Dejerine–Sottas syndrome grown to maturity: overview of genetic and morphological heterogeneity and follow-up of 25 patients

Abstract

Dejerine–Sottas syndrome (DSS) is an early onset demyelinating motor and sensory neuropathy with motor nerve conduction velocities below 12 m s⁻¹. The phenotype is genetically heterogeneous, and autosomal dominant (AD) as well as autosomal recessive (AR) inheritance is described. Nerve pathology is highly variable. It is generally presumed that clinical course is severe, leading to wheelchair dependency at an early age. In this study we documented the clinical and pathological features in 25 patients with a DSS and we evaluated the clinical course. In our series 14 patients had an AD mutation and six were probably affected by an AR disorder. In three patients inheritance mode was unknown and two patients obviously suffered from an acquired disorder. The clinical course in all patients was documented. Nine of the 25 patients showed a moderate handicap in adult life; walking distance was still at least 1 km. Age at last investigation of the ambulant patients ranged from 22 to 62 years (mean 38.6 years), and ambulant patients were found in all genetic subgroups. We conclude that DSS, although in general denoting a more serious neuropathy than CMT1, does not imply a severe disability or wheelchair dependency in adult life.

Introduction

The early concept of Dejerine–Sottas disease/HMSN type III

In 1886, Charcot and Marie in Paris, and Tooth in London, described a familial syndrome of progressive peroneal muscular atrophy (Charcot & Marie, 1886; Tooth, 1886). The concept of a separate childhood entity followed the publication of a paper ‘Sur la névrite interstitielle, hypertrophique et progressive de l’enfance’ (Dejerine & Sottas, 1893). These authors described hypertrophic nerves in a brother and sister with a severe progressive sensorimotor neuropathy, born from apparently unaffected parents. In the girl symptomatic onset was in infancy; in the boy symptoms were recorded for the first time at age 14 years during a hospital stay for treatment of a kyphoscoliosis. Several authors concluded therefore that in the boy the disorder had started in adolescence. However, as his mother had already noticed spinal deformities at the age of 4–5 years, it is most likely that disease onset in the boy was also at an early age (Dejerine & Sottas, 1893). In the following decades, numerous reports described patients with peroneal muscular atrophy, who showed different modes of inheritance and a substantial variation in clinical features.

In the late 1950s, a new area was exposed with the introduction of nerve conduction studies and the investigation of nerve biopsies. Dyck (1975) correlated conduction changes with nerve biopsy findings and presented an innovative classification of the peroneal muscular atrophy syndrome, since usually termed hereditary motor and sensory neuropathies (HMSNs) after the non-specific term introduced by Thomas et al. (1974). Dyck divided the milder HMSN variant Charcot–Marie–Tooth disease (CMT) into a demyelinating form with marked nerve conduction slowing (HMSN type I/CMT1) and a neuro-axonal form with normal or only modest slowing of nerve conduction (HMSN type II/CMT2). Both HMSN I and II are usually autosomal dominantly (AD) inherited, but autosomal recessive (AR) transmission was well documented by Harding & Thomas (1980a).

Dyck designated a third form: HMSN type III or Dejerine–Sottas (DS) disease defined as a severe demyelinating neuropathy, presenting in infancy with delayed motor development, very slow nerve conduction velocities (< 10–12 m s⁻¹) and usually elevated CSF protein. Progression was severe and walking was lost early. Inheritance was thought to be AR (Dyck, 1975). Hypomyelination of the peripheral nerves was seen as a pathological hallmark (Dyck et al. 1971; Ouvrier et al. 1987). Onion bulbs of concentric thin Schwann cell lamellae (‘classic’ onion bulbs) and myelin breakdown products were considered evidence of an ongoing demyelinating process.

However, the concept of HMSN type III or DS disease provoked much controversy, and several authors expressed doubt as to whether it existed as a separate entity (Brust et al. 1978; Westerberg, 1982). In Westerberg’s opinion, the HMSN type III concept was merely based on the selection of more severe cases of HMSN type I. Others have emphasized the morphological heterogeneity of HMSN type III (Gabreëls-Festen et al. 1994). Some authors suggested restricting HMSN type III/DS disease to cases of early onset demyelinating HMSN with MNCV < 6 m s⁻¹ (Benstead et al. 1990; Gabreëls-Festen et al. 1994) or to a distinct morphological variant, namely congenital hypomyelination (vide infra) (Harding & Thomas, 1980b).

Kennedy and colleagues had discerned a morphologically distinct variant: congenital hypomyelination neuropathy (CHN), a severe neuropathy with a congenital or early infantile onset and in peripheral nerve biopsy no or extremely thin myelin sheaths and ‘onion bulbs’ of mainly basal membranes (Kennedy et al. 1977). An earlier description of congenital hypomyelination was given by Lyon (1969). Also, one of the index cases of the HMSN type III concept of Dyck probably falls in this category (Dyck & Gomez, 1968). Several cases with a similar pathology have been described subsequently (Anderson et al. 1973; Joosten et al. 1974; Kasman et al. 1976; Koto et al. 1978; Moss et al. 1979; Ulrich et al. 1981; Guzzetta et al. 1982; Ono et al. 1982; Pagès et al. 1983; Tachi et al. 1984; Harati & Butler, 1985; Lütschg et al. 1985; Vital et al. 1987; Ouvrier et al. 1990; Routon et al. 1991; Boylan et al. 1992; Gabreëls-Festen et al. 1994; Nara et al. 1995; Sawaishi et al. 1995; Bornemann et al. 1996; Tyson et al. 1997; Phillips et al. 1999; Pomerance et al. 2000). Most cases occurred sporadically; in a number of cases clear evidence of AR inheritance was present.

Kennedy and colleagues considered CHN a non-progressive disorder of myelin formation, but in an earlier publication they did not exclude the possibility of a disturbance in myelin maintenance as ‘so little myelin was formed that these last findings (i.e. demyelination) cannot be expected’ (Kennedy et al. 1971). Some authors also considered CHN distinct from DS disease: CHN being a non-progressive disorder with a defect in myelin formation and DS disease a progressive demyelinating disorder (Harati & Butler, 1985; Nara et al. 1995; Phillips et al. 1999). However, in several cases of CHN signs of myelin breakdown have been observed (Lyon, 1969; Joosten et al. 1974; Moss et al. 1979; Ulrich et al. 1981; Lütschg et al. 1985; Vital et al. 1987). Others considered congenital hypomyelination/amyelination to be the severest end of the spectrum of HMSN III (Guzzetta et al. 1982).

Five reports have described patients with a near total amyelination of the peripheral nerves, but with normal myelination in the central nervous system (Karch & Urich, 1975; Palix & Coignet, 1978; Hakamada et al. 1983; Seitz et al. 1986; Charnas et al. 1988). Patients were floppy at birth, often with arthrogryposis, and showed respiratory distress and swallowing difficulties, leading to death in days or months. This congenital amyelination is probably the most severe expression of CHN (Gabreëls-Festen & Gabreëls, 1993).

In 1977 Joosten and colleagues reported another morphological phenotype of DS disease, distinct from the types with classic onion bulbs and congenital hypomyelination/amyelination in showing many focally folded myelin sheaths (FFM) (Joosten et al. 1977). In later years several authors described morphologically identical cases, although many patients showed a later disease onset and less severe slowing of nerve conduction. AD as well as AR cases were reported (Dayan et al. 1968; Lütschg et al. 1985; Vallat et al. 1987; Vital et al. 1987; Ohnishi et al. 1989; Gabreëls-Festen et al. 1990; Malandrini et al. 1992; Peudenier et al. 1993; Umehara et al. 1993; Barbieri et al. 1994; Sabatelli et al. 1994; Schenone et al. 1994; James et al. 1995; Tyson et al. 1997). Ben Othmane et al. (1993) designated the AR variants of this morphological subtype as CMT4B. The early onset forms of the variant with FFM and MNCV < 12 m s⁻¹ may be considered as a morphological subtype of the DS syndrome (DSS).

Recent developments

The remarkable advances in the elucidation of the molecular defects of HMSN during the last decade of the 20th century have changed many traditional views and concepts on this subject. Recent genetic investigations showed that several of the earlier published cases of DSS resulted from (de novo) heterogeneous dominant point mutations of the Peripheral Myelin Protein 22 gene (PMP22) (Roa et al. 1993; Gabreëls-Festen et al. 1995; Valentijn et al. 1995), or the Protein zero gene (P0) (Hayasaka et al. 1993; Nakagawa et al. 1999). More recently it was demonstrated that also AD mutations of the Early Growth Response protein 2 gene (EGR2) might result in DSS (Warner et al. 1998). Furthermore, an AR inherited form of DSS may result from mutations in the PRX gene (Boerkoel et al. 2001). In addition, one mutation in the inhibitory domain of EGR2 and at least one PMP22 mutation in the C-terminal intracellular domain of the protein are inherited as an AR trait (Warner et al. 1998; Parman et al. 1999). These mutations are silent in the heterozygous parents, but cause a DSS or CHN in the homozygous children (Parman et al. 1999). Two specific mutations of P0 are leading to a DSS phenotype in the homozygous state, but result in a mild CMT1 or CMT2 phenotype in the heterozygous state (Ikegami et al. 1996; Pareyson et al. 1999). The phenotypic expression of the few reported cases of a homozygous PMP22 duplication may range from DSS to CMT1, like in the heterozygous state (LeGuern et al. 1997; Sturtz et al. 1997). This suggests that solely the number of PMP22 gene copies does not determine disease severity.

The other presently known AR inherited genetic defects that cause a demyelinating neuropathy generally express a CMT1 phenotype and not a DSS phenotype because onset is not in infancy and upper limb MNCVs are > 12 m s⁻¹. However, the clinical course may be severe, leading to early wheelchair dependency in several cases (Kalaydjieva et al. 1998; Houlden et al. 2001). Now it is evident that the phenotypic expression of different mutations in a myelin gene might range from a severe, early-onset demyelinating neuropathy (DSS or CHN) to a later-onset demyelinating CMT1 phenotype and even to an axonal neuropathy, clinically presenting as CMT2. As a result the traditional classification of HMSN, which was based on clinical features, nerve conduction studies, nerve pathology and mode of inheritance, is being supplemented and partially replaced by a genetic classification. However, for the practising clinician this has resulted in a classification system that is becoming more and more complex, confusing and inadequate. One might ask whether it has diagnostic value to retain the phenotypic designation Dejerine–Sottas syndrome. Is there, for example, a greater chance to deal with certain genetic defects in cases of DSS? Does DSS indicate a distinct group of neuropathies with a more severe clinical course, leading to early wheelchair dependency, as usually presumed? Are distinct genetic defects or pathological features indicative for a more progressive course?

To answer these questions we present an overview of the clinical and the pathological features of 25 patients with a DSS resulting from different genetic defects. We documented the clinical course in our patients and compared our data with literature reports.

Patients

We have documented a group of 25 patients with a DSS who had a sural nerve biopsy that was evaluated in our department between 1970 and 1995. Most patients were described previously (Gabreëls-Festen et al. 1990, 1992, 1993, 1994, 1995, 1996; Hoogendijk et al. 1993; Meijerink et al. 1996). Selection criteria were as follows: (1) chronic motor and sensory neuropathy; (2) infantile onset of (motor) symptoms manifested by congenital hypotonia, delayed walking (> 18 months) or early complaints about motor performances (< 2 years); and (3) median motor nerve conduction velocities (MNCV) of 12 m s⁻¹ or less. Information about the present clinical status of patients was obtained by recent examination or detailed inquiry by telephone. Stages of disability were graded as follows: 1 = walking moderately impaired; 2 = maximum walking distance over at least 1 km, minimal impairment hand function; 3 = walking with braces over limited distances, disturbed hand function; 4 = wheelchair use but indoor walking possible, claw hands; 5 = fully wheelchair dependent, only limited hand function left.

DNA was available from 21 patients. Testing for the 1.5 Mb duplication was performed as described previously (Hensels et al. 1993). After exclusion of the duplication, mutation screening of PMP22, P0, EGR2 and Myotubularin-related protein-2 gene (MTMR2) was performed in selected cases (Valentijn et al. 1992; Nelis et al. 1994; Warner et al. 1998; Bolino et al. 2000).

Sural nerve biopsy was performed in 24 patients. Nerves were prepared for light and electron microscopical investigation, including teased fibre studies, using previously described techniques (Gabreëls-Festen et al. 1992; Lenssen et al. 1998).

Results and discussion

Twenty-five patients with a DSS were subdivided into six groups depending on the genetic defect or, if no genetic defect could be established, on morphological characteristics (Table 1).

Table 1.

Twenty-five patients with DSS classified according to genetic defects or pathological features

Genetic/pathological subgroups of DSS
AD PMP22 point mutations	5 patients
AD 17p11.2 duplication	3 patients
AD P0 mutations	6 patients
AR congenital hypomyelination	3 patients
AR neuropathy with FFM	3 patients
Genetically unclassified DS syndrome	3 patients
CIDP	2 patients

In Tables 2–5 the main clinical, electrophysiological and morphological findings and the disability stage at last investigation are summarized. The age at first clinical and electrophysiological examination usually coincided with the age at biopsy.

Table 2.

Clinical and morphological features in DSS patients with PMP22 mutations: five patients with PMP22 missense mutations and three patients with the 17p11.2 duplication

Patient	1	2	3	4	5	6	7	8
Gender/age	m/30 y	m/17 y	f/8 y	f/8 y	f/4.5 y	f/7 y	f/5 y	f/3.5 y
Mutation	Leu16Pro	Leu16Pro	Leu16Pro	Leu80Arg	Leu105Arg	duplication	duplication	duplication
Inheritance	AD	AD	AD	dn AD	dn AD	AD	dn AD	AD
Age at onset	‘early’	< 2 y	12 m	< 2 y	2 m	< 1 y	neonatal	< 2 y
Age of walking	normal	normal	normal	delayed	4 y*	22 m	22 m	normal
Scoliosis	+	+	–	+	+	–	–	–
Wheelchair (age)	–	–	+ (21 y)	+ (14 y)	+ (4 y)	–	–	+ (8 y)
CSF protein (mg L−1)a	1070	nd	nd	nd	1840	280	nd	nd
Median MNCV (m s−1)b	10	7	0	8	0	9	12	12
Age at last exam	62 y	42 y	31 y	16 y	17 y	35 y	25 y	11 y
Disability stage	2	2	5	4	5	2	2	3–4
Biopsy at age	30 y	17 y	8 y	8 y	4.5 y	7 y	5 y	3.5 y
MF density/normalc	15%	20%	15%	24%	17%	19%	68%	50%
TTFA/normald	nd	448%	458%	nd	393%	nd	118%	182%
OBs/100 MF	99%	97%	90%	92%	100%	89%	37%	27%
Mean g-ratioe	0.81	0.80	0.88	0.88	0.97	0.62	0.55	0.59

^*Supported walking; dn = de novo; nd = not determined; OBs = onion bulbs; † = deceased.

^aCSF protein normal below 400 mg L⁻¹.

^bMedian MNCV 0 = no response.

^cMF density = myelinated fibre density, mean of age-matched controls: 14170 (2–5 years, n = 6); 13180 (6–10 years, n = 11); 10530 (11–20 years, n = 5); 9760 (31–50 years, n = 5).

^dTTFA = total transverse fascicular area, mean of age-matched controls: 0.61 (2–5 years, n = 4); 0.72 (6–10 years, n = 5); 0.87 (11–30 years, n = 5); 1.16 (31–50 years, n = 3).

^eMean of age-matched controls: 0.73 (2–5 years, n = 2); 0.67 (6–20 years, n = 5); 0.66 (21–50 years, n = 3).

Table 5.

Clinical and morphological features in five DSS patients: three genetically unclassified patients and two CIDP patients

Patient	21	22	23	24	25
Gender/age	f/47 y	f/45 y	m/5 y	m/29 y	f/15 y
Mutation	no DNA	no DNA	no DNA	no duplic	no DNA
Inheritance	sporadic	sporadic	sporadic	sporadic	sporadic
Age at onset	< 1 y	< 1 y	2 m	< 1 y	< 1 y
Age of walking	delayed	5 y	20 m*	18 m	26 m
Scoliosis	+	+	+	+	+
Wheelchair (age)	+ (50 y)	+ (15 y)	+ (2 y)	–	+ (16 y)
CSF protein (mg L−1)a	1520	nd	960	2460	1370
Median MNCV (m s−1)b	3	< 10	0	12	7.5
Age at last exam	63 y (†)	50 y	6 y (†)	34 y	16 y (†)
Disability stage	4	4–5	4	2	5
Biopsy at age	47 y	51 y (†)	5 y	29 y	15 y
MF density/normalc	4.5%	9%	12%	24%	13%
TTFA/normald	50%	157%	279%	272%	148%
OBs/100 MF	100%	49%	100%	100%	nd
Pathology type	FFM	FFM	classic OBs	CIDP	CIDP

^*Supported walking; dn = de novo; nd = not determined; OBs = onion bulbs; † = deceased.

^aCSF protein normal below 400 mg L⁻¹.

^bMedian MNCV 0 = no response.

^cMF density = myelinated fibre density, mean of age-matched controls: 14170 (2–5 years, n = 6); 13180 (6–10 years, n = 11); 10530 (11–20 years, n = 5); 9760 (31–50 years, n = 5).

^dTTFA = total transverse fascicular area, mean of age-matched controls: 0.61 (2–5 years, n = 4); 0.72 (6–10 years, n = 5); 0.87 (11–30 years, n = 5); 1.16 (31–50 years, n = 3).

^eMean of age-matched controls: 0.73 (2–5 years, n = 2); 0.67 (6–20 years, n = 5); 0.66 (21–50 years, n = 3).

Autosomal dominant PMP22 mutations

Five patients from three families showed a PMP22 missense mutation and three patients had the 17p11.2 duplication (Table 2). Patients 1–3 are from the same family (Gabreëls-Festen et al. 1992). CSF protein was markedly increased in the two patients in whom it was examined. Only patient 1 had palpably thickened nerves. Patient 3 was by far the most severely affected person compared to the other affected members of this family.

Disability stage in the adult patients with the PMP22 missense mutations ranged from 2 to 5 (fairly well ambulant – fully wheelchair dependent), even within one family. A daughter of patient 3 who had inherited the Leu16Pro mutation could hardly walk independently at the age of 4 years. Median MNCV in the child was decreased to 8 m s⁻¹ at the age of 11 months.

Although onset of walking was markedly delayed in two of the three patients with a PMP22 duplication, the progression and disability stage of these two patients in adulthood was less severe than in the third duplication patient. On the contrary, patient 8 walked in time, but has been using a wheelchair for longer distances since 8 years of age. The mother of this girl (median MNCV 25 m s⁻¹) had hardly been aware of being affected. Moreover, the mother of patient 6 (median MNCV 16 m s⁻¹) had experienced hardly any limitations in daily activities up to age 54 years, but her grandchild, the affected daughter of patient 6, did not walk until age 24 months; and at present, at the age of 7 years, she has serious difficulty in walking. This marked variability in phenotypic expression of PMP22 duplication cases, even within one family, was well documented before (Birouk et al. 1997; Thomas et al. 1997).

Although there is a marked clinical overlap between the two types of PMP22 mutations, pathological features are essentially different, indicating a different pathogenesis. Myelin sheaths in PMP22 missense mutation cases are remarkably thin and are a sign of a generalized, primary hypomyelination, expressed by a high g-ratio (axon diameter/total fibre diameter), probably resulting from a disturbed myelin formation due to a dominant negative effect of the mutated protein (Gabreëls-Festen et al. 1995; D’Urso et al. 1998). Myelin is probably unstable, because from an early age on many large onion bulbs are present, composed of thin Schwann cell lamellae or basal membranes or both as sign of an ongoing process of demyelination and remyelination (Gabreëls-Festen et al. 1995; Ouvrier, 1996). Total transverse fascicular area (TTFA) is markedly increased to the highest values documented in our series of demyelinating HMSN, due to a massive increase in collagen fibres (Fig. 1). Few cases have been designated CHN, favouring an even more severe disturbance in myelin formation (Simonati et al. 1999; Fabrizi et al. 2001).

Fig. 1.

Nerve biopsy of patient 5 with PMP22 missense mutation. Large onion bulbs of Schwann cell lamellae around all demyelinated or hypomyelinated fibres. There is a massive increase in endoneurial collagen. Scale bar = 5 μm.

In PMP22 duplication cases pathology is age-related. In the first years of life there is an active demyelinating process, which subsides in the first decade (Fig. 2). Initially, onion bulbs are small and scanty, but from the second half of the first decade well-developed onion bulbs appear (Gabreëls-Festen et al. 1992, 1995; Sander et al. 1998). This is consistent with the observation that MNCV slowing is evolving during the first years of life and remains constant after the age of 3–5 years (Garcia et al. 1998). Clinical deterioration in later stages of the disease is probably correlated to secondary loss of large calibre motor and sensory axons, expressed by reduced compound motor and sensory action potentials (Roy et al. 1989; Berciano et al. 2000; Krajewski et al. 2000). From an early age, myelinated fibre size histograms show a loss of large and small myelinated fibres. The lack of small fibres is especially remarkable. We could demonstrate this phenomenon most convincingly in myelinated axon histograms of our youngest patients with a duplication, who are aged 2–3 years (Fig. 3). A hypothetical explanation is that, due to the overexpression of PMP22, Schwann cells might leave the proliferation phase prematurely, before all future myelinated axons, especially the smaller ones, have reached a 1 : 1 relationship with the Schwann cells. Myelin sheaths of nerve fibres in duplication cases are in general too thick in comparison with axonal diameters, expressed by a lower mean g-ratio than normal, despite the fact that several fibres are demyelinated or thinly remyelinated (Gabreëls-Festen et al. 1992, 1995; Fabrizi et al. 1998; Sander et al. 1998). This feature is already seen in nerves biopsied at 2–3 years of age. Axonal atrophy secondary to the demyelinating process as a potential cause of this hypermyelination is refuted, because otherwise an increase of small myelinated axons would be expected. However, early duplication cases show a lack of the smallest myelinated fibre population. Therefore, the increased myelin thickness may be a direct consequence of the increased PMP22 gene dosis (Gabreëls-Festen et al. 1995).

Fig. 2.

Nerve biopsy of patient 8 with PMP22 duplication, showing small onion bulbs around a small number of fibres. Demyelinated fibres (arrow). Most myelinated fibres show a relatively thick myelin sheath. Scale bar = 10 μm.

Fig. 3.

Diameter histogram of myelinated axons of patient 8, with 17p11.2 duplication (white area), in comparison with age-matched control (shaded area). There is a lack of small diameter axons.

Autosomal dominant P0 mutations

Six patients of six different families in our series with a DSS phenotype showed a missense mutation in the P0 gene (Table 3). Two non-related patients displayed the same mutation. Parents were unaffected; all mutations occurred de novo. At the last investigation, the level of disability ranged from 2 to 5. In some patients sensory ataxia was a major problem. Four patients were using a wheelchair. Patient 9 had palpably thickened nerves. He showed a mild hearing loss and pupillary response to light was decreased (Adie syndrome). These symptoms have been noted before in a few P0 mutations, usually presenting with an axonal phenotype (CMT2) (De Jonghe et al. 1999; Misu et al. 2000). CSF protein in patient 9 was markedly increased. Repeated neurophysiological testing at the age of 45 years showed conduction blocks over both the median and the ulnar nerves, which might suggest a superimposed CIDP (Donaghy et al. 2000). Patient 10, 36 years old, is still walking well. She has a daughter who inherited the mutation. The girl’s motor development was markedly delayed; at the age of 4.5 years she could take only a few steps unsupported. Patient 11 at the age of 22 years showed a mild to moderate handicap. She could still walk over a distance of 2 km. Patient 12 at 15 years of age, showed unchanged MNCVs at a second electrophysiological examination. Patient 14 had a normal development until the age of 6 months when motor development stopped. Her motor functions progressively deteriorated after the age of 10 months. A diagnosis of Guillain–Barré was considered, but CSF protein was normal. She died at the age of 22 months.

Table 3.

Clinical and morphological features in six DSS patients with AD P0 missense mutations

Patient	9	10	11	12	13	14
Gender/age	m/35 y	f/14 y	f/2 y	m/3 y	f/3 y	f/10 m
Mutation	Thr34Ile	Ser54Cys	Ile135Leu	Lys130Arg	Lys130Arg	Arg98Cys
Inheritance	dn AD	dn AD	dn AD	dn AD	dn AD	dn AD
Age at onset	< 1 y	6 m	neonatal	< 1 y	< 1 y	6 m
Age of walking	3 y	2.5 y	2.5 y	2.5 y	2.5 y	no
Scoliosis (age)	+ (12 y)	+ (10 y)	–	–	–	–
Wheelchair (age)	+ (43 y)	–	–	+ (15 y)	+ (8 y)	–
CSF protein (mg L−1)a	1600	650	345	410	300	170
Median MNCV (m s−1)b	6	8	7	11	10	8.5 (ulnar)
Age at last exam	50 y	36 y	22 y	15 y	11 y	22 m
Disability stage	5	2	2	4	4–5	†
Biopsy at age	35 y	14 y	2 y	3 y	3 y	10 m
MF density/normalc	12%	44%	45%	60%	38%	46%
TTFA/normald	276%	92%	180%	159%	139%	nd
OBs/100 MF	50%	nd	21%	13%	9%	7%
Mean g-ratioe	0.72	nd	nd	nd	nd	0.75
Pathological type	UCM	FFM	FFM	FFM	FFM	UCM

^*Supported walking; dn = de novo; nd = not determined; OBs = onion bulbs; † = deceased.

^aCSF protein normal below 400 mg L⁻¹.

^bMedian MNCV 0 = no response.

^cMF density = myelinated fibre density, mean of age-matched controls: 14170 (2–5 years, n = 6); 13180 (6–10 years, n = 11); 10530 (11–20 years, n = 5); 9760 (31–50 years, n = 5).

^dTTFA = total transverse fascicular area, mean of age-matched controls: 0.61 (2–5 years, n = 4); 0.72 (6–10 years, n = 5); 0.87 (11–30 years, n = 5); 1.16 (31–50 years, n = 3).

^eMean of age-matched controls: 0.73 (2–5 years, n = 2); 0.67 (6–20 years, n = 5); 0.66 (21–50 years, n = 3).

Nerve biopsies of our patients showed two distinct types of nerve pathology as described earlier in patients with P0 mutations (Gabreëls-Festen et al. 1996). Most mutations show a chronic demyelinating process with formations of onion bulbs, but ultrastructural examination reveals either uncompacted myelin (UCM) (Fig. 4) or a normal compact myelin structure, but nearly all fibres show focally folded myelin sheaths (FFM) (Fig. 5) (Gabreëls-Festen et al. 1996). The occurrence of UCM is in agreement with the known function of P0 as a homophilic adhesion molecule, but the functional significance of FFM in P0 mutations is unexplained. CHN was reported in a minority of P0 mutations, suggesting a disturbed myelin formation (Warner et al. 1996; Tachi et al. 1998; Mandich et al. 1999; Phillips et al. 1999).

Fig. 4.

Nerve biopsy of patient 9 with P0 mutation, showing myelinated fibres with uncompacted myelin. Scale bar = 5 μm.

Fig. 5.

Nerve biopsy of patient 10 with P0 mutation, showing myelinated fibres with focally folded myelin. Scale bar = 5 μm.

Autosomal recessive congenital hypomyelination neuropathy

Three patients from two families, both with healthy and distantly related parents, presented with an early onset neuropathy and extremely reduced nerve conduction velocities (Table 4). Patients 15 and 16 are sister and brother. Patient 16 has two healthy children. Sensory ataxia was a main complaint in the three patients. Thickened nerves were palpable in patients 16 and 17. CSF protein was markedly increased in all. A second electrophysiological investigation in the three patients at or shortly before the last examination showed the same values for median MNCV as in earlier registrations. Disability at the last examination ranged from 2 to 5. The marked discrepancy in disability between the two siblings is striking. Patient 16, at the age of 40 years, could walk distances of more than 1 km. His sister became wheelchair dependent at 12 years. She died at age 40 years from breast carcinoma.

Table 4.

Clinical and morphological features in AR DSS: three patients with CHN and three patients with the FFM variant

Patient	15	16	17	18	19	20
Gender/age	f/14 y	m/12 y	f/3 y	m/28 y	f/26 y	f/3 y
Mutation	unknown	unknown	unknown	unknown	unknown	unknown
Inheritance	AR	AR	AR	AR	AR	sporadic
Age at onset	20 m	18 m	neonatal	< 1 y	neonatal	< 1 y
Age of walking	20 m	18 m	4 y	2 y	4 y	3 y
Scoliosis	+ (12 y)	–	–	+ (12 y)	+ (13 y)	+ (4 y)
Wheelchair (age)	+ (12 y)	–	–	–	+ (12 y)	+ (12 y)
CSF protein (mg L−1)a	1790	870	568	nd	nd	increased
Median MNCV (m s−1)b	5	4	11	10	2	4 (ulnar)
Age at last exam	40 y (†)	40 y	27 y	51 y	48 y	24 y
Disability stage	5	2	3	1–2	5	4
Biopsy at age	14 y	12 y	3 y	no biopsy	26 y	3 y
MF density/normalc	16%	29%	25%		7%	decreased
TTFA/normald	nd	nd	166%		184%	nd
Mean g-ratioe	0.93	0.83	0.95		nd	nd
Pathology type	CHN	CHN	CHN		FFM	FFM

^*Supported walking; dn = de novo; nd = not determined; OBs = onion bulbs; † = deceased.

^aCSF protein normal below 400 mg L⁻¹.

^bMedian MNCV 0 = no response.

^cMF density = myelinated fibre density, mean of age-matched controls: 14170 (2–5 years, n = 6); 13180 (6–10 years, n = 11); 10530 (11–20 years, n = 5); 9760 (31–50 years, n = 5).

^dTTFA = total transverse fascicular area, mean of age-matched controls: 0.61 (2–5 years, n = 4); 0.72 (6–10 years, n = 5); 0.87 (11–30 years, n = 5); 1.16 (31–50 years, n = 3).

^eMean of age-matched controls: 0.73 (2–5 years, n = 2); 0.67 (6–20 years, n = 5); 0.66 (21–50 years, n = 3).

Nerve pathology in all three patients revealed very thin or absent myelin sheaths, and nearly all fibres were enclosed by onion bulbs mainly composed of paired basal membranes, consistent with CHN (Fig. 6). Only one thin Schwann cell lamella could be seen, sometimes at the outer ring of the basal membranes. Notably, pathology was essentially the same in both siblings. Some authors consider CHN a disorder of myelin formation without signs of demyelination. However, several authors observed myelin breakdown products in nerve biopsies of CHN patients. We were able to demonstrate clear signs of demyelination and remyelination in teased fibres of one of our CHN patients (Fig. 7).

Fig. 6.

Nerve biopsy of patient 17 with CHN. Axons are enclosed in Schwann cells without myelin and surrounded by onion bulbs of basal lamina and an occasional small and thin Schwann cell fragment. Scale bar = 5 μm.

Fig. 7.

Teased fibre of case 17 with CHN. Two consecutive segments of a fibre with demyelinated (D) and remyelinated (R) internodes. Scale bar = 30 μm. (Reprinted from Neuromuscular Disorders 1994; 4: 63–69, Gabreëls-Festen et al. The status of HMSN type III, by permission from Elsevier Science.)

Most striking is the difference in handicap between the siblings. They showed similar degrees of nerve conduction slowing, and nerve pathology in both was exactly the same. This emphasizes that the severity of the demyelinating process, on which the DSS concept is based, is not the determining factor of clinical severity. Axonal degeneration, secondary to the myelin disorder, probably largely determines the clinical outcome (Krajewski et al. 2000). The mechanisms that relate demyelination to axonal degeneration are still largely unknown. Presumably this relation is modulated by factors in the genetic background, but also by exogenous influences (Garcia et al. 1995).

Warner and colleagues demonstrated a homozygous mutation of the inhibitory domain of the EGR2 gene in three brothers with congenital hypomyelination. Their non-symptomatic, consanguineous parents were heterozygous carriers of the mutation (Warner et al. 1998). A few de novo AD mutations of the PMP22 and the P0 gene were found also to result in CHN (Warner et al. 1996; Ikegami et al. 1998; Mandich et al. 1999; Simonati et al. 1999; Fabrizi et al. 2001). DNA investigation of our patients 15–17 excluded mutations in the PMP22, P0 and the EGR2 gene. This implies that other, yet unknown gene(s) may be responsible for autosomal recessive CHN.

AR neuropathy with focally folded myelin

We classified three patients in this group because of comparable pathological findings, although AR inheritance was demonstrated only in two (Table 4). Patients 18 and 19 are brother and sister; their parents were shown to be unaffected by clinical and electrophysiological examination. We included the sporadic patient 20 in this group as the AD inherited gene that can cause this morphological phenotype, notably P0, was excluded as a causative gene.

CSF protein was increased in the one patient in whom it was examined. Scoliosis was an early symptom in all, and two patients became wheelchair dependent in late childhood. A second electrophysiological investigation in patient 18 when he was 51 years old showed a largely unchanged median MNCV of 9.3 m s⁻¹. In patient 20 motor responses over upper and lower limb motor nerves were no longer obtainable at the age of 21 years. Disability at last examination in the three patients ranged from 1–2 to 5. Again, the discrepancy in severity of the neuropathy in the two siblings was striking. Patient 18, at 51 years old, was functioning in a full-time management job. He can still walk several kilometres, although in the last two years fatigue after physical exercise is increasing. His sister has been wheelchair dependent since the age of 12 years. The striking difference in the severity of disease progression underlines the belief that other factors, probably secondary to the demyelinating process, play a role in determining final clinical outcome.

Nerve biopsy was performed in two patients; no biopsy was performed in the brother of patient 19. Nerve fibre density was markedly decreased, especially in patient 19. Many demyelinated fibres were seen, some within large onion bulbs. Pathology was dominated by the frequent occurrence of focal thickenings of the myelin sheath (FFM).

Autosomal recessive demyelinating neuropathy with FFM, designated CMT4B by some authors (Ben Othmane et al. 1993), is genetically heterogeneous. The first locus was mapped on chromosome 11q23 (CMT4B1), and recently the myotubularin-related protein-2 gene (MTMR2) was identified as the causative gene (Bolino et al. 2000; Houlden et al. 2001). Most patients with MTMR2 mutations show normal early motor milestones and MNCVs are usually > 12 m s⁻¹. Onset in the second or third year of life is followed by a fast progressive clinical course, leading to severe disability in the second decade (Quattrone et al. 1996; Tyson et al. 1997). Several patients died towards the end of the third or forth decade from respiratory distress. This fast and severe progression may discriminate this genetic entity from the other AR forms with FFM.

Recently it was shown that at least two additional disease loci exist for AR neuropathy with FFM, one localized on chromosome 11p15 (CMT4B2), and a third one on a yet unknown locus. Patients from these families had CMT1 phenotypes with onsets beyond early childhood (Gambardella et al. 1998; Ben Othmane et al. 1999).

DNA investigation in our three patients excluded P0, PMP22 and MTMR2 mutations as causative genes. Our patients were not tested for PRX mutations. The patients in whom a PRX mutation was demonstrated complained of a more severe sensory impairment including pain than usually perceived in patients with DSS or CMT1. Nerve pathology shows onion bulbs around thinly myelinated or naked axons and hypermyelinated regions (Boerkoel et al. 2001; Guilbot et al. 2001). None of our patients complained of pronounced sensory disturbances or dysesthesia.

Genetically unclassified DS syndrome

Three patients from our records with a DSS died before genetic testing became available. Parents were unaffected by history (patients 21 and 22) or by clinical and electrophysiological examination (patient 23).

Patient 21 had used a wheelchair from the age of 50 years. She died from cardiopulmonary failure at the age of 63 years. CSF protein was markedly elevated. She showed a perceptive hearing loss and a decreased pupillary response to light. Notably, the latter symptoms were observed also in our patient 9 with a Thr34Ile mutation in the P0 gene and were previously described in patients with certain P0 mutations (De Jonghe et al. 1999; Misu et al. 2000). A nerve biopsy of patient 21 showed a near total loss of myelinated fibres, but the remaining fibres showed many focal myelin thickenings in teased fibre preparations. Large onion bulbs were mainly denervated or surrounded the few remaining fibres. DNA investigation was not possible, but in view of the hearing loss, the Adie syndrome and the pathological features (de novo) AD mutation of the P0 gene might be a possible genetic defect.

Patient 22 showed a marked delay in motor development. She had a severe kyphoscoliosis. Nerve thickening was palpable. In the last years she complained of progressive sensory loss in fingers and distal legs. She died at age 51 years from a progressive respiratory insufficiency and heart failure. Sural nerve obtained at post-mortem examination demonstrated a severe loss of myelinated fibres. Many fibres were demyelinated; several fibres showed thick and folded myelin sheaths. There were only a few small onion bulbs present (Fig. 8). Some denervated Schwann cell structures impressed as bands of Büngner rather than denervated onion bulbs. No DNA analysis could be performed.

Fig. 8.

Nerve biopsy of patient 22 with genetically unclassified DSS. There are many demyelinated fibres (arrow) and fibres with thick and folded myelin (arrowhead), but few onion bulbs. Scale bar = 5 μm.

Patient 23 had feeding problems in the neonatal period. He walked with help at 20 months but was never able to walk unsupported. Clinical and electrophysiological characteristics and sural nerve pathology closely resembled the features in our patient 5 with a PMP22 mutation. In addition one small endoneurial infiltrate and incidental macrophage-induced demyelination was observed (Gabreëls-Festen et al. 1994). The boy died at 6 years of age following a progressive course, before DNA investigation was available. The precise cause of death is not known, but we assume that a (superimposed) CIDP might have played a part in the severe progression.

Chronic inflammatory demyelinating polyneuropathy (CIDP)

In two sporadic cases of early onset demyelinating motor and sensory neuropathy and severe MNCV slowing, compatible with DSS, an initial diagnosis of demyelinating HMSN was made. DNA investigation in patient 24 was negative for the 17p11.2 duplication. CSF protein was markedly elevated in both patients. Sural nerve biopsy showed an active demyelinating process with many endoneurial and epineurial lymphocytic infiltrates in both patients. Onion bulbs were small in patient 25, but strikingly large in patient 24, exceeding the size of onion bulbs usually seen in HMSN type I (Fig. 9). Both patients responded to corticosteroid therapy, but response in patient 25 was short-lived. She deteriorated progressively, became wheelchair dependent and died at age 16 years. No DNA investigation could be performed.

Fig. 9.

Nerve biopsy of patient 24 with CIDP. There are large onion bulbs around myelinated or demyelinated (arrow) fibres, and some denervated onion bulbs. Scale bar = 5 μm.

It may be extremely difficult to distinguish CIDP starting in infancy from inherited neuropathies, as the evolution of symptoms is usually identical to HMSN, and a preceding normal period cannot be established with certainty (Sladky et al. 1986; Bird & Sladky, 1991). No arguments for an inherited disorder will be found in cases of de novo dominant mutations or a first presentation of a recessive genetic defect. Mild inflammatory signs may be present in HMSN cases, which was documented in 17p11.2 duplication cases (Gabreëls-Festen et al. 1993). However, the possibility of an inflammatory autoimmune component, superimposed on a genetic disorder, cannot be excluded. Many reports have mentioned a positive response of proven HMSN on immunosuppressive therapy (Dyck et al. 1982; Mitchell et al. 1987; Antoine et al. 1989; Bird & Sladky, 1991; Donaghy et al. 2000; Gabriel et al. 2002).

Summary

In our series of 25 DSS patients, 14 were affected by an AD mutation and six patients by a likely AR genetic defect. In three patients no DNA investigation could be performed. Two patients suffered from an acquired disorder.

Clinical disability in adulthood showed a conspicuous variation. Three patients died before adulthood. In some patients decease might be related to a (superimposed) CIDP. The remaining 21 patients could be followed up into the second decade or longer. Age at last investigation ranged from 11 to 63 years (mean 33.3 years). Nine of the 25 patients were still ambulant at last investigation (mean age 38.6 years, range 22–62 years) and four of them, who had reached the age of 40 years, could even walk more than 1 km.

A more or less severe handicap was neither restricted to AD- or AR-inherited disorders nor to certain genetic defects. The degrees of motor development delay, conduction slowing or CSF protein increase were not predictive of a more severe clinical course. Most striking was the difference in handicap between the two pairs of siblings suffering from AR-inherited CHN and from an AR variant with FFM (CMT4B), respectively. This emphasizes that the severity of the demyelinating process, on which the DSS concept is based, is not the major determining factor of clinical severity but, most presumably, the (secondary) axonal loss (Krajewski et al. 2000).

Conclusions

DSS, defined as an early onset demyelinating motor and sensory neuropathy with upper limb MNCV below 12 m s⁻¹, is genetically heterogeneous and morphological features are highly variable. DSS is usually caused by AD mutations of the P0 or the PMP22 gene and less frequently by AD mutations of the EGR2 gene or AR mutations of some known and a number of yet unknown genes. CIDP may present also as an early onset, demyelinating neuropathy, resembling DSS.

Clinical outcome of DSS is markedly variable: nearly one-third of the patients in our series who had reached a mean age of 39.9 years are still ambulant and able to walk over at least 1 km. A relatively mild course is seen in AD as well as AR forms and occurs in all genetic subtypes of DSS. There is a marked intrafamilial difference in clinical severity in adulthood. This again emphasizes that the severity of the demyelinating disorder on which the DSS concept is based is not the major determining factor for clinical disability.

We conclude that DSS, although in general denoting a more serious neuropathy than CMT1, does not imply severe disability or wheelchair dependency in adult life.