Dementia and visual hallucinations associated with limbic pathology in Parkinson's disease

Parkinsonism and Related Disorders 15 (2009) 196e204 www.elsevier.com/locate/parkreldis

Dementia and visual hallucinations associated with limbic pathology in Parkinson’s disease M.E. Kalaitzakis a, L.M. Christian a, L.B. Moran a, M.B. Graeber b, R.K.B. Pearce a, S.M. Gentleman a,* a

Department of Clinical Neuroscience, Neuropathology Unit, Division of Neuroscience and Mental Health, Imperial College Healthcare NHS Trust, London, UK b The Athenaeum, Pall Mall, London, UK Received 12 February 2008; received in revised form 3 April 2008; accepted 1 May 2008

Abstract The pathological basis of dementia and visual hallucinations in Parkinson’s disease (PD) is not yet fully understood. To investigate this further we have conducted a clinico-pathological study based on 30 post-mortem PD brains. PD cases were stratified into groups according to clinical characteristics as follows: (1) cognitively intact (n ¼ 9); (2) cases with severe dementia and visual hallucinations (n ¼ 12); (3) cases with severe dementia and no visual hallucinations (n ¼ 4); and (4) cases with severe visual hallucinations and no dementia (n ¼ 5). The extent of a-synuclein (aSyn), tau and amyloid b peptide (Ab) deposition was then examined in the CA2 sector of the hippocampus and in neocortical and subcortical areas known to subserve cognitive function. We find that dementia in PD is significantly associated with aSyn in the anterior cingulate gyrus, superior frontal gyrus, temporal cortex, entorhinal cortex, amygdaloid complex and CA2 sector of the hippocampus. Ab in the anterior cingulate gyrus, entorhinal cortex, amygdaloid complex and nucleus basalis of Meynert is also associated with dementia as is tau in the CA2 sector of the hippocampus. aSyn burden in the amygdala is strongly related to the presence of visual hallucinations but only in those PD cases with concomitant dementia. Statistical analysis revealed that aSyn burden in the anterior cingulate gyrus could differentiate demented from non-demented PD cases with high sensitivity and specificity. We conclude that aSyn in limbic regions is related to dementia in PD as well as to visual hallucinations when there is an underlying dementia. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: a-Synuclein; Parkinson’s disease; Clinico-pathological; Visual hallucinations; Dementia; Alzheimer’s disease

1. Introduction The pathological hallmark of Parkinson’s disease (PD) is the aggregation of a-synuclein (aSyn) in Lewy bodies (LBs) and in the form of Lewy neurites (LNs). While the disease is initially characterized by motor deficits, later on dementia and visual hallucinations (VH) are commonly present [1e7] although the specific anatomical and pathological correlates of these clinical features remain unclear. The presence of LBs in limbic and neocortical regions has been suggested as * Corresponding author. Tel./fax: þ44 20 8846 7680. E-mail address: [email protected] (S.M. Gentleman). 1353-8020/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.parkreldis.2008.05.007

a cause for cognitive impairment and dementia but this does not account for co-existent Alzheimer’s disease (AD)-type changes that are often found alongside aSyn pathology. In cases with mixed pathology it is difficult to determine whether LBs or AD-type pathology are responsible for dementia. In addition, to date there have been few neuropathologic studies of PD patients with VH [8e10] and the anatomico-pathological correlate(s) for VH is/are still unknown. In this study, we assessed the association between dementia and VH with aSyn and AD-type pathology (amyloid b peptide (Ab) and tau) using an experimental design that enabled us to isolate clinical features of interest and study their anatomical and pathological basis.

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

2. Materials and methods 2.1. Clinical assessment and selection of cases Clinical data of cases were compiled retrospectively from hospital records into summaries by a movement disorder neurologist (RKBP). Only subjects who had been evaluated by a clinician within 2 years prior to death and had complete clinical histories were included. The clinical diagnoses of PD, PDD and dementia with Lewy bodies (DLB) were based on published criteria [11e13]. PD was considered to be present if the patient had at least 2 of the 4 cardinal symptoms (rigidity, hypokinesia, resting tremor and postural instability) and exhibited a positive response to levodopa [12]. Patients with PD who developed late dementia (>2 years after motor symptoms) were classified as PD with dementia (PDD) [11]. The diagnosis of DLB was made if dementia preceded extrapyramidal symptoms by 2 years or developed together within a 12-month period [13]. On the basis of these guidelines, all cases that fulfilled the clinical criteria for DLB were excluded from further analysis. All cases in our cohort initially presented with motor deficits while cognitive and psychiatric complications developed later on in the disease course (range 2e34 years). The diagnosis of dementia satisfied DSM-IV [14] and ICD-10 clinical criteria. The diagnosis of VH, anxiety, depression, delusions and paranoia were made by the treating neurologist and were documented in the medical files. As our study was based on retrospective data, the clinical severity of symptoms was measured on a semi-quantitative global impression scale (0, absent; 1, mild; 2, moderate; and 3, severe) by 2 investigators (MEK and RKBP). Retrospective analysis from movement disorder specialists is an accepted method of case ascertainment and has been frequently used in clinico-pathological work studying both dementia and parkinsonism [15e17]. For the purpose of the present study, PD cases were stratified into 4 dementia status groups according to clinical characteristics. We identified 9 PD cases without dementia and without VH (with the exception of 1 case with very mild VH) and with no or only mild co-morbid neuropsychiatric features (PD(H) group); 12 PD cases with moderate to severe dementia and with VH (range 2e 3; PDD(Hþ) group); 4 PD cases with moderate to severe dementia alone without VH (range 2e3; PDD(H) group) and 5 PD cases with moderate to severe VH alone without dementia (range 2e3; PD(Hþ) group) (Table 1). Thus, a total of 30 PD cases were included in this study. For these cases gender, age at PD onset, age at death, duration of disease, as well as the clinical severity of the individual neuropsychiatric symptoms assessed are presented in Table 1. All, apart from possibly 1 case, were L-Dopa responsive (Table 1). Details of drug treatment including potential aggravating medication and cognitive enhancing agents are presented in Table 1.

2.2. Neuropathological assessment Neuropathological diagnosis was based on aSyn, tau and Ab immunohistochemistry of superior frontal gyrus, hippocampus, and midbrain. Confounding pathology was assessed on haematoxylin and eosin stained slides from 18 brain tissue blocks. Tissue was collected and processed according to an established protocol [18]. Neuropathological diagnosis was performed using international neuropathological consensus criteria for the definite diagnosis of PD (http:// www.ICDNS.org). AD pathology of isocortical and/or entorhinal types was also assessed using the generic grading system posted at http://www.ICDNS. org as well as the Braak and Braak staging to stage tangle pathology using AT8 immunohistochemistry [19]. In the present study all cases with a clinicopathological diagnosis of DLB or AD were excluded.

2.3. Immunohistochemistry Immunohistochemistry was performed using standard protocols [20]. The primary antibodies used in this study were: aSyn for identification of LBs and LNs (Becton-Dickinson, Oxfordshire, UK, at a dilution of 1:1000), tau for visualization of neurofibrillary changes (Autogen Bioclear, Wiltshire, UK, PHF-Tau (clone AT8) at a dilution of 1:800) and 1E8 for visualization of Ab plaques (Ab) (courtesy of GlaxoSmithKline, Middlesex, UK, at a dilution of 1:1000). The selection of the aSyn antibody used in the present research was based on its consistency and strong immunoreactivity with a variety of

197

aSyn-containing structures shown in a comparative study of different aSyn Abs performed by Croisier et al. [21] that was also confirmed by a more recent study by the BrainNet Europe Consortium [22] (Fig. 1). The tau antibody (clone AT8) used is routinely applied across many different centres for diagnostic purposes [23] and it has also recently been used as the antibody of choice for staging AD neurofibrillary pathology according to the Braak and Braak scheme [19] (Fig. 1). The 1E8 antibody has been raised against residues 18e22 Ab peptide and labels the full range of Ab-related pathology; diffuse, cored plaques and deposits in the walls of cerebral blood vessels [24,25] (Fig. 1).

2.4. Semi-quantitative assessment of aSyn and AD lesions For each case, 3 representative sections from the superior frontal gyrus, anterior cingulate gyrus (BA 24; ACG), temporal cortex, entorhinal cortex, hippocampus CA2 sector, subiculum, amygdaloid complex and nucleus basalis of Meynert (NBM) were assessed for the occurrence and extent of aSyn, Ab and tau deposition. Sections were screened in their entirety at 10 primary magnification for overall deposit burden. Assessment of pathology was carried out using a CERAD-based [26] visual impression of the numerical density of deposits ranging from 1 to 3 corresponding to sparse, moderate and frequent. A value of 0.5 was assigned to those fields where only a single lesion was present. In a few cases, where the density did not clearly fit into one of the main categories, intermediate values were assigned. The occipital lobe was also examined in the same manner for the PD(H) and PD(Hþ) groups only. Sections were graded by 2 investigators (MEK and LMC). Cohen’s kappa statistics revealed an inter-rater reliability of 0.8 for aSyn, 0.9 for tau and 0.86 for Ab.

2.5. Statistical analysis The association of a-synuclein (aSyn), tau and Ab with dementia and VH plus the interaction between these 2 factors was tested using a 2-way ANOVA in each anatomical region. Although these measures are on an ordinal scale, the use of a parametric ANOVA is appropriate (as opposed to a non-parametric ordinal regression) because the number of items in the scale is large. The extent of aSyn, tau and Ab burden in the occipital lobe between the PD(H) and PD(Hþ) cases was examined with the use of the non-parametric ManneWhitney U-test. The interaction between the clinical variables age at death and duration of disease with the neuropathological variables (aSyn, tau and Ab) was assessed using Spearman’s 2-tailed correlation analysis (non-parametric). Ordinal regression was applied to detect relationships between age at disease onset, age at death and duration of disease with dementia and VH in our cohort. Cohen’s kappa statistic was used to test inter-rater reliability for the aSyn, tau and Ab semi-quantitative assessment between the 2 investigators. Statistical analyses were performed using the SPSS programme version 15.0 for Windows XP (SPSS, Inc., Chicago, IL). p values <0.05 were considered significant. The diagnostic performance of aSyn and Ab deposition in predicting dementia (regardless of the presence of VH) was calculated with measures of sensitivity, positive predictive value (PPV), specificity and negative predictive value (NPV) in each region studied. The methodology was adopted from Harding et al. [27]. Briefly, sensitivity refers to the probability of having moderate to severe aSyn and/or Ab deposition with high clinical scores for dementia. Specificity refers to the probability of not having high scores for dementia if moderate to severe aSyn and/or Ab burden is not reached. PPV refers to the probability of having high scores for dementia in the presence of moderate to severe aSyn and/or Ab deposition. NPV refers to the probability of not having high scores for dementia in the presence of moderate to severe aSyn and/or Ab burden.

3. Results For the 30 PD cases included in this study (Table 1) age at disease onset and age at death did not differ between groups, however PD(Hþ) cases had a statistically significant longer disease duration ( p ¼ 0.03). All PD cases had received

198

Table 1 Clinical data for the PD groups Case Gender Age Age at Duration of Neuropathological ICDNS AD grade number at onset death disease diagnosis (http:// (http://www.ICDNS. (years) (years) (years) www.ICDNS.org) org)

58

6

PD

2

m

76

85

9

PD

3

m

47

74

27

PD

4

m

75

82

7

PD

5 6

m f

68 65

78 75

10 10

PD PD

7

m

70

77

7

PD

8

m

86

89

3

PD

9

m

67

77

10

PD

67

77

PDD(HD) group 10 f 57

80

23

PD

11

m

77

82

5

PD

12

m

56

70

14

PD

13

m

72

81

9

PD

14

m

57

73

16

PD

15

m

71

82

11

PD

16

m

75

84

9

PD

17

f

67

85

18

PD

18

m

74

87

13

PD

19

m

54

78

24

PD

20

f

73

84

11

PD

21

m

78

84

6

PD

Mean

D

VH

A

Other medication

De Pr Dl

Isocortical 0eentorhinal I Isocortical IIeentorhinal 0 Isocortical Ieentorhinal 0 Isocortical Ieentorhinal II na Isocortical 0eentorhinal 0 Isocortical 0eentorhinal I Isocortical 0eentorhinal I Isocortical 0eentorhinal I

IeII

0

0

1e2 0

0

0



Madopar



IeII

0

0

1

0

0

0

þ



IeII

0

0

0

1

0

0

þ

Sinemet, Madopar, Pergolide Sinemet, Pergolide

IeII

0

0

0

1

0

0

þ

Sinemet



IeII 0

0 0

0 0

0 0

0 0

0 0

0 0

þ þ

Amantadine Madopar

 

IeII

0

1

0

0

0

0

þ

Madopar



IeII

0

0

0

0

0

0

þ

Madopar



IeII

0

0

0

0

0

0

þ

Madopar



Isocortical IIeentorhinal III Isocortical Ieentorhinal II Isocortical IVeentorhinal III Isocortical 0eentorhinal I Isocortical Ieentorhinal II Isocortical Ieentorhinal II Isocortical 0eentorhinal I Isocortical IIeentorhinal I Isocortical Ieentorhinal II Isocortical 0eentorhinal I Isocortical IVeentorhinal III Isocortical

IeII

2

3

0

0

0

0

þ



0

2e3 3

1

1

0

0

þ

Sinemet, Amantadine Madopar

III

2e3 2

1

0

0

1

þ

Sinemet, Madopar

Olanzapine

IeII

2e3 2

1e2 0

0

0

þ

Sinemet, Madopar



IeII

2

2

0

0

0

0

þ

Olanzapine

IeII

2

2e3 0

1

0

1e2

þ

Benzhexol, Sinemet, Madopar Ropinirole, Sinemet

Rivastigmine

0

2

2

3

2

0

0

þ

Pergolide, Sinemet

Donepezil, Risperidone

IeII

2e3 2e3 1

1

0

0

þ



IeII

2

3

0

0

0

0

þ

Bromocriptine, Madopar, Sinemet Pergolide, Madopar

IeII

2e3 3

2

0

0

0

þ

Olanzapine, Melleril

III

3

2

3

0

0

0

þ

Sinemet, Amantadine, Cabergoline Sinemet

Diazepam, Olanzapine

IeII

2

2

0

0

0

0

þ

Pergolide, Sinemet,

Donepezil, Olanzapine



9.8

Donepezil



M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

PD(HL) group 1 f 52

L-Dopa Antiparkinsonian responsiveness medication

Braak AD Clinical severity stages scores of neuropsychiatric features

IIeentorhinal II Mean

80

13.2

PDD(HL) group 22 m 62

71

9

PD

23

m

61

80

19

PD

24

m

35

54

19

PD

25

m

67

79

12

56

71

14.7

PD(HD) group 26 f 67

80

27

f

59

28

m

29 30

Isocortical IIIeentorhinal III Isocortical 0eentorhinal III Isocortical 0e entorhinal I

IeII

2

0

0

0

0

0

þ

Sinemet



III

2e3 0

2

2

0

1

þ

Sinemet



0

2

1e2 0

0

0

0

þ



PD

Isocortical IVeentorhinal III

IeII

3

1

0

0

0

0

þ

Bromocriptine, Sinemet, Pergolide, Madopar Madopar

13

PD

IeII

0

3

0

0

0

0

þ

Sinemet



79

20

PD

IeII

0

2

1e2 1

0

0

þ

62

20

PD

0

0

2e3 0

0

0

0

þ

m

41

75

34

PD

IeII

0

2

0

1

0

0

þ

m

48

76

28

PD

IeII

0

2e3 1

0

0

0

þ

Sinemet, Amantadine, Cabergoline Pergolide, Ropinirole, Pramipexole Pergolide, Sinemet, Artane Amantadine, Pergolide, Sinemet, Madopar

Risperidone

42

Isocortical 0eentorhinal I Isocortical 0eentorhinal I Isocortical 0eentorhinal 0 Isocortical Ieentorhinal II Isocortical 0eentorhinal I

51

74

23

Mean

Mean

Ropinirole



 Quatiapine 

PDD(H), PD cases with high clinical scores for dementia and no VH; PD(Hþ), PD cases with high clinical scores for VH and no dementia; PDD(Hþ), PD cases with high clinical scores for both dementia and VH; PD(H), cognitively intact PD cases; D, dementia; VH, visual hallucinations; A, anxiety; De, depression; Pr, paranoia; Dl, delusions; PD, Parkinson’s disease; m, male; f, female. Semi-quantitative clinical severity scores: 0, absent; 1, mild; 2, moderate; 3, severe; þ, yes; , no; , possibly; na, not available.

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

67

199

200

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

Fig. 1. Illustrations of aSyn (a), Ab (b) and tau (c) immunostaining. Magnification 400. Abbreviations: aSyn, a-synuclein; Ab, amyloid beta peptide.

medication with the potential to exacerbate hallucinations (Table 1). Statistical analysis revealed a significant association between dementia in PD and aSyn burden in the ACG ( p ¼ 0.001), superior frontal gyrus ( p ¼ 0.02), temporal cortex ( p ¼ 0.02), entorhinal cortex ( p ¼ 0.002), amygdaloid complex ( p ¼ 0.007) and CA2 sector of the hippocampus ( p ¼ 0.01) (Table 2). Although aSyn burden did not associate with VH when present in isolation (PD(Hþ) group), a strong association was found between aSyn burden in the amygdaloid complex ( p ¼ 0.01) and VH in PD cases with concomitant dementia (PDD(Hþ)) (Fig. 2). Of note is that aSyn load in the NBM did not associate with dementia and was severe in all groups studied independent of clinical phenotype.

A strong association was found between dementia and Ab load in the ACG ( p ¼ 0.01), entorhinal cortex ( p ¼ 0.002), amygdaloid complex ( p ¼ 0.004) and NBM ( p ¼ 0.02) whereas no association was detected between Ab burden in PD cases with VH in isolation or with concomitant dementia (Table 2). Although tau pathology occurred mildly in all groups, tau burden in the CA2 sector of the hippocampus significantly associated with the presence of dementia ( p ¼ 0.01) (Table 2). The associations observed did not correlate with age at death or duration of disease. Neither aSyn nor tau was detected in the occipital lobe of PD(H) or PD(Hþ) cases while the occipital Ab burden was mild in both groups. In order to determine if there was an anatomico-pathological feature that would differentiate cases with or without

Fig. 2. Immunostaining for aSyn in the anterior cingulate gyrus of a PD case without dementia or VH (a) in contrast to a case exhibiting both dementia and VH (b). Immunostaining for aSyn in the amygdala of a PD case without VH (c) and a case with VH associated with dementia (d). Magnification 200. Abbreviations: PD, Parkinson’s disease; VH, visual hallucinations; aSyn, a-synuclein.

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204 Table 2 Statistical associations between pathology in neocortical and subcortical regions with dementia and VH in PD patients Brain area

Dementia (n ¼ 16)

Visual hallucinations (n ¼ 17)

aSyn Superior frontal gyrus Anterior cingulate gyrus Temporal cortex Entorhinal cortex Hippocampus CA2 sector Subiculum Amygdaloid complex Nucleus basalis of Meynert

0.02 0.001 0.02 0.002 0.01 0.08 0.007a 0.4

0.8 0.6 0.4 0.9 0.2 0.6 0.9 0.5

Ab Superior frontal gyrus Anterior cingulate gyrus Temporal cortex Entorhinal cortex Hippocampus CA2 sector Subiculum Amygdaloid complex Nucleus basalis of Meynert

0.08 0.01 0.07 0.002 0.1 0.07 0.004 0.06

0.5 0.3 0.4 0.9 0.4 0.7 0.7 0.9

Tau Superior frontal gyrus Anterior cingulate gyrus Temporal cortex Entorhinal cortex Hippocampus CA2 sector Subiculum Amygdaloid complex Nucleus basalis of Meynert

0.2 0.5 0.7 0.4 0.01 0.3 0.1 0.07

0.8 0.8 0.6 0.5 0.9 0.5 0.6 0.3

a

In PD cases with dementia and concomitant visual hallucination p ¼ 0.01. PD, Parkinson’s disease; VH, visual hallucinations.

dementia (irrespective of the presence of VH; i.e. PDD(Hþ) and PDD(H) vs. PD(H) and PD(Hþ) groups) the sensitivity PPV, specificity and NPV were calculated for each area. The sensitivity and PPV for moderate to severe aSyn burden in cases with dementia (PDD(Hþ) and PDD(H) group) was 75e85% in the ACG (Table 3). The sensitivity and PPV for Ab deposition, was 25 and 80% respectively. This indicates that aSyn in the ACG is a better pathological marker for detecting cases with dementia with or without VH Fig. 2. Furthermore, the specificity and NPV for cases without dementia (i.e. PD(H) and PD(Hþ) groups) was high in the range of 75e85%. 4. Discussion Dementia is a common feature in PD but its anatomical correlates remain a matter of controversy. In our cases dementia was associated with significant degenerative changes in the limbic system. While the number of our cases with dementia was relatively small (n ¼ 16), pathological changes in the amygdaloid complex, entorhinal cortex, ACG and CA2 sector of the hippocampus seem likely responsible for dementia in PD. Other recently published findings demonstrate a relationship between LB densities in the parahippocampal gyrus and amygdala [27e31] and LNs in the CA2 sector of the

201

Table 3 Two by two contingency table for calculation of sensitivity, PPV, specificity and NPV for aSyn burden in the anterior cingulate gyrus in demented PD cases regardless of visual hallucinations (PDD(Hþ) and PDD(H)) group Anterior cingulate gyrus aSyn burden

2e3 0e1 Total cases

Clinically demented

Total number of cases

Yes

No

12 4 16

2 12 14

14 16 30

aSyn, a-synuclein; PDD(Hþ), cases with high clinical scores for both dementia and visual hallucinations; PDD(H), cases with high clinical scores for dementia but no visual hallucinations; PPV, positive predictive value; NPV, negative predictive value. Sensitivity of demented PD cases (12/16100), 75%; PPV for demented PD cases (12/14100), 85%; specificity for cases without dementia and VH (12/14100), 85%; NPV for cases without dementia and VH (12/16100), 75%.

hippocampus, with dementia in PD [27,29]. Our data are in agreement with these studies, pointing to the presence of limbic dysfunction as a central basis of dementia in PD. It is also of interest that the extent of aSyn pathology in the NBM was high in both demented and cognitively intact cases highlighting a universal cholinergic deficit in PD, with or without clinically evident cognitive decline. Some studies have demonstrated that increased numbers of cortical LBs correlate with cognitive decline independently of or in addition to AD-type pathology [11,32,33]. The results of our study suggest aSyn pathology as an important correlate of dementia in PD. This is in agreement with a recent prospective community-based study indicating that LB pathology and not AD histopathological changes drive the progression of cognitive impairment in PD [34]. However, AD-type pathology was also associated with dementia in our cohort. This finding demonstrates that AD-type pathology in cortical and limbic regions may have a permissive effect in precipitating dementia in PD. AD pathology in other brain areas may be important for the development of overt dementia. We have recently reported a strong association between dementia in PD and striatal b-amyloid burden [35]. In addition, recent studies employing both immunohistochemical and biochemical techniques show that LBs in the cerebral cortex in PD are significantly increased in brains with a high Ab load pointing to an interaction between these 2 pathologies at a molecular level [36,37]. The pathological and clinical significance of LNs in PD is still uncertain. The CA2 sector of the hippocampus shows a preferential susceptibility to LN-type pathology whereas the neighbouring hippocampal sectors are relatively spared. Although studies have focused on cortical and limbic LBtype pathology as the basis of dementia, in this report a strong association was observed between aSyn burden (i.e. LNs) in the CA2 sector of the hippocampus and dementia. Earlier studies have reported a positive correlation between LN burden in the CA2 sector with the presence of dementia in PD [10], however this correlation has been viewed with caution as ubiquitin immunohistochemistry was used. In this study we confirm with the use of aSyn immunohistochemistry the existence of

202

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

prominent LN-type pathology in the CA2 sector of the hippocampus and strengthen previous findings of an association between LNs in the CA2 sector of the hippocampus with dementia in PD. Tau burden in the CA2 sector was also significantly associated with dementia in PD further supporting a primary role of this anatomical region for the presence of dementia in PD. Previous studies have shown that high LB densities in the cingulate cortex are observed in other conditions such as DLB and Alzheimer’s disease [27,38e40]. However, our results indicate that aSyn burden in the ACG could differentiate cases with dementia from those without with high sensitivity and specificity. The positive predictive value (i.e. the probability of having high scores for dementia in the presence of moderate to severe aSyn burden in the ACG) was up to 85%, further supporting a primary role of aSyn in the pathological processes of cognitive dysfunction in PD as well as pointing to the ACG as an important focus of damage in PDD. Our results are supported by imaging studies demonstrating a decreased metabolic pattern in the ACG of PDD subjects when compared to PD and control cases [41]. Given that the ACG functionally contributes to cognitive and attentional processes as well as to emotional states [42e45], it is not difficult to envision that pathological changes in this area will play a primary role in the development of cognitive deficits and dementia in PD. We have recently reported a robust association between striatal pathology and dementia in PD [35]. The striatum processes motor (i.e. putamen) as well as cognitive (i.e. caudate nucleus) and emotional or motivational information (i.e. ventral striatum) and is reciprocally connected with the limbic system. Specifically, there are reciprocal connections between the dorsal caudate nucleus and the prefrontal cortex [46]. The ventral striatum receives allo-mesocortical input and projects to anterior cingulate cortex and medial orbitofrontal cortices [46,47]. It has also been shown that the shell of the nucleus accumbens is closely linked to or continuous with the anterior extension of the amygdala [48]. Therefore, these structures reside in circuitry important in cognitive function. Further larger studies are required to determine which areas have the most significant effects for the manifestation of dementia. Recent surveys have revealed that anywhere from 6 to 60% of PD patients experience VH [6,7]. Risk factors for VH include higher age, duration of disease, cognitive impairment, medication, depression and sleep disturbances. In our group of patients characterized by VH but no dementia (PD(Hþ)) we observed an association with longer duration of disease which is in agreement with previous reports [49e52]. The most frequently incriminated drugs for the occurrence of VH in PD patients are dopaminergics [53e55], anticholinergics [56,57] and amantadine [58]. Interestingly, however, all cases in this study (with and without VH) received drugs with the potential to exacerbate hallucinations. Data from a community-based study involving 235 PD patients demonstrated an association of hallucinations with advanced age, cognitive decline and depression but not with anti-parkinsonian medication [2]. In another study including 102 PD patients, VH were

associated with lower cognitive scores, lower visual acuity, higher Unified Parkinson Disease Rating scores and higher Global Deterioration Scale scores [59]. However, there was no association found with the use of dopaminergic drugs, other medications, or L-Dopa dosage or duration of treatment [59]. Therefore, the findings in our study support previous reports [2,59,60] of no simple direct causeeeffect connection between VH and medication. Thus, it appears that medication per se could not account (at least directly) for the presence of VH. This underscores that a pathological substrate has a primary role in development of VH with a permissive effect of drug administration. In our cohort aSyn burden in the amygdaloid complex significantly associated with VH in demented PD cases. Interestingly, VH in non-demented PD subjects (PD(Hþ)) did not associate with aSyn burden in the amygdaloid complex. Previous studies by Harding et al. have shown that LB burden in the amygdala positively associates with VH in both demented and non-demented PD patients [8,10]. Recently, Papapetropoulos and colleagues have also reported an increased LB burden in the amygdala, frontal, temporal and parietal cortices in PD subjects with VH [9]. The findings of the present study are in line with clinico-pathological studies pointing to aSyn in the amygdala as an important pathological substrate of VH in demented PD patients. The absence of association of aSyn deposition in the amygdala with VH in non-demented patients may demonstrate that dementia, as an index of a general degradation of information processing, predisposes for the occurrence of VH and therefore an increased burden of aSyn in amygdala may not be related to the presence of VH per se but an underlying dementia is a significant determinant for it. It is therefore not surprising that significant positive correlations of cognitive impairment and VH have been reported in previous studies [61]. However, the number of cases in the PD(Hþ) group was small which may account for the discrepancy between this report and that of Harding et al. [10]. Further studies are needed to unravel the pathological basis of VH in both demented and non-demented PD patients. Our report has several strengths including the relatively large number of cases studied (n ¼ 30) and the availability of detailed clinical documentation. However, it has those limitations commonly seen in retrospective medical note analysis and it is possible that some patients in the PD(H) group had cognitive impairment or VH which were not recorded, leading to an underestimate of their incidence. The results of the present study indicate a clinical relevance of aSyn lesions in the limbic system with respect to the occurrence of dementia and VH in PD. The significance of the observed associations requires further elucidation.

Acknowledgements This work was funded by the UK Parkinson’s Disease Society, registered charity 948776. SMG is funded in part by NIH grant AG12411. Tissue samples were supplied by the Parkinson’s Disease Society Tissue Bank at Imperial College

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

London, funded by the Parkinson’s Disease Society of the United Kingdom. The help of tissue bank members is greatly appreciated. The authors would also like to thank Dr Federico Turkheimer for his statistical advice. We express our deepest appreciation to the donors and their families for donating human brain tissue for research.

References [1] Aarsland D, Tandberg E, Larsen JP, Cummings JL. Frequency of dementia in Parkinson disease. Arch Neurol 1996;53:538e42. [2] Aarsland D, Larsen JP, Cummins JL, Laake K. Prevalence and clinical correlates of psychotic symptoms in Parkinson disease: a communitybased study. Arch Neurol 1999;56:595e601. [3] Aarsland D, Larsen JP, Lim NG, Janvin C, Karlsen K, Tandberg E, et al. Range of neuropsychiatric disturbances in patients with Parkinson’s disease. J Neurol Neurosurg Psychiatry 1999;67:492e6. [4] Brown RG, Marsden CD. How common is dementia in Parkinson’s disease? Lancet 1984;2:1262e5. [5] Friedman A, Barcikowska M. Dementia in Parkinson’s disease. Dementia 1994;5:12e6. [6] Diederich NJ, Goetz CG, Stebbins GT. Repeated visual hallucinations in Parkinson’s disease as disturbed external/internal perceptions: focused review and a new integrative model. Mov Disord 2005;20:130e40. [7] Cummings JL. Behavioral complications of drug treatment of Parkinson’s disease. J Am Geriatr Soc 1991;39:708e16. [8] Harding AJ, Broe GA, Halliday GM. Visual hallucinations in Lewy body disease relate to Lewy bodies in the temporal lobe. Brain 2002; 125:391e403. [9] Papapetropoulos S, McCorquodale DS, Gonzalez J, Jean-Gilles L, Mash DC. Cortical and amygdalar Lewy body burden in Parkinson’s disease patients with visual hallucinations. Parkinsonism Relat Disord 2006; 12:253e6. [10] Harding AJ, Stimson E, Henderson JM, Halliday GM. Clinical correlates of selective pathology in the amygdala of patients with Parkinson’s disease. Brain 2002;125:2431e45. [11] Hurtig HI, Trojanowski JQ, Galvin J, Ewbank D, Schmidt ML, Lee VM, et al. Alpha-synuclein cortical Lewy bodies correlate with dementia in Parkinson’s disease. Neurology 2000;54:1916e21. [12] Daniel SE, Lees AJ. Parkinson’s Disease Society Brain Bank, London: overview and research. J Neural Transm 1993;39(Suppl):165e72. [13] McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, et al. Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology 1996;47:1113e24. [14] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Press; 1994. [15] Litvan I, MacIntyre A, Goetz CG, Wenning GK, Jellinger K, Verny M, et al. Accuracy of the clinical diagnoses of Lewy body disease, Parkinson disease, and dementia with Lewy bodies: a clinicopathologic study. Arch Neurol 1998;55:969e78. [16] Papapetropoulos S, Lieberman A, Gonzalez J, Mash DC. Can Alzheimer’s type pathology influence the clinical phenotype of Parkinson’s disease? Acta Neurol Scand 2005;111:353e9. [17] Papapetropoulos S, Gonzalez J, Lieberman A, Villar JM, Mash DC. Dementia in Parkinson’s disease: a post-mortem study in a population of brain donors. Int J Geriatr Psychiatry 2005;20:418e22. [18] Vonsattel JP, Aizawa H, Ge P, DiFiglia M, McKee AC, MacDonald M, et al. An improved approach to prepare human brains for research. J Neuropathol Exp Neurol 1995;54:42e56. [19] Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, DelTredici K. Staging of Alzheimer disease-associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol (Berl) 2006; 112:389e404.

203

[20] Parkkinen L, Kauppinen T, Pirttila T, Autere JM, Alafuzoff I. Alpha-synuclein pathology does not predict extrapyramidal symptoms or dementia. Ann Neurol 2005;57:82e91. [21] Croisier E, MRes DE, Deprez M, Goldring K, Dexter DT, Pearce RK, et al. Comparative study of commercially available anti-alpha-synuclein antibodies. Neuropathol Appl Neurobiol 2006;32:351e6. [22] Alafuzoff I, Parkkinen L, Al-Sarraj S, Arzberger T, Bell J, Bodi I, et al. Assessment of alpha-synuclein pathology: A study of the BrainNet Europe Consortium. J Neuropathol Exp Neurol 2008;67:125e43. [23] Alafuzoff I, Pikkarainen M, Al-Sarraj S, Arzberger T, Bell J, Bodi I, et al. Interlaboratory comparison of assessments of Alzheimer disease-related lesions: a study of the BrainNet Europe Consortium. J Neuropathol Exp Neurol 2006;65:740e57. [24] Gentleman SM, Greenberg BD, Savage MJ, Noori M, Newman SJ, Roberts GW, et al. A beta 42 is the predominant form of amyloid beta-protein in the brains of short-term survivors of head injury. Neuroreport 8 1997;8:1519e22. [25] Shepherd CE, Grace EM, Mann DM, Halliday GM. Relationship between neuronal loss and ’inflammatory plaques’ in early onset Alzheimer’s disease. Neuropathol Appl Neurobiol 2007;33:328e33. [26] Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, et al. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part II. Standardization of the neuropathologic assessment of Alzheimer’s disease. Neurology 1991;41:479e86. [27] Harding AJ, Halliday GM. Cortical Lewy body pathology in the diagnosis of dementia. Acta Neuropathol (Berl) 2001;102:355e63. [28] Apaydin H, Ahlskog JE, Parisi JE, Boeve BF, Dickson DW. Parkinson disease neuropathology: later-developing dementia and loss of the levodopa response. Arch Neurol 2002;59:102e12. [29] Churchyard A, Lees AJ. The relationship between dementia and direct involvement of the hippocampus and amygdala in Parkinson’s disease. Neurology 1997;49:1570e6. [30] Kovari E, Gold G, Herrmann FR, Canuto A, Hof PR, Bouras C, et al. Lewy body densities in the entorhinal and anterior cingulate cortex predict cognitive deficits in Parkinson’s disease. Acta Neuropathol (Berl) 2003;106:83e8. [31] Mattila PM, Rinne JO, Helenius H, Roytta M. Neuritic degeneration in the hippocampus and amygdala in Parkinson’s disease in relation to Alzheimer pathology. Acta Neuropathol (Berl) 1999;98:157e64. [32] Mattila PM, Rinne JO, Helenius H, Dickson DW, Roytta M. Alpha-synuclein-immunoreactive cortical Lewy bodies are associated with cognitive impairment in Parkinson’s disease. Acta Neuropathol (Berl) 2000; 100:285e90. [33] Mattila PM, Roytta M, Torikka H, Dickson DW, Rinne JO. Cortical Lewy bodies and Alzheimer-type changes in patients with Parkinson’s disease. Acta Neuropathol (Berl) 1998;95:576e82. [34] Aarsland D, Perry R, Brown A, Larsen JP, Ballard C. Neuropathology of dementia in Parkinson’s disease: a prospective, community-based study. Ann Neurol 2005;58:773e6. [35] Kalaitzakis ME, Graeber MB, Gentleman SM, Pearce RK. Striatal betaamyloid deposition in Parkinson disease with dementia. J Neuropathol Exp Neurol 2008;67:155e61. [36] Pletnikova O, West N, Lee MK, Rudow GL, Skolasky RL, Dawson TM, et al. Abeta deposition is associated with enhanced cortical alpha-synuclein lesions in Lewy body diseases. Neurobiol Aging 2005;26:1183e92. [37] Lashley T, Holton JL, Gray E, Kirkham K, O’Sullivan SS, Hilbig A, et al. Cortical alpha-synuclein load is associated with amyloid-beta plaque burden in a subset of Parkinson’s disease patients. Acta Neuropathol (Berl) 2008;115:417e25. [38] Braak H, Braak E, Yilmazer D, de Vos RA, Jansen EN, Bohl J. Pattern of brain destruction in Parkinson’s and Alzheimer’s diseases. J Neural Transm 1996;103:455e90. [39] Lippa CF, Schmidt ML, Lee VM, Trojanowski JQ. Antibodies to alphasynuclein detect Lewy bodies in many Down’s syndrome brains with Alzheimer’s disease. Ann Neurol 1999;45:353e7. [40] Rezaie P, Cairns NJ, Chadwick A, Lantos PL. Lewy bodies are located preferentially in limbic areas in diffuse Lewy body disease. Neurosci Lett 1996;212:111e4.

204

M.E. Kalaitzakis et al. / Parkinsonism and Related Disorders 15 (2009) 196e204

[41] Yong SW, Yoon JK, An YS, Lee PH. A comparison of cerebral glucose metabolism in Parkinson’s disease, Parkinson’s disease dementia and dementia with Lewy bodies. Eur J Neurol 2007;14:1357e62. [42] Bush G, Vogt BA, Holmes J, Dale AM, Greve D, Jenike MA, et al. Dorsal anterior cingulate cortex: a role in reward-based decision making. Proc Natl Acad Sci U S A 2002;99:523e8. [43] Carter CS, Macdonald AM, Botvinick M, Ross LL, Stenger VA, Noll D, et al. Parsing executive processes: strategic vs. evaluative functions of the anterior cingulate cortex. Proc Natl Acad Sci U S A 2000;97:1944e8. [44] Devinsky O, Morrell MJ, Vogt BA. Contributions of anterior cingulate cortex to behaviour. Brain 1995;118(Pt 1):279e306. [45] Paus T. Primate anterior cingulate cortex: where motor control, drive and cognition interface. Nat Rev Neurosci 2001;2:417e24. [46] Parent A. Extrinsic connections of the basal ganglia. Trends Neurosci 1990;13:254e8. [47] Parent A, Hazrati LN. Functional anatomy of the basal ganglia. I. The cortico-basal gangliaethalamo-cortical loop. Brain Res Brain Res Rev 1995;20:91e127. [48] Heimer L, Alheid GF, de Olmos JS, Groenewegen HJ, Haber SN, Harlan RE, et al. The accumbens: beyond the coreeshell dichotomy. J Neuropsychiatry Clin Neurosci 1997;9:354e81. [49] Barnes J, David AS. Visual hallucinations in Parkinson’s disease: a review and phenomenological survey. J Neurol Neurosurg Psychiatry 2001;70:727e33. [50] Fenelon G, Mahieux F, Huon R, Ziegler M. Hallucinations in Parkinson’s disease: prevalence, phenomenology and risk factors. Brain 2000; 123(Pt 4):733e45.

[51] Graham JM, Grunewald RA, Sagar HJ. Hallucinosis in idiopathic Parkinson’s disease. J Neurol Neurosurg Psychiatry 1997;63:434e40. [52] Sanchez-Ramos JR, Ortoll R, Paulson GW. Visual hallucinations associated with Parkinson disease. Arch Neurol 1996;53:1265e8. [53] Moskovitz C, Moses H, Klawans HL. Levodopa-induced psychosis: a kindling phenomenon. Am J Psychiatry 1978;135:669e75. [54] Serby M, Angrist B, Lieberman A. Mental disturbances during bromocriptine and lergotrile treatment of Parkinson’s disease. Am J Psychiatry 1978;135:1227e9. [55] Feigenson JS, Sweet RD, McDowell FH. Piribedil: its synergistic effect in multidrug regimens for parkinsonism. Neurology 1976;26: 430e3. [56] de Smet Y, Ruberg M, Serdaru M, Dubois B, Lhermitte F, Agid Y. Confusion, dementia and anticholinergics in Parkinson’s disease. J Neurol Neurosurg Psychiatry 1982;45:1161e4. [57] Goetz CG, Tanner CM, Klawans HL. Pharmacology of hallucinations induced by long-term drug therapy. Am J Psychiatry 1982;139:494e7. [58] Harper RW, Knothe BU. Coloured Lilliputian hallucinations with amantadine. Med J Aust 1973;1:444e5. [59] Holroyd S, Currie L, Wooten GF. Prospective study of hallucinations and delusions in Parkinson’s disease. J Neurol Neurosurg Psychiatry 2001;70: 734e8. [60] Williams DR, Lees AJ. Visual hallucinations in the diagnosis of idiopathic Parkinson’s disease: a retrospective autopsy study. Lancet Neurol 2005;4:605e10. [61] Poewe W. Psychosis in Parkinson’s disease. Mov Disord 2003;18(Suppl. 6):S80e7.