Acta Neuropathologica

, Volume 119, Issue 6, pp 715–722

α-Synuclein pathology in the spinal cord autonomic nuclei associates with α-synuclein pathology in the brain: a population-based Vantaa 85+ study

Authors

    • Department of Pathology, Helsinki University Central HospitalUniversity of Helsinki
    • Department of Neurosurgery, Helsinki University Central HospitalUniversity of Helsinki
  • Anders Paetau
    • Department of Pathology, Helsinki University Central HospitalUniversity of Helsinki
  • Liisa Myllykangas
    • Department of Pathology, Helsinki University Central HospitalUniversity of Helsinki
  • Irma-Leena Notkola
    • Finnish Information Centre for Register Research
  • Hannu Kalimo
    • Department of Pathology, Helsinki University Central HospitalUniversity of Helsinki
  • Tuomo Polvikoski
    • Institute for Ageing and Health, Campus of Ageing and VitalityNewcastle University
Original Paper

DOI: 10.1007/s00401-009-0629-6

Cite this article as:
Oinas, M., Paetau, A., Myllykangas, L. et al. Acta Neuropathol (2010) 119: 715. doi:10.1007/s00401-009-0629-6

Abstract

In most subjects with Parkinson’s disease and dementia with Lewy bodies, α-synuclein (αS) immunoreactive pathology is found not only in the brain but also in the autonomic nuclei of the spinal cord. However, neither has the temporal course of αS pathology in the spinal cord in relation to the brain progression been established, nor has the extent of αS pathology in the spinal cord been analyzed in population-based studies. Using immunohistochemistry, the frequency and distribution of αS pathology were assessed semiquantitatively in the brains and spinal cord nuclei of 304 subjects who were aged at least 85 in the population-based Vantaa 85+ study. αS pathology was common in the spinal cord; 102 (34%) subjects had classic αS pathology in the thoracic and/or sacral autonomic nuclei. Moreover, 134 (44%) subjects showed grain- or dot-like immunoreactivity in neuropil (mini-aggregates) without classic Lewy neurites or Lewy bodies (LBs). The latter type of αS accumulation is associated with age, but also the classic αS pathology was found more often in the oldest compared to the youngest age group. The severity of αS pathology in the spinal cord autonomic nuclei is significantly associated with the extent and severity of αS pathology in the brain. Of the subjects, 60% with moderate to severe thoracic αS pathology and up to 89% with moderate to severe sacral αS pathology had diffuse neocortical type of LB pathology in the brain. αS pathology exclusively in the spinal cord was rare. Our study indicates that in general αS pathology in the spinal cord autonomic nuclei is associated with similar pathology in the brain.

Keywords

α-SynucleinSpinal cordDementia with Lewy bodiesParkinson’s diseasePopulation-based studyαS pathologyLewy-related pathology

Introduction

α-Synuclein (αS) is the major protein component of Lewy bodies (LBs) and Lewy neurites (LNs) [24] found in Parkinson’s disease (PD) and dementia with Lewy bodies (DLB), and also present in the lower brainstem nuclei and anterior olfactory structures of some neurologically unimpaired subjects [8]. The etiologies behind the aggregation of this natively unfolded pre-synaptic nerve terminal protein to form αS pathology/Lewy-related pathology remain unknown. In PD, αS pathology is most common in the dopaminergic projection cells of the substantia nigra (SN) pars compacta resulting in the devastation of these neurons, a process which eventually is sufficient to cause rigidity and bradykinesia [7]. When this pathological process extends to limbic and neocortical areas it associates with the clinical syndrome of PD dementia (PDD). Thus, the distribution of αS pathology in PD is assumed to follow a topographic sequence where the αS pathology proceeds to the rostral upper stages after the brainstem has become completely affected [6]. In DLB, αS pathology in the brainstem is required for the neuropathologic diagnosis, but limbic and diffuse neocortical distribution of αS pathology associates better with the clinical syndrome of DLB [16]. Furthermore in DLB, the most severe αS pathology and neuron loss in the SN have been detected in association with diffuse neocortical distribution of αS pathology [18].

In addition to the brain, the autonomic nuclei of the spinal cord are often affected in PD and DLB [11, 25], as well as in neurologically asymptomatic individuals with LBs in brain stem nuclei termed incidental LB disease (iLBD) [14]. αS pathology in the spinal cord may provide the pathological basis for autonomic failure in α-synucleinopathies [12, 23]. Moreover, αS immunoreactive aggregates have been found in peripheral autonomic plexuses in one-tenth of the population aged 44–84 years who underwent a surgical resection of an abdominopelvic organ suggesting that these aggregates may represent an early stage of αS pathology and a developing α-synucleinopathy [17]. αS pathology in the form of LNs and LBs has been detected in spinal cord autonomic nuclei in 17.3% of neurologically unimpaired elderly subjects [4]. However, it is not known whether the αS pathology occurs first in the peripheral autonomic nervous system or centrally, and whether the site of the first occurrence may vary between the α-synucleinopathies.

In this study, we evaluated the frequency of αS pathology in the thoracic and sacral spinal cord, and examined the relationship between these findings and our previously reported brain αS pathology [18] in a very elderly population-based cohort.

Materials and methods

The Vantaa 85+ study cohort of 601 individuals includes all residents of Vantaa, a town in Southern Finland, who were aged at least 85 in 1991. The detailed study design has been described earlier [1, 22]. Results of neuropathologic examinations of consented autopsies were available in 304 (54%) of the 565 eligible study subjects, who died during the 10-year follow-up from April 1, 1991 to April 1, 2001. Their ages at the time of death ranged from 85 to 106 years (92.38 ± 3.7, mean ± SD) and the male to female ratio was 1:5 (52 men, 252 women). This prospective, population-based cohort study has been approved by the ethics committee of the Health Centre of the city of Vantaa and the tissue retention by the National Authority for Medicolegal Affairs, Helsinki, Finland.

All brain and spinal cord specimens were fixed in phosphate buffered 4% formaldehyde for at least 2 weeks before sampling. Brain samples were obtained following recommendations of the First DLB consortium international workshop [15], and assessed for changes in αS pathology [16] and neurofibrillary Alzheimer’s disease (AD) [5] as described earlier [18, 22]. Sections of the SN and hippocampus with transentorhinal cortex were stained with the haematoxylin and eosin (H&E) method and with antibodies against αS and screened for αS pathology. If any LBs or LNs were detected in the screened areas, the immunohistochemistry for αS was performed on cortical samples. The type of αS pathology (none, brainstem-predominant, limbic, diffuse neocortical) was determined for every subject [18]. A semiquantitative grading of the cell loss/atrophy in the ventrolateral tier of SN pars compacta was determined from none (0) to severe (3), as reported earlier [18].

Tissue samples from the upper third of the thoracic (level T3–4) spinal cord and from the sacral (level S1–2) spinal cord of each subject were embedded in paraffin. Following examination by conventional H&E staining, αS pathology assessed in 4 μm thick sections from thoracic and sacral spinal cord were immunostained manually using mouse monoclonal αS antibody (Transduction Laboratories, Lexington, KY, USA, clone 42, diluted 1:800). Antigen retrieval was carried out by microwave heating in citrate buffer (pH 6) for 20 min, followed by immersion in 100% formic acid for 5 min. More details regarding αS immunohistochemistry has been described earlier [19]. The false positive and negative staining was avoided using appropriate positive and negative control tissues.

The presence and load of αS pathology were detected in the sympathetic intermediolateral columns of thoracic spinal cord and in the sacral parasympathetic nuclei. The most severely affected side of the tissue sample was taken for the grading. The immunoreactivity of αS was assessed semiquantitatively in microscopic fields magnified at ×200 and scored as none (negative), mini-aggregates (+; immunoreactive aggregates as grains or dots in neuropil and/or cytoplasm, but no LNs or LBs), mild (1; at least one LN and/or LB), moderate (2; some LNs and/or LBs), and severe (3; many LNs and/or LBs) (Fig. 1). Thus, in addition to the classic αS pathology, grain- or dot-like αS immunoreactivity was recorded as an independent finding. This “mini-aggregate” category includes subjects who do not have unequivocal classic αS pathology, LNs and LBs, but do have αS immunopositivity in neuropil and/or cytoplasm, while the more severe forms have at least some classic αS pathology. All immunostained spinal cord sections were evaluated for αS pathology independently by two investigators (MO and TP) with excellent degree of agreement (p < 0.001, linear-by-linear exact test), and some cases were reviewed blindly by additional investigators (AP or HK).
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0629-6/MediaObjects/401_2009_629_Fig1_HTML.jpg
Fig. 1

A semiquantitative assessment of the αS pathology in the spinal cord. The figures illustrating the increasing severity of the immunoreactivity in the intermediolateral column of thoracicspinal cord from negative (a), to mini-aggregates with low magnification (b), mini-aggregates (c), mild (d), moderate (e) and severe (f) forms. αS immunostaining: magnification, 400× a and cf 200×; bScale bars, a,cf 40 μm and b 80 μm

Statistical methods

Statistical analyses were conducted with the SPSS 15.0 program for Windows (SPSS Inc., Chicago, IL). For comparisons of different groups of αS pathology and the degree of SN cell loss (categorical variables) the Chi-squared or Fisher exact tests or Chi-squared or exact tests for linear trend were used. Predictive factors for the positive αS pathology in the spinal cord autonomic nuclei were assessed by multivariate logistic regression analysis considering age at death, gender and αS brain pathology as independent variables. The significance level was set at p < 0.05.

Results

Frequency of αS pathology in the spinal cord

The mild, moderate and severe forms of spinal cord αS pathology with classic LB or LN structures were seen in autonomic nuclei of the thoracic cord in 78 (26%) subjects and sacrally in 76 (25%) subjects (Table 1). In these, the most common structures were dots or short and stubby LNs, whereas longer curvilinear LNs and LBs were infrequent and most often seen in the severe form (Fig. 1). In addition to the classic αS pathology, grain- or dot-like αS mini-aggregates were seen in the thoracic sympathetic intermediolateral columns in 153 (50%) and in the sacral parasympathetic nuclei in 115 (38%) of the 304 subjects (Table 1), thus making it the most common form of the detected αS immunoreactivity (Fig. 1). Forty-five subjects had thoracic αS immunoreactivity (mini-aggregates to moderate) without any αS immunoreactivity in the sacral spinal cord, whereas there were only five subjects with sacral immunoreactivity (mini-aggregates or mild) without any thoracic αS immunoreactivity (Table 1). Classic αS pathology was found at least at one level of the autonomic nuclei (thoracic or sacral) in 102 (34%) subjects, and αS mini-aggregates in 134 (44%) subjects (Table 2). The mean age at death was not significantly different for those with classic pathology (92.5 years; SD ± 4.0 years), with mini-aggregate (93.2 years; SD ± 3.6 years) or without any (90.6 years; SD ± 3.0 years) αS immunoreactivity. When the subjects were divided into three age groups on the basis of the age at death (85–89, 90–94 and ≥95 years), the spinal cord αS immunoreactivity (classic together with mini-aggregates) was most frequent in the oldest age group (Table 2), in which αS immunopositivity was found over five times more often in the autonomic nuclei than in the youngest age group (OR 5.39; 95% CI 2.2–13.3; p < 0.001, adjusted for gender and age at death). In the oldest age group, only 9% did not show any αS immunoreactivity, compared to the 35% in the youngest age group (p < 0.001, Table 2). Although, the association with age was mostly based on the mini-aggregate αS immunoreactivity in the spinal cord (Table 2), the classic αS pathology was significantly more common in the oldest than in the youngest age group (OR 7.54; 95% CI 2.13–26.67; p = 0.002, even when mini-aggregate immunoreactivity was excluded, Table 3). This association with age was independent, i.e. found regardless of the αS pathology in the brain (Table 3). The spinal cord αS immunoreactivity or the classic αS pathology alone was not associated with gender (Table 3).
Table 1

α-S immunoreactivity in the sympathetic intermediolateral cell column of thoracic spinal cord (n, %) according to the αS immunoreactivity in the sacral parasympathetic nucleus

Sacral αS immunoreactivity score

Thoracal αS immunoreactivity score

Negative

Mini-aggregates

Mild

Moderate

Severe

Total

Negative

68 (60%)

35 (31%)

9 (8%)

1 (1%)

0 (0%)

113 (100%)

Mini-aggregates

1 (1%)

98 (85%)

11 (10%)

3 (3%)

2 (2%)

115 (100%)

Moderate

4 (7%)

20 (35%)

8 (14%)

18 (31%)

8 (14%)

58 (100%)

Severe

0 (0%)

0 (0%)

0 (0%)

4 (40%)

6 (60%)

10 (100%)

Very severe

0 (0%)

0 (0%)

0 (0%)

2 (25%)

6 (75%)

8 (100%)

Mild, moderate and severe forms represent the classic αS pathology

Table 2

Distribution of αS pathology in the intermediolateral cell column of thoracic spinal cord and/or in the sacral parasympathetic nucleus according to the age at death and gender

Age at death

Type of αS pathology

None

Mini-aggregatesa

Mild–severeb

All

 Total (n = 304)

68 (22%)

134 (44%)

102 (34%)

 Men (n = 52)

12 (23%)

17 (33%)

23 (44%)

 Women (n = 252)

56 (22%)

117 (46%)

79 (31%)

85–89 years

 Total (n = 82)

29 (35%)

24 (29%)

29 (35%)

 Men (n = 15)

7 (47%)

2 (13%)

6 (40%)

 Women (n = 67)

22 (33%)

22 (33%)

23 (34%)

90–94 years

 Total (n = 146)

32 (22%)

71 (49%)

43 (29%)

 Men (n = 23)

4 (17%)

11 (48%)

8 (35%)

 Women (n = 123)

28 (23%)

60 (49%)

35 (28%)

≥95 years

 Total (n = 76)

7 (9%)

39 (51%)

30 (39%)

 Men (n = 14)

1 (7%)

4 (29%)

9 (64%)

 Women (n = 62)

6 (10%)

35 (56%)

21 (34%)

aGrain- or dot-like

bClassic αS pathology

Table 3

Associations of age at death, gender and αS pathology in the brain with αS pathology in the spinal cord

Type of αS immunoreactivity/pathology in the spinal cord

ORa

95% CI

p

Mini-aggregatesb

 Age at death 85–89 years

1

  

 90–94 years

3.31

1.58–6.97

0.002

 ≥95 years

9.04

3.17–25.80

<0.001

 Male

1

  

 Female

1.53

0.61–3.83

0.36

 No αS pathology in the brain

1

  

 Brainstem-predominant

1.74

0.18–16.73

0.63

 Limbic

6.66

0.81–54.87

0.08

 Diffuse Neocortical

5.68

0.64–50.48

0.12

Classic αS pathologyc

 Age at death 85–89 years

1

  

 90–94 years

2.02

0.74–5.48

0.17

 ≥95 years

7.54

2.13–26.67

0.002

 Male

1

  

 Female

0.98

0.32–3.03

0.98

 No αS pathology in the brain

1

  

 Brainstem-predominant

4.59

0.38–54.98

0.23

 Limbic

70.47

8.87–559.60

<0.001

 Diffuse Neocortical

89.73

11.38–707.413

<0.001

Classic αS pathologyd

 Age at death 85–89 years

1

  

 90–94 years

0.91

0.44–1.89

0.80

 ≥95 years

1.55

0.68–3.55

0.30

 Male

1

  

 Female

0.60

0.28–1.33

0.21

 No αS pathology in the brain

1

  

 Brainstem-predominant

3.23

0.72–14.4

0.13

 Limbic

13.20

6.07–28.7

<0.001

 Diffuse Neocortical

26.19

11.02–62.21

<0.001

Logistic regression analysis. OR odds ratio; 95% CI 95% confidence interval

aThe ORs are adjusted mutually for age at death, gender and type of αS pathology in the brain

bDichotomous variable; grain- or dot-like immunoreactive mini-aggregates detected in the intermediolateral cell column of thoracic spinal cord and/or in the sacral parasympathetic nucleus

cOnly classic αS pathology was considered positive and the mini-aggregates were excluded from the analysis

dOnly classic αS pathology was considered positive and the mini-aggregates were included as negative

Association between cerebral and spinal cord αSyn pathology

αS pathology was present at least in the SN and/or hippocampal–transentorhinal region in 36% of the autopsied 304 subjects, and the brain αS pathology was categorized as follows: 8 (3%) had brainstem-predominant, 42 (14%) had limbic and 47 (15%) had diffuse neocortical type of αS pathology [18]. Thirteen (4%) subjects had αS pathology confined to the hippocampal–transentorhinal region [18].

The extent of brain αS pathology was significantly associated with the classic αS pathology in the spinal cord; subjects with diffuse neocortical type of LB pathology had almost 90 times more often at least mild αS pathology in the spinal cord compared to the subjects without αS pathology in the brain (OR 89.73; 95% CI 11.4–707.4; p < 0.001, adjusted for gender and age at death, Table 3). This association remained significant irrespective of whether the mini-aggregate αS immunoreactivity group was excluded from the analysis or included in the negative group (Table 3). Of the 50 subjects with moderate or severe αS pathology in the thoracic intermediolateral column, 30 (60%) had diffuse neocortical, and of the 18 subjects with moderate or severe αS pathology in the sacral parasympathetic nucleus 16 (89%) had diffuse neocortical type of brain LB pathology (Fig. 2). On the other hand, among the 47 subjects with diffuse neocortical type of LB pathology, only one did not have any spinal αS immunoreactivity. Furthermore, αS pathology in the spinal cord autonomic nuclei showed a significant association with the severity of SN neuron loss (p < 0.001), but this association was lost when adjusted for brain αS pathology. Thus, this association was most likely based on αS pathology in the brain. Eight subjects with the brainstem-predominant type of LB pathology and 13 subjects with the brain αS pathology confined to the hippocampal–transentorhinal region had similar, mostly negative or mini-aggregate αS immunoreactivity in the spinal cord autonomic nuclei [18], and therefore these two groups have been combined in Fig. 2. The spinal cord classic αS pathology did not associate with the neurofibrillary Braak stage (p = 0.67).
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0629-6/MediaObjects/401_2009_629_Fig2_HTML.gif
Fig. 2

The distribution of αS pathology in the brain associates with the severity of αS pathology (a) in the intermediolateral column of thoracic spinal cord (p < 0.001, the SPSS linear-by-linear exact test), and (b) in the sacral parasympathetic nucleus (p < 0.001, linear-by-linear exact test). Of the spinal cord categories, moderate and severe αS pathologies have been combined in the same group, as have been done to the brain categories brainstem-predominant and asterisk indicates αS pathology restricted to the transentorhinal cortex

Discussion

To our knowledge, this is the first study where the extent of αS pathology in the spinal cords has been analyzed in a population-based cohort. Our study indicates that αS pathology is common in the autonomic nuclei of the spinal cord in elderly individuals. One-third of the study subjects showed classic αS pathology and in addition almost half of them had grain- or dot-like αS immunopositive deposits in the thoracic and/or sacral autonomic nuclei. Because the mean age at death in our study is 92 years, it is understandable that previous studies of younger, neurologically asymptomatic individuals have demonstrated lower frequencies of αS pathology than the 34% detected in this study. In two recent studies, spinal cord αS pathology was detected in 8% of individuals over 60 years of age [14] and in 17.3% of individuals with mean age of 80.7 years [4].

Frequency of the spinal cord αS immunoreactive alterations increased with age, but interestingly, only in a minority (22% of thoracic, 9% of sacral) these alterations reached the level of moderate or severe pathology. Actually, the age-related increase in the spinal αS immunoreactive alterations was mainly based on the age-associated increase of the mini-aggregates form (Table 3). We propose that there are two separate processes in the spinal cord: (1) age-related, possibly not pathological, accumulation of small, grain- or dot-like αS mini-aggregates, mostly in neuropil; and (2) real α-synucleinopathy-related pathology, i.e. occurrence of classic αS immunoreactive LNs and LBs. The most severe brain and spinal cord αS pathologies are most often closely linked to each other, although, in our study population there are also subpopulations with extensive αS pathology in the brain without corresponding spinal cord pathology and vice versa. The presence of these subpopulations challenges the hypothesis that there is only one induction site with one progression pathway in α-synucleinopathies, a finding supported by a few other previous reports [9, 20, 26]. Also, the severity of αS pathology in the spinal cord was identical in subjects with no brain αS pathology and in those with only limited brain αS pathology (brainstem-predominant or hippocampal–transentorhinal region) (Fig. 2), suggesting that limited brain αS pathology and the spinal cord αS pathology are not interrelated.

Although the brain αS pathology was associated with the Braak neurofibrillary stage [18], no such association was seen with the spinal cord αS pathology. Yet, the AD pathology in the brain may contribute to the pathological disease progression of αS pathology.

Our results of the αS pathology distribution in the spinal cord autonomic nuclei suggest that αS pathology proceeds along a descending pathway, i.e. affecting the thoracic spinal cord before the sacral: 19% of αS pathology positive subjects had thoracic without sacral αS pathology and only 2% of them had sacral without thoracic αS pathology (Table 1). Furthermore, αS pathology was more frequent and more severe in the thoracic than in the sacral cord (Fig. 2). Alternative explanation for the discrepancy between the thoracic and sacral αS pathologies might be that the sympathetic nervous system is more prone to be affected by αS pathology than the parasympathetic system. In line with this, it has been suggested that αS aggregates in peripheral sympathetic neurons prior to parasympathetic ones [17]. This applies to the mini-aggregates as well, which were also more common in the thoracic than in the sacral spinal cord (Table 1). As the αS mini-aggregates are also associated with age, one could speculate that ageing occurs earlier in the sympathetic than in the parasympathetic nervous system.

Interestingly, among the subjects with moderate or severe αS pathology in the autonomic spinal cord nuclei three did not have αS pathology in the brain. Of course we cannot exclude that they have brain αS pathology in areas what we have not studied including the olfactory bulbs [2]. On the other hand, these three may be exceptional and represent the concept of pure autonomic failure (PAF). It is known that striking αS pathology can exist in the peripheral autonomic system without αS pathology in the brain [21]. However, although PAF arises from peripheral autonomic neurons it may progress into the central nervous system [13]. Based on our findings, in most cases with obvious αS pathology in the spinal cord there is also αS pathology in the brain. Thus, the pure “peripheral” form of severe α-synucleinopathy appears to be rare.

The strength of the present study is the exceptional material of the autopsied subpopulation from a population-based cohort of very elderly individuals. However, there are also some limitations in the present study. First, our sampling strategy does not cover the whole length of spinal parasympathetic and sympathetic nuclei. Second, according to the previously proposed hypothesis [6, 10], αS pathology proceeds from the brainstem towards the neocortex. Therefore, we screened for αS pathology using samples of the brainstem and hippocampal–transentorhinal region. We assumed that there is no (significant) αS pathology in the neocortex without αS pathology in the samples screened. Thus, we have not investigated all brain areas in all cases with immunohistochemistry. This screening system might have created a bias in the correlation between the brain and spinal cord αS pathology, as there might have been more individuals with limited αS pathology in the brain. However, because not only the distribution of the αS pathology but also its severity appeared to indicate similar disease progression in the brain and spinal cord, our results are likely to be reliable. Third, the significance of the mini-aggregate type of αS immunoreactivity detected is debatable. The grain- and dot-like staining may only represent age-related, fixation time influenced non-specific accumulation of αS in presynaptic terminals. A recent methodological paper has suggested the use of proteinase K pretreatment to avoid this [3]. In our material, the proteinase K antigen retrieval resulted in disappearance of all αS pathology, including the classic LBs and LNs, and thus could not be used.

Results of this population-based analysis suggest that there may be more than one pathologic process which result in accumulation of αS positive deposits, and the development of the αS pathology does not always follow a standard topographically predictable progression.

Acknowledgments

The authors thank Mrs Tuija Järvinen for excellent technical assistance. This study was financially supported by the Finnish Medical Foundation and the Helsinki University Central Hospital competitive research fund (EVO).

Conflict of interest statement

The authors report no conflicts of interest.

Copyright information

© Springer-Verlag 2009