Locus Coeruleus Neurofibrillary Degeneration in Aging, Mild

Neurobiology of Aging 28 (2007) 327–335

Locus coeruleus neuroﬁbrillary degeneration in aging, mild cognitive impairment and early Alzheimer’s disease Aneta Grudzien a, Pamela Shaw a, Sandra Weintraub a, Eileen Bigio a, Deborah C. Mash b, M. Marsel Mesulam a,∗

a Cognitive Neurology and Alzheimer’s Disease Center, Northwestern University, Feinberg School of Medicine, 320 East Superior Street, Searle 11-453, Chicago, IL 60611, USA b Department of Neurology, University of Miami School of Medicine, 1501 N.W. 9th Ave., Miami, FL 33136, USA

Received 15 July 2005; received in revised form 22 December 2005; accepted 4 February 2006 Available online 29 March 2006

Abstract

Neurofibrillary degeneration in the nucleus basalis and a loss of its cortical cholinergic projections are prominent components of the neuropathology in Alzheimer’s disease (AD). The AD brain is also associated with a degeneration of the noradrenergic projections arising from the nucleus locus coeruleus (LC), but the time course of this lesion is poorly understood. To determine whether the LC displays neurofibrillary abnormalities early in the course of events leading to AD, we examined tissue specimens from seven cognitively normal controls and five subjects at the stages of mild cognitively impairment (MCI) or early AD. Tyrosine hydroxylase immunochemistry was used as a marker of LC neurons while AT8 immunolabeling visualized abnormal tau associated with neurofibrillary tangles and their precursors. Thioflavine-S was used as a marker for fully developed tangles. We found that AT8-positive labeling and thioflavine-S positive tangles were present in both groups of specimens. However, the percentage of neurons containing each of these markers was significantly higher in the cognitively impaired group. The MMSE scores displayed a negative correlation with both markers of cytopathology. These results indicate that cytopathology in the LC is an early event in the age-MCI-AD continuum and that it may be listed among the numerous factors that mediate the emergence of the cognitive changes leading to dementia. © 2006 Elsevier Inc. All rights reserved.

Keywords: Alzheimer’s disease; Mild cognitive impairment; Locus coeruleus; Norepinephrine; Neuroﬁbrillary tangles

1. Introduction the time course of transmitter-specific pathologies helps to differentiate the degenerative changes that are part of the Alzheimer’s disease (AD) leads to severe cognitive and initial pathogenetic cascade from those that may reflect behavioral impairments that compromise independent daily end-stage events. This is an important distinction since activities. Neurofibrillary tangles (NFTs) and ␤-amyloid changes that can be traced to the early and prodromal stages plaques in cortical and subcortical areas are its prominent of the disease become potential targets for therapies aimed at histopathological markers. Other significant features include disease prevention. Addressing this question requires post- neuronal death, synaptic loss and depletion of transmitters mortem examination of subjects who have had longitudinal such as acetylcholine, serotonin and noradrenaline (NA) neuropsychological characterization until shortly before [15]. The loss of neurotransmitters has attracted a great deal death. of attention because it is the one aspect of the disease most We recently conducted such an investigation on choliner- amenable to pharmacological intervention. Determining gic pathways in a sample of 12 rigorously characterized subjects. We found that cytoskeletal abnormalities in the nucleus ∗ Corresponding author. Tel.: +1 312 908 9339; fax: +1 312 908 8789. basalis (NB), the origin of cholinergic innervation of the cere- E-mail address: [email protected] (M.M. Mesulam). bral cortex, begin during the course of normal aging and that

0197-4580/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.neurobiolaging.2006.02.007 328 A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 they become significantly more prominent at a prodromal could therefore contribute to the deterioration of mood as stage of the disease known as mild cognitive impairment or well as cognition in AD. MCI. We also determined that the extent of cytopathology in the NB was significantly correlated with performance on a delayed memory test, suggesting that pathological alter- 2. Methods ations in cortical cholinergic pathways may contribute to the declining memory performance seen in the course of the age- 2.1. Participants MCI-AD continuum [26]. The present study aimed to determine whether a sim- Postmortem brainstem tissue was obtained from partici- ilar conclusion can be reached in the case of the nora- pants who enlisted in a longitudinal study at the University of drenergic projections originating in the nucleus locus Miami and who were cognitively normal at the time of entry. coeruleus (LC). This nucleus displays severe cytopathology All subjects agreed to annual neuropsychological testing and in advanced stages of AD, leading to a prominent loss of consented to brain donation. From the study’s data bank, we noradrenergic markers in the cerebral cortex [3,18,22,46]. identified all subjects who had converted from normalcy to However, the status of the LC at the earlier stages of impairment and who had at least two full neuropsychological the aging-MCI-AD continuum has yet to be investigated tests within the enrolment period. The criteria also required in a neuropsychologically well characterized sample of that one of these examinations be within 18 months of death. individuals. We excluded specimens with significant neuropathological The LC is the primary source of cortical, thalamic, cere- changes other than plaques and tangles. Subjects who had bellar, brain stem and spinal cord NA. Its ascending pro- medical diseases or medication intake that could interfere jections branch extensively and innervate all parts of the with cognitive function were also excluded. Five individu- cerebral cortex [14,27,35]. Investigations in several animal als (three women, two men; mean age, 92.4) fulfilled the models have demonstrated that the activity of LC neurons and criteria and were included in the present study. Four individ- the resultant release of NA influence attention, arousal and uals were at the stage of MCI and one at the early stages of memory [39]. This noradrenergic innervation is thought to AD (8-CI–12-CI). Seven subjects (four women, three men; optimize the signal-to-noise ratio at the stage of information mean age 89.7) for whom we had definite evidence of cog- processing and to influence response selection at the stage nitive normalcy throughout the follow-up period comprised of decision making [8,11]. Increasing LC neuronal activity the control group (1-CN–7-CN). Serial neuropsychological mediates changes from EEG synchronization to low voltage examinations on six of these subjects and Telephone Inter- fast activity, from slow wave sleep to REM, and from behav- views of Cognitive Status (TICS) for one subject (7-CN) ioral drowsiness to vigilance. In humans, clinical experience served as evidence for cognitive normalcy [29]. Other salient suggests that a deficiency of cortical monoamines, including individual characteristics can be found in Table 1. All of these NA, promotes symptoms of depression [7,45,46]. A dysfunc- subjects, except 7-CN, were also reported in a study investi- tion of LC neurons and its resultant effects on cortical NA gating the NB [26].

Table 1 Subject characteristics Subject no. MMSE (mpd)a Age at death, sex Autopsy Braak stage CERAD Brain weight (g) LC (A6) AT8 LC (A6) interval (h)b (Ent/STS) (%)c tangles (%)c 1-CN 25 (12) 88, F 9 2 S/S 1104 12/218 (5.5) 3/218 (1.4) 2-CN 29 (10) 89, F 9 1–2 S/S 1180 16/554 (2.9) 3/554 (0.54) 3-CN 30 (13) 90, M 2.5 2 S/S 1155 14/252 (5.6) 3/252 (1.2) 4-CN 28 (18) 91, M 12 2 S/S 1309 4/196 (2.0) 3/196 (1.5) 5-CN 28 (15) 92, M 7 3 S/S 1290 14/581 (2.4) 11/581 (1.9) 6-CN 28 (4) 95, F 3.2 3 M/F 1096 15/267 (5.6) 7/267 (2.7) 7-CN NAd (11) 83, F 13 2 S/S 1175 14/415 (3.4) 8/415 (1.9) 8-CI 24 (16) 89, F 4.5 2 M/S 1280 15/217 (6.9) 9/217 (4.1) 9-CI 24 (1) 90, M 3 3 M/S 1380 15/253 (5.9) 10/253 (4.0) 10-CI 27 (9) 92, F 3.5 3–4 S/S 1084 24/548 (4.4) 22/548 (4.0) 11-CI 23 (11) 92, M 4.5 5 F/F 1100 58/323 (18) 65/323 (20) 12-CI 27 (1) 99, F 5 3–4 S/M 1060 27/353 (7.6) 13/353 (3.7) Tangle densities of the entorhinal, transentorhinal, fusiform, and inferotemporal corticies were used to determine the Braak & Braak stage. CERAD classification indicates sparse (S), moderate (M) and frequent (F) distribution of plaque densities in the entorhinal cortex (Ent) and the superior temporal sulcus (STS). TICS: Telephone Interview of Cognitive Status; MMSE: Mini-Mental State Examination; CERAD: Consortium to Establish a Registry for Alzheimer’s Disease. a Scores from the final testing session in the months prior to death (mpd). b Time elapsed between death and autopsy. c The numerator denotes the number of tangles (or AT8-positive cells) while the denominator reflects the number of locus coeruleus (LC[A6]) neurons from an adjacent section. d TICS was used instead of MMSE testing. Score = 38/50. A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 329

2.2. Neuropsychology NA The psychological testing assessed memory, language, 45 0 NA NA NA NA visuospatial skills and executive functions of the subjects 27 27 10 23

[1,2,13,30,43]. We also had dementia severity scores based 101 on the Mini-Mental State Examination. A complete list of 52 31 27 36 33

tests and the subjects’ respective scores during the ﬁrst and 407 300 last sessions is presented in Table 2. Determining what is considered normal performance for those above 89 years of age 27 1 23331 13 27 29 is a complex process as normative ranges above this age are 19 23 12 4 5 10 rarely available and can vary by age, gender and education 5 level. In this study, the normative values were determined 0 132 14 10 13 13 13 14 using the eldest available group data, the 70–89-year-old age 24 26 Wiley-Liss. 6 14 NA 15 15 14 12 14 12 67 NA 83 80 93 range. Additionally, individual scores were compared against 30 normative scores of 50–59-year-old (the exception being t’s age. Reprinted from Mesulam MM, Shaw P, Mash D, 2 12 24

the CERAD 15-item BNT where we used the 50–69-year- 113

old age range for normative scores) [9,10,19,20,40]. Based were determined using literature cited in the methods. Subject on available age- and education-related ranges, test scores were converted into a non-linear six-point scale (Fig. 1). A grade of 1 and 2 indicated impairment, 3 and 4 signified average performance and a grade of 5 and 6 denoted supe- 677 rior performance (i.e. performance that is considered normal for the 50–59-year-old range). Although only the results of 3 the first and last testing sessions are presented (Table 2), all sessions were considered for determining the individual’s cognitive status. Behavioral questionnaires and family interviews yielded data on daily living activities. Cognitive impairment that involved memory but that was not severe enough to significantly interfere with independent daily living activities led to a MCI diagnosis while the presence of more severe impairments and corresponding disruptions of daily living activities led to a diagnosis of early AD [31,36,37]. 11 11 11 9 11 9 11 11 11 9 10 NA 11 7 2.3. Histology 28 29 29 30 30 28 28 28 29 28 29 Fixation of tissue occurred following a postmortem inter- 5687 66 85 11 10 val of 2.5–13 h as formerly described [24]. Available brain- 25 stem tissue was sectioned coronally into 1–2 cm blocks, 7 35 N5 456334NA4 53555 6NA58 4 frozen on dry ice, cut into 40 ␮m sections using a sliding F, 12 years F, 16 years M, 16 years M, 16 years M, 16 years F, 12 years F, 18 years M, 14 years F, 12 years M, 16 years F, 13 years microtome and collected in phosphate buffer. The identification of the LC was based on adjacent sections stained with (1) cresyl violet, (2) tyrosine hydroxylase (TH) using stan- dard methods and a modified (3) acetylcholinesterase (AChE) procedure [24]. Cresyl violet staining was used to iden- tify the characteristically magnocellular and neuromelanin- rich LC neurons while TH immunoreactivity was used as a marker for the noradrenergic nature of these neurons [35]. The AChE histochemistry yielded the characteristic AChE- First and last testing in mpdAge at test 78 12 35 10 81 49 87 13 86 93 88 18 86 84 89 15 82 33 89 85 4 90 91 16 92 74 95 1 82 77 88 9 84 90 22 86 11 92 12 1 90 91 98 99 rich staining pattern of LC neurons and also permitted the macroscopic delineation of the LC from other major struc- 9/8 CERAD constructions (total = 11) 2 BVRT (total = 10) 3 Logical Memory II-Delay (total = 50) 7 7 21 11 25 7 17 30 19 16 7 13 14 13 20 15 CERAD list acquisition (total = 30) 17 20 19 21 15 21 18 18 17 16 20 21 21 16 21 4/5 CERAD list delayed recall (10) 13 CERAD BNT (total = 15) 14 15 15 14 15 15 15 15 15 15 15 15 13 8/16 CFL Word Fluency (uncorrected) 17 16 33 44 55 52 46 44 31 32 33 41 33 27 44 20 31 249 Trail Making Part B 143 123 182 191 175 225 89 149 93 143 166 104 66 231 214 NA 154 213 93 Trail Making Part A 62 43 74 63 42 49 50 57 46 68 48 44 64 tures at that level of the brainstem [21]. In keeping with 27 MMSE (total = 30) 27 ≥ ≥ ≥ ≥ ≥ ≥ ≥ ≤ ≤ numerous reports in the literature, these three stains pro- ≥ vided unequivocal identification of the LC, also known as the A6 cell group [46]. Thioflavine-S histofluorescence iden- 9 4 15 20 7 13 20/23 128 42 29 ≥ ≥ ≥ ≥ ≥ ≥ ≥ ≤ ≤ ≥ Normative cutoffs (M/F)50–59 years 70–89 years Subject no./gender, education 1-CN 2-CN 3-CN 4-CN 5-CN 6-CN 8-CI 9-CI 10-CI 11-CI 12-CI tified fully formed NFTs whereas AT8 (Pierce Endogen) Table 2 Neuropsychological scores The table presents the quantitative test7-CN scores was from evaluated using the TICS; first herWeintraub and S. data last Cholinergic are testing Nucleus not session Basalis included for Tauopathy above. 11 Emerges Bold, of Early italicized the in numbers the subjects indicate Aging-MCI-AD included test Continuum. in Ann scores this Neurol that study. Normative 2004;55:815–828. are cutoffs With lower permission than from normal for the subjec 330 A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 immunohistochemistry recognized NFTs as well as abnormally phosphorylated tau that precedes tangle formation [5]. Most of the AT8 immunolabeling was performed using the brown DAB reaction product. However, in order for the AT8 immunoreactivity to be differentiated from neuromelanin with even greater clarity, some labeling was based on an alkaline phosphatase blue reaction product. Immunohisto- chemical and histofluorescence procedures used in this study are described elsewhere [16]. The AT8 and TH primary anti- bodies were substituted with an irrelevant IgG in negative controls. Brainstem sections from two cognitively intact and neurologically healthy 35-year-old were utilized as positive controls. The anteroposterior extent of the LC is approxi- mately 1 cm. We confined our observations to the anterior 0.5 cm of the LC for which we had tissue available in all subjects.

2.4. Neuropathology

Thioflavine-S staining of entorhinal and superior temporal sulcus (STS) areas determined the Braak and Braak NFT [6] and CERAD plaque staging [28]. A Nikon Eclipse E800 scope with a computer driven stage and Stereoinvestigator- linked software (Microbrightfield Inc., Colchester, VT) was used to view, trace and digitize stained sections that were then viewed on a computer screen. For examining NFT distribution, thioflavine-S sections were scanned under a violet filter (380–420 nm) with a 20× objective and 10× ocular. The stage was guided automatically in a frame-by-frame manner; digital markers were placed on the identified NFTs, prevent- ing them from being counted twice. Using the same approach, brightfield illumination was used in viewing cresyl violet, AChE, and TH sections (10× objective and 10× ocular) for the purpose of determining the boundaries of the LC (by cresyl violet and AChE staining) and counting TH-positive LC Fig. 1. Neuropsychological profiles. First and last testing session scores from a representative normal control and cognitively impaired individual neurons within these boundaries. These values were used to are presented in a semi-quantitative six-point nonlinear scale. Striped bars calculate the percentage of TH-positive (that is, noradrener- depict the first testing session, whereas the black vertical bars represent the gic) neurons that contained fully formed NFT or AT8-positive final session prior to death. The horizontal line (2) depicts the cut-off for age tauopathy. Digitized files of adjacent thioflavine-S and AT8 70–89 years, and the horizontal line (4) indicates performance comparable sections were merged and aligned with matching fields in to the 50–59 (50–69 years for CERAD BNT) age range. mpd: months prior to death. the TH-labeled sections for the purpose of illustrating the anatomical relationships among these markers. Table 1 illus- trates the quantitative results. of CN and CI neuropsychological profiles are depicted in Fig. 1.

3. Results 3.2. Neuropathology

3.1. Neuropsychological subject characterizations When examined with the stains noted above, the specimens displayed no major neuropathological lesions (such as Tables 1 and 2, respectively, provide a summary of the infarction, neoplasm or infection) other than amyloid plaques pathological and neuropsychological characteristics of the and NFTs. The cognitively normal (CN) control group dis- cognitively normal (CN) and cognitively impaired (CI) sub- played tangles largely confined to medial temporal structures, jects. In all five cases with cognitive impairment, there was corresponding to Braak stages 1–3. Six of the seven subjects indication of a decline in cognitive performance so that the in the CN group had a very light density of mature plaques approximate the time of conversion from normal to MCI in the entorhinal area and the cortex of the superior temporal or from MCI to early AD could be determined. Examples sulcus (STS). In the MCI cases, the Braak NFT stage ranged A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 331 from 2 to 4. Subject 11-CI, who had the clinical findings of immunoreacted sections. Sections treated with a non-specific early AD, displayed a Braak stage of 5, indicating a definite IgG instead of the TH antibody did not yield immune active spread of NFT to neocortical areas. Neuritic plaque densities perikarya in that region. As neurons of the LC are also AChE- also tended to be higher in the CI group. However, there was rich, the enzymatic histochemical procedure was particularly also considerable overlap between the two groups: 10-CI had useful for the macroscopic identification of the LC as well distinctly fewer plaques than 6-CN and 8-CI had fewer NFTs as the delineation of other nuclear groups and their adjacent than 5-CN or 6-CN. fiber fascicles. The LC was cytoarchitectonically differentiated from the nucleus subcoeruleus, which also contains 3.3. Cytoarchitecture TH-positive perikarya, on architectonic grounds. As all LC neurons are homogeneously TH-positive [42], the number of The LC was identified by its characteristic hyperchromic TH-positive neurons was used as the denominator in calcu- magnocellular and melanin-containing neurons in cresyl lating the percentage of LC neurons containing NFTs or AT8 violet sections and TH-reactive perikarya in the matching immunoreactivity.

Fig. 2. Neuroﬁbrillary tangles in the locus coeruleus. The green circles depict the melanin-rich noradrenergic locus coeruleus neurons. The red stars indicate the neuroﬁbrillary tangles. For each subject, the illustrated section contained the greatest number of tangles. 332 A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335

Fig. 3. AT8 immunoreactive tauopathy in the nucleus locus coeruleus of subjects in the stages of cognitively normal (CN) old age, MCI and early AD. Black arrows point to examples of cell bodies with tauopathy, red arrows to neurites with tauopathy, green arrows to melanin-containing locus coeruleus neurons without tauopathy. In (B), alkaline phosphatase has been used to obtain a blue AT8 reaction product, in the other panels DAB was used to obtain a brown AT8 reaction-product. The less intense naturally-occurring brown hue of melanin is easily distinguished from the DAB, and especially from the alkaline phosphatase reaction product. Magniﬁcation: 50× at the ﬁlm plane.

3.4. Locus coeruleus tangles Both CN and CI individuals displayed AT8 immunoreactivity in perikarya and the neuropil (Fig. 3). The percentage of To determine NFT percentages in the LC, thioflavine-S AT8 immunopositive neurons ranged from 2 to 5.6 in the CN stained sections were matched to their adjacent TH sections. group and from 6.9 to 18% in the CI group. The only overlap The percentages in each case were derived from an exami- occurred in subject 10-CI who had 4.4% AT8-positive LC nation of two or three sections through the LC, depending neurons, falling within the range of the CN group. As in the on the availability of tissue. The maps in Fig. 2 were gener- case of NFTs, subject 11-CI was the outlier with the highest ated by identifying TH-positive perikarya (as green circles) percentage (18%) of AT8 neurons (Table 1). In general, the and NFTs (as red stars) in separate but matching sections distribution of AT8 phosphotau labeling was more extensive and then overlapping the landmarks visible in both sec- that the thioflavine-S labeled NFTs, supporting the conjec- tions. In the CN group, the percentage of NFT-containing ture that AT8 recognized abnormally phosphorylated tau not LC neurons ranged from 0.54 to 2.7, whereas in the CI group only in NFTs but also in their precursors. the range extended from 3.7 to 20%. There was no over- At the pontine levels that were examined, the LC (and lap between the two groups. The one outlier, subject 11-CI its ventral subcoeruleus extension) was the only nucleus that with 20% of its LC neurons containing NFT, was also the consistently displayed AT8-positivity in all specimens. The only subject in the sample with a diagnosis of early AD serotonergic raphe nuclei (NR) were not always present in the (Table 1). sections that we examined. However, in the few cases where it was available, it was the only other nucleus at this level of the 3.5. Locus coeruleus AT8-immunopositive neurons brainstem to contain AT8-positive neurons, with the cognitively impaired subjects clearly exhibiting greater degrees of The AT8 antibody recognizes abnormally phosphorylated abnormality (Fig. 4). Underscoring the remarkable selectiv- tau not only in NFTs but also at stages of tauopathy preceding ity in the distribution of tauopathy early in the age-MCI-AD NFT formation. Using an approach similar to that described continuum, vast regions of the pons, containing the pontine, in the previous paragraph, we determined the percentage of pedunculopontine, laterodorsal tegmental, parabrachial and AT8 containing perikarya in the LC (Table 1; Figs. 3 and 4). reticular nuclei were without NFT or AT8 immunolabeling. Negative control sections showed that processing with an irrelevant IgG (instead of the AT8 antibody) did not generate 3.6. Quantitative analysis immunoreactivity in the LC. Positive controls were obtained in two 35-year-old specimens with no known neurological The nonparametric Mann–Whitney rank-sum test demon- disease where immunolabeling with AT8 exhibited charac- strated a significant difference between the CN and CI teristic staining in the LC but did not reveal definite AT8 group for the percentage of LC neurons containing NFT immunoreactivity. (median = 4.0 for CI and 1.5 for CN, p = 0.004) and AT8 A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 333

Fig. 4. The stars depict AT8-positive cells in the brainstem. LC: locus coeruleus; L sub. C: locus subcoeruleus; LL: lateral lemniscus; ML: medial lemniscus; MLF: medial longitudinal fasciculus; PVG: periventricular grey; PCS: pedunculus cerebelli superior; Rn: raphe nuclei [34]. immunoreactivity (median = 6.9 for CI and 3.4 for CN, of CSF NA levels in AD [12,33,41], suggesting that there p = 0.018). The Wilcoxon matched-pairs signed rank test may be a compensatory up-regulation of NA synthesis in the revealed that the percentage of AT8-immunopositive neu- unaffected LC neurons [38] or possibly a reduced reuptake rons was higher than that of NFT-containing neurons capacity in the pathologically compromised axons [12]. (median = 5.5 for AT8 and 2.3 for NFT, p = 0.010). The Spear- Nearly all previous investigations on this subject were man rank order correlation showed that the MMSE scores based on patients with fully established clinical and patho- (n = 11) possessed a strong negative correlation with both logical AD. Although one of these investigations attempted NFT-containing (rs = −0.827; p = 0.002) and AT8-positive to separate the specimens into early and late stages [3,4], the (rs = −0.628; p = 0.038) neurons. groups were formed on the basis of neuropathological rather than neuropsychological evidence. Our sample is relatively unique because of the longitudinal neuropsychological char- 4. Discussion acterization of the subjects and the focus on MCI, a diagnosis that reflects a prodromal stage of AD when made on the basis Pathology in the LC and loss of cortical noradrenergic of memory impairments in subjects where the only signifi- markers in AD have been reported previously [15]. For exam- cant neuropathology is confined to NFT and amyloid plaques ple, Zweig and colleagues reported significant neuronal loss [31,37]. and NFTs in the LC of patients with AD [46]. Furthermore, Our results show that the LC develops fully formed neu- Mann et al. found a correlation between LC atrophy and fre- rofibrillary tangles and their precursors in the course of quency of NFTs in the cortex [23] while Bondareff et al. normal aging but that these abnormalities become signifi- reported a positive correlation of NFTs with the duration and cantly more prominent at the stages of MCI and early AD. severity of the dementia [4]. In a study of 47 autopsied speci- The extent of this LC cytopathology was correlated with mens from subjects with AD, Zubenko and colleagues linked overall cognitive state as measured by the MMSE, suggest- the noradrenergic pathology in the LC to the emergence of ing that noradrenergic dysfunction may be listed among the clinical symptoms of depression [45]. More recently, lum- numerous factors mediating the onset of cognitive impair- bar punctures revealed an apparently paradoxical increase ments in the aging-MCI-AD continuum. Several lines of 334 A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 evidence have shown that even early stages of neurofibrillary beneficial value in modifying disease progression. Combina- degeneration can interfere with normal cellular function- tions of cholinergic, noradrenergic and serotonergic therapies ing [17]. Such changes can have particularly far reaching may therefore warrant closer investigation. consequences when they involve the LC, a small nucleus, containing no more than 20,000 neurons in the normal human brain [32], each of which sends widely divergent axons cov- Note added in proof ering extensive cortical and thalamic regions. Cytopathology in even a small percentage of LC neurons can therefore affect Heneka et al. reported that experimentally induced LC noradrenergic physiology within widespread cortical and tha- degeneration promoted amyloid plaque deposition and neu- lamic territories. The presence of fully formed NFT and their ronal loss in an APP23 transgenic mouse model of AD, precursors in the LC even at very low Braak & Braak stages leading to the suggestion that LC degeneration may con- also establishes that the LC is one of the few regions in tribute to the development of AD neuropathology. Heneka the brain with a very high susceptibility to neurofibrillary MT, Ramanathan M, Jacobs AH, Dumitrescu-Ozimek L, degeneration. We did not measure the volume of the LC and Bilkei-Gorzo A, Debeir T, Sastre M, Galldiks N, Zimmer A, therefore cannot make a statement about atrophy. The pres- Hoehn M, Heiss WD, Klockgether T, Staufenbiel M. Locus ence of atrophy in the MCI and early AD brains would not coeruleus degeneration promotes Alzheimer pathogenesis in have influenced our conclusions since our results are based amyloid precursor protein 23 transgenic mice. J Neurosci on the percentage of LC neurons containing abnormal inclu- 2006 Feb;26(5):1343–54. sions rather than on their densities. A comparison to our observations in the cholinergic nucleus basalis (NB) shows that the LC has an NFT suscep- Acknowledgements tibility that is equivalent to that of the NB. The one common denominator of these two nuclei is that they send extensively This work was supported in part by Alzheimer’s Disease branching projections to the cerebral cortex, an anatomical Core Center grant P30AG-13854 from the National Institute feature that seems to determine vulnerability to neurofibril- of Aging to Northwestern University. The authors are grateful lary degeneration among subcortical nuclei [25]. Zarow and to Dr. Rademaker for his assistance in biostatistics and the colleagues demonstrated that LC neuronal loss in AD can University of Miami Brain Endowment Bank for the tissue be more extensive than the cholinergic cell loss in the NB specimens. and that it correlates more strongly with the estimated duration of illness. However, their study was based on neuronal loss, which is a late marker of neuronal pathology, and also on subjects who had fully established AD [44]. Based on References our observations, it would be reasonable to conclude that [1] Benton AL, Hamsher KdS. Multilingual Aphasia Examination. Iowa cytopathology within the cell groups that give rise to the cor- City: University of Iowa; 1989. tical cholinergic and noradrenergic innervations are equally [2] Benton AL. Revised visual retention test. New York: Psychological early events in the course of the neuropathological cascade Corporation; 1974. that leads to AD. [3] Bondareff W, Mountjoy CQ, Rossor RM, Iversen LL, Reynolds GP, In the brainstem, the anatomical selectivity of LC pathol- Hauser DL. Neuronal degeneration in the locus ceruleus and cortical correlates of Alzheimer’s disease. Alzheimer Dis Assoc Discord ogy was quite striking. At the pontine levels that were exam- 1987;1(4):256–62. ined, only neurons located within the boundaries of the LC [4] Bondareff W, Mountjoy CQ, Roth M, Rossor MN, Iversen LL, (including its subcoeruleus extension) invariably contained Reynolds GP. Age and histopathologic heterogeneity in Alzheimer’s NFT or AT8 immunoreactivity in all samples. The numerous disease: evidence for subtypes. Arch Gen Psychiat 1987;44(5):412–7. surrounding nuclei were all uniformly free of such pathol- [5] Braak E, Braak H, Mandelkow EM. A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles and neuropil threads. ogy (Fig. 4). The only other nucleus with similar pathology Acta Neuropathol 1994;87:554–67. was the serotonergic raphe nucleus, which notably is the [6] Braak H, Braak E. Evolution of the neuropathology of Alzheimer’s only other pontine nucleus with extensive axonal projections disease. Acta Neurol Scand 1996;165(Suppl.):3–12. directed to the cerebral cortex. However, few of our speci- [7] Chan-Palay V, Asan E. Alterations in catecholamine neurons of the mens contained the raphe nuclei in their entirety. Our initial locus coeruleus in senile dementia of the Alzheimer type and in Parkin- son’s disease with and without dementia and depression. J Comp Neurol impression, based on these limited observations, was that the 1989;287(3):373–92. raphe cytopathology occurred in the cognitively impaired [8] Clayton EC, Rajkowski J, Cohen JD, Aston-Jones G. Phasic activation subjects but not in those that were cognitively normal for of monkey locus ceruleus neurons by simple decisions in a forced- age. This preliminary observation, however, requires further choice task. J Neurosci 2004;24(44):9914–20. documentation. [9] Craik FIM, Salthouse T. The handbook of cognitive aging. Mahwah, NJ: Lawrence Erlbaum; 2000. In summary, our study conclusively demonstrated that LC [10] Crum RM, Anthony JC, Bassett SS, Folstein MF. Population-based tauopathy is a component of MCI and early stages of AD, sup- norms for the Mini-Mental State Examination by age and educational porting the possibility that noradrenergic therapies may have level. JAMA 1993;269:2386–91. A. Grudzien et al. / Neurobiology of Aging 28 (2007) 327–335 335

[11] Devilbiss DM, Waterhouse BD. The effects of tonic locus ceruleus (CERAD). Part II. Standardization of the neuropathologic assessment output on sensory-evoked responses of ventral posterior medial tha- of Alzheimer’s disease. Neurology 1991;41:479–86. lamic and barrel field cortical neurons in the awake rat. J Neurosci [29] Monteiro IM, Boksay I, Auer SR, Torossian C, Sinaiko E, Reisberg 2004;24(48):10773–85. B. Reliability of routine clinical instruments for the assessment of [12] Elrod R, Peskind ER, DiGiacomo L, Brodkin KI, Veith RC, Raskind Alzheimer’s disease administered by telephone. J Geriatr Psychiat Neu- MA. Effects of Alzheimer’s disease severity on cerebrospinal fluid rol 1998;11(1):18–24. norepinephrine concentration. Am J Psychiat 1997;154:25–30. [30] Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum [13] Folstein M, Folstein S, McHugh PR. Mini-mental state: a practical G, et al. The consortium to establish a registry for Alzheimer’s dis- method for grading the cognitive state of patients for the clinician. J ease (CERAD). Part I. Clinical and neuropsychological assessment of Psychiat Res 1975;12:189–98. Alzheimer’s disease. Neurology 1989;39(9):1159–65. [14] Gatter KC, Powell TPS. The projection of the locus coeruleus upon the [31] Nestor PJ, Scheltens P, Hodges JR. Advances in the early detection of neocortex in the macaque monkey. Neuroscience 1977;2:441–5. Alzheimer’s disease. Nat Med 2004;10(Suppl. 5):S34–41. [15] Geula C, Mesulam MM. Cholinergic systems in Alzheimer’s disease. [32] Ohm TG, Busch C, Bohl J. Unbiased estimation of neuronal num- In: Terry RD, Katzman R, Bick KL, Sisodia SS, editors. Alzheimer bers in the human nucleus coeruleus during aging. Neurobiol Aging Disease. 2nd ed. Philadelphia: Lippincott Williams &Wilkins; 1999. p. 1997;18(4):393–9. 269–92. [33] Oishi M, Mochizuki Y, Yushihashi H, Takasu T, Nakano E. Laboratory [16] Guillozet AL, Smiley JF, Mash DC, Mesulam MM. Butyryl- examinations correlated with severity of dementia. Ann Clin Lab Sci cholinesterase in the life cycle of amyloid plaques. Ann Neurol 1996;26:340–5. 1997;42:909–18. [34] Olszewski J, Baxter D. Cytoarchitecture of the human brain stem. 2nd [17] Hatanpaä¨ K, Brady DR, Stoll J, Rapoport SI, Chandrasekaran K. Neu- ed. Basel: Karger; 1982. ronal activity and early neurofibrillary tangles in Alzheimer’s disease. [35] Pearson J, Goldstein M, Brandeis L. Tyrosine hydroxylase immunohis- Ann Neurol 1996;40:411–20. tochemistry in human brain. Brain Res 1979;165:333–7. [18] Ishii T. Distribution of Alzheimer’s neurofibrillary changes in the [36] Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern brain stem and hypothalamus of senile dementia. Acta Neuropathol Med 2004;256:183–94. 1966;6:181–7. [37] Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen [19] Ivnik RJ, Malec JF, Smith GE, Tangalos EG, Petersen RC. Neuropsy- E. Mild cognitive impairment: clinical characterization and outcome. chological tests’ norms above age 55: COWAT, BNT, MAE Token, Arch Neurol 1999;56(3):303–8. WRAT-R Reading, AMNART, Stroop, TMT, and JLO. Clin Neuropsy- [38] Raskind MA, Peskind ER, Holmes C, Goldstein DS. Patterns of chol 1996;10:262–78. cerebrospinal fluid catechols support increased central noradrener- [20] Ivnik RJ, Malec JF, Smith GE, Tangalos EG, Peterson RC, Kokmen E, gic responsiveness in aging and Alzheimer’s disease. Biol Psychiat et al. Mayo’s older Americans normative studies: WMS-R norms for 1999;46(6):756–65. ages 56 through 94. Clin Neuropsychol 1992;6(Suppl.):49–82. [39] Robbins TW, Everitt BJ. Central norepinephrine neurons and behav- [21] Knight DP. Histochemical demonstration of the catecholamines and ior. In: Bloom FE, Kupfer DJ, editors. Neuropsychopharmacology: the acetylcholine esterase in the same cell bodies in the locus ceruleus (rat fourth generation of progress. New York: Raven Press; 1995. hind brain). Proc R Microsc Soc 1970;6:26. [40] Spreen O, Strauss E. A compendium of neuropsychological tests. New [22] Mann DMA, Yates PO, Hawkes J. The pathology of the human locus York: Oxford University Press; 1998. coeruleus. Clin Neuropathol 1983;2:1–7. [41] Tohgi H, Ueno T, Takahashi S, Nozaki Y. Concentrations of [23] Mann DMA, Yates PO, Marcyniuk B. Changes in Alzheimer’s dis- monoamines and their metabolites in the cerebrospinal fluid from ease in the magnocellular neurons of the supraoptic and paraventricular patients with senile dementia of the Alzheimer’s type and vascular nuclei of the hypothalamus and their relationship to the noradrenergic dementia of the Binswanger’s type. J Neural Transm Park Dis Dement deficit. Clin Neuropathol 1985;4(3):127–34. Sect 1992;4:69–77. [24] Mesulam MM, Geula C. Nucleus basalis (Ch4) and cortical cholinergic [42] Valentino RJ, Aston-Jones GS. Physiological and anatomical determi- innervation in the human brain: observations based on the distribu- nants of locus coeruleus discharge: behavioral and clinical implications. tion of acetylcholinesterase choline acetyltransferase. J Comp Neurol In: Bloom FE, Kupfer DJ, editors. Psychopharmacology: the fourth 1988;275:216–40. generation of progress. New York: Raven Press; 1995. [25] Mesulam MM. Neuroplasticity failure in Alzheimer’s disease: bridging [43] Wechsler D. WAIS-R manual. New York: Psychological Corporation; the gap between plaques and tangles. Neuron 1999;24(3):521–9. 1981. [26] Mesulam MM, Shaw P, Mash D, Weintraub S. Cholinergic nucleus [44] Zarow C, Lyness SA, Mortimer JA, Chui HC. Neuronal loss is basalis tauopathy emerges early in the aging-MCI-AD continuum. Ann greater in the locus coeruleus than nucleus basalis and substantia nigra Neurol 2004;55:815–28. in Alzheimer and Parkinson diseases. Arch Neurol 2003;60(3):337– [27] Mesulam MM, Van Hoesen W, Pandya DN, Geschwind N. Limbic 41. and sensory connections of the inferior parietal lobule (area PG) in the [45] Zubenko GS, Moossy J, Kopp U. Neurochemical correlates of major rhesus monkey: a study with a new method for horseradish peroxidase depression in primary dementia. Arch Neurol 1990;47(20):209–14. histochemistry. Brain Res 1977;136:393–414. [46] Zwieg RM, Ross CA, Hedreen JC, Steele C, Cardillo JE, Whitehouse [28] Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, PJ, et al. The neuropathology of aminergic nuclei in Alzheimer’s dis- et al. The consortium to establish a registry for Alzheimer’s disease ease. Ann Neurol 1988;24:233–42.