Human Subjects and Tissue Source

In our prior studies on PD pharynx [22,23,24,25], 20 whole-mount (larynx-pharynx-tongue-soft palate) specimens and psoas major muscles (control) were obtained from autopsied subjects with clinically diagnosed and neuropathologically confirmed PD. In the present study, the larynges from the 20 whole-mount PD specimens and 8 age-matched healthy controls were used to detect the abnormalities that may exist in the laryngeal nerves, muscles and mucosa. The PD specimens were provided by the Arizona Study of Aging and Neurodegenerative Disorders (AZSAND) and the Brain and Body Donation Program (BBDP) [26] at Banner Sun Health Research Institute. The healthy autopsied specimens without known systemic neuromuscular disorders were obtained from Department of Pathology at Mount Sinai Medical Center in New York City.

Inclusion/Exclusion Criteria for Specimen Procurement

We developed an inclusion/exclusion plan for specimen procurement. Specimens meeting the criteria as described below were accepted. Inclusion criteria for PD specimens were defined as follows: (1) The PD specimens from the donors must be clinicopathologically diagnosed as PD; (2) PD subjects with or without SSV disorders were included; (3) Ages ranging from 55 to 85 were included; and (4) No ethnic backgrounds were excluded. Exclusionary criteria for PD specimens included: (1) cases without an established diagnosis of PD; (2) parkinsonian symptoms caused by other diseases; (3) SSV disorders caused by neurological disorders other than PD; and (4) other confounding conditions such as a history of malignancy, local trauma, surgery, radiotherapy or chemotherapy at the head and neck region. Age-matched controls were selected only when the subjects had no neuromuscular disorders affecting functions of the upper aerodigestive tract and fulfilled the exclusion criteria.

PD and SSV Evaluations

The AZSAND/BBDP provided detailed clinical and neuropathological data for each of the autopsied PD subjects (Table 1). All PD patients received annual standardized cognitive testing (neuropsychological test battery) and movement neurological examinations. Disease severity was clinically rated using the Hoehn and Yahr (H&Y) Scale [27] and disability and motor impairment using the Unified Parkinson’s Disease Rating Scale (UPDRS) [28]. Specific clinicopathological diagnostic criteria for PD were used [29].

Table 1 Demographic, clinical features, and PNS and CNS pathology severity in PD subjects and normal controls

SSV impairments in PD patients were assessed subjectively using UPDRS ratings (i.e., swallowing and speech scores) (Table 1). A swallowing score (0–4) was given by using item 7 of the UPDRS Part II Scale (i.e., swallowing score of 0 = normal; 1 = rare choking; 2 = occasional choking; 3 = requires soft food; and 4 = requires nasogastric tube or percutaneous endoscopic gastrostomy feeding). A speech score (0–4) was given by using items 5 and 18 of the UPDRS Parts II and III Scales (i.e., speech score of 0 = normal; 1 = mildly affected, no difficulty being understood, slight loss of expression, diction and/or volume; 2 = moderately impaired, monotone, slurred but understandable; 3 = marked impairment, difficult to understand; 4 = unintelligible). In this study, several PD patients also received some objective SSV assessments, including modified barium swallow (MBS)—performed by radiologists with speech pathologist assistance—and fiberoptic flexible laryngoscopy or laryngeal videostroboscopy performed by otolaryngologists.

Brain Neuropathologic Assessments

One of the objectives of this study was to determine the relative contributions of PNS vs. CNS pathology to SSV disorders in PD. In this series, all PD subjects received brain neuropathologic assessments after death. Brain pathology severity was rated by an experienced neuropathologist (Dr. Thomas G. Beach), who provided a detailed neuropathologic report for each subject. Specifically, all autopsied PD cases had standardized brain PAS semi-quantitative density grading in 10 standard brain regions (total possible score of 40) (Table 1) and staging of brain PAS distribution using the Unified Staging System for Lewy Body disorders (USSLB) developed by Beach et al. [30]. The reliability and validity of the USSLB has been confirmed by subsequent independent studies [31,32,33]. For statistical purposes, semi-quantitative microscopic lesion density estimates were converted to numerical scores from 0 to 4 for Lewy-type α-synucleinopathy. For correlational analyses, the PAS density scores of the 10 brain regions are summed to a single global score for each subject. In addition, all autopsy cases received a substantia nigra pigmented neuron loss (SNPNL) score (Table 1), a measure of CNS dopamine depletion for correlation analyses. The SNPNL score is a semi-quantitative score (0–3) that has been validated by strong and significant correlation with striatal tyrosine hydroxylase ELISA [30].

Laryngeal Tissue Sampling and Preparation

The larynx specimens were obtained from 1 to 2 days after death (Table 1); this postmortem interval does not hinder reliable morphological and histochemical analysis of autopsied tissues when the body has been stored in a refrigerated area [22,23,24,25, 34,35,36,37].

Each larynx was bisected in the midline, forming a left and a right specimen. In this study, the left semi-larynges from PD subjects 1, 3, 5, 7, 9, 11, 13, 15, 17, 19 and right semi-larynges from PD subjects 2, 4, 6, 8, 10, 12, 14, 16, 18, 20 were investigated. For normal control (NC) group, left semi-larynges from NC 1, 3, 5, 7 and right semi-larynges from NC 2, 4, 6, 8 were studied.

For each larynx, the studied tissues included nerves, muscles, and mucosa. The nerves included recurrent laryngeal nerve (RLN), external superior laryngeal nerve (ESLN), and internal superior laryngeal nerve (ISLN). A 10-mm-long segment of each nerve was sampled before they enter the larynx (Fig. 1A), fixed with 10% neutral buffered formalin overnight, sectioned longitudinally (5 µm thick), and prepared for PAS immunohistochemistry. Three muscle samples (10 × 10 mm/each) were taken from RLN-innervated thyroarytenoid (TA; adductor of the vocal fold) and posterior cricoarytenoid (PCA; abductor of the vocal fold) muscles, and ESLN-innervated cricothyroid (CT; tensor of the vocal fold) muscle (Fig. 1B). Three mucosa samples (10 × 10 mm/each) were taken from laryngeal surface of the epiglottis (LSE), aryepiglottic fold (AEF) (Fig. 1C), and true vocal fold (TVF) (Fig. 1D). These regions of mucosa were chosen because they have rich sensory nerve terminals that are responsible for eliciting swallowing and laryngeal closure to protect airway during swallowing. The muscle and mucosa samples were frozen in isopentane cooled by dry ice and sectioned longitudinally (40 µm thick) and/or transversely (10 µm thick) on a cryostat (Reichert-Jung 1800, Mannheim, Germany) at − 25 °C. The sections were stored at − 80 °C until staining was performed.

Fig. 1

Schematics and photographs, showing sampling sites of the laryngeal nerves (boxed regions in (A), muscles (B), and mucosa (boxed regions in C and D). The photograph from inside of the larynx (D) shows the sampling site of the mucosa overlaying the true vocal fold (TVF). AEF, aryepiglottic fold; CT, cricothyroid muscle; ESLN, external superior laryngeal nerve; HB, hyoid bone; ISLN, internal superior laryngeal nerve; LSE, laryngeal surface of the epiglottis; PCA, posterior cricoarytenoid muscle; RLN, recurrent laryngeal nerve; SLN, superior laryngeal nerve; TA, thyroarytenoid muscle; TC, thyroid cartilage

Staining Methods

The longitudinal nerve sections were immunostained for PAS to identify PAS-immunoreactive (PAS-ir) axons. The longitudinal muscle and mucosa sections were stained using immunohistochemical methods, including neurofilament (NF) staining to label all the axons and immunostaining for PAS to identify PAS-ir axons. The muscle cross-sections were stained with hematoxylin and eosin to examine myofiber morphology and immunostained to detect fiber type-grouping, atrophied fibers, and denervated myofibers, as described below.

NF Staining to Label All the Axons

Some longitudinal muscle and mucosa sections were immunostained with monoclonal antibody SMI-31 against NF (Covance Research Products, Berkeley, CA) (Fig. 2) to label nerve fascicles, twigs, and axon terminals, as described [25, 38, 39]. Briefly, the sections were (1) blocked in phosphate-buffered saline (PBS) containing 0.3% Triton and 2% bovine serum albumin (BSA) for 30 min; (2) incubated with primary antibody SMI-31 (1:800; Covance Research Products, Berkeley, CA) in PBS containing 0.03% Triton at 4 ºC overnight; (3) incubated with anti-mouse biotinylated secondary antibody (1:1000; Vector Laboratories, Burlingame, CA) for 2 h; (4) treated with a VectaStain ABC kit (1:1000; ABC Elite, Vector); and (5) treated with diaminobenzidine-nickel as chromogen to visualize peroxidase labeling. Control sections were incubated without primary antibody.

Fig. 2

Photomicrographs of longitudinal sections of laryngeal muscles (A–C) and different regions of laryngeal mucosa (D–F) from a PD subject (PD no. 12). The sections were immunostained with monoclonal antibody SMI-31 against neurofilaments. Note that the RLN-innervated laryngeal muscles (i.e., TA, PCA and CT) and ISLN-innervated mucosa regions (i.e., TVF, LSE and AEF) examined have numerous intramuscular motor and intramucosal sensory nerve fascicles, twigs, and axons (dark staining). Magnification: 200 × for A through F

PAS Immunohistochemistry

Some longitudinal sections of the laryngeal nerves, muscles and mucosa were immunostained for PAS (Figs. 3, 4, 5), as described [22,23,24,25, 40, 41]. In brief, the tissue sections were (1) pretreated with proteinase K (1:100; Enzo Life Sciences, Farmingdale, NY) diluted in 0.1 mol/L PBS at 37 °C for 30 min; (2) immersed in 1% H2O2 in 0.1 mol/L PBS with 0.3% Triton X-100 (PBS-TX) at pH 7.4 for 30 min; (3) incubated in anti-PAS monoclonal antibody (psyn no. 64; 1:1000; Wako, Richmond, VA) in PBS-TX at 4 °C overnight; (4) incubated in biotinylated anti-mouse IgG (1:000; VectaStain kit, Vector) in PBS-TX for 2 h at room temperature (RT); (5) treated with avidin–biotin complex (ABC; Vector), with A and B components of the kit both at 1:1000 dilution for 30 min; (6) treated with 3,3׳-diaminobenzidine (Sigma, St. Louis, MO) (5 mg/100 ml) with added saturated nickel ammonium sulfate (2/100 mL) and H2O2 (5 μL/100 mL of 1% H2O2) in the dark for 30 min. The treated sections were washed 3 times in PBS-TX between staining steps. Control sections were incubated without a primary antibody.

Fig. 3

Photomicrographs of the longitudinal sections of the RLN from a PD patient (PD no. 5, male, age: 80) (A) and an age-matched normal control (NC no. 2, male, age: 80) (B). The sections were immunostained with monoclonal anti-phosphorylated α-synuclein antibody (psyn 64). Note that there are numerous PAS-ir axons (darkly stained threads and dots) in the RLN of the PD subject (A), whereas there are no PAS-ir axons in the control (B). Magnification: 200 × for A and B

Fig. 4

A–E Photomicrographs of the longitudinal sections of motor nerves (i.e., RLN and ESLN) (A, B) and their innervating muscles studied (i.e., TA, PCA, and CT) (C–E) from the same PD subject (PD no. 12) as that described in Fig. 2, who had severe SSV disorders (see Table 1). The sections were immunostained with monoclonal anti-phosphorylated α-synuclein antibody (psyn 64). Note that there are numerous PAS-ir axons (darkly stained threads and dots) in the laryngeal motor nerves and muscles. Magnification: 200 × for A through E. (A’–E’) The images in (A–E) were opened using ImageJ software and converted to 8-bit (binary) images, color thresholded, and particle analyzed to calculate the number and percent area of the PAS-ir axons in the peripheral laryngeal motor nervous system. For this subject, the mean number and mean area of PAS-ir axons in the laryngeal motor nerves and muscles examined were calculated to be 214 and 0.50, respectively

Fig. 5

A–D Photomicrographs of the longitudinal sections of sensory nerve (i.e., ISLN) (A) and its innervating laryngeal mucosa (i.e., TVF, LSE, and AEF) (B–D) from the same PD subject (PD no. 12) as that described in Figs. 2 and 3. The sections were immunostained with monoclonal antibody psyn 64. Note that there are numerous PAS-ir axons (darkly stained threads and dots) in the ISLN and different mucosa regions examined. Magnification: 200 × for A through D. A’–D’ The images in (A–D) were converted to 8-bit (binary) images by the use of ImageJ software to compute the number and percent area of the PAS-ir axons in the peripheral laryngeal sensory nervous system. For this subject, the mean number and mean area of PAS-ir axons in the ISLN and laryngeal mucosa regions were calculated to be 396 and 0.62, respectively

Immunostaining to Identify Fiber Type-Grouping and Atrophied Myofibers

Some cross-sections of the laryngeal muscles examined were immunostained with monoclonal antibody NOQ7-5-4D to label slow type I muscle fibers for identifying fiber type-grouping and atrophied myofibers (Fig. 6A–C) caused by partial denervation, as described [24]. In brief, muscle sections were (1) fixed in 4% paraformaldehyde for 10 min; (2) blocked in 2% BSA and 0.1% Triton X-100 for 20 min; (3) incubated with monoclonal antibody NOQ7-5-4D (1:1000; Sigma, St. Louis, MO) for 1 h at RT; (4) incubated with an anti-mouse IgG (ATCC, Rockville, MD) for 1 h; (5) reacted in ABC reagent for 1 h; and (6) processed with DAB substrate kit (SK-4100; Vector) for 10 min. Control sections were stained without primary antibody.

Fig. 6

Photomicrographs of the cross-sections of the TA muscles from PD no. 12 (A, A’), PD no. 15 (B, B’), and NC no. 2 (C, C’). The sections were stained with monoclonal antibody NOQ7-5-4-D (A–C) to label slow type I fibers (dark staining) and stained for N-CAM (A’–C’) to identify denervated muscle fibers (bright staining). Fiber type-grouping (arrow in A) and atrophied myofibers (A, B) were identified in the TA muscle of the PD subjects. N-CAM immunostaining showed that there were more denervated muscle fibers in the PD muscles (A’ B’) than in the controls. There were a few atrophied and denervated myofibers in the control muscles (C, C’). Magnification: 100 × for A–C and 200 × for A’–C’

Neural Cell Adhesion Molecule (N-CAM) Immunohistochemistry to Detect Denervated Myofibers

N-CAM is a molecular marker of muscle fiber denervation. It is abundant on the surface of early embryonic myotubes, decreases in level as development proceeds, almost disappears in the adult muscle, and reappears when adult muscles are denervated [42,43,44]. Therefore, N-CAM immunostaining is widely used for detecting denervated myofibers [44,45,46,47]. In this study, immunofluorescent labeling of N-CAM was performed (Fig. 6A’–C’) as described in our publications [24, 48]. Briefly, cross-sections of the laryngeal muscles were (1) fixed with methanol at − 20 °C for 20 min; (2) blocked with 5% donkey serum (Sigma) in PBS for 30 min; (3) incubated with a primary rabbit anti-rat N-CAM antibody (Chemicon, Temecula, CA) for 2 h; and (4) incubated with a secondary CY3-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 h. The sections were washed extensively in PBS between steps. Control sections were stained without primary antibody. The stained sections were mounted with Vectashield mounting medium (Vector Laboratories), kept in the dark at 4 °C, and photographed.

Image Acquisition and Analysis

The stained sections of the laryngeal nerves, muscles and mucosa were examined under a Zeiss photomicroscope (Axioplan-1; Carl Zeiss, Gottingen, Germany) and photographed using a USB 3.0 digital microscope camera (Infinity 3-3URC; Lumenera Corp., Ottawa, Ontario, Canada). Images at a magnification of 200 × were imported into an image-processing program (ImageJ v. 1.45 s; National Institutes of Health, Bethesda, MD) to compute the number and area fraction of the PAS-ir axons within a section area (1.0 mm2).

Assessments of PNS Pathology Severity

For a given tissue sample (i.e., laryngeal nerves, muscles, and mucosa), three sections at different spatial levels stained for PAS were examined to select one with the greatest number of PAS-ir axons. A microscopic field with the highest density of PAS-ir axons in the selected section was identified and photographed at 200 × magnification to estimate the number and area fraction of the PAS-ir axons with ImageJ.

On the basis of our prior studies on the PD pharynx [22, 23, 25], we developed a total PNS pathology score (TPPS) to indicate the PAS lesion severity. The TPPS consisted of a motor PAS score and a sensory PAS score. A motor PAS score was derived from the mean number of the PAS lesions (MNPL) in the motor nerves (i.e., RLN and ESLN) and their innervating muscles (i.e., TA, PCA, and CT). A sensory PAS score was derived from the MNPL in the sensory nerve (i.e., ISLN) and its innervating regions of the laryngeal mucosa (i.e., TVF, LSE, and AEF). For each PD subject, the motor and sensory PAS scores were summed and averaged to yield a TPPS, a measure of PAS pathology severity in the larynx by using a 4-tier scale (0–3): 0 = no lesions, 1 = 1 to 100 lesions (mild), 2 = 101 to 200 lesions (moderate), 3 = more than 200 lesions (severe) as calculated by ImageJ.

Statistical Analyses

The data gathered in the study were analyzed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA) or R software version 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria). Continuous random variables were summarized as mean or median (interquartile range) depending on whether the data are from a normal distribution. Categorical variables were presented as count (percentages) and examined using Fisher’s exact test or Pearson’s Chi-square test, as appropriate. Correlation between swallowing and speech scores and disease severity levels (i.e., H&Y stages and motor UPDRS), PNS and CNS pathology severity levels (i.e., TPPS, TBS, and SNPNL) were evaluated using Spearman’s rank correlation. Scatterplots with Locally Estimated Scatterplot Smoothing (LOESS) fit were produced to depict the relationship between speech score, swallowing score, SS score (defined as the sum of swallowing score and speech score), and TPPS. LOESS is a non-parametric regression technique that fits a smooth curve to a set of data points correlation between both PD durations (shorter vs. longer) and H&Y stages (H&Y stages 2, 3 vs. 4, 5), along with age at PD onset and sex, against swallowing score, speech score, TPPS, TBS, and SNPNL were examined using ordinal logistic regression and simple logistic regression, as appropriate. Statistical significance was set at p < 0.05.