OUP user menu

Clinical, pathological and genetic characterization of hereditary sensory and autonomic neuropathy type 1 (HSAN I)

Henry Houlden, Rosalind King, Julian Blake, Mike Groves, Seth Love, Cathy Woodward, Simon Hammans, James Nicoll, Graham Lennox, Dominic G. O'Donovan, Carolyn Gabriel, P. K. Thomas, Mary M. Reilly
DOI: http://dx.doi.org/10.1093/brain/awh712 411-425 First published online: 19 December 2005


Hereditary sensory and autonomic neuropathy type I (HSAN I) is the most frequent type of hereditary neuropathy that primarily affects sensory neurons. The genetic locus for HSAN I has been mapped to chromosome 9q22.1–22.3 and recently the gene was identified as SPTLC1, encoding serine palmitoyltransferase, long chain base subunit-1. Sequencing in HSAN I families have previously identified mutations in exons 5, 6 and 13 of this gene. We analysed the SPTLC1 gene for mutations in 8 families with HSAN I, 60 individuals with sporadic sensory neuropathy, 6 HSAN II families, 20 Charcot-Marie-Tooth type I families and 20 families with Charcot-Marie-Tooth type II. Six HSAN I families and a single sporadic neuropathy case had an identical SPTLC1 mutation. No mutations were found in the other groups. Genetic haplotyping across the HSAN I critical region in 5 families and the sporadic case suggested a common founder. Several characteristics, previously not widely recognized were identified, including lack of penetrance of the SPTLC1 mutation in some individuals, variability in age of onset along with an earlier age of onset in younger generations, in some patients surprisingly early and often severe motor involvement and an earlier onset characterized by motor involvement with demyelinating features in males compared to females in 4 families. The sensory findings were often disassociated with prominent pain and temperature loss. Neurophysiology mainly showed a sensory axonal neuropathy but in many individuals there was electrical evidence of demyelination. Sural nerve biopsies from six affected individuals and the post-mortem findings in 1 case showed mainly axonal loss. This in depth study on the phenotype of HSAN I in 6 families and a single sporadic case with a common founder identifies a number of poorly recognized features in this disorder and highlights the clinical heterogeneity both within and between families suggesting the influence of other genetic and acquired factors.

  • phenotype
  • peripheral nervous system
  • neuropathology
  • mutation
  • hereditary neuropathy
  • AD = autosomal dominant
  • CMT = Charcot-Marie-Tooth
  • HSAN I = hereditary sensory and automatic neuropathy type 1


The hereditary sensory and autonomic neuropathies (HSAN) are clinically and genetically a heterogeneous group of disorders, which are currently classified into five types (Houlden et al., 2004a). HSAN type 1 (HSAN I) is an autosomal dominant (AD) neuropathy characterized by marked sensory involvement, variable motor involvement and minimal autonomic involvement. Single cases and families compatible with HSAN I were originally described in the early French literature (Leplat, 1846; Nélaton, 1852). In 1922, Hicks (1922) described a remarkable London HSAN I family, 10 of whom suffered from perforating ulcers of the feet, lancinating shooting pains and deafness. Jughenn (Jughenn et al., 1949) and Denny-Brown (1951) demonstrated that the pathological process underlying the distal ulceration and mutilating arthropathy seen in these conditions was a neuropathy, rather than an anatomical disorder such as syringomyelia, as had been previously suggested.

Since then, many other HSAN I families have been reported with similar clinical features to that of Hicks' patients, although deafness appeared to be a rare feature (Thèvenard, 1942, 1953; Jackson, 1949; Campbell and Hoffman, 1964; Wallace, 1965, 1970; Whitaker et al., 1974; Dyck and Ohta, 1975; Dubourg et al., 2000).

The HSAN I locus was mapped using genetic linkage analysis to the chromosome 9q22.1–q22.3 region (Nicholson et al., 1996; Bejaoui et al., 1999). The ubiquitously expressed SPTLC1 gene was mapped to within this region and three missense mutations were identified in exons 5 and 6 (Bejaoui et al., 2001; Dawkins et al., 2001). More recently an additional mutation has been identified in exon 13 (Verhoeven et al., 2004). Haplotype analysis in an Australian and two British families with the same C133W mutation identified a common haplotype, suggesting a British founder (Nicholson et al., 2001).

We have identified 8 British families with AD HSAN I, 60 individuals with sporadic sensory neuropathy, 6 families with HSAN II, 20 families with AD Charcot-Marie-Tooth (CMT) type II and 20 families with AD CMT type I. Two of the AD HSAN I families showed linkage to the chromosome 9q22.1–22.3. To ascertain the degree of genetic heterogeneity in our group of HSAN I families and the frequency of mutations in a population of patients with sporadic sensory neuropathy, we screened the SPTLC1 gene in one affected member from each family and each individual case. Six HSAN I families and one sporadic sensory neuropathy case were identified with the same mutation in SPTLC1. None of the families with SPTLC1 mutations have been reported in the past or are part of other published series (Dawkins et al., 2001). We looked for a common ancestry by carrying out haplotype analysis with flanking markers, SNPs and di- and tri-nucleotide repeat markers located within the introns and exons of SPTLC1. We have also performed a detailed clinical, neurophysiological and pathological analysis (six nerve biopsies and one post-mortem) in affected members of these families.



Approval to perform this study was obtained from the joint medical and ethics committee at The National Hospital for Neurology and Neurosurgery. The clinical and electrophysiological features of all affected cases are presented in Fig. 1 and Tables 1, 2 and 3.

Fig. 1

HSAN I family trees. Arrow indicates the proband in each family. Squares, male; circles, female; filled symbols, affected; and line through symbols, deceased. +/−, heterozygous mutation and −/−, normal genotype.

View this table:
Table 1

Clinical details of HSAN I families

FamilySexCurrent age (Y)Age of onset (Y)First presentationPositive sensory symptomsSignsUlcersAmputationsAreflexia
MotorSuperficial sensoryDeep sensoryUpper limb
Family 1
    III-3MDied 5018Decreased sensation feetPain and burning in the feetNo+NoNoLeft footLeft great toeNo
    III-4F5347Decreased sensation and blisters on feetParaesthesia in feetNo++NoFeetNoReduced AJs
    III-6M6358Cold feetPain and burning in feet+++NoRight great toeRight great toeNormal
    III-8F6247Paronychia that failed to healShooting and burning pains in feet and hands+++++Right great toeRight great toeAbsent AJs
    III-14M6930Pain in the feet and burningPain in hands and feet. Paraethesia in hands++++++++++Ulcers on feetNoDepressed KJs, absent AJs
    IV-3M3321Blisters on right footNoNo+N/ANoFoot and great toeNoNo
    IV-8M3629Pain and temperature loss in the feetShooting pains in feetNoNoNoNoNoNoAll sluggish
    IV-11F4322Decreased left foot sensationPain, left leg and hands, burning and paraesthesia hands++NoNoNoNoNo
Family 2
    II-2FDied 6560Sores on feet, slow to healN/A++N/ANoFeetNoN/A
    III-2F5224Decreased sensation in feet then burnsShooting and jumping pains+++++++Painless burnsCharcot joint in footAbsent AJs
    III-4M5420Blisters on feetNeedle pains in legs++++++FootNoAbsent AJs
    IV-1M2312Decreased sensation in feetShooting pain in feet, paraesthesia in hands++++++++Great toes ulcer, nail injuryNoAreflexic
    IV-2 TwinF2017Pain in feetPins and needles, pain in feet++NoNoNoNoNo
    IV-3 TwinF2017Pain in feetPins and needles, pain in feet++NoNoNoNoNo
Family 3
    II-2MDied 5915Ulcer on big toeShooting pains, severe stump pain+++++N/A++Repeated ulceration in both legsLegs amputatedAbsent lower limbs
    III-2F2816Painless blisterParaesthesia in hands+++No+FeetNoAbsent AJs
    III-5F3520Painless injury on right halluxNoNoNoNoNoNoNoNo
Family 4
    III-1MDied 80N/AN/AN/A+N/AN/ANoLegsBoth legs amputatedN/A
    IV-3MDied 62N/ADecreased sensation in feetN/AN/A+N/AN/ANoNoN/A
    IV-4F94Unaffected clinicallyNo symptomsNoNoNoNoNoNoNoNo
    IV-6MDied 8130SensoryNo++N/AN/AFeetToe amputatedN/A
    IV-8F9070Decreased sensation in feetNo++++N/A++FootNoN/A
    V-1M7120Cold handsPain in the hands and feet+++++++++++Foot and handPart of hand amputatedAll absent
    V-3M6112Decreased sensation in fingersNo+++++NoNoAbsent AJs
    V-7F6730Decreased sensation in feetPain in the hands and feet++++N/A++NoNoReduced UL and LL. Absent AJs
    VI-1M4118Decreased sensation in feetNo+++++++++++FeetCharcot joint right ankleReduced UL Absent LL
    VI-3M2512Pain and paraesthesia in both feetPain and paraesthesia in feet+++++NoNoAbsent AJs
    VI-11M4015Decreased sensation feetNo++N/ANoHeel ulcerPart of foot amputatedN/A
Family 5
    II-2FDied 7240Paraethesia in feetShooting pain in the legs and hands, paraesthesia in the legs, burning in the hands++++++++++++Hands and toesToes amputatedAbsent lower limbs
    III-3F7118Decreased sensation in feetBurning and shooting pains in the feet+++++++++++Ulcers on feet and toesToes amputatedAll absent
    IV-2F2135Painless injuryPain in feet and legsNo+N/ANoNoNoNo
Family 6
    II-1M6860Tingling both handsNumb to mid calf. Pain feet legs, burning in hands+++++++On fingers and toesNoAreflexic apart from triceps jerks
    II-6F5850AtaxiaPins and needles in the hands and feet, pain and numbness in the feetNo++No++NoNoNo
    II-11M4835Ulcers on soles of feetSore feet, numbness mid calf, tingling fingers++++++Ulcers on feetToes amputated left footAbsent AJ's
    III-3M3227Numb feetNumb, burning, pins and needles, sharp pains feet, burning hands+++++++Ulcers on feetNoAbsent Right biceps, supinator and both AJ's
    III-4M4128Numbness in fingersNumbness and shooting pains in hands and feet++++++NoNoAbsent in the lower limbs
Sporadic case (S1)
    II-1M4520Blister on his footNo++++++++++On fingers and toesNoTriceps jerks only present
  • AJs = ankle jerks, KJs = knee jerks, LL = lower limbs, UL = upper limbs, N/A = not available. +++ = severely affected (motor: proximal motor weakness, sensory: abnormality of the limbs and the trunk, upper limbs: proximal involvement), ++ = moderately affected (motor: distal weakness, sensory: abnormalities of the limbs, upper limbs: distal involvement, + = mildly affected (motor: mild distal weakness, sensory: mild distal sensory abnormalities, upper limbs: mild distal involvement), Y = years, superficial sensory = pin prick, deep sensory = joint position sense, vibration.

View this table:
Table 2

Motor nerve conduction studies in HSAN I families

FPatientSexAge at testMedian nerve: SE on APBUlnar nerve: SE on ADMC peroneal nerve: SE on EDBPosterior tibial nerve: SE on AH
DML (ms)CV (m/s)Dist amp (mV)DML (ms)CV (m/s)Dist amp (mV)DML (ms)CV (m/s)Dist amp (mV)DML (ms)CV (m/s)Dist amp (mV)
F 1III-3M31N/A3.848.36.537.7N/A
IV :8M3235113.22.6498.9AbsentAbsent
IV :11F343566.326211.36.3180.57.5260.5
F 2III-2F28N/A3.2535.64.3251.05.4260.1
III-2F485.9254.63.9N/A2.8 prox dAbsentN/A
F 3II-2M1650260.1
II-2M31AbsentAbsentto TA 6.3290.2Absent
F 4V-1M60AbsentAbsentAbsentAbsent
F 5II-2F62N/A4.731N/AN/A
F 6II-1M622.9432.93128.7230.8N/A
II-1M655.7213.3 prox cb5.2232.6 prox dN/AN/A
II-1M675.5320.9 prox d5.7262.1 prox cbAbsentAbsent
prox c504.2prox c571.3 d
S1II-1M37AbsentAbsent, 3.433to FCUAbsentAbsent
  • F = Family; S1 = sporadic case; N/A = not available; SE = surface electrode; APB = abductor pollicis brevis; ADM = abductor digiti minimi; FCU = flexor carpi ulnaris; EDB = extensor digitorum brevis; AH = abductor hallucis; prox = proximal; d = dispersion; c = conduction; cb = conduction block; U = unobtainable. zTA = Tibialis anterior, ‘Normal’ = reported as normal.

View this table:
Table 3

Sensory nerve conduction studies in HSAN I families

FPatientSexAge at testMedian nerve (FII-W)Ulnar nerve (FV-W)S Peroneal nerveSural nerve
CV (m/s)Amp (μV)CV (m/s)Amp (μV)CV (m/s)Amp (μV)CV (m/s)Amp (μV)
F 1III-3M317N/AAbsentN/A
F 2III-2F28Lat. 2.9 ms37Lat. 2.6 ms18N/AAbsent
III-4M27N/ALat. 2.5 ms13N/AAbsent
F 3II-2M16Lat. 2.3 ms8N/AN/AAbsent
F 4V-1M60AbsentAbsentN/AAbsent
F 5II-2F62AbsentAbsentN/A
III-3F58Radial: normal: 55 uV, 61 m/sAbsent
F 6II-1M62432.9442N/AAbsent
  • N/A = not available, S1 = sporadic case, FII-W = finger 2 to wrist, FV-W = finger 5 to wrist, velocity to onset latency, S = superficial. Lat = latency, ‘Normal’ = reported as normal.

Genetic sequencing

DNA was extracted from blood samples obtained with informed consent from affected and unaffected individuals. The chromosome 17 duplication and TTR gene mutations were excluded in appropriate cases (Thomas et al., 1997).

In the AD HSAN I families, all 15 exons of the SPTLC1 gene and flanking introns were sequenced. In the other HSAN cases, only exons 5 and 6 were sequenced, as the first three described mutations in SPTLC1 had been identified in these two exons. The 15 exons and flanking intronic regions of SPTLC1 were amplified by standard polymerase chain reaction (PCR) using TaqGold polymerase and the same primers were used for PCR and sequencing (PCR method and primer sequenced available on request). PCR product was purified and sequenced with BigDye Terminator cycle sequencing kit (Perkin-Elmer). Sequencing was performed on an ABI377 automated sequencer.

Clinical neurophysiology

Nerve conduction studies are detailed in Tables 2 and 3 and were performed by standard methods (De Lisa et al., 1994).

Sensory conduction in the upper limbs was recorded orthodromically with ring-stimulating electrodes on the fingers and recording the evoked responses with surface electrodes over the median or ulnar nerves at the wrist. An antidromic recording technique was used in the lower limbs, stimulating the superficial peroneal and sural nerves above the ankle and recording over the nerves at the ankle.

Motor conduction was studied by recording compound muscle action potentials (CMAP) with surface electrodes over abductor policis brevis (APB) for the median nerve and adductor digiti minimi (ADM) for the ulnar nerve and stimulating both at the wrist and the elbow for each case. Peroneal motor conduction was measured recording from extensor digitorum brevis (EDB) and stimulating the peroneal nerve at the ankle and at the fibular head. Posterior tibial nerve motor conduction was assessed, recording over abductor hallucis (AH) stimulating at the ankle and popliteal fossa.

Sural nerve biopsy

Sural nerve fascicular biopsies were obtained from a standard retromalleolar site. The nerve tissue was divided and portions fixed in 10% neutral buffered formaldehyde overnight and embedded in paraffin wax or fixed in 3% glutaraldehyde in 0.05 M sodium cacodylate buffer pH 7.4, post-fixed in 1% osmium tetraoxide and embedded in Araldite CY212 resin. Paraffin sections were stained with haematoxylin and eosin, and 1 μm resin sections cut and stained with methylene blue, azure II and basic fuchsin. 70 nm resin sections of areas of interest were cut and contrasted with uranyl acetate and lead citrate for electron microscopy (Sievers, 1971).


An autopsy with full consent for the retention of tissue was performed (JN and SL) on one of the patients with HSAN I (Family 4, IV-6) after he died of bronchopneumonia at the age of 80 years. The brain, spinal cord, multiple spinal nerve roots and dorsal root ganglia, several blocks of skeletal muscle and peripheral nerve, including the sural nerve at the level of the lateral malleolus, were fixed for histology, the brain and spinal cord by suspension in 20% formalin. The blocks of sural nerve were fixed in cacodylate-buffered glutaraldehyde and processed to Araldite for semithin and ultrathin sectioning. The semithin sections were stained with toluidine blue and paraphenylene diamine. The ultrathin sections were contrasted with lead citrate and uranyl acetate. The remaining blocks were fixed in formalin, embedded in paraffin wax and sections stained with haematoxylin and eosin, haematoxylin/van Gieson, phosphotungstic acid/haematoxylin, Palmgren silver impregnation, luxol fast blue/cresyl violet and solochrome cyanin. Sections of muscle were, in addition, immunostained for fast and slow myosins.



Sequencing of SPTLC1 in the AD families with HSAN I revealed that six of the eight families analysed had the same heterozygous mutation of T to G at nucleotide position 399, causing the substitution of cysteine (C) to tryptophan (W) at amino acid 133. In one sporadic case (S1) of sensory neuropathy the same mutation was identified. The family history is unknown in this case as the patient is adopted. The C133W mutation was not present in 200 UK controls. A further three polymorphisms were identified in controls: SPTLC1 exon 6 +16 bp G/C, exon 8 poly(A) repeat and exon 14 +52 bp A/G. We analysed the intronic region of SPTLC1 for the polymorphisms (SNPs and di- and tri-nucleotide repeats) listed in Table 4. These polymorphisms were used to obtain disease haplotypes in families with mutations, to look for a founder effect (Table 4). A common haplotype was identified in the families analysed (F1–F5) and the individual case, strongly suggesting that the disease gene was derived from an English ancestor common to all C133W mutated HSAN I families.

View this table:
Table 4

Clinical details

The HSAN I families with the same SPTLC1 mutation could not be genealogically linked. They originated from London, South and South Western regions of England. The pedigrees of the six HSAN I families with mutations are given in Fig. 1. All families showed evidence of AD inheritance. The clinical features of the families are presented in Table 1. The mean age of onset for affected individuals from all families was 29 years, the age of onset range was 12–70 years; one individual was not affected at age 89. There was variability in the age of onset within families. In Family 4 cases V-3 and VI-3 had onset at 12 years compared with their aunt/great aunt IV-8 with an age of onset of 70 years and their mother/grandmother IV-4 who was clinically unaffected at the age of 89 and had reportedly normal nerve conduction studies at the age of 81. In Family 1, individuals II-4 and II-6 also were not reported to be affected but must have possessed the SPTLC1 mutation. The mean age at death was earlier than the general UK population at 67 years in affected HSAN I individuals; probably reflecting the complications of ulcers, amputations and infection as in the case of Family 3, II-2 who died at age 59 years.

The most frequent presenting feature of HSAN I was decreased sensation in the feet. Limb pain, painless blisters and ulcers on the feet were also commonly encountered. Positive sensory symptoms of shooting and burning pain and paraesthesia, most often in the feet, were seen in all families although not all individuals had these symptoms. The sensory loss had a glove and stocking distribution and was dissociated. Negative sensory symptoms, involving the loss of superficial sensation was found in the majority of family members. Pain and temperature were affected first and most severely with light touch less affected. Joint position sense and vibration were also affected but much less and later than superficial sensation. In some individuals, such as Family 1, III-14 and Family 3, II-2, this sensory loss also involved the abdomen and shoulders. Painless ulcers that were slow to heal and often became infected were found in all families. This led to a number of individuals having Charcot joints and amputations. Motor signs were surprisingly frequent and severe in our HSAN I families. Several patients had severe motor involvement with proximal and distal weakness and wasting that was the predominant clinical feature (Table 1). Within families motor weakness varied from no involvement to severe involvement. These cases were also thoroughly investigated for an additional acquired neuropathy. The HSAN I families had relatively few autonomic abnormalities (Table 1).

In Family 5, the affected grandmother, II-2 and mother, III-3 (maximum inspiratory pressure −6 kPa) had slowly progressive breathing problems, starting when they were in their early 50s. The mother also had a hoarse voice and was shown to have diaphragmatic weakness. The daughter (IV-2) is affected but in her 30s shows no diaphragmatic problems. In Family 1, two individuals (III-14 and III-6) had early cognitive decline. In case III-14 the cognitive decline commenced at age 55 and was associated with dysarthria and depressed corneal reflexes. This patient also had a history of alcohol overuse and was found to have a high CSF protein of 0.60 g/l but a sural nerve biopsy was typical for HSAN I (Fig. 3A and B). Another case (III-3) from Family 1 had a CSF protein in the normal range of 0.30 g/l. In case III-6 the onset was at age 60 with associated dysphasia. Family 1 case IV-3 had facial motor tics and left ptosis; the tics preceded the neuropathy. One member of Family 6, the proband II-1, had HSAN I with evidence of superimposed chronic inflammatory demyelinating polyneuropathy (CIDP) (details given below).

An interesting observation in the HSAN I families is the later age of onset and often reduced severity in females versus males. The mean female age of onset was 34 years (n = 19) and that for male was 25 years (n = 15). Analysing the age of onset data with a paired t-test gave a P-value of 0.076, just below the significance level. The reduced disease severity in females is particularly apparent in Families 2, 3, 4 and 6. These male/female differences may reflect protection from the extra X-chromosome or other factors such as hormones. Another observation is the apparent decrease in age of onset and often increase in disease severity with successive family generations, this may reflect closer monitoring of these individuals or the influence of other unidentified risk factors.

Family 6, proband (II-1) is an unusual case of HSAN I with probable superimposed CIDP. This 69-year-old man was well until the age of 60 when he developed a rapidly progressive painful motor and sensory neuropathy (including skin ulcers), culminating in him being wheelchair bound. There was electrophysiological evidence of a rapidly progressive demyelinating neuropathy with slow conduction velocities down to 21 m/s in the right median, definite conduction block and temporal dispersion on repeated testing (see Tables 2 and 3), a high CSF protein of 1.93 g/l and a nerve biopsy showing a devastated nerve with only one fibre present but no inflammatory cells (Fig. 2A and B). He was diagnosed with CIDP and eventually responded well (left with mainly sensory signs and distal motor weakness) to cyclophosphamide, having failed to respond to IVIG, steroids and plasma exchange.

Fig. 2

A 1 μm stained resin section (A) and electron micrograph (B) of the sural nerve biopsy from Family 6 II-1. The single myelinated fibre in A (arrow) is visible at the top of B; many more unmyelinated axons (arrows) remain than myelinated. Scale bar = 10 μm (A) and 5 μm (B). A 1 μm stained resin section (C) and teased fibre preparation of the sural nerve biopsy from Family 6 II-11 (D and E). There are around 50 myelinated fibres in the whole biopsy. At least two of them appear to have internodes that seem to be demyelinating in a segmental pattern (between arrows D and E). Scale bar = 10 μm (C) and 20 μm (D and E).

He initially denied any family history until his son (III-3) and brother (II-11) presented with a neuropathy. Other affected individuals were then identified (Fig. 1 and Table 1). The males presented earlier and more severely (especially motor involvement) with demyelination electrically. The female case (II-6) presented later with a painful axonal neuropathy. One male had taken excess alcohol in the past (III-4) but had an identical neuropathy, another male (II-11) also had a nerve biopsy (Fig. 2C–E) that showed significant loss of myelinated axons and axonal degeneration. All affected family members including the index case were found to have the C133W SPTLC1 mutation.


Details of electrophysiology in the HSAN I families is shown in Tables 2 and 3.

Variability of abnormalities was seen within families and between families. Overall, the findings showed an axonal motor and sensory peripheral neuropathy. Sensory potentials were usually absent in the lower limbs but were often recordable in the upper limbs where they were surprisingly normal in some individuals, as seen in female members of Families 1 and 2. Motor conduction showed a much greater variability between individuals. Motor response amplitudes were usually very small or absent in the lower limbs but amplitudes in the hands were frequently normal or only moderately reduced. Motor conduction velocities ranged from normal (≥50 m/s) through intermediate slowing (35–50 m/s) to slowing unequivocally in the demyelinating range (≤35 m/s). Conduction slowing in the demyelinating range was seen with well-preserved motor response amplitudes. This was most clearly seen in the male members of Family 6, one of whom also showed motor conduction block and dispersion of motor response (Family 6 II-1); this patient was also thought to have CIDP (see above). Subject Family 6 III-3 was shown to have dispersion of ulnar motor response to proximal stimulation as well as slowing of motor conduction in distal nerve segments with normal conduction velocity measured proximally, suggesting some patchiness. In Family 2, subject IV-1 progressed rapidly over a short time going from normal median nerve motor conduction velocity (56 m/s) at age 15 to ‘demyelinating’ conduction velocity (30 m/s) aged 22. Needle electromyography of distal limb muscles was performed (not detailed in Tables 2 and 3) and showed moderate to severe chronic denervation in keeping with significant motor axon loss.

Nerve biopsy results

Sural nerve biopsies were carried out on the following cases: Family 6 II-1 (Fig. 2A and B), Family 6 II-11 (Fig. 2C–E) and Family 6 III-4, Family 1 III-14 (Fig. 3A and B), Family 3 II-2 (Fig. 3C) and Family 4 VI-1. The sural nerve biopsy findings in these cases were very similar. In these severely affected nerves, very few myelinated fibres remained (Fig. 3C) but electron microscopy showed a reasonable number of unmyelinated axons (Fig. 3A) although the presence of stacks of flattened Schwann cell processes suggested unmyelinated axon loss (Fig. 3A and B). There was very little to suggest regenerative activity. A teased nerve fibre preparation of the few myelinated fibres left in the sural nerve from (Family 6 II-11) showed two fibres that had internodes apparently undergoing primary (segmental) demyelination (Fig. 2D and E). The biopsy findings in the patient initially diagnosed as having CIDP (Family 6 II-1, Fig. 2A and B) were also similar; only two myelinated fibres remained and no inflammatory cells or onion bulbs were present. Electron microscopy showed occasional evidence of recent axonal degeneration, and numerous unmyelinated axons remained. No infiltrating cells were found and there were no abnormal inclusions in any cell type.

Fig. 3

Nerve biopsy from Family 1 III-14 (A). EM shows numerous unmyelinated axons (arrowheads) but occasionally they are abnormally pale suggesting degeneration (arrow). Bar = 10 μm. (B) Shows flattened Schwann cell sheets in the same specimen as A; there are no axons in this field. Bar = 1 μm. (C) Shows a nerve biopsy from Family 3 II-2. A single fascicle containing very few, small, myelinated fibres. Bar = 50 μm. Lumbar dorsal root ganglion from the autopsy case Family 4 case IV-6 (D) showing nodules of Nageotte (arrows). Haematoxylin and eosin stain, Bar = 100 μm. (E) Shows the lumbar dorsal root ganglion from Family 4 case IV-6 showing axonal swellings (arrowheads) and some fine axonal sprouts (arrows), Palmgren's silver stain. Bar = 100 μm. (F) Shows the interosseous muscle from Family 4 case IV-6, immunostained for fast myosin and showing obvious fibre-type grouping. Bar = 200 μm.

Autopsy findings

Autopsy was carried out on Family 4 case IV-6 (Fig. 3D–F). General examination confirmed the bronchopneumonia. There was severe, generalized muscular wasting. No ulcers or trophic skin changes were evident. The brain and spinal cord appeared macroscopically normal. No histological abnormalities were noted in the brain. There was mild, symmetrical loss of fibres from the gracile funiculi of the thoracic and cervical spinal cord but not at lumbar levels. No other abnormalities were seen in the spinal white matter. Anterior horn cells appeared to be well preserved at all levels, as were intermediolateral column neurons and Clarke's column in the thoracic segments. The dorsal spinal nerve roots were mildly fibrotic and depleted of myelinated fibres. Ventral roots were relatively well preserved. The dorsal root ganglia showed evidence of mild to moderate loss of ganglion cells. Sites of ganglion cell degeneration were marked by nodules of Nageotte and clusters of fine axonal sprouts (Fig. 3D). Scattered axonal swellings were present within the ganglia (Fig. 3E).

No significant abnormalities were detected on paraffin histology of the ulnar or femoral nerve. The vagus, sympathetic trunk and superior cervical and coeliac ganglia appeared normal. The sural nerve was moderately fibrotic and contained only a few remaining myelinated fibres (∼40 per mm2), most of which were of relatively large calibre. Electron microscopy showed these few myelinated fibres to have a normal ultrastructural appearance, and also revealed moderate loss of unmyelinated axons. No other significant ultrastructural abnormalities were identified.

Sections of skeletal muscle revealed scattered moderately atrophic fibres, most of which contained fast myosin and were therefore presumably of histochemical type 2. Very little fibre-type grouping was evident in sections from the larger proximal muscles such as sternomastoid, deltoid, biceps brachii, psoas and quadriceps femoris. Extensive fibre-type grouping was, however, noted in sections of interosseous muscle, suggesting that there had been some previous denervation and reinnervation (Fig. 3F).


Sequencing the SPTLC1 gene identified six HSAN I families (HSAN I Families 1–6, Fig. 1) and one sporadic case of sensory neuropathy (S1) with the same C133W mutation (the British mutation). This mutation, reported in the past (Dawkins et al., 2001), is not present in 200 UK controls and this amino acid of the SPTLC1 gene is highly conserved in many species.

Haplotype analysis in multiple affected individuals in Families 1–5 and the sporadic case showed a common haplotype across the SPTLC1 gene and in flanking markers in these individuals, strongly suggesting a founder effect (Table 4). A founder effect has previously been observed in Australian and some British families with the same C133W SPTLC1 mutation, we were unable to trace our HSAN I families back to each other or to Australian ancestors by genealogy (Nicholson et al., 2001). In general, the size of a conserved haplotype for a particular mutation provides an estimate of the age of the mutation. In our group of SPTLC1 families the conserved haplotype spanned 1.57 cM between markers D9S1815 and D9S1803 but there was evidence that the ancestral haplotype extended further out between D9S1796 and D9S196 over 2.41 cM (http://cedar.genetics.soton.ac.uk). If the UK population expanded at a constant exponential rate, the Luria-Delbruck method (Luria and Delbruck, 1943) estimates the number of generations (g) that have passed since the mutation was first introduced. That is, the proportion of chromosomes carrying the mutation in which recombination with the conserved haplotype has occurred equals1 − eg(𝛉), where 𝛉 is the recombination fraction in morgans (Anikster et al., 2001; Austerlitz et al., 2003). From our haplotype data and the recombination rate between families, we estimate the mutation occurred ∼900–1600 years, ago. The Luria-Delbruck equation (Luria and Delbruck, 1943) often gives an underestimate of the true mutation age and this may reflect in the figures given (Colombo, 2000a, b; Austerlitz et al., 2003). Analysis of the entire group of published families with the C133W mutation (Bejaoui et al., 2001; Dawkins et al., 2001) would give a better estimate given that we have only looked at a small group from the South and South West of England.

SPTLC1 mutations are rare in cases of sporadic neuropathy but common in families with hereditary sensory neuropathy. No mutations were identified in the recessive HSAN II families, CMT I or II families. Two families with HSANI I had no SPTLC1 mutations. The first family (HSAN I F7) had at least four affected individuals over three generations with the disease inherited as an AD trait. The age at onset of the neuropathy was in the early 30s. It mainly involved sensory nerves, with positive and negative features, but later also involved the motor system. Nerve conduction showed a predominantly sensory neuropathy. The most distinctive feature in this family was the autonomic involvement of the gastrointestinal tract, leading to marked weight loss, abdominal pain and diarrhoea. Two individuals required PEG feeding in the latter years of their lives due to the severe GI problems. No nerve biopsy was carried out in this family. This family has atypical features that may well involve a mutation in a different gene. The second family (HSAN I F8) had four affected members over three generations; the age of onset was in the 20s with marked sensory symptoms and later limb ulcers. A nerve biopsy carried out in the proband from family HSAN I F8 showed the same features as the HSAN I families with SPTLC1 mutations. A mutation in the Rab7 gene has recently been identified in this family (Houlden et al., 2004b).

The clinical features (Table 1) of these families with HSAN I and a mutation in SPTLC1 show early dissociated sensory involvement affecting pain and temperature with preservation of vibration and joint position sense. There were frequent positive sensory symptoms of severe shooting or burning pains in the limbs. Due to this sensory involvement many affected individuals developed neuropathic ulcers and Charcot joints and needed amputations. There was prominent and often early motor involvement in most of the affected members of these families. In some cases this was severe and the most prominent feature of the disease, and often these individuals were clinically diagnosed as having CMT.

In comparison to the five (Houlden et al., 2004a) previously reported families with SPTLC1 mutations and one family with linkage to chromosome 9 region (Table 5), the clinical manifestations of the HSAN I families reported here are similar. We have additionally identified several poorly recognized features including the variable age of onset, the lack of disease penetrance, the dissociated sensory loss and in some families the more severe and earlier onset in males, with significant motor involvement and demyelinating motor conduction velocities. There were few autonomic features but the more severely affected individuals tended to have some involvement. Family 5 had prominent diaphragmatic weakness in cases II-2 and III-3. The affected daughter (IV-2) is still below the age of onset when respiratory signs were seen in other family members. The possible cause of these problems in Family 5 is phrenic nerve involvement affecting severe cases of HSAN I, but this has only been identified in the one family.

View this table:
Table 5

Clinical comparison between published HSAN I families linked to the SPTLC1 gene or with known mutations

FamilyAffectedSPTLC1 mutationMean age of onset (years)First symptomsPositive sensory featuresNegative sensory featuresWeaknessSensory lossFoot ulcerationUpper limb involvement
The E family or Family 1 (Australian) (Wallace, 1970; Dawkins et al., 2001)41 (3 NP)C133W65 (range 14–60) (n = 38)Ulceration, blister, painless injury7 cases (n = 27)24 cases (n = 32)4 cases (n = 38)32 cases (n = 33) (less Pr)35 cases (n = 40)14 cases (n = 38)
English Family X (Campbell and Hoffman, 1964; Dawkins et al., 2001)12C133W20–40 years (n = 12)Hyperkeratosis then blister or ulcerN/AN/A0 cases (n = 6)6 cases (n = 6) (less LT)5 cases (n = 7)1 cases (n = 12)
Whitaker Canadian family (Whitaker et al., 1974; Bejaoui et al., 2001)13 (1 NP)C133Y20 (range 6–39) (n = 8)Decreased foot sensation, weak ankles1 case (n = 7)1 case (n = 7)6 cases (n = 7) (severe in Cases 3 and 4)7 cases (n = 7) (later Pr involved)6 cases (n = 7)7 cases (n = 7)
Belgian family (Verhoeven et al., 2004)2G387A14 (range 13–15) (n = 2)Gait problems2 cases (n = 2)2 cases (n = 2)2 cases (n = 2)2 cases (n = 2)2 cases (n = 2)2 cases (n = 2)
French HSAN family (Dubourg et al., 2000)5 (1 NP)Linked to ch9 region33 (range 20–52) (n = 5)Sensory loss, toe weakness, injury1 case (n = 5)1 case (n = 5)5 cases (n = 5)5 cases (n = 5) (less Pr)4 cases (n = 5)5 cases (n = 5)
  • N = Number of cases, NP = number of individuals where the gene showed lack of penetrance, LT = light touch, N/A = not available, ch = chromosome, Pr = proprioception.

In keeping with the clinical heterogeneity the neurophysiological findings showed considerable intra- and inter-familial variability. There was evidence for a predominantly axonal and sensory peripheral neuropathy. A significant motor neuropathy was also clearly present in most patients. Motor conduction velocities were often normal but intermediate slowed velocities hinting at a demyelinating process were seen. This degree of slowing could be explained by the loss of large fast-conducting axons or the presence of significant numbers of regenerating poorly myelinated fibres with short internodes. Some individuals (Family 2 III-2 and IV-1 and Family 6 II-11, III-3 and III-4), however, showed motor conduction velocities in the frankly demyelinating range with well-preserved motor response amplitudes suggesting a demyelinating contribution to the neuropathy in these individuals. The possibility of demyelination was most convincing in subject Family 6 II-1 in whom motor conduction block and motor response dispersion was seen, although this patient probably had superimposed CIDP as discussed above. An alternative explanation for this striking motor slowing could be that of abnormal axonal function and altered excitability rather than a defect of myelin itself as only little and equivocal evidence of demyelination has been seen on nerve biopsies. The variable neurophysiological findings and the presence of motor slowing were noted in the previously reported HSAN I families (Table 6) but did not receive detailed comment in the published reports. Our findings also highlight a possible sex difference with males being more likely to show demyelinating features and females relatively less large fibre sensory involvement electrophysiologically. The significance of these observations is unclear and requires further investigation.

View this table:
Table 6

Neurophysiological and pathological comparison between published HSAN I families linked to the SPTLC1 gene or with known mutations

Ulnar MNCV (m/s)Ulnar CMAP (mV)Ulnar DML (ms)Median MNCV (m/s)Median CMAP (mV)Median DML (ms)Peroneal MNCV (m/s)Anterior Tibial MNCV (m/s)Sensory SNAP (μV)Nerve biopsy
Whitaker Canadian family Cases 1–5, respectively (Whitaker et al., 1974)27, 31, 31, 33, 56N/AN/AN/AN/AN/AN/AND, NR, 24, 27, 314 ulnar absent, 1 ulnar normalIn Case 3 (sural). No myelinated fibres seen, many axis cylinders. Increased fibrous tissue
French HSAN family. Cases 1–6, respectively (Dubourg et al., 2000)37, 50, 61, 42, 45, 310.4, 11.3, 5, 0.6, 0.6, 2.64.2, 2.3, 3.8, 4, 4.2, 4NR, 54, 52, 38, NR, 320.1, 7.8, 3, 0.3, NR, 0.98.2, 3.3, 3.6, 5.4, NR, 4.3Case 2—43, Cases 1, 3–6 NRN/ACases 1, 4–6 absent, Case 3 reduced ulnar, median and suralN/A
  • MNCV = motor nerve conduction velocity, SNAP = sensory nerve action potential, CMAP = compound muscle action potential, DML = distal motor latency, m/s = meters per second, ms = milliseconds, m/V = millivolts, mV = microvolts, N/A = not available, ND = not done.

The electrophysiological findings mirrored the significant loss of myelinated axons seen on nerve biopsy and at post-mortem examination although there was no pathological evidence of demyelination. Family 3, case II-2 had an age of onset of 15 years and died at the age of 59. He had a sural nerve biopsy at age 27 years which showed very similar changes to Family 1, III-14 who was biopsied at the age of 40 years, and Family 4, IV-6 whose sural nerve was examined post-mortem at age 80 years. In none of these cases was there evidence of nerve fibre regeneration within the sural nerve. There were also few signs of active degeneration. So few fibres remained that meaningful morphometry was not possible. The degenerative changes at the level of the dorsal root ganglia in Family 4, IV-6 were less extensive than those described in the single previously reported comprehensive post-mortem study of HSAN I reported by Denny-Brown (1951). Despite our finding of frequent motor involvement in HSAN I, post-mortem examination of several muscles in Family 4, IV-6 was noteworthy for the paucity of grouped fibre atrophy in most muscles examined, suggesting that motor nerve fibre degeneration and regeneration had been limited in this particular patient.

The HSAN I families identified here have the same SPTLC1 mutation and in the families analysed are likely to all have the same common founder. We would expect the defective glucosyl ceramide synthesis to be the same in all affected individuals and thus to lead to similar disease onset and clinical features. The similar peripheral nerve pathology between families also suggests that the genetic defect causes a similar degenerative process but once the nerve is severely affected, comparison is difficult. In fact, there was marked clinical and electrophysiological heterogeneity within and between families. The extent of this is unusual but makes an interesting comparison with the CMT 1a families with the chromosome 17 duplication (Thomas et al., 1997). If the heterogeneity was only between families, then the involvement of other neuropathy-modulating genes would be a possibility. This could account for the familial tendencies for prominent involvement of the diaphragm and the early cognitive decline in two individuals. The differences between affected individuals within families suggest that acquired as well as genetic factors could also influence the clinical variability.

In Family 6 patient II-1 may have superimposed CIDP that could have contributed to his severe and relatively rapid-onset neuropathy. Superimposed CIDP has been observed in other forms of genetic neuropathy (Ginsberg et al., 2003) but further observations of HSAN I families are needed to confirm this. In particular, the high CSF protein may suggest an inflammatory neuropathic component but could be part of the C133W SPTLC1 phenotype as we have seen this in two cases. The difference between males and females in some families is fascinating and raises the possibility of hormonal influences or sex-linked genetic factors. The mechanism of action of SPTLC1 mutations is not fully understood. When this process is defined it will give valuable information on the mechanism of peripheral nerve degeneration as well as some insight into the broad clinical heterogeneity in HSAN I.

The identification of previously poorly recognized clinical features in these families is very important to help clinicians identify affected families in the future. Although clinicians will consider this diagnosis in typical families, HSAN I secondary to the C133W (and perhaps other) SPTLC1 mutations should be considered in patients with a motor and sensory neuropathy with suggestive demyelinating features especially if there is marked sensory involvement and a suggestion of a family history.


We are grateful to The Wellcome Trust, The Medical Research Council and The Mason Medical Research Foundation for their support. We are grateful to Prof. Nick Wood, Prof. Richard Hughes, Dr Wren, Dr Mwenda, Dr Jestico, Dr McLean, Dr Murphy, Dr Rajabally and Dr Damien for referring cases and families. We also thank the families for their assistance with this work.


View Abstract