Dejerine–Sottas disease in childhood—Genetic and sonographic heterogeneity

Abstract Introduction The nerve sonographic features of Dejerine‐Sottas disease (DSD) have not previously been described. Methods This exploratory cross‐sectional, matched, case–control study investigated differences in nerve cross‐sectional area (CSA) in children with DSD compared to healthy controls and children with Charcot–Marie–Tooth disease type 1A (CMT1A). CSA of the median, ulnar, tibial, and sural nerves was measured by peripheral nerve ultrasound. The mean difference in CSA between children with DSD, controls, and CMT1A was determined individually and within each group. Results Five children with DSD and five age‐ and sex‐matched controls were enrolled. Data from five age‐matched children with CMT1A was also included. Group comparison showed no mean difference in nerve CSA between children with DSD and controls. Individual analysis of each DSD patient with their matched control indicated an increase in nerve CSA in three of the five children. The largest increase was observed in a child with a heterozygous PMP22 point mutation (nerve CSA fivefold larger than a control and twofold larger than a child with CMT1A). Nerve CSA was moderately increased in two children—one with a heterozygous mutation in MPZ and the other of unknown genetic etiology. Conclusions Changes in nerve CSA on ultrasonography in children with DSD differ according to the underlying genetic etiology, confirming the variation in underlying pathobiologic processes and downstream morphological abnormalities of DSD subtypes. Nerve ultrasound may assist in the clinical phenotyping of DSD and act as an adjunct to known distinctive clinical and neurophysiologic findings of DSD subtypes. Larger studies in DSD cohorts are required to confirm these findings.

The term was initially used to describe a clinical phenotype characterized by symptom onset in the first two years of life, delayed motor development, hypotonia, and extremely slow nerve conduction (median motor nerve conduction velocity, MNCV, of 12 m/s or less; Gabreels-Festen, 2002;Ouvrier, McLeod, & Conchin, 1987;Yiu & Ryan, 2012). Other common clinical features include areflexia, muscle wasting and weakness, foot deformity, and sometimes enlarged nerves on clinical examination (Ouvrier et al., 1987). The terminology used for the early-onset demyelinating neuropathies can be unclear.
High-resolution peripheral nerve ultrasound allows rapid, noninvasive imaging of peripheral nerves (Goedee et al., 2013). Few dedicated pediatric nerve ultrasound studies have been published. Yiu et al. (2015) found a two-to threefold increase in nerve cross-sectional area (CSA) in children with CMT1A compared to controls. Nerve ultrasound may play a role in the diagnosis of pediatric inherited neuropathies, and in the era of next-generation sequencing, knowledge of specific nerve sonographic features, in addition to clinical and neurophysiologic findings, may assist in the interpretation of genetic testing results.
We evaluated nerve CSA in children with DSD in comparison both with age-and sex-matched healthy controls and age-matched children with CMT1A.

| Study objectives
The primary objective of this study was to investigate differences in nerve CSA in children with DSD as measured by peripheral nerve ultrasound compared to healthy controls. Given the more severe clinical course and neurophysiologic abnormalities in DSD compared to CMT1A, we hypothesized that nerve CSA would be significantly greater in children with DSD compared to their healthy controls, and greater than matched children with CMT1A. The secondary objective was to define the clinical features of DSD in this cohort.

| Study design and recruitment
This cross-sectional, matched, case-control study was conducted at The Royal Children's Hospital, Melbourne (RCH). Children with DSD were recruited from the neuromuscular clinic. Inclusion criteria for children with DSD included (1) symptom onset in the first 2 years of life; (2) delayed motor development; and (3) median nerve MNCV ≤ 12 m/s. Healthy controls were enrolled at a ratio of 1:1 and were age-matched (within one year either side) and sex-matched.
They were recruited from friends and family members of RCH staff and were screened for symptoms and signs of neuromuscular disorders. Data from matched children with CMT1A from a previous study of nerve ultrasound (Yiu et al., 2015) were also used for comparison.
For children with CMT1A, age match (within 1 year) was chosen in preference to sex match.

| Nerve ultrasound
Nerve ultrasound images were obtained in the dominant upper and lower limb at the following sites: (1) median nerve at the midhumerus (two-thirds from the lateral tip of the acromion to the lateral epicondyle of the humerus), at the elbow (in the cubital fossa), at the forearm (lower third), and at the wrist (at the level of the lunate); (2) ulnar nerve at the midhumerus, just distal to the elbow (one-fifth forearm distance distal to the medial epicondyle), and at the midforearm; (3) tibial nerve at the ankle; and (4) sural nerve at the ankle.
All nerves were transversely imaged with the probe perpendicular to the nerve, providing the most accurate and smallest CSA.
Nerve CSA was measured by tracing the nerve just inside the hyperechoic rim. Three images were taken and averaged at each nerve site.

| Clinical assessment
During the same study visit as the ultrasound assessment, weight and height were recorded and body mass index (BMI) was calculated.
Children with DSD also underwent a neurologic assessment, including history, physical examination, and assessment using the Charcot-Marie-Tooth Pediatric scale (CMTPedS), a clinical rating tool designed to measure disease severity in children with CMT (Burns et al., 2012).
This 11-item rating tool provides a total score between 0 and 44, with a higher score reflecting greater impairment. Individual results were also plotted graphically. Sample size was convenience-based.

| Standard protocol approvals, registrations, and patient consents
This study was approved by the RCH Human Research and Ethics Committee (HREC number 34242). Informed consent was obtained for all participants.

| RESULTS
Five children with DSD were enrolled-one boy and four girls, with a mean age of 7.8 (SD 3.9) years. Key clinical and genetic features are shown in Table 1. Case 5 has been previously reported (Yiu et al., 2017). Five age-and sex-matched controls were recruited with a mean T A B L E 1 Clinical and genetic findings in five children with Dejerine-Sottas disease age of 8.2 (SD 4.0) years. The greatest difference in age between each child with DSD and their matched control was 9.8 months (case 5control 9.8 months older than DSD case). There was no significant difference in height, weight, or body mass index (BMI) between children with DSD and controls.
Clinical features of the five children with DSD are summarized in Table 1. The mean age at onset was 6.6 (SD 5.4) months. All had delayed motor development. Four children presented with delayed motor milestones, hypotonia, and/or hip dysplasia. The fifth child (case 2) presented at birth with hypotonia, hyporeflexia, aspiration, and feeding problems and also had motor delay in the first two years of life. At presentation, three had hyporeflexia or areflexia.
At the time of the study, all children had gait abnormalities. Four were unable to walk or stand independently, wore ankle-foot orthoses (AFO's), and had foot deformity and weakness of the hand muscles.
The fifth child was independently ambulant but had frequent trips and falls. Orthopedic complications were common-four children had scoliosis and three bilateral hip dysplasia. Three children had respiratory insufficiency, with reduced vital capacity on spirometry and/or recurrent respiratory infections requiring hospital admissions. One child received nocturnal noninvasive ventilation (case 1). Three children had bulbar dysfunction.
All children had distal lower limb muscle wasting-two mild, two moderate, and one with severe wasting extending above the knee. Wasting of the thenar, hypothenar, and intrinsic hand muscles was moderate in one child and severe in two. Weakness was most marked distally; however, most children also had proximal muscle weakness of both upper and lower limbs. The CMTPedS scores reflect the overall severity of the neuropathy, with a mean of 31.0 (SD 11.8).

| Neurophysiologic, neuropathologic, and genetic findings
Nerve conduction studies were performed in all children before four years of age. Median nerve MNCV was 12 m/s or less in all cases ( T A B L E 2 Nerve cross-sectional area in five children with Dejerine-Sottas disease compared to controls chip. We assume there may be another, at present, unidentified gene mutation contributing to her DSD phenotype. Coverage of the coding regions of PRX (case 3) and PMP22 (case 1) was effectively complete.

| Comparison of nerve CSA between children with DSD, healthy controls, and children with CMT1A
Whilst nerve CSA was greater in children with DSD, there was no significant difference in nerve CSA between children with DSD and healthy controls at any site when analyzed using paired t tests (Table 2). Due to the genetic heterogeneity and small size of the cohort, nerve CSA was also individually compared between each child with DSD and their control. The difference in nerve size between children with DSD and controls differed markedly between participants. Nerve CSA was similar to matched controls in cases 1 and 2. Nerve CSA was markedly increased in cases 4 (PMP22 point mutation) and 5 (MPZ point mutation) compared to control values in all nerves studied. Case 3 showed variable increase in nerve CSA between sites, with enlargement seen mainly in the median and tibial nerves compared to the matched control ( Figure 1 and Table 3).
An additional comparison of nerve CSA in DSD was made with age-matched children with CMT1A from a previous study (Yiu et al., 2015). There was less than 7 months difference in age between each child with DSD and their matched CMT1A subject. Three children were not sex-matched. There was no significant difference between height or weight between children with DSD and CMT1A (data not shown). Figure 1 and Table 3  control. Nerve size in cases 3 and 5 was enlarged to a similar degree to that seen in CMT1A.

| DISCUSSION
This single-center study confirms the genetic heterogeneity of the early-onset demyelinating neuropathies of childhood, is the first to describe the high-resolution nerve ultrasound features of this cohort of children, and demonstrates a marked variability in the degree of nerve enlargement in DSD. The latter likely reflects the genetic heterogeneity of DSD and subsequent downstream pathobiologic processes.
We previously showed a marked (two-to threefold) and widespread increase in nerve CSA in a genetically homogeneous cohort of children with CMT1A (Yiu et al., 2015).  (Abe et al., 2010;Al-Thihli et al., 2008). Given the lack of nerve enlargement in this case compared to the child with a PMP22 point mutation, one could hypothesize that this child either has a mutation in another gene or, if she does have a second, as yet undetected mutation in PMP22, that this reflects mutation-specific effects of different PMP22 mutations on nerve enlargement.
The increase in nerve CSA seen on nerve ultrasound parallels the increase in total transverse fascicular area (TTFA) seen in neuropathologic studies of CMT. In CMT1A, TTFA is 1.5-to twofold that of healthy controls (Gabreels-Festen, 2002), reflecting the increase in nerve CSA seen by nerve ultrasound (Yiu et al., 2015). Similarly, Gabreels-Festen documented a TTFA/normal ratio of 393%-458% in children with PMP22 missense mutations and a 92%-276% in individuals with MPZ missense mutations; Gabreels-Festen (2002) findings in keeping with the changes seen in nerve CSA in the present study. The marked increase in TTFA is thought to be secondary to an increase in collagen fibers and endoneurial extracellular matrix (Gabreels-Festen, 2002;Yiu et al., 2015). Large numbers of onion bulbs may also contribute to increased fascicular size in DSD. with specific genetic subtypes. If we postulate that the main contributor to nerve enlargement is increased volume of endoneurial extracellular matrix and perhaps onion bulb formations, the variable nerve ultrasound findings in this study may reflect the effect of different gene mutations on extracellular matrix production and onion bulb formation, either directly or via abnormal Schwann cellextracellular matrix interplay (Colognato & Tzvetanova, 2011;Yiu et al., 2015).
In CMT1A, nerve CSA shows a modest correlation with disability (Yiu et al., 2015). Although limited by the small sample size, the degree of nerve enlargement did not appear to correlate with clinical severity in this study, as measured by the CMTPedS total score. Four children had marked neurologic disability; three of the four children old enough to perform the CMTPedS had scores greater than 30, indicating severe impairment.
This study has a number of limitations. Due to low prevalence of DSD, the sample size was small, limiting statistical analysis. The ultrasonographer was not blinded to the clinical status of the patient.
The workup of early-onset neuropathies involves clinical history and examination, neurophysiologic studies, and genetic testing. In the era of next-generation sequencing, specific nerve sonographic features may act as an adjunct to clinical and neurophysiologic phe- Center, Nijmegen, the Netherlands; and all participants of this study.

CONFLICT OF INTEREST
None declared.