Pathologic Classification of a Late Onset Peripheral Neuropathy in a Spontaneous Labrador Retriever Dog Model

Abstract

Late-onset peripheral neuropathy (LPN) is a heritable canine neuropathy commonly found in Labrador retrievers and is characterized by laryngeal paralysis and pelvic limb paresis. Our objective was to establish canine LPN as a model for human hereditary peripheral neuropathy by classifying it as either an axonopathy or myelinopathy and evaluating length-dependent degeneration. We conducted a motor nerve conduction study of the sciatic and ulnar nerves, electromyography (EMG) of appendicular and epaxial musculature, and histologic analysis of sciatic and recurrent laryngeal nerves in LPN-affected and control dogs. LPN-affected dogs exhibited significant decreases in compound muscle action potential (CMAP) amplitude, CMAP area, and pelvic limb latencies. However, no differences were found in motor nerve conduction velocity (MNCV), residual latencies, or CMAP duration. Distal limb musculature showed greater EMG changes in LPN-affected dogs. Histologically, LPN-affected dogs exhibited a reduction in the number of large-diameter axons, especially in distal nerve regions. In conclusion, LPN in Labrador retrievers is a common, spontaneous, length-dependent peripheral axonopathy that is a novel animal model of age-related peripheral neuropathy that could be used for fundamental research and clinical trials.

Graphical Abstract

The pathologic features of late-onset peripheral neuropathy (LPN) were classified in Labrador retrievers via motor nerve conduction study, electromyography (EMG), and histologic analysis of nerves in affected and control dogs. Labrador retriever LPN is a length-dependent peripheral axonopathy and a potential large animal model of human age-related peripheral neuropathy.

Introduction

Late-onset peripheral neuropathy (LPN) is a common life-limiting, progressive canine disease typically affecting older, medium- to large-breed dogs. LPN is clinically characterized by paralysis of the intrinsic laryngeal muscles and progressive pelvic limb weakness [1–5]. Labrador retrievers account for up to 70% of LPN cases [6–8]. The onset of clinical signs in Labrador retrievers is typically 10 years of age [5]. Based on a genome wide association study and variant calling from whole genome sequencing, there is strong evidence of a substantial genetic component to LPN in the Labrador retriever [9].

Similar to many human peripheral neuropathies, LPN in dogs predominantly affects the longest nerves [6–8,10–11]. In dogs, the recurrent laryngeal nerve is proportionally longer than in humans [12]. Loss of recurrent laryngeal nerve function results in upper airway obstruction; changes in bark, inspiratory stridor, and exercise intolerance are typical, with severe signs including dyspnea, cyanosis, syncope, and heat stroke [1, 3, 13–14]. Pelvic limb weakness, decreased pelvic limb postural reactions, and pelvic limb hyporeflexia are also typical [6–8,11]. Electrodiagnostic studies of thoracic limbs in affected dogs also indicate neurodegeneration [7]; however, clinical signs of thoracic limb weakness are less commonly detected clinically.

Hereditary peripheral neuropathies in humans are relatively common (e.g., Charcot-Marie-Tooth (CMT) disease affecting ~1 in 2500 individuals), and likewise, LPN in the Labrador retriever is suggested to be a genetic disease [6], although causal genetic variants are not known. Moreover, multiple breed-specific genetic associations have been identified in early-onset versions of neurodegenerative conditions associated with laryngeal paralysis [15–18].

Spontaneous canine models of neurodegenerative diseases are valuable resources for investigation of disease pathogenesis and for trials of disease-modifying therapies. Examples of successful large animal neurodegenerative spontaneous disease models include canine amyotrophic lateral sclerosis, epilepsy, muscular dystrophy, and canine cognitive dysfunction [19–22]. Peripheral neuropathy is common in dogs, and Labrador retrievers have the highest breed prevalence of LPN [23]. LPN is thus a promising large animal model for human inherited peripheral neuropathies, such as CMT, distal hereditary motor neuropathy (DHMN) and other late-onset genetic neuropathies.

Across species, inherited peripheral neuropathies are caused by disturbances in myelination or axonal degeneration [24]. Nerve conduction studies (NCS) and electromyography (EMG) are diagnostic tools for specific neurologic disease classification [25–29]. Similarly, electrodiagnostic findings in dogs can differentiate between axonopathy and diseases of myelination [30]. In addition, histological analysis of peripheral nerve biopsies can be used to help determine the underlying pathologic cause of neuropathy [31–34].

Clinically, LPN appears to be a peripheral neuropathy affecting motor nerves, although sensory involvement has not been thoroughly examined [6–8, 10–11]. Prior qualitative work of LPN in multiple breeds of dogs has shown a generalized loss of axons in the tibial and ulnar nerves [7], with decreases in NCV and CMAP amplitude when compared to historic controls [6–7], as well as increased EMG abnormalities in distal limb musculature [6]. Given the diverse findings in LPN in multiple breeds, we undertook a more focused study on the pathologic features of LPN in the Labrador retriever utilizing motor nerve electrodiagnostic study and quantitative spinal cord and peripheral nerve pathology. Here, we show that Labrador retrievers clinically diagnosed with LPN have electrodiagnostic and histologic features consistent with a length-dependent peripheral axonopathy.

Materials and Methods

For the electrodiagnostic study, dogs were recruited from the University of Wisconsin Veterinary Care Hospital between May 2021 and December 2021. For the neuropathology study, tissues were collected from dogs as part of a body donation program through the University of Wisconsin-Madison School of Veterinary Medicine from dogs euthanized for reasons unrelated to this study between July 2018 and January 2022. For both studies, procedures were conducted with the approval of the Institutional Animal Care and Use Committee, School of Veterinary Medicine, University of Wisconsin-Madison (V1070). Dogs were recruited, or bodies used postmortem, with informed written consent from each owner.

Electrodiagnostic Selection Criteria

Electrodiagnostic studies were performed on 2 groups of dogs, an LPN group and an age- and height-matched control group. The LPN group included pure-bred Labrador retrievers older than 10 years of age diagnosed with LPN by either a board-certified veterinary neurologist (HR) or board-certified veterinary surgeon (SJS). For inclusion, dogs had to be over 10 years of age and present with clinical signs of inspiratory stridor consistent with laryngeal paralysis and a neurological examination indicative of peripheral neuropathy. Pure-bred status was determined using pedigree information. For the control group, healthy age- and height-matched dogs were used. Withers height was measured as previously described [35]. Dogs were deemed healthy with normal physical and neurologic examinations, complete blood counts and serum biochemistries. Labrador retrievers and Golden retrievers less than 13.5 years of age were excluded from the control group to prevent inadvertent inclusion of preclinically affected dogs. For both groups, dogs were excluded if serum biochemistry values were abnormal or if dogs had any known conditions associated with peripheral neuropathy, such as diabetes, a history of chemotherapy or steroid use, or other uncontrolled endocrinopathies. Each dog’s weight, height, sex, and neuter status were recorded. The neurologic examination performed included cranial nerve evaluation, spinal reflex evaluation (biceps brachii, triceps, patellar, gastrocnemius, withdrawal reflexes, anal reflex and cutaneous trunci reflex), paw replacement tests and gait analysis.

Electrodiagnostic Study Sedation Protocol

A subset of dogs were premedicated at home the day of the study with oral gabapentin (8.45 ± 1.78 mg/kg; Zoetis Inc., Kalamazoo, MI) and/or oral trazadone (8.68 ± 0.70 mg/kg; Zoetis Inc., Kalamazoo, MI) to prevent acute respiratory distress from anxiety or excitement or to minimize anxiety as deemed necessary based on individual clinical history. Dogs were sedated by a board-certified veterinary anesthesiologist (RJ) using an intravenous (IV) bolus of butorphanol (0.24±0.09 mg/kg; Zoetis Inc., Kalamazoo, MI) and an IV bolus followed by a bolus of dexmedetomidine followed by a constant rate infusion (2.2±0.7 μ/kg and 0.66±0.43 μg/kg/hr, respectively; Zoetis Inc., Kalamazoo, MI). Sedation was individually titrated so that dogs maintained significant palpebral reflexes, jaw tone and laryngeal (swallow) reflexes, yet tolerated study procedures. Oxygen was provided via a face mask. Dogs were monitored for respiratory rate, pulse rate and hemoglobin oxygen saturation (SpO₂ from a transmittance probe placed on the lip; data not shown). At study completion, rectal body temperature was obtained, and atipamezole (0.038±0.012 mg/kg; Zoetis Inc., Kalamazoo, MI) was administered intramuscularly. Postprocedural rectal temperatures were within normal limits for all dogs.

Nerve Conduction Study and Electromyography

A routine motor nerve conduction study (mNCS) was completed using an evoked potential system (UltraPro S100, Natus Neurology Inc, Middleton, WI). Needle electrodes were used [6–7, 36]. The sciatic nerve was stimulated at the level of the hip (proximal sciatic), the common peroneal branch was stimulated at the stifle joint (mid sciatic) and superficial fibular branch (distal sciatic) was stimulated at the level of the tarsus; the recording electrodes were placed over the extensor digitalis brevis lateralis muscle. The ulnar nerve was stimulated at the level of the elbow joint and the carpus with the recording electrodes placed superficial over the digital interosseus muscle. Recordings of CMAP amplitude, CMAP duration, CMAP area, MNCV, latency, and residual latency were undertaken for both sciatic and ulnar nerves as described previously [30]. Residual latency was calculated from the formula Residual latency = L – D/V, where L is the observed distal latency (ms), D is the distance (mm) from the distal stimulating to the distal recording electrode, and V is the calculated distal NCV (m/s). Motor nerve conduction velocity was calculated by measuring the length of the nerve between the proximal and distal stimulating sites and dividing the length by the latency of the proximal CMAP - distal CMAP. In a few dogs, not all electrodiagnostic measurements were completed due to the dogs not tolerating the procedure and/or concerns regarding prolonged sedation.

After the nerve conduction study was completed, routine needle EMG was undertaken as previously described [7] (UltraPro S100, Natus Neurology Inc., Middleton, WI). In the pelvic limb, the muscles assessed included the metatarsal interosseous, cranial tibial, gastrocnemius, semitendinosus, quadriceps, gluteus medius, and biceps femoris. In the thoracic limb, the muscles assessed included the metacarpal interosseous, flexor carpi ulnaris, extensor carpi ulnaris, flexor carpi radialis, extensor carpi radialis, biceps brachii, triceps, deltoid, supraspinatus, and infraspinatus. The cervical, thoracic, and lumbar epaxial muscles were also evaluated. Spontaneous activity, including positive sharp waves (PSWs) and fibrillation potentials, was recorded and graded as 0, 1+, 2+, 3+ or 4+. A score of 0 meant an absence of the abnormality under question, 1+ meant the abnormality was persistent but not sustained in more than 2 areas, 2+ meant moderate numbers of abnormalities in at least three areas of the muscle tested, 3+ meant the abnormality was present in all areas of the muscle tested, and 4+ meant the baseline signal was obliterated by the abnormal signal [37].

Neuropathology Inclusion Criteria

The neuropathology evaluation and the electrodiagnostic studies comprised distinct sets of dogs. Tissues from three groups of dogs were included: an LPN group, an aged control group, and a young control group. The LPN group included tissues from dogs over 10 years of age diagnosed with LPN by either a board-certified veterinary neurologist (HR) or board-certified veterinary surgeon (SJS) prior to euthanasia based on the clinical signs detailed above. The aged control group included tissues from medium to large breed dogs over 12 years of age that did not have clinical evidence or a history of neurodegenerative disease; Labrador and Golden retrievers were only considered for this group if they were over 12.75 years of age with no clinical signs or known familial history of LPN. The young control group included tissues from healthy, medium-to large-breed dogs aged 2 and 3 years.

Neuropathology Study

Nerve tissue was harvested within one hour of euthanasia. For peripheral nerve analysis, samples were taken from the RLN midway between the thoracic inlet and the distal aspect of the larynx and three levels of the sciatic nerve, including the sciatic nerve at the level of the hip (proximal sciatic), the common peroneal nerve at the level of the stifle (mid-sciatic) and the superficial fibular nerve at the level of the hock (distal sciatic). For analysis of motor neurons in the ventral horns of the spinal cord, the spinal cord was collected at the vertebral L5-L6 level, wherein the L7-S1 nerve roots and the sciatic nerve arise in the dog.

Peripheral nerves were fixed in 2.5% glutaraldehyde and 2% paraformaldehyde in a 0.1M phosphate buffer and subsequently resin embedded as previously described with minor modifications in that 1% osmium tetroxide was used post-fixation, and resin concentrations were increased over a 2-day period, after which samples were polymerized at 60°C for 24 hours [38]. Half micrometer sections were made and stained with toluidine blue. Images were taken at 60x magnification and blinded prior to analysis. A moving mean was used to pinpoint the minimum number of fields of view needed for LPN-affected, aged control, and young control group morphometric evaluation. Seven fields of view were imaged and evaluated for LPN-affected dogs. Five fields of view were imaged and evaluated for aged and young control dogs. Morphometric evaluation was performed using ImageJ [39]. Cross-sectional images were manually assessed to determine the g-ratio and axon diameter. The g-ratio was calculated as the ratio of the inner axonal diameter to the total outer diameter, as previously described [40]. Axon diameters were calculated by counting all myelinated axons. To evaluate group differences, axons were binned into 2 μm groups from <2 μm to >6 μm.

For transmission electron microscopy (TEM), peripheral nerves were fixed, post-fixated, and resin embedded as described in the previous paragraph. One-micrometer sections were cut, and TEM was performed on a Philips CM120 transmission electron microscope. The images were captured with a BioSprint 12 series digital camera using AMT Image Capture Engine V700.

Spinal cord sections were collected within 1 hour of euthanasia and stored in 10% formalin prior to paraffin embedding. Serial deparaffinized samples from each dog were stained for Nissl substance using cresyl violet acetate (26089–20; Electron Microscopy Sciences, USA) and immunostained for homeobox transcription factor 9 (HB9), a motor neuron specific transcription factor (1:100; ABN174; Sigma‒Aldrich, USA). ABN174 was raised against KLH-conjugated linear peptide corresponding to human MNX1 and was evaluated by Western Blot in PANC-1 cell lysate with a 50 kDa band observed (manufacturer datasheet). The primary antibody was followed by anti-rabbit HRP secondary antibody (760–4311; Roche, USA) and counterstaining with hematoxylin. Images were blinded prior to analysis. One spinal cord section was used for HB9 immunostaining and analysis, and the next cut section was used for Nissl staining and analysis. The number of motor neurons in each section were counted for both HB9 immunostaining and Nissl staining. Morphometric evaluation consisting of counting ventral horn motor neurons was undertaken using ImageJ [39].

Data Analysis

Data were analyzed using GraphPad Prism 9.2.0 software. Data were tested for normality using the Shapiro‒Wilk test. For the electrodiagnostic study, Student’s t test was used to compare mean age, weight, height, and age between groups, and the Mann‒Whitney U test was used to compare differences in temperature after testing between groups. A 2-way ANOVA with a Šídák multiple comparisons post hoc test was used to analyze differences between LPN-affected and control groups at different locations for motor nerve conduction outputs. Student’s t test was used to compare the means of latency, residual latency, and thoracic MNCV results. For peripheral nerve histopathology analysis, the Mann‒Whitney U test was used to evaluate differences in group age between the LPN and aged control groups. Differences in median axon diameter and g-ratio between groups at each location were analyzed with a Kruskal‒Wallis test. Differences in axon diameter and g-ratio between groups for each level at which tissue was collected were analyzed with either a one-way ANOVA or a Kruskal‒Wallis test, as appropriate. For spinal cord analysis, motor neurons were counted in both ventral horns and compared between groups using one-way ANOVA. The results are presented as the mean ± standard deviation or median (range), as appropriate. The results were considered significant at P<0.05.

Results:

Electrodiagnostic Study Population:

For the electrodiagnostic study, a total of 18 dogs were recruited, including 10 LPN-affected dogs and 8 control dogs. One affected dog was removed from the study after physical examination due to concern for cachexia; further evaluation provided evidence of metastatic cancer on thoracic radiographs. The LPN-affected group (n=9) was composed of pure-bred Labrador retrievers, including 1 intact male, 4 castrated males, and 4 spayed females. The control group (n=8) included Labrador retrievers (n=2), Golden retrievers (n=2), an American Staffordshire Terrier (n=1), an English Setter (n=1), a Siberian Husky (n=1), and a Standard Poodle (n=1), of which there was 1 intact male, 4 spayed females, and 3 castrated males. There were no significant differences in age (P=0.72) or height (P=0.22) between groups, although the LPN-affected group had a significantly increased weight compared to the control group (P<0.01) (Table 1).

Table 1.

Electrodiagnostic study population phenotypic variables

Variable	LPN Affected Group	Control Group	P value
Age (Years)	11.72±1.21 (n=9)	11.93±1.21 (n=8)	P=0.72
Height (cm)	59.43±3.31 (n=7)	57.02±3.90 (n=8)	P=0.22
Weight (kg)	36.21±6.76 (n=9)	27.55±4.540 (n=8)	P<0.01

LPN – late-onset peripheral neuropathy

Neurologic Examination:

For the electrodiagnostic study, all dogs in the LPN-affected group had evidence of LPN, including increased upper respiratory noise indicative of laryngeal paralysis (100%), reduced pelvic limb withdrawal reflexes in both pelvic limbs (77%), absent pelvic limb paw replacement tests (100%), absent thoracic limb paw replacement test (66%), paraparesis (55%) or tetraparesis (33%). Of the control group, 75% had an absent paw replacement test in the pelvic limbs; all other aspects of the neurologic examination were normal in this group.

Electrodiagnostic Findings:

Overall, sciatic nerve CMAP amplitude was decreased in the LPN-affected group compared to the control group, although the decreases were only significant when the mid and distal sciatic nerves were stimulated (Table 2, Figure 1A/B, Figure 2A). Similarly, ulnar nerve CMAP amplitude was significantly decreased in the LPN-affected group compared to the control group when the distal ulnar nerve was stimulated, while the apparent difference with stimulation at the proximal ulnar nerve was not significantly different between groups. (Table 2, Figure 2B).

Table 2.

CMAP Amplitude, Duration and Area

		CMAP Amplitude (mV)			CMAP Duration (ms)			CMAP Area (mVms)
Nerve	Level	LPN Affected Group	Control Group	P value	LPN Affected Group	Control Group	P value	LPN Affected Group	Control Group	P value
Sciatic	Proximal (Hip)	1.60 ± 1.33 (n=9)	5.56 ± 4.22 (n=8)	P=0.13	9.34 ± 3.52 (n=9)	8.05 ± 2.65 (n=8)	P=0.63	2.54 ± 1.78 (n=9)	8.61 ± 6.65 (n=8)	P=0.09
	Mid (Stifle)	1.40 ± 0.89 (n=8)	6.35 ± 3.61 (n=8)	P<0.05	8.50 ± 2.84 (n=8)	7.49 ± 1.06 (n=8)	P=0.77	3.12 ± 2.76 (n=8)	11.20 ± 7.43 (n=8)	P<0.05
	Distal (Hock)	2.05 ± 1.20 (n=9)	10.99 ± 8.03 (n=8)	P<0.001	7.02 ± 2.32 (n=9)	6.52 ± 0.72 (n=8)	P=0.96	3.20 ± 2.39 (n=9)	14.20 ± 9.20 (n=8)	P<0.001
Ulnar	Proximal (Elbow)	3.12 ± 1.35 (n=7)	10.39 ± 7.79 (n=7)	P=0.24	8.17 ± 2.66 (n=7)	7.00 ± 1.83 (n=7)	P=0.54	4.33 ± 2.37 (n=7)	11.76 ± 8.86 (n=7)	P=0.26
	Distal (Carpus)	4.3 7± 4.01 (n=9)	17.00 ± 14.80 (n=8)	P=0.01	6.14 ± 2.02 (n=9)	5.75 ± 2.14 (n=8)	P=0.91	5.01 ± 5.71 (n=9)	17.36 ± 15.26 (n=8)	P<0.05

LPN – late-onset peripheral neuropathy; CMAP - compound muscle action potential

Figure 1:

Representative mNCS and EMG traces. The representative image for Control mNCS is from a 10.4-year-old English Setter (A), and the representative image for the LPN-Affected is from an 11.0-year-old Labrador retriever clinically diagnosed with LPN showing decreased CMAP amplitudes compared to control (B). Two EMG abnormalities noted were positive sharp waves (C) and fibrillation potentials (D). Representative positive sharp waves measured from quadricep muscle of 12.3-year-old Golden retriever control (C). Representative fibrillation potentials measured from cranial tibial muscle of 12.4-year-old Labrador retriever clinically diagnosed with LPN (D). LPN = late-onset peripheral neuropathy; EMG = electromyography; mNCS = motor nerve conduction study; CMAP = compound muscle action potential.

Figure 2:

CMAP waveform analysis between groups. The LPN-Affected group had significantly decreased CMAP amplitude compared to the Control group in the sciatic nerve mid and distal sites (A) and in the ulnar nerve distal site (B). There were no significant differences in CMAP duration between groups in the sciatic nerve (C) or the ulnar nerve (D). The LPN-Affected group had a significant decrease compared to the Control group in the sciatic nerve at mid and distal sites (E) and the distal site of the ulnar nerve (F). Mean +/− SEM. * P<0.05, **P<0.01, ***P<0.001. LPN = late onset peripheral neuropathy; CMAP = compound muscle action potential. Sample sizes for each measurement are reported in Table S3.

Sciatic nerve CMAP duration was not significantly different between the LPN-affected group and the control group at any stimulation site (Table 2, Figure 2C). Ulnar nerve CMAP duration was not significantly different between groups at either stimulation site (Table 2, Figure 2D).

The sciatic nerve CMAP area was significantly decreased in the LPN-affected group compared to the control group when the distal/mid sciatic nerve and distal ulnar nerve were stimulated, but the apparent differences between groups for proximal stimulation of either the sciatic nerve or ulnar nerve did not reach significance (Table 2, Figure 2E/F).

Canine LPN has been associated with decreased nerve conduction velocity [6–7], but no difference in sciatic nerve MNCV was observed in affected Labrador retrievers (Table 3, Figure 3A). Within the LPN-affected group, the mid-distal MNCV was significantly decreased compared to the proximal-mid MNCV (P=0.03) (Figure 3A). There was no difference between the mid-distal MNCV and proximal-mid MNCV segments in the control group (P=0.16). There was no significant difference between groups for ulnar nerve MNCV (Table 3, Figure 3B).

Table 3:

Motor Nerve Conduction Velocity

Nerve	Segment	LPN Affected Group MNCV (m/s)	Control Group MNCV (m/s)	P value
Sciatic	Proximal-Mid	49.50 ± 6.78 (n=9)	53.50 ± 13.59 (n=8)	P=0.70
	Mid-Distal	38.35 ± 6.52 (n=9)*	45.50 ± 6.00 (n=8)	P=0.23
	Proximal–Distal	43.44 ± 5.55 (n=9)	49.13 ± 7.99 (n=8)	P=0.40
Ulnar	Proximal–Distal	40.86 ± 4.85 (n=7)	48.29 ± 15.29 (n=7)	P=0.26

^*P>0.05 versus the relevant proximal-mid segment.

LPN – late-onset peripheral neuropathy; MNCV – motor nerve conduction velocity

Figure 3:

MNCV analysis between groups. No differences were observed between groups for sciatic (A) or ulnar (B) MNCV. Sciatic nerve MNCV in the LPN-Affected group was slower in the proximal-mid segment than in the mid-distal segment (A). Mean +/− SEM. * P<0.05. LPN – late onset peripheral neuropathy; MNCV = motor nerve conduction velocity. Sample sizes for each measurement are reported in Table S3.

Sciatic nerve latency was significantly prolonged in the LPN-affected group at the proximal, mid and distal levels compared to the control group (Table 4, Figure 4A). There was no difference in sciatic nerve residual latency between the LPN-affected (1.52 ± 0.39) and control (1.21 ± 0.30) groups (P=0.10) (Figure 4C). Ulnar nerve latency was not significantly different between groups (Table 4, Figure 4B). There was no difference in ulnar nerve residual latency between the LPN-affected (1.51 ± 0.22) and control (1.30 ± 0.64) groups (P=0.43) (Figure 4D).

Table 4:

Latency values from sciatic and ulnar nerves

Nerve	Location	LPN Affected Group Latency (ms)	Control Group Latency (ms)	P value
Sciatic	Hip	11.51 ± 1.35 (n=9)	9.66 ± 1.25 (n=8)	P=0.01
	Stifle	7.08 ± 0.99 (n=9)	5.79 ± 0.49 (n=8)	P=0.005
	Hock	2.61 ± 0.46 (n=9)	2.06 ± 0.28 (n=8)	P=0.01
Ulnar	Elbow	7.05 ± 0.73 (n=7)	6.37 ± 0.98 (n=7)	P=0.17
	Carpus	2.84 ± 0.51(n=9)	2.57 ± 0.43 (n=8)	P=0.25

Figure 4:

Latency analysis between groups. The LPN-Affected group had significantly increased distal latency at all levels of the sciatic nerve compared to the Control group (A); no difference in distal latency was seen in the ulnar nerve (B). No difference in residual latency was observed between groups for either the sciatic (C) or ulnar (D) nerves.

A routine EMG study was completed for all dogs. The LPN-affected dogs had a higher frequency of abnormalities in distal musculature, such as the gastrocnemius and cranial tibial. The EMG findings are summarized in supplementary Table S2.

Neuropathology of LPN-affected Labradors

For dogs from which tissues were collected for neuropathologic evaluation, all dogs in the LPN-affected group had evidence of peripheral neuropathy at least 6 months prior to euthanasia, with signs including increased upper respiratory noise indicative of laryngeal paralysis (100%), paraparesis (73%) or tetraparesis (18%), absent pelvic limb paw replacement tests (100%), and reduced pelvic limb withdrawal reflexes (72%). In the aged control group, abnormalities on neurologic examination included paraparesis (75%) and absent pelvic limb paw replacement (75%), with one dog also having absent thoracic limb paw replacement (12%) and another dog having mild reduction in pelvic limb withdrawal reflex; all other aspects of the neurologic examination were unremarkable in this group. For the young control group, neurologic assessment was deemed normal based on observation; due to animal disposition, a full neurologic examination was not possible (Table S1).

For the neuropathology studies, postmortem tissues were collected from a total of 25 dogs. Group assignment, studies performed, age and pertinent neurologic exam findings are summarized in Table S1. For peripheral nerve evaluation, tissues from 18 dogs were used. The LPN-affected group (n=8) included tissues from purebred Labrador retrievers with a median age of 13.2 years (range: 10.3–14.5). The aged control group (n=5) included 4 pure-bred Labradors and 1 Basset Hound, with a median age of 15.7 years (range 13.9–15.9). The young control group (n=5) included 4 Pitbull crosses and 1 Rottweiler, with an estimated age range of 2 years based on physical and dental examination. It should be noted that the aged control group was significantly older than the LPN-affected group (P=0.003).

Qualitative assessment between groups showed evidence of a generalized decrease in axon density and a loss of large diameter myelinated axons in the LPN-affected group and aged controls, most evidently at the distal most aspects of nerves evaluated (Figure 5A–H), compared to the young control group (Figure 5I–L). There is also evidence of axon degeneration in both groups. Quantitative analysis revealed a significant decrease in median diameters in the LPN-affected vs young control groups at all levels of the sciatic (proximal, mid and distal) and recurrent laryngeal nerves (P=0.01, P=0.03, P=0.005, P=0.003, respectively) (Figure 6A). The significant decrease in median myelinated axon diameters in the LPN-affected vs young control groups also appeared to be present in the aged control group, but the apparent differences only reached significance at the distal sciatic nerve (P=0.04) (Figure 6A). No difference was seen between groups at any nerve or level for median g-ratio (Figure 6B).

Figure 5:

Peripheral nerve histology. Overall, a generalized loss of axons in the LPN-affected group can be appreciated along the length of the sciatic nerve (A-C), which is also evident in the aged controls (E-G) when compared to a healthy young control dog (I-K). In the recurrent laryngeal nerve, both the LPN-affected group (D) and the aged control group (H) had evidence of loss of large-diameter axons with an increased number of small-diameter axons when compared to the young control (L); here, as in the distal sciatic nerve, the LPN-affected group had a much more severe loss of overall axons compared to the aged control group. LPN = late-onset peripheral neuropathy. Scale bars = 10 μm.

Figure 6:

Median axon diameters and g-ratios from 3 levels of the sciatic nerve and the recurrent laryngeal nerve (RLN). The LPN-Affected group had significantly decreased median axon diameters at all levels of the sciatic nerve and the RLN compared to the Young Control group; notably, this difference was most significant at the most distal aspect of the sciatic nerve (A). The aged control group also had a significantly decreased median axon diameter compared to the young control group at the distal sciatic nerve segment (A). There were no differences between groups for median g-ratios (B). Mean +/− SEM. * P<0.05, **P<0.01. LPN = late onset peripheral neuropathy. LPN Affected group, n=8; Aged Control, n=5 for all levels and groups except the proximal sciatic, where n=4; Young Control, n=5.

Evaluation of axon diameter distributions between groups showed consistently greater differences between the LPN-affected group and the young control group. In the proximal sciatic nerve, the LPN-affected group had a significantly greater percentage of axons <2 μm in diameter (P=0.01) and significantly decreased proportions of larger diameter (>2 μm) axons (P=0.008, P=0.04, respectively) compared to the young control group (Figure 7A). There were similar results in the mid and distal sciatic nerves, with the aged control group following a similar trend as the LPN-affected group (Figure 7B/C). In the recurrent laryngeal nerve, both the LPN-affected and aged control groups had a significantly greater proportion of axons <2 μm in diameter (P<0.0001, P<0.0001, respectively) but a significantly decreased percentage of larger diameter axons (2–4 μm: P=0.0006, P=0.01, axons 4–6 μm: P=0.002, P=0.005, respectively) compared to the young control group; for axons >6 μm in diameter, only the LPN-affected group had a significantly decreased percentage of axons (P=0.01) compared to the young control group (Figure 7D).

Figure 7:

Quantification of axon diameter distribution and g-ratio. The mean +/− SEM for each group is presented for percent total axons across axon diameters (A-D) and g-ratio (E-H). Axon diameters are binned into 2 μm groups for all graphs. Comparisons of axon diameter show that the LPN-Affected group had a significantly greater proportion of smaller diameter axons than the Young Control group than the Aged Control group, despite the LPN-Affected group being significantly younger than the Aged Control group (P=0.003). Additionally, the LPN-Affected group had a consistently lower percentage of axons over 2 μm in diameter (A-D). In contrast, few differences in the g-ratio were observed between groups (E-H). LPN = late-onset peripheral neuropathy.

While overall myelin thickness (as measured by the g-ratio) was unchanged between groups (Figure 6B), the proximal sciatic level of the LPN-affected group had mild hypermyelination of axons <2 μm in diameter (P=0.04) and a significantly increased g-ratio for axons >6 μm in diameter (P=0.03), while the aged control group had a significantly increased g-ratio for axons 4–6 μm in diameter (P=0.006) and for axons >6 μm in diameter (P=0.008) (Figure 7E). No differences in the g-ratio were observed between groups at the mid-sciatic or distal sciatic levels (Figures 6F, 6G). Overall, there were no significant changes in g-ratio distributions observed in comparing LPN-affected to aged controls.

TEM of two LPN-affected dog sciatic nerves confirmed the light microscopic observations that there was a major loss of myelinated axons and profound endoneurial fibrosis (Figure 8). Major loss of myelinated axons was not observed in the two aged control dog sciatic nerves analyzed. Importantly, scattered denervated Schwann cells (Bands of Büngner) were seen throughout the LPN-affected nerves, although they were not present in large numbers, suggesting their loss with time (Figure 8A–D). Bands of Büngner were not observed in aged control dog nerves.

Figure 8:

Four examples of denervated Schwann cell (Büngner) bands from the sciatic nerve of an LPN-affected dog. In each case, the cell process is surrounded by a basal lamina. The numerous Büngner bands confirm the extensive loss of axons. Excess collagen is seen in A-C, and the processes are extending around ‘pockets’ of collagen. Scale bar = 400 nm for A-C; Scale bar = 1 μm for D. LPN = late-onset peripheral neuropathy.

For spinal cord analysis, postmortem tissues from 17 dogs were used. The LPN-affected group (n=7) included pure-bred Labrador retrievers with a median age of 12.3 years (range: 10.3–14.1). The aged control group (n=5) included pure-bred Labradors with a median age of 14.0 years (range: 12.8–15.9). The young control group (n=5) included 4 Pitbull crosses and 1 Rottweiler, with an estimated age of 2 years based on physical and dental examination. Both the LPN-affected group and aged control group were significantly older than the young control group (P=0.01, P=0.02, respectively). There was no significant difference between the age of the LPN-affected group and the aged control groups (P=0.12). Spinal cord motor neuron quantification did not show any significant differences in motor neurons between groups for either Nissl staining (P=0.36) or HB9 (P=0.32) (Figure 9).

Figure 9:

Spinal cord histology with HB9 and Nissl staining. The ventral horns from L5-L6 spinal cord sections from dogs in the LPN affected (A, D), aged control (B, E) and young control groups (C, F) are shown. Analysis indicated that quantification of motor neurons was not different between groups for either the HB9 stain (G) or the Nissl stain (H). LPN = late-onset peripheral neuropathy.

Discussion:

Our results indicate that LPN in the Labrador retriever is a peripheral neuropathy that can be characterized as a length-dependent axonopathy. Our study agrees with prior qualitative studies that have indicated that LPN is a polyneuropathy that has features consistent with an axonopathy [6–7]. Our findings are also consistent with a distal neuropathy, as we analyzed the spinal cord and determined that motor neuron quantity did not differ between LPN-affected dogs and control dogs. In combination with the histologic and electrodiagnostic alterations identified, our results support the prior clinical assumption that LPN is a length-dependent peripheral axonopathy.

Some previous reports of LPN in Labradors and other breeds showed evidence of reduced nerve conduction velocity, which is normally associated with altered myelination [6–7]. However, in this Labrador retriever-focused study, we did not find alterations in MNCV between the LPN-affected and control groups. In contrast, CMAP amplitude and CMAP area were both significantly decreased in the LPN-affected group in the sciatic and ulnar nerves when compared to the control group. Moreover, histopathologic evaluation and g-ratio quantitation indicated that myelination was not substantially affected in the LPN-affected group compared to either control group. There was a significant decrease in the percentage of large-diameter myelinated axons, particularly in distal nerve portions, in the LPN-affected group compared to the young control group. This decrease was not evident when the LPN-affected group was compared to the aged control group. When analyzed with TEM, a prominent loss of myelinated axons and presence of denervated Schwann cells was observed in the LPN-affected group but not the aged controls. Since the LPN-affected group was, on average, 2.5 years younger than the aged control group, these results are in support of premature large diameter myelinated axon degeneration in the LPN-affected group. Inclusion of an age-matched control group is needed to make conclusions about this potential trend. It is also likely that alterations associated with axonal degeneration are influenced by molecular alterations that could escape detection via morphologic observation. Further exploration of the genetic makeup of LPN, coupled with subsequent in vitro investigations, could yield additional insights not discerned in this study. Overall, the results indicate that LPN is an axonopathy due to the pronounced distal loss of axons.

LPN in Labrador retrievers is a potential model for human conditions. The most common collection of autosomal inherited peripheral neuropathies in humans are forms of CMT, of which CMT2 is typically an axonopathy. These conditions typically present as a mixed motor and sensory neuropathy, and there are a large proportion of human cases of axonal CMT in which a genetic cause has yet to be identified [41]. This is particularly the case for late-onset neuropathy, where only some genetic causes have been identified, such as mutations in the metalloprotease neprilysin (MME) gene [42], and in general, these are sometimes classified as idiopathic neuropathy. DHMN is also a set of inherited peripheral neuropathies but is characterized by loss of motor nerve function [43].

There is currently no established large animal model for inherited peripheral neuropathy, although a canine model of giant axonal neuropathy has been established [44]. While rodent models have helped advance the understanding of disease, there are several types of CMT that do not have good rodent models [45]. Therefore, LPN in Labradors may be a very useful model of late-onset idiopathic neuropathies given that affected Labrador retrievers are very common. For instance, more pathways of axon degeneration have been identified recently, and pharmacological approaches to treat axon degeneration are becoming available. LPN has strong potential as an axonopathy model not only from a clinical treatment perspective but also as a means of identifying novel gene pathways associated with axonopathy.

From a clinical standpoint, LPN presents with signs of peripheral motor dysfunction. The most concerning clinical feature of LPN is dysfunction of the recurrent laryngeal nerve, leading to paralysis of laryngeal musculature and risk for life-threatening asphyxiation. In carnivora, due to a longer cervical length compared to primates, the recurrent laryngeal nerve is comparatively a substantially longer nerve [12], and thus at higher risk for disease in length-dependent axonopathies. Interestingly, from a clinical perspective, the left larynx is typically paralyzed first, followed within months of the right side becoming paralyzed. This supports a length-dependent mechanism, as the left RLN is longer than the right and is also a feature noted in horses affected with laryngeal paralysis [46]. The other most clinically recognized feature of LPN in Labrador retrievers is hindlimb weakness, which is most evident as a decrease in hock flexion during gait analysis, also supporting a length-dependent disease.

There are several limitations to this study. Dogs were diagnosed with LPN via clinical history, physical examination, and neurologic examination. Ideally, all dogs would have a complete diagnostic work-up, including magnetic resonance imaging (MRI), to rule out any degenerative disc disease. However, MRI imaging in dogs requires general anesthesia, which was deemed too risky for the LPN patient population. The control groups for these studies had variable heterogeneity of the populations with respect to breed, although careful size matching was undertaken. Currently, there is no evidence that alterations in peripheral nerve aging exist in the breeds of size-matched dogs, but further work is needed to evaluate this potential. The TEM conducted in this study was strictly qualitative. The use of TEM for quantitative analysis of peripheral nerve sections would be beneficial in future studies. The neuropathology studies were dependent on a body donation program, which was limited to only including medium- to large-breed dogs. However, age matching was not feasible due to the LPN-affected Labrador retrievers having a decreased survival time compared to unaffected aged control dogs [5]. Consequently, our aged control group was significantly older than our LPN-affected group for the peripheral nerve histology study. Ultimately, identification of the genetic underpinning(s) of LPN in the Labrador retriever would be necessary to create a breed-, age- and size-matched control group. Such discovery work would also enable further understanding of the genetic underpinnings that result in LPN’s pathologic profile.

Conclusions

In conclusion, this study provides evidence that LPN in the Labrador retriever breed is the result of a length-dependent axonopathy. Overall, LPN in the Labrador retriever is an easily accessible spontaneous large animal model for the study of peripheral neuropathy across species and thus has substantial value as a one-health model of disease.

Key Points:

1) Late-onset peripheral neuropathy in Labrador retrievers length-dependent peripheral axonopathy, 2) Late-onset peripheral neuropathy in Labrador retrievers is a common, spontaneous, novel animal model of age-related peripheral neuropathy that could be used for fundamental research and clinical trials.

List of abbreviations:

LPN

Late-onset peripheral neuropathy

EMG

Electromyography

CMAP

Compound muscle action potential

MNCV

motor nerve conduction velocity

CMT

Charcot-Marie-Tooth

DHMN

Distal hereditary motor neuropathy

NCS

Nerve conduction study

IV

Intravenous

PSWs

positive sharp waves

MME

metalloprotease neprilysin

MRI

Magnetic resonance imaging

TEM

Transmission electron microscopy

Availability of data and materials:

The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.