Vocal fold atrophy in an Alzheimer’s disease rat model

Abstract

Objectives:

Advanced Alzheimer’s disease (AD) is associated with impaired voice, oropharyngeal dysphagia, and aspiration pneumonia. Vocal fold atrophy is hypothesized to contribute to these changes. While vocal fold atrophy is observed in AD, advanced age is co-prevalent and a confounding variable. The apolipoprotein E4 knock-in (ApoE4-KI) rat model is commonly used to study cognitive and motor changes in AD. This study aims to compare vocal fold atrophy in adult AD rats vs. age-matched controls.

Methods:

Vocal folds were visualized endoscopically in 4-month ApoE4-KI rats (n=12) and age-matched controls (n=8) following anesthesia. The thyroarytenoid (TA) muscle was collected for histology. Vocal fold bowing index (BI) was calculated from endoscopic still images by two blinded reviewers using digital software. Digital myofiber analysis was used to calculate TA myofiber average cross-sectional area (CSA) and minimum Feret diameter (MFD).

Results:

Mean BI was significantly higher in AD rats compared to controls (10.7±2.9 vs. 6.4±2.6, difference 4.3 with 95%CI 1.6-6.9, p<0.01). TA myofiber average CSA and MFD were significantly higher in AD rats compared to controls (average CSA (μm²), 586.3±76.9 vs. 435.8±49.4, difference 150.5 with 95%CI 91.2-209.9, p<0.01; MFD (μm), 23.0±1.1 vs. 19.8±1.3, difference 3.2 with 95%CI 2.0-4.4, p<0.01).

Conclusion:

Non-aged AD rats had significantly higher BI and TA myofiber size than controls. BI was strongly correlated with histologic metrics. These findings suggest that AD may lead to greater vocal fold atrophy in non-aged animals. Additional studies are needed to determine the mechanism(s) for vocal fold atrophy in this model.

Introduction

Alzheimer’s disease (AD) affects 7.2 million Americans over the age of 65 and is the sixth-leading cause of death in that age group.¹ AD is a neurodegenerative disorder characterized by a progressive decline in cognitive, behavioral, and motor function. Additionally, oropharyngeal dysphagia is prevalent in AD, affecting 84-93% of patients with moderate-severe AD.² Dysphagia contributes to poor feeding and weight loss in AD patients; more significantly, it also increases the risk of aspiration pneumonia and pneumonia-associated death.^3,4 Physiologic changes like increased swallow reflex latency and decreased swallow volume are predictive of aspiration and aspiration pneumonia in AD.^5–7 AD-related oropharyngeal dysphagia is also characterized by a prolonged oral phase, delayed hyolaryngeal elevation, decreased chewing, and decreased tongue movements.⁸ However, the underlying mechanism of swallow dysfunction in AD is poorly understood.

AD is associated with decreased activation of cortical swallowing centers including the pre- and post-central gyri and the Rolandic and frontal opercula.⁹ Additionally, autonomic peripheral neuropathies are associated with swallow dysfunction in advanced-stage AD.¹⁰ Our laboratory has performed electrophysiologic studies of the superior laryngeal nerve (SLN) in an AD rat model.¹¹ AD rats were found to have increased compound motor action potential (CMAP) latency and duration, increased sensory nerve action potential (SNAP) duration, and decreased swallow frequency.¹¹ In total, these results suggest that afferent and efferent conduction delays may contribute to a delayed swallow reflex.

Our laboratory has also investigated changes in laryngeal musculature associated with aging.^12–16 21-month-old aged rats were found to have increased recurrent laryngeal nerve (RLN) CMAP latency compared to 4-month-old non-aged rats. Aged rats also had increased vocal fold bowing compared to non-aged rats.^{12,13,15–17} Most notably, aged rats exhibited histologic laryngeal changes including increased thyroarytenoid myofiber cross-sectional area and minimum Feret diameter.¹⁵ There was also a trend toward decreased RLN myelination, but this difference did not reach statistical significance.¹⁵ These results demonstrate that aging rats have quantifiable changes in laryngeal innervation, musculature, and framework.

In this study, we aimed to quantify vocal fold bowing and laryngeal muscular changes in an AD rat model. Motor dysfunction in general is frequently observed in AD patients, and is associated with higher rates of cognitive decline.^18–20 The mechanism of motor dysfunction in AD is thought to be multifactorial, with potential contributors including mitochondrial dysfunction,^21–24 amyloid accumulation,^25–27 and neuroinflammation.²⁸ In particular, the apolipoprotein E ε4 (ApoE4) allele, a genetic polymorphism that predisposes individuals to AD, is associated with poor mobility and more rapid motor decline.^29–31

Given findings in patients with ApoE4 polymorphisms, we set out to measure vocal fold bowing and perform thyroarytenoid (TA) myofiber analyses in an ApoE4 knock-in (ApoE4-KI) rat model. This model exhibits neurodegeneration, tauopathy, and cognitive deficits as observed in human AD.^32–36 Additionally, the ApoE4-KI rat model exhibits laryngeal electrophysiologic changes as described above.¹¹ By examining both endoscopic vocal fold bowing and laryngeal histologic changes, we aimed to develop a mechanistic understanding of vocal fold atrophy in AD. We hypothesized that ApoE4-KI rats would exhibit vocal fold atrophy, characterized by increased bowing and laryngeal muscular changes, which may in turn contribute to dysphagia and aspiration.

Materials and Methods

This study was conducted in adherence with guidelines by the Public Health Service policy on Humane Care and Use of Laboratory Animals, and the Animal Welfare Act (7U.S.C. et seq.). The animal use protocol was approved by the Institutional Animal Care and Use Committee. Rats were housed in a restricted-access facility with 12-hour cycles of light and dark. The welfare of animals was assessed by facility staff daily. Two cohorts of male Sprague-Dawley (SD) rats were used for the study, a wild-type cohort with n=8 (Charles River Laboratories, Wilmington, MA) and an ApoE4-KI cohort with n=12 (Inotiv, Lafayette, IN).

Surgical Preparation

Surgical preparation for this study was performed as previously described by our laboratory.^{11,14,15,37–39} A weight-based cocktail of ketamine (50 mg/kg) and xylazine (5 mg/kg) was injected intraperitoneally for anesthesia. The rat was then placed supine on an intubation stand (Kent Scientific, Torrington, CT). All equipment was secured on a vibration control table (TMC Vibration Contral Air Table, Peabody, MA). An 11.5 Fr, 30° rigid endoscope (Karl Storz, Tuttlingen, Germany) was advanced transorally to visualize the bilateral true vocal folds. Approximately 10 seconds of high-definition video (60 frames per second) was recorded using a digital video capture system (Stryker Corp., Kalamazoo, MI). A midline incision was made extending from the hyoid bone to the sternal notch. At the end of the experiment, the rat was euthanized with intraperitoneal injection of pentobarbital (200 mg/kg) and the larynx, including the intrinsic musculature, was harvested for histology.

Endoscopic Measurement of Bowing Index

Vocal fold morphology was quantified using still images from endoscopic videos. Endoscopic stills were identified from each video at the point of maximal vocal fold abduction. The anterior commissure and bilateral vocal processes were identifiable in all stills (Figure 1A). Vocal fold length and bowed distance were measured bilaterally for all rats, and total Bowing Index (BI) was calculated as displayed in Figure 1B. Measurements were performed independently by two blinded raters using digital software (Glottal Image Capture, GlottICv1.0)¹⁷. Intra-rater reliability was tested using duplicate videos presented to each rater. Inter-rater reliability was tested using intraclass correlation calculated between raters.

Figure 1.

Measurement and calculation of vocal fold bowing index bilaterally. A. Endoscopic image representing the (x) bilateral true vocal folds, (y) bilateral arytenoid cartilages, and (*) laryngeal surface of the epiglottis. Blue lines represent the vocal fold length from the vocal process to the anterior commissure. Pink lines represent the bowed distance. B. Calculation of bowing index. Vocal length (L₁ and L₂) is displayed in blue and bowed distance (d₁ and d₂) is displayed in pink. Dark blue curved lines approximate the free margins of the true vocal folds.

Tissue Processing and Histology

Larynges, including the intrinsic musculature, were harvested and fixed in 10% buffered formalin solution. Samples were embedded in paraffin and sectioned into 4 μm coronal slices using a rotary microtome. Sections were then mounted on glass sides and treated with a series of xylene and ethanol washes. Hematoxylin and eosin Y (H&E) were applied sequentially. Slides were imaged using an optical microscope with 20x air and 60x oil lenses and the TA muscle was identified (Figure 2A). Images were processed with ImageJ (version 2.3, Bethesda, MD) and each TA myofiber was manually highlighted (Figure 2B). Myofiber number, average cross-sectional area (CSA), and average minimum Feret diameter (MFD) were calculated. Minimum Feret diameter represents the smallest cross-sectional diameter of each myofiber.⁴⁰

Figure 2.

Thyroarytenoid myofiber analysis using ImageJ. A. H&E staining of the thyroarytenoid muscle. B. Digital myofiber identification and isolation. Cross-sectional area and minimum Feret diameter are calculated for each displayed myofiber.

Data Analysis and Statistics

Descriptive statistics and figure production were performed in GraphPad Prism (version 10, San Diego, CA). Mean and standard deviation were used to summarize variables, and violin plots were created for data visualization. Two-tailed t-tests were used to determine p-values with α=0.05. Mean difference and 95% confidence intervals (95%CI) were also calculated. To compare metrics, Pearson’s correlation coefficient with 95%CI was calculated. Scatter plots with linear regression were created for data visualization.

Results

Bowing Index

Total BI was calculated bilaterally for all rats. Mean total BI was significantly higher in AD rats compared to controls (10.7±2.9 vs. 6.4±2.6, difference 4.3 with 95%CI 1.6 to 6.9, p<0.01) (Figure 3). There was a moderate intraclass correlation between raters (r=0.50, 95%CI −0.01 to 0.82). Two duplicate measurements were performed by each rater. For the first set of duplicates, rater 1 scored 12.71 and 14.76; rater 2 scored 11.03 and 11.67. For the second set of duplicates, rater 1 scored 3.12 and 4.57; rater 2 scored 8.11 and 9.34.

Figure 3.

Comparison of total bowing index between wild-type (WT) and ApoE4-KI Alzheimer’s disease rats. Bowing index is significantly higher in ApoE4-KI rats compared to controls (10.7±2.9 vs. 6.4±2.6, difference 4.3 with 95%CI 1.6-6.9, p<0.01).

Muscle Histology

Myofiber number, average CSA, and average MFD were calculated for all rats. There was no significant difference in fiber number between AD rats and controls (2291±225 vs. 2086±257, difference 205 with 95%CI −35 to 445, p=0.09) (Figure 4A). Average CSA (in μm²) was significantly higher in AD rats compared to controls (586.3±76.9 vs. 435.8±49.4, difference 150.5 with 95%CI 91.2 to 209.9, p<0.01) (Figure 4B). Average MFD (in μm) was significantly higher in AD rats compared to controls (23.0±1.1 vs. 19.8±1.3, difference 3.2 with 95%CI 2.0 to 4.4, p<0.01) (Figure 4C).

Figure 4.

Myofiber analysis of wild-type (WT) and ApoE4-KI Alzheimer’s disease rats. A. Comparison of fiber number between WT and ApoE4-KI rats. There is no significant difference in fiber number between ApoE4-KI rats and controls (2291±225 vs. 2086±257, difference 205 with 95%CI −35 to 445, p=0.09). B. Comparison of average cross-sectional area (CSA) between WT and ApoE4-KI rats. Average CSA (μm²) is significantly higher in ApoE4-KI rats compared to controls (586.3±76.9 vs. 435.8±49.4, difference 150.5 with 95%CI 91.2 to 209.9, p<0.01). C. Comparison of average minimum Feret diameter (MFD) between WT and ApoE4-KI rats. Average MFD (μm) is significantly higher in ApoE4-KI rats compared to controls (23.0±1.1 vs. 19.8±1.3, difference 3.2 with 95%CI 2.0 to 4.4, p<0.01).

Correlation coefficients were calculated to compare BI to myofiber metrics. There was no significant correlation between BI and fiber number (r=0.37, 95%CI −0.09 to 0.70, p=0.11) (Figure 5A). There was a strong correlation between BI and average CSA (r=0.71, 95%CI 0.39 to 0.88, p<0.01) (Figure 5B). There was a strong correlation between BI and average MFD (r=0.73, 95%CI 0.42 to 0.88, p<0.01) (Figure 5C). Additionally, correlation coefficients were calculated to compare fiber number with other myofiber metrics. There was no significant correlation between fiber number and average CSA (r=0.34, 95%CI −0.12 to 0.68, p=0.14) (Supplemental Figure 1A). There was a moderate correlation between fiber number and average MFD (r=0.46, 95%CI 0.02 to 0.75, p=0.04) (Supplemental Figure 1B).

Figure 5.

Correlation between bowing index and myofiber average cross-sectional area (CSA) and minimum Feret diameter (MFD). Wild-type (WT) and ApoE4-KI Alzheimer’s disease rats are displayed in blue and purple, respectively. A. There is no significant correlation between bowing index and fiber number (r=0.37, 95%CI −0.09 to 0.70, p=0.11). B. There is a strong correlation between bowing index and average CSA (r=0.71, 95%CI 0.39 to 0.88, p<0.01). C. There is a strong correlation between bowing index and average MFD (r=0.73, 95%CI 0.42 to 0.88, p<0.01).

Discussion

In this study, we quantified vocal fold changes in adult AD rats vs. aged-matched controls. Prior studies have performed similar quantification in rodent models of aging,^15,41 but this is the first study to our knowledge to investigate laryngeal structural changes in AD. We found that AD rats had increased vocal fold bowing and histologic muscular changes, specifically, increased TA myofiber CSA and MFD. Bowing index was strongly correlated with CSA and MFD, suggesting a mechanistic relationship between increased muscle fiber area and vocal fold bowing. These results provide evidence that AD leads to changes in vocal fold structure and muscle fiber morphology in non-aged animals. In tandem with our neurophysiologic studies in AD rats,¹¹ we demonstrate that the ApoE4-KI rat model has observable anatomic and physiologic changes in the larynx compared to controls. The ApoE4-KI rat model may have utility in testing treatments for vocal atrophy and oropharyngeal dysphagia, although additional studies are necessary to determine the underlying mechanisms of these abnormalities.

The results of this study mirror our findings in the F344XBN aging rat model.¹⁵ In that study, aged rats were found to have increased total BI, myofiber CSA, and myofiber MFD. Histologic findings in both studies were surprising given that aging and AD are associated with decreased muscle mass and body weight.^{18,27,28,42–44} An explanation for these findings is the possibility that myofiber cross-sectional area and diameter do not adequately reflect overall muscle mass or functionality. A characteristic of aging muscle is fat infiltration around and within fibers, which may present as enlarged myofibers that have abnormal contractility.^45–47 Fat infiltration may be occurring in the TA muscle with aging and neurodegenerative disease, leading to enlarged TA myofibers in the context of increased vocal fold bowing. Aging is also associated with irregular myofiber shape, which may affect CSA and MFD at varying degrees depending on the cross-sectional location that these metrics are calculated.⁴⁸ Future studies may attempt to measure TA contraction force in addition to myofiber morphology, which may provide a more comprehensive investigation of the muscular changes occurring in AD.

The underlying cause of TA muscle dysfunction in AD is unclear, but may be related to intrinsic muscle changes, central nervous dysfunction, and/or peripheral nervous dysfunction. Prior studies have suggested that skeletal muscle may exhibit abnormal mitochondrial function in AD, increasing oxidative stress and ultimately resulting in muscle loss.^21–24 Future studies may attempt to quantify mitochondrial function in TA myofibers to test this hypothesis and potentially develop targeted treatments. Muscle dysfunction may also reflect abnormalities in the central or peripheral nervous system, and atrophy may be a result of reduced neural inputs. Prior studies demonstrating decreased activation of cortical swallowing centers and autonomic peripheral neuropathies in AD patients provides some evidence for this hypothesis.^9,10 Additionally, we have demonstrated that AD rats have increased CMAP latency and duration when measuring from the SLN.¹¹ Aging rats have similar physiologic changes in the RLN and also demonstrate decreased RLN myelination.¹⁵ Given these findings, we anticipate that AD rats will exhibit increased RLN latency and decreased myelination; we plan on performing these evoked response and histologic measurements in future studies.

In humans, aging is associated with vocal atrophy, vocal fold bowing, and TA muscle changes.^42–44 Cadaveric studies have demonstrated a decrease in TA cross-sectional area and myofiber size in elderly individuals compared to younger adults.^42,43 However, in an MRI study with human subjects, Ziade et al. found no significant difference in TA muscle volume between elderly (>65 years old) and non-elderly (<65 years old) adults.⁴⁹ In various histologic studies of the aging TA in humans, no consistent change in muscle size was found, but the aging TA did have altered fiber types, increased connective tissue content, and increased variability of fiber sizes.^50–52 Additionally, there are prominent age-related changes in the vibratory vocal fold mucosa that are not captured in TA muscle histology and may have a significant contribution to observed bowing.^53,54 In summary, although vocal fold bowing is associated with TA muscle changes in humans, it is unlikely that TA atrophy alone is responsible for age-related bowing. Given the similarities between aging and AD rat models, AD patients are likely to have similar laryngeal changes as aging humans. Future studies are necessary to determine if these changes develop independent of age in human subjects. Notably, we have demonstrated that patients with Parkinson’s disease exhibit a greater degree of vocal fold bowing than age-matched controls.¹³

There are several limitations to this study. First, the endoscopic and histologic metrics obtained in this study do not necessarily correlate with changes in laryngeal function. Future studies may attempt to investigate laryngeal physiology through quantification of aspiration, RLN function, and TA contraction force. Second, this study does not test effects of sex differences or aging in the ApoE4-KI rat model. Future studies will incorporate female ApoE4-KI rats and follow rats longitudinally over the course of months. Finally, there may be utility in studying additional rat models of AD, including the McGill-R-Thy1-APP model, which exhibits prominent motor deficits.⁵⁵

Conclusion

We find that non-aged ApoE4-KI rats have significantly higher BI and TA myofiber CSA and MFD than age-matched controls. BI was strongly and positively correlated with both CSA and MFD. These findings suggest that AD may lead to vocal fold atrophy independent of advanced age. Future studies will determine the mechanism of vocal fold atrophy in this model, quantify the physiologic effects of vocal fold atrophy, and investigate interventions to treat laryngeal dysfunction in AD.