Acta Neuropathologica

, 118:777

Diminished tyrosine hydroxylase immunoreactivity in the cardiac conduction system and myocardium in Parkinson’s disease: an anatomical study

Authors

    • Institute of Clinical NeuroanatomyGoethe University Frankfurt
    • Laboratory of Neuropathology, Center for Clinical ResearchUniversity of Ulm
    • Anatomy and Developmental Biology, School of Biomedical SciencesUniversity of Queensland
  • Kelly Del Tredici
    • Clinical Neuroanatomy, Department of Neurology, Center for Clinical ResearchUniversity of Ulm
  • James W. Langston
    • The Parkinson’s Institute
  • Heiko Braak
    • Clinical Neuroanatomy, Department of Neurology, Center for Clinical ResearchUniversity of Ulm
Original Paper

DOI: 10.1007/s00401-009-0596-y

Cite this article as:
Ghebremedhin, E., Del Tredici, K., Langston, J.W. et al. Acta Neuropathol (2009) 118: 777. doi:10.1007/s00401-009-0596-y

Abstract

Clinical and autopsy studies have consistently reported cardiac sympathetic dysfunction in the left ventricular wall in patients with Parkinson’s disease (PD). Whether the nerve fibers of the cardiac conduction system or the atrial walls are equally affected in this disease process has not yet been well documented. Therefore, the aim of this study was to investigate sympathetic nerves in the cardiac conduction system as well as in the walls of all four heart chambers in patients with PD, in incidental Lewy body disease (iLBD), and in controls. Heart tissue from five PD patients, two iLBD cases, and seven controls were investigated immunohistochemically using antibodies directed against tyrosine hydroxylase (TH) and α-synuclein (syn-1). A marked diminution of TH immunoreactivity (IR) within nerve fibers was observed in four PD patients and in both individuals with iLBD. In contrast, all control subjects displayed dense TH-IR nerve structures. The depletion in TH-IR involved not only the ventricles, but also the conduction system and the atrium showing a global change within cardiac TH-IR nerve fibers in the course of PD. In conclusion, the alterations in cardiac sympathetic nerves of patients with PD or in individuals with iLBD are homogeneous and global within the heart. The clinical implications related to this complete cardiac sympathetic dysfunction, including clinical correlates, diagnostic implications, and treatment, however, remain to be determined in a larger autopsy-controlled cohort of prospectively followed individuals.

Keywords

Cardiac conduction systemLewy body diseaseParkinson’s diseaseSympathetic dysfunctionTyrosine hydroxylase

Introduction

The cardinal symptoms of Parkinson’s disease (PD) are movement-related disturbances related largely to the lesion of the substantia nigra. In the course of the disease, however, the clinical spectrum of PD commonly involves non-motor symptoms (for review, see [6, 29]), and it has been shown that PD-related neuropathology is not confined to the brain or central nervous system but is distributed throughout the entire nervous system [2, 17]. In particular, autonomic failure is regarded as a core feature of PD [5, 27, 28], and it has been recognized that the peripheral autonomic nerves of various end organs are pathologically affected in the course of PD [16, 17, 19].

The pathological involvement of the noradrenergic sympathetic nervous system is frequently observed in PD, as evidenced by the down-regulation of tyrosine hydroxylase (TH) enzyme in the heart of PD patients [20, 25, 26]. TH catalyzes the rate-limiting step in the biosynthesis of noradrenaline [22] and, thus, a reduction in TH could indicate either neuronal fiber loss or a significant down-regulation of TH in nerve cells. In either scenario, reduction of TH in the heart, as shown by TH immunostaining of heart tissue, could be interpreted as a sign of dysfunction within the cardiac sympathetic system. As such, we employ the terminus cardiac “sympathetic dysfunction.”

Scintigraphic studies have demonstrated that cardiac sympathetic dysfunction is frequently present in PD patients with autonomic failure [4, 12, 13, 23]. These findings have also been corroborated by autopsy-based studies which consistently reported a purported decrease in left ventricular cardiac TH-immunoreactive nerve fibers in PD patients [20, 25, 26]. Moreover, similar observations have also been reported in individuals with incidental Lewy body disease (iLBD) [9, 11, 15, 24]. This has resulted in speculation that cardiac sympathetic dysfunction, sometimes referred to as “denervation,” could be an early phenomenon in the course of PD [13, 23, 30]; if so, investigation of sympathetic nerves in the heart might lead to the development of a useful diagnostic marker in the differential diagnosis of early PD [10].

Previous autopsy-based studies, however, have only investigated the involvement of left ventricular sympathetic nerve fibers in PD patients, but sympathetic nerve fibers are abundantly found not only in the left ventricle, but also in the conduction system of the heart [7, 8]. However, the cardiac conduction system and the atria in PD patients have not been documented so far. Thus, this study aimed to investigate sympathetic nerves in the cardiac conduction system as well as in the walls of all four heart chambers. For this purpose, hearts from PD and iLBD cases as well as controls were compared.

Materials and methods

Study population

Tissue samples from the heart were obtained at autopsy from five PD patients (three males; age range 70–86 years; mean age ± SD 79.3 ± 6.6 years), two cases with incidental Lewy body disease (iLBD; both males, aged 70 and 80 years), and seven controls (three males and four females; age range 62–85 years; mean age ± SD 71.6 ± 8.4 years). Consent for autopsy was obtained for all patients as well as controls and the study was approved by the university ethics review committee.

Clinical data

Sample demographics and relevant clinical data are shown in Table 1. PD diagnosis had been made during life and was confirmed at autopsy (mean disease duration, 5 ± 2.83 years). Four of the five PD patients displayed symptoms compatible with autonomic failure (orthostatic hypotension and/or urinary incontinence; Table 1). Neither the iLBD cases nor the controls suffered from neurological disease, and autonomic failure was not documented in their clinical data.
Table 1

Demographic and clinicopathological features of the individuals with PD and iLBD as well as controls

Case

Diagnosis

Age

Sex

Disease duration

PD stage

TH-IR

α-Syn-1

H–Y

L-dopa

OH

UI

Diabetes

HBP

1

PD

70

F

6

3

3

0

3

2

PD

86

M

8

5

2

2

4

+

+

3

PD

81

M

5

4

1

1

3

+

+

4

PD

81

M

2

5

1

3

4

+

+

+

5

PD

77

F

7

5

1

0

5

+

+

+

6

iLBD

70

M

2

1

1

7

iLBD

82

M

3

1

2

8

Control

67

F

0

2

0

9

Control

81

M

0

3

0

10

Control

68

F

0

3

0

11

Control

66

M

0

2

0

12

Control

62

F

0

2

0

13

Control

85

F

0

2

0

14

Control

72

M

0

3

0

Age and disease duration are given in years; α-Syn-1, severity of α-synuclein pathology was rated as follows: score 0 = pathology absent, score 1 = mild pathology, score 2 = moderate pathology, and score 3 = marked pathology

HBP high blood pressure, H–Y Hoehn and Yahr score, iLBD incidental Lewy body disease, L-dopa L-dopa treatment, OH orthostatic hypotension, PD stage stage of Parkinson’s disease (PD) neuropathology [3], TH-IR density of tyrosine hydroxylase immunoreactivity (TH-IR) in heart sections was rated as follows: score 0 = absent, score 1 = marked diminution of TH-IR, score 2 = moderately intact TH-IR, and score 3 = strongly preserved TH-IR; UI urinary incontinence

Neuropathological analysis of the brain

All autopsies of PD patients fulfilled published criteria for macroscopic and microscopic diagnoses of PD, and all iLBD cases as well as controls underwent neuropathological examination. Brains were fixed in 4% buffered aqueous solution of formaldehyde. Extensive analysis of brain pathology was performed on 100 μm polyethylene glycol (PEG) sections from standardized brain regions to facilitate neuropathological PD staging [3].

Sampling and sectioning of the heart tissue

Heart tissue was fixed and stored in 4% buffered aqueous solution of formaldehyde. Initially, macroscopic and histological examinations of each heart were performed to exclude cases with ischemic heart disease. To facilitate the investigation of the cardiac autonomic nerve system, including the innervation of the conduction system, a set of four tissue blocks was systematically sampled from each heart (Fig. 1). As shown in Fig. 1, the regions of interest included: the anterior ventricular walls, sinus node, and posterior intercaval atrial area. The region of the sinus node, which is located at the anterior and lateral atrial junction with the superior caval vein, was dissected perpendicular to the terminal groove (for review see [1]). This tissue block included portions of the superior vena cava and that of the atrium together with the terminal crest (Fig. 2a, b). In two PD patients, one iLBD case and two controls, the atrioventricular (AV) node together with the penetrating AV bundle (His bundle), the principal Tawara bundles, and the Purkinje fibers were also investigated. The interatrial septum encompassing the Koch’s triangle was removed en bloc to permit study of both the AV node and the His bundle. The tendon of Todaro, the annular attachment line of the septal leaflet of tricuspid valve, the orifice of the coronary sinus, and the atrioventricular component of membranous septum (as part of the central fibrous body) in the right atrium served as anatomic landmarks of Koch’s triangle [1]. The AV node is located near the apex of the Koch’s triangle, whereas the His bundle traverses the central fibrous body (right fibrous trigone) and runs towards the apex of the heart immediately beneath the interventricular membranous septum before branching. To study the Tawara bundles, a tissue block from the muscular and central part of the ventricular septum was cut horizontally perpendicular to the anterior interventricular artery, and another tissue block containing the septomarginal trabecula (moderator band) was removed to investigate the Purkinje fibers.
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0596-y/MediaObjects/401_2009_596_Fig1_HTML.gif
Fig. 1

Schematic diagram of the human heart. Shaded areas illustrate locations of sites for tissue samples. a Sites for the sinus node and the ventral regions of the left and right ventricles. b Site for the posterior intercaval area of the right atrium

https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0596-y/MediaObjects/401_2009_596_Fig2_HTML.jpg
Fig. 2

a Masson’s trichrome staining employs light green for selective stain of the connective tissue of the sinus node in a case with incidental Lewy body disease (iLBD, case 7, Table 1). The boundary of the sinus node is demarcated by a dotted line. The epicardial side of the node is filled with adipose tissue (arrowhead). A nodal artery is located in the middle of the node (asterisk). b Masson’s trichrome staining employing aniline blue for selective stain of the connective tissue of the sinus node in a control (case 8, Table 1). The node is located along the terminal crest (arrow) and the epicardial side of the node is filled with adipose tissue (arrowhead). A ramifying nodal artery is located in the middle of the node (asterisks). Scale bars 250 μm

Histologically, a conspicuous feature of the cardiac conduction system is a combination of relatively small specialized myocardial cells and abundant fibrous connective tissue, mainly collagen and fibroblasts [1]. To confirm the presence of conducting tissue, paraffin sections (10–20 μm) were stained with Masson’s trichrome method (Fig. 2a, b) to differentiate the specialized myocardial cells from the connective tissue components.

Processing of the heart tissue blocks

All tissue blocks were divided into two blocks for PEG embedding of unconventionally thick sections (100 μm) and relatively thin serial cryosections of 30 μm thickness. Since most intrinsic cardiac ganglia and nerve plexuses are located in the atrial walls, tissue blocks from the posterior intercaval atrial wall were used to evaluate as much nerve fibers as possible. For this purpose, atrial tissue block was first flattened and plane thick sections (100 μm) through the epicardium were cut parallel to the outer surface of atrium. In general, to ensure accuracy in topographic orientation within the intact heart during tissue processing, all tissue blocks were marked with a solvent-resistant black dye on the inferior surface and right edge of the tissue. Complementary electronic copies from each tissue surface of interest were also made for better recognition and orientation during assessment.

Immunohistochemical staining

The primary antibodies used were a monoclonal mouse antibody against α-synuclein (anti-syn-1; 1:2,000; Clone number 42, BD Bioscience, Mountain View, CA, USA) as a marker of PD-related pathology and a polyclonal rabbit antibody against tyrosine hydroxylase (anti-TH; 1:2,000; AB152: Chemicon/Millipore, Schwalbach, Germany) as a marker for noradrenergic nerve fibers. Additionally, a polyclonal rabbit anti-protein gene product 9.5 (anti-PGP 9.5; 1:2,000; RA95101: UltraClone Limited; purchased through Biozol, Eiching, Germany) was used as a pan neuronal marker to estimate inter-individual variability in innervation of the heart.

Following pretreatment with hydrogen peroxide and bovine serum albumin to inhibit endogenous peroxidase and prevent nonspecific binding, and after additional pretreatment with formic acid to facilitate the immunoreaction, incubation with primary antibody took place for 18 h. Subsequent to processing with biotinylated corresponding secondary antibodies for 2 h, reactions were visualized with the avidin–biotin complex (ABC, Vectastain) and a chromogen of choice as follows: for Syn-1 immunostaining, the 3,3′-diaminobenzidine tetrahydrochloride (DAB) was used, giving a brown color, whereas the Vector SG (SK-4700; Vector) chromogen (blue/gray) was used for TH immunostaining. Positive and negative control sections were routinely included to confirm specificity of the staining.

Semiquantitative assessment

The intensity of TH-immunoreactive nerve fibers in cardiac sections from both patients and controls were assessed according to a semiquantitative four-point rating scale: absent (0) or marked diminution (1), moderately intact (2), and strongly preserved (3). Severity of α-synuclein pathology was rated as follows: pathology absent (0), mild pathology (1), moderate pathology (2), marked pathology (3).

Statistical analysis

Between-group comparisons were made using the Spearman’s rho test for rank order correlation and Mann–Whitney U test. Calculations were performed with the aid of SPSS, release 17 (SPSS Inc., 2008, Chicago, IL). Significance was set at P value of ≤0.05.

Results

The cardiac conduction system was densely innervated by TH-immunoreactive (IR) nerve fibers (Figs. 3g, 4a). Apart from conduction system, it was noticeable that most TH-positive nerve fibers were found in the epicardium, including the adjoining subepicardial area of the myocardium, and concomitant to blood vessels (Fig. 3a–h). Normal nerve bundles had a consistent caliber and showed uniformly TH immunostained neuronal processes in control hearts (Fig. 3e). By comparison, in PD and iLBD cases, nerve bundles were mostly negative for TH staining or showed focal TH-positive swellings in their neuronal processes (Fig. 4e). Some intrinsic cardiac neuronal somata were also positive for TH immunostaining in controls but not in PD patients (Fig. 4f).
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0596-y/MediaObjects/401_2009_596_Fig3_HTML.jpg
Fig. 3

Immunoreaction (IR) for tyrosine hydroxylase (TH) in heart tissue sections: left ventricle (a, b), right ventricle (c, d), right atrium (e, f), and sinus node (g, h). b, d, f, and h represent PD patients with diminished TH-IR, whereas a, c, e, and g denote the controls with preserved TH-IR. Thick arrows point to nerve fascicles in the epicardium, thin arrows to dense perimysial nerve fibers. Scale bars in ac, e, and f 100 μm; scale bars in d, g, and h 50 μm

https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0596-y/MediaObjects/401_2009_596_Fig4_HTML.jpg
Fig. 4

a The boundary of the sinus node is demarcated by a dotted line in one case with incidental Lewy body disease (iLBD, case 7, Table 1). The node is located along the terminal crest (arrow) and the epicardial side of the node is filled with adipose tissue (arrowhead). A ramifying nodal artery is located in the middle of the node (asterisks). A detailed view of the squared area is shown in b. b Syn-1-immunoreactive pathology (thin arrows) in nerve fibers within or close to the sinus node. The thick arrow denotes the terminal crest. c Syn-1-immunoreactive pathology in nerve fibers (arrows) of the left ventricular wall from a PD patient (case 4, Table 1). d Double-labeling with TH (blue/gray) and Syn-1 (brown) antibodies in the right atrial tissue section. Both antibodies immunolabel the same nerve fibers (arrows) in a PD patient (case 2, Table 1). e TH-immunoreactive focal swellings (arrows) in neuronal processes occur in degenerative nerve bundles in atrial epicardium. f Intrinsic cardiac ganglia in atrial epicardium with two TH-immunoreactive nerve cell bodies (arrows) and their processes in a control case. Scale bar in a corresponds to 500 μm; scale bar in c and f correspond to 100 μm; scale bars in b, d and e correspond to 50 μm

As shown in Figs. 3 and 4, a marked diminution in TH-IR was observed in four of five PD patients as well as in both individuals with iLBD. In contrast, all control subjects displayed dense TH-IR nerve structures, thereby indicating a statistically significant difference when compared with those of the PD/iLBD cases (P = 0.02). The decrease in TH-IR involved all cardiac sympathetic nerve fibers, i.e., the small nerve fibers of the heart muscles, the vessel-associated neural network, and the large nerve bundles in the epicardium. These changes were observed not only in both ventricles, but also in the right atrium as well as in the conduction system of the heart, thereby indicating a global diminution of TH immunoreactivity in nerve fibers.

No significant correlation existed between the abnormal cardiac sympathetic nerves and disease duration (P = 0.5), or Hoehn–Yahr scores (P = 0.2). The presence of orthostatic hypotension and other autonomic dysfunctions was documented in three of the five PD patients (Table 1), each of whom displayed considerable loss of cardiac sympathetic nerves. One PD patient with a clinical history of urinary incontinence in the absence of orthostatic hypotension showed moderate density of TH-IR nerve fibers. The remaining PD patient lacked any clinical history of autonomic dysfunction and had preserved density of TH-IR nerve fibers.

PD-related pathology as shown by the presence of α-synuclein-IR could be detected in three of five (60%, see Table 1 and Fig. 4c) PD and in both iLBD cases (with PD stages 2 and 3, see Table 1 and Fig. 4b) but not in controls. Presence of α-synuclein immunostaining in neuronal processes was frequently colocalized with TH staining (Fig. 4d).

Discussion

To date, both clinical and pathological studies have provided strong evidence for the pathological involvement of the cardiac sympathetic nerves of the left ventricle in PD patients [4, 9, 1113, 15, 20, 2326]. It remains unclear, however, whether sympathetic nerves of other portions of the heart are equally affected by the disease process. Our study was primarily designed to investigate the pathological changes in sympathetic nerves in different regions of the heart, including the conduction system, which has been neglected until now.

The present findings corroborate studies reporting substantial alterations in cardiac sympathetic nerve fibers as measured by loss of TH immunoreactivity in the left ventricular wall of PD patients [20, 25, 26]. More importantly, however, our results extend previous reports demonstrating that hearts of PD or iLBD cases do not show a limited (i.e., regional) distribution pattern of the cardiac sympathetic dysfunctional process but, instead, a uniform involvement of the heart chambers, including the conduction system. These changes seem to occur independently of age, gender, and duration of disease.

PD-related pathology as shown by the presence of α-synuclein-IR was found in three of five PD (60%) and in both iLBD cases but not in controls. Consistent with previous observations [21, 25, 26], the presence of α-synuclein pathology in neuronal processes was frequently colocalized with TH staining, thereby indicating a role for α-synuclein protein misfolding as a possible cause of the cardiac sympathetic dysfunction.

Sympathetic nerves displayed normal TH immunoreactivity in the one PD patient without orthostatic hypotension, whereas the remaining four PD patients with orthostatic hypotension as well as both iLBD cases showed markedly decreased sympathetic innervation. These results are similar to previous observations [21, 25] and indicate that cardiac sympathetic dysfunction is not inevitably associated with PD. Moreover, these results suggest that there may be a relationship between cardiac sympathetic dysfunction and orthostatic hypotension in patients with PD, consistent with some previous reports [4, 12, 13] and contrary to others [10, 14, 18]. A caveat in this study is that it is retrospective and, as such, we cannot rule out inconsistent symptom ascertainment regarding the diagnosis and the precise course of the autonomic dysfunction.

Furthermore, it has been speculated that cardiac sympathetic dysfunction could be an early phenomenon in the course of PD [9, 11, 13, 15, 23, 24, 30]. If so, the investigation of sympathetic nerves in the heart could serve as an early marker of the disease. Our finding that sympathetic nerves in iLBD (viewed by many as pre-motor or prodromal) are altered supports this viewpoint. A significant correlation between the PD-related neuropathological processes in the heart and brain, however, could not be established. This could have been due to a floor effect owing to the presence of a marked depletion of TH immunoreactivity in sympathetic nerve fibers in almost all PD and iLBD cases.

Classically, intrinsic cardiac ganglia are considered to belong to the cholinergic parasympathetic system. Emerging evidence, however, suggests that a major subpopulation (40–50%) of the cardiac ganglion cells possess a dual cholinergic-noradrenergic phenotype [31], although the pharmacological and physiological relevance of such nerve cells still has to be determined. Consistent with this observation, we observed TH-positive neuronal cell bodies in some epicardial atrial ganglia. It has been suggested that these intrinsic cholinergic nerves may exercise full noradrenergic function by virtue of possessing all necessary noradrenergic properties [31]. The fact that TH-positive neuronal cell bodies or nerve fibers in both PD and iLBD were rarely observed suggests that nerve cells with dual phenotype are also affected in the course of PD. A more comprehensive and systematic study is required to draw firmer conclusions in this regard.

In summary, the present study shows that the pattern of cardiac sympathetic dysfunction in PD and iLBD is homogeneous and involves not only the vessel-associated neural network and working myocardium, but also the conduction system. Moreover, the results suggest that intrinsic cardiac nerve cells with dual cholinergic-noradrenergic properties might also be affected in the course of PD. The clinical implications related to this complete cardiac sympathetic dysfunction, however, remains to be determined, but if ways to routinely and inexpensively assess these changes clinically could be developed, it is possible that they could be incorporated into a diagnostic battery for Lewy body diseases. Finally, it will be important to determine which, if any, clinical manifestations are associated with these pathological changes. If such correlations can be established, these observations could have important therapeutic implications.

Acknowledgments

This study was funded by the Michael J. Fox Foundation for Parkinson’s Research (New York City, USA). The authors wish to thank Prof. Martin-Leo Hansmann, Director of the Dr. Senckenberg Institute for Pathology, Goethe University Frankfurt am Main, and Dr. Susanne Braun, Chief Pathologist, Offenbach Clinic, for providing autopsy material. They are also grateful to Mr. Mohamed Bouzrou (Goethe University Frankfurt) and Ms. Gabriele Ehmke (University of Ulm) for immunostaining, Ms. Inge Szász-Jacobi, who assisted with the graphics, and the Braak Collection (Goethe University Frankfurt).

Copyright information

© Springer-Verlag 2009