) genotype, Braak staging, sex, and age. Regression analysis was complemented with analysis of variance (ANOVA) by using SPSS for Windows (version 8.0; SPSS, Chicago, Ill). Statistical significance was assumed at P<.05. Bonferroni correction for multiple testing was applied.

Brain tissue levels of NGF were significantly elevated in the group with AD, as compared with the control group, in the hippcampus and frontal cortex (P<.001, and P<.001, respectively) (Figure 1). Nerve growth factor concentrations in the hippocampus of the group with AD amounted to 32.2±4.4 pg/mg (mean±SEM, n=9), compared with 5.6±1.1 pg/mg for the control group (n=9). Nerve growth factor levels in the frontal cortex of the group with AD were 13.8±1.2 pg/mg (n=19), compared with 3.5±0.7 pg/mg for the control group (n=9). Nerve growth factor levels in the parietal cortex were slightly increased compared with the control group, but this difference did not reach statistical significance (Table 2). Nerve growth factor levels in the cerebellum were not different between the group with AD and the control group (Table 2). In the hippcampus and parietal cortex, levels of BDNF were significantly reduced in the group with AD as compared with the control group (P<.001, and P<.01, respectively). Brain-derived neurotrophic factor concentrations in the hippocampus of the group with AD were 41.4±3.8 pg/mg (n=9), compared with 71.2±5.5 pg/mg in the control group (n=9). Brain-derived neurotrophic factor levels in the parietal cortex of the group with AD were 10.4±1.5 pg/mg (n=8), compared with 21.9±3.1 pg/mg in the control group (n=5). Brain-derived neurotrophic factor levels in the frontal cortex were slightly decreased for the group with AD compared with the control group, but this difference did not reach statistical significance, and BDNF levels in the cerebellum were not different between the group with AD and the control group (Table 2). Levels of NT-4/5 were slightly decreased in the hippocampus and cerebellum in the group with AD compared with the control group (AD, 42.0±6.1 pg/mg, n=9; control, 65.2±6.9 pg/mg, n=9; P<.05 [t test], P=.06 [Mann-Whitney test]; AD, 29.8±2.4 pg/mg, n=5; control, 41.8±3.5 pg/mg, n=7; P<.05) (Table 2). Levels of NT-3 were unchanged in all brain regions investigated.

Nerve growth factor/NT-3 ratios were significantly elevated in hippocampus and frontal cortex in the group with AD compared with the control group (0.34±0.06 and 0.06±0.01, P<.001; 0.54±0.07 and 0.19±0.0, P<.001) (Figure 2 and Table 3). In contrast, BDNF/NT-3 ratios were were significantly reduced in frontal and parietal cortices in the group with AD group compared with the control group (0.67±0.09 and 1.08±0.14, P<.05; 0.30±0.04 and 0.71±0.14, P<.01) (Table 3). Brain-derived neurotrophic factor/NT-3 ratio in the hippocampus was apparently decreased in the group with AD compared with the control group, but this difference did not reach statistical significance (P=.056, t test; P=.058, Mann-Whitney test). Finally, NT-4/NT-3 ratio was significantly reduced in the AD-affected hippocampus compared with the control group (0.39±0.03 and 0.66±0.05, P<.001) (Table 3).

Further analyses revealed that there was no apparent correlation of brain tissue levels of neurotrophins or neurotrophin/NT-3 ratio with APOE genotype, age, sex, or neuropathological staging.

This study demonstrates decreased levels of BDNF in hippocampus and parietal cortex, and increased levels of NGF in hippocampus and frontal cortex in AD-affected brain tissue compared with control brain tissue measured by sensitive and specific ELISAs. We selected these brain regions because they are differentially affected by the histopathological changes during neurodegeneration in AD.²⁴ Usually, the hippocampal area, the perforant path, and the entorhinal cortex exhibit neurofibrillary tangles earliest in the course of AD. Among cortical association areas, the parietal cortex shows deposition of β-amyloid plaques, neurofibrillary tangles, and neuronal loss earlier than the frontal cortex, whereas the cerebellum is usually preserved and may serve as a negative control. We used rapid-autopsy–derived brain tissue to minimize postmortem proteolytic modifications and assayed for the 4 neurotrophins in parallel under well-standardized conditions. In addition, to control for potential effects of alterations in tissue wet weight on neurotrophin concentrations, we calculated neurotrophin/NT-3 ratios. We used NT-3 as the denominator, because NT-3 levels did not show changes in brain distribution in this study.

Significant decreases of BDNF protein levels were detected in the hippocampus and parietal cortex in AD-affected brain tissue compared with control brain tissue. Brain-derived neurotrophic factor/NT-3 ratio analysis confirmed the observed decrease in the parietal cortex and indicated a further decrease in the frontal cortex in the group with AD compared with the control group. Brain-derived neurotrophic factor/NT-3 ratio in the hippocampus was also decreased in the group with AD compared with the control group, but this difference did not reach statistical significance in that type of analysis. Brain-derived neurotrophic factor protein and its signaling receptor trkB are widely expressed in the developing and adult central nervous system²⁵ in a number of neuronal subpopulations including basal forebrain cholinergic and γ-aminobutyric acid neurons, mesencephalic γ-aminobutyric acid neurons, substantia nigra dopaminergic neurons, cerebellar granule cells, as well as central nervous system neurons of the striatum, hippocampus, and trigeminal mesencephalic nucleus. Brain-derived neurotrophic factor not only acts on hippocampal pyramidal and dentate granule cells, but these neurons also produce BDNF to act on innervating basal forebrain cholinergic neurons. Similar to NGF, BDNF protects basal forebrain cholinergic neurons from degenerative changes after axotomy in the adult brain.²⁶ In addition to the basal forebrain cholinergic system, neurons containing 5-hydroxytryptamine are also reduced in the AD-affected brain and 5-hydroxytryptamine deficiencies were associated with depression and behavioral disturbances in AD.²⁷ Our finding is in line with a previous report of reduced BDNF messenger RNA levels in the hippocampus, measured by in situ hybridization,¹⁶ with reduced BDNF expression in the parietal cortex measured by room temperature -polymerase chain reaction²⁸ and with reduced BDNF protein levels in the entorhinal cortex in the brain of patients with AD.¹⁷ Thus, diminished BDNF production may contribute to a reduced neurotrophic support of cholinergic and 5-hydroxytryptamine-containing neurons during neurodegeneration in AD.

Our finding of increased levels of NGF as well as NGF/NT-3 ratios in the hippocampus and frontal cortex is in agreement with previous reports where NGF protein levels were shown to be increased in cortical and subcortical brain areas including the frontal and parietal cortex and the hippocampus.^9-11 In contrast to BDNF, NGF messenger RNA levels were reported to be unchanged in AD-affected cortex, hippocampus, and septum/nucleus basalis area.²⁹ Increased brain tissue levels of NGF in various brain regions in AD may be caused by both reduced uptake and retrograde transport of NGF to NGF-sensitive cell bodies. Recently, it was shown that expression and protein levels of the specific NGF high-affinity receptor trkA were reduced in target regions of basal forebrain cholinergic neurons, such as the cortical association areas.^12-15 In addition, inflammatory signals may induce an additional production of NGF in activated microglial cells.³⁰

In agreement with previous reports, we found no differences in NT-3 levels between patients with AD and controls in regions affected by the characteristic neuropathology of AD, including the hippocampus, entorhinal, and parietal cortex.^11,31 In contrast, NT-4 measurements and calculation of NT-4/NT-3 ratios indicated a slight decrease of NT-4 in the AD-affected hippocampus. Surprisingly, we also found a significant reduction in NT-4/5 levels in the cerebellum of patients with AD. It is unclear in which cell types these changes are manifest. Recently, Skaper et al³² reported that neurotrophins, including BDNF and NT-4/5, rescued cerebellar granule neurons from oxidative stress-mediated apoptotic death. However, a potential regulatory role of NT-4 in the hippocampus and cerebellum of patients with AD remains to be investigated.

We demonstrated decreased levels of BDNF and increased levels of NGF in hippocampus and cortical areas in AD-affected brain tissue compared with control brain tissue. Decreased levels of BDNF may be associated with lack of trophic support and may contribute to the degeneration of specific neuronal subpopulations in the AD-affected brain, including the basal forebrain cholinergic system. Elevated levels of NGF may reflect a reduced uptake and retrograde transport by the NGF high-affinity receptor trkA.

Accepted for publication December 27, 1999.

This work was supported by grants 3100-049397.96/1 from Schweizerischer Nationalfonds, Bern, Switzerland (Drs Hock and Otten), SFB505 from Deutsche Forschungsgemeinschaft, Bonn, Germany (Dr Otten), and P50-AG05128 from the National Institute on Aging, National Institutes of Health, Bethesda, Md (Dr Hulette and Ms Rosenberg).

We thank K. Kräuchi, PhD, for help with statistical analyses.

Corresponding author: Christoph Hock, MD, Department of Psychiatry Research, University of Zürich, Lenggstrasse 31, CH-8029 Zürich 8, Switzerland (e-mail: chock@bli.unizh.ch)