The hereditary sensory and autonomic neuropathies (HSANs) are rare. Dyck et al classified HSANs into five types according to their clinical presentation1: HSAN I is dominant, starting in the second to fourth decade with sensory loss and foot ulcers; HSAN II is recessive, starting in infancy with lack of many sensory modalities; HSAN III (Riley–Day syndrome) is recessive, has autonomic dysfunction and loss of pain and temperature sensation; HSAN IV is recessive congenital insensitivity to pain with anhidrosis; and HSAN V (OMIM 608654), also called congenital insensitivity to pain without anhidrosis, is the rarest of the five types. The symptoms of HSAN V are similar to type IV with selective loss of pain and temperature sensation, but with intact sweat function and few other autonomic deficits.1^–3 Loss of pain perception results in painless fractures, bone necrosis, osteochondritis and joint destruction. The endpoint is severe Charcot arthropathy.

The classification of HSAN is changing as new genetic mutations are being described. Gene loci have been identified for most of the HSAN types. At least eight loci are associated with HSAN and six genes have been identified to date.4 5 Type IV is caused by a mutation in the nerve growth factor receptor gene (NTRK1).6 7 We recently reported on a Swedish family, clinically best described as HSAN type V, having a point mutation in the nerve growth factor beta (NGFβ) gene on chromosome 1p11.2-p13.2.8 The severe form of the disease is inherited in an autosomal recessive mode. We previously described three homozygous patients with the severe type of the disease starting in childhood and three heterozygous adult patients with Charcot arthropathy.9

The aim of the present study was to describe clinical symptoms and signs in relation to neurophysiological and neuropathological findings in patients with the NGFβ mutation in the heterozygous form.

PATIENTS AND METHODS

Patients

Starting with three severely affected homozygous patients treated at Gällivare Hospital, Sweden, we screened the pedigree of 105 individuals and identified 40 heterozygous subjects. We excluded one patient with both an NGFβ mutation and Charcot–Marie–Tooth disease (CMT1A). One patient with Charcot ankle arthropathy declined examination. Among the previously published heterozygote patients,9 two had died and one old man with severe heart disease refused examination. Six others also declined participation and we excluded two patients with diabetes mellitus and one with Guillain–Barre syndrome. Of the remaining heterozygotes, 26 (12 men, mean age 50 (23) years; range 13–90) were available for examination (table 1). Another 27 relatives (15 men, mean age 42 (22) years (range 15–84)) from different family branches of the pedigree, but without the mutation, served as controls in the clinical examination part. All studied cases and controls gave their informed consent to participate. The local ethics committee approved the study.

Clinical examination

The Michigan Neuropathy Screening Instrument (MNSI) consisting of a self-administered questionnaire and a brief physical examination protocol10 were used for all 53 cases (26 heterozygous cases and 27 controls). The same neurologist (MF) performed a clinical neurological examination, including monosynaptic and corneal reflexes, muscle strength, test for ataxia and test for pain (pinprick), heat, cold, tactile and vibration sensation, in a blinded fashion. Autonomic function was evaluated with an orthostatic test. The tilt test was performed with passive tilting from a supine position after 5 min of rest to 60° upright. Heart rate and blood pressure were recorded before tilting, directly after tilting and 3 min after tilting.

An orthopaedic examination, including radiographs of the knees, ankles and lumbar spine, was done. Neuropathic joint disease was classified according to Koshino (stadium I–III).11 Clinical signs of carpal tunnel syndrome (CTS) were tested with the Tinel and Phalen tests in cases with neurophysiological signs of CTS. Any presence of thenar atrophy was noted.

Genetic analysis

Genomic DNA was prepared from whole blood using standard salt methods. For the mutation analysis, exon 3 of the NGFB gene was amplified by standard PCR using the primers, forward 5′-GAA TTC TCG GTG TGT GAC AGT GTC AGC GTG-3′ and reverse 5′-CCC CTC CCT ACC TCA ACC TGT AAA TTA TTT T-3′. PCR products were purified using multiscreen PCR plates (Millipore) according to the manufacturer’s instructions. Sequencing was performed using Big Dye 3.0 kit (Applied Biosystems, Foster City, California, USA) according to the manufacturer’s instructions using the primers 5′-GTG AAC ATT AAC AAC AGT GTA TTC-3′ or 5′- AAA ATA ATT TAC AGG TTG AGG TAG GG-3′. Labelled products were resolved through 36 cm capillary arrays using POP-4 polymer on an ABI 3100 DNA sequencer and analysed using the Sequencing Analysis 3.7 software (Applied Biosystems).

Neurophysiological examinations

Motor and sensory nerve conductions were measured using conventional neurophysiological techniques (Viking Select/Quest; Nicolet Biomedical, Madison, Wisconsin, USA) in 23 of the heterozygote patients. Care was taken that skin temperature was maintained at or above 30°C at the base of the second digit in the hand and at or above 29°C at the dorsum of the foot. Motor conduction velocity and amplitude were measured in the median, peroneal and tibial nerves, as were F wave latencies in the median and tibial nerves. Sensory nerve conduction velocity and amplitude were recorded in the median, radial and sural nerves. In nine patients, additional recordings were performed in the ulnar nerve and/or in the contralateral median nerve. For each recorded parameter the deviation from the laboratory’s reference value (correlated to age and/or height) was expressed in SD, and an index consisting of the average of the SDs of 12 parameters was calculated.12 If a single parameter was clearly more deviant than the others, this was excluded and a new index based on the remaining parameters was calculated. Similarly, if electrophysiological signs of CTS were present (prolonged distal motor and/or sensory latency in the median nerve), a new index based on all parameters except those from the median nerve was calculated. An index of less than –2.5/√N (N = number of parameters), together with at least two abnormal values (<−2.5 SD) from different nerves, was interpreted as an electrophysiological sign of peripheral neuropathy.

Temperature thresholds for warm and cold stimuli were also measured in 23 of the heterozygote patients according to the method of levels.13 With the same equipment and at the same areas, the thresholds for heat pain and cold pain were examined. The procedure was repeated five times and the mean value was taken as the pain threshold. If the technician judged the results as too variable, a new session was performed. The maximal and minimal temperatures were 50°C and 0°C, respectively. If the patient did not signal pain, the temperature automatically returned to baseline after reaching these levels.

In seven patients the sympathetic skin response (SSR) was recorded. Single electrical shocks, well above the motor threshold level, were delivered over the right median nerve and the SSR was recorded over the dorsum and palm of the left hand.14 The subjects were informed about the procedure, but not told when the stimulation was to be performed. If SSR was absent, the result was judged as abnormal.

Neuropathological studies

Biopsies from the sural nerve were sampled under local anaesthesia from behind the lateral malleolus from six selected heterozygous patients, three with neuropathy/Charcot arthropathy and three without any symptoms. Two to four nerve fascicles were dissected and a 3–4 cm long piece was sampled for analysis. One part was snap frozen and 6–8 μm thick cryostat sections were used for immunohistochemical analyses. One part was fixed in 2.5% glutaraldehyde in Sorensen’s phosphate buffer, pH 7.4, and sampled for paraffin and plastic embedding as well as for teasing. Paraffin longitudinal and transverse sections (3 μm thick) were stained with haematoxylin–eosin and Luxol fast blue. Semithin (1 μm thick) plastic sections were stained with toluidine blue. Furthermore, ultrathin sections for electron microscopy were processed. Morphometric analyses were performed using a Nikon Eclipse 600 Microscope equipped with a 100× (fibre measurements) and 20× (nerve fascicle measurements) objective and a Nikon D-eclipse C1 digital camera (Teknooptik AB, Stockholm, Sweden). The images were transferred to an hp 2025 Compact computer (Teknooptik AB, Stockholm, Sweden). The transverse area of the selected nerve fascicles was photographed and the photocopies were used for identifying myelinated nerve fibres. The perimeter of each myelinated nerve fibre was measured using Eclipse Net software (V.1.20, Laboratory-imaging; Teknooptik AB, Stockholm. Sweden). All myelinated nerve fibres in the selected nerve fascicles were measured. The perimeter of the nerve fibres was recalculated to a circular diameter. Approximately 1000 (range 953–1265) myelinated nerve fibres were measured in each biopsy. The perimeter of each separate nerve fascicle was measured and fibre density (number of myelinated nerve fibres/mm2) was calculated.

Statistical methods

Mean (SD) values for fibre density and circular diameter were calculated for each biopsy separately. The grand mean (SD) value of the six sural nerve biopsies were compared with our normal material, consisting of sural nerve biopsies from 10 normal individuals, with a median age of 46 (range 18–61) years using the Student’s t test and Mann–Whitney U test.15

RESULTS

The clinical presentation and neurological findings for the heterozygous patients are summarised in table 1, sorted by age.

MNSI questionnaire and physical assessment

We found no differences in heart, lung, cerebral or other neurological symptoms or other diseases between the heterozygous patients and controls. Seventeen heterozygous patients and seven controls had osteoarthritis. In the heterozygous group, the MNSI history score was 2.00 (1.53) compared with 1.37 (1.88) in controls (NS). The MNSI status score was 1.79 (1.70) in the heterozygous cases compared with 1.28 (1.55) in controls (NS). However, when the patients and controls were dichotomised for age, the oldest (>54 years old) had more joint manifestations and pathological MNSI scores with MNSI history 2.18 (1.66) (p = 0.06) and MNSI status 2.90 (1.75) (p = 0.10).

Neurological and orthopaedic investigation

Seven heterozygous patients and two controls showed clinical signs of peripheral neuropathy. Patients with neuropathy had a mean age of 78 (9) years compared with the mean age of all heterozygous patients of 50 (23) years. Six heterozygous patients had a pathological orthostatic test compared with only one in the control group. The corneal reflex was normal in all patients. Sensation for pinprick, touch, heat and cold was normal in the young heterozygous patients, and decreased in a few of the old heterozygous patients. Sensation for vibration in the great toe was reduced in six and absent in 13 heterozygous cases. Ten patients had clinical symptoms typical of CTS, including positive Tinel and Phalen tests. Charcot arthropathy was diagnosed in eight heterozygous patients, but in none in the control groups. One illustrative case is shown in fig 1. Mean age for those with Charcot joints was 69 (15) years (range 51–90) compared with 44 (22) years (range 13–77) in those without Charcot arthropathy (p = 0.17). Localisation of Charcot joints was eight knees (two bilateral), one ankle, one metatarsophalangeal joint and one spine. One joint were classified as stadium I, two joints as stadium II and eight joints as stadium III, according to Koshino (fig 1A, 1B).11 Three patients had been treated with total knee replacements and one with ankle artrodesis. No mutilating symptoms, autoamputations or painless fractures were found. Lumbar disk degeneration was present in 10 patients and in eight controls while two patients had scoliosis. In total, 10 heterozygous patients had clinical symptoms or signs of neuropathy/CTS and/or Charcot arthropathy, 10 were asymptomatic and six had CTS exclusively. The asymptomatic patients were, as a rule, younger. Mean age of the asymptomatic patients was 38 (19) years compared with patients with Charcot arthropathy and/or neuropathy 69 (13) years.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Case No 29 with Charcot arthropathy in the left knee (Koshino stadium III), joint destruction and varus deformity. (A) Frontal and (B) side knee radiogram.

Neurophysiology

Twelve of the 23 patients had various nerve conduction abnormalities (table 2). Four patients, all older than 65 years, had neurophysiological signs suggesting polyneuropathy. One 90-year-old man (case No 26) had a severe neuropathy (index of –6.92) while one had moderate and two slight changes (index −2.35, −1.67 and –1.61). These changes included reduced motor and sensory amplitudes (most marked for the sural nerve) and reduced conduction velocities and/or prolonged F wave latencies. Three of the patients with electrophysiological signs of polyneuropathy also had alterations suggesting CTS, as had eight further patients without signs of polyneuropathy. Pathological quantitative temperature tests were found in five patients. The threshold for cold perception in the foot was increased in four patients. The level for heat pain was increased in 2/23 and for cold pain in 5/26. SSR was normal in all examined patients.

Nerve biopsies

A moderate loss of both thin myelinated nerve fibres (fig 2) and unmyelinated fibres (fig 3) was observed in all six biopsied heterozygous cases (mean age 64 years (range 45–84)). Mean fibre density was 6.4×103 (1.2×103) fibres/mm2 and mean circular diameter was 7.5 (0.6) μm. Ten sural nerve biopsies considered normal, from 10 patients (seven men, three women; mean age 46 years (range 18–61)) who were given diagnoses other than polyneuropathy on follow-up examinations served as controls.15 Corresponding values from controls were 9.6×103 (1.5×103) fibres/mm2 and 5.8 (0.6) μm. Statistical analysis gave p values of 0.001 for fibre density and 0.0001 for circular diameter (table 3). The myelinated fibre density was low (below −2 SD of the 10 healthy controls) in all patients except for the youngest, a 45-year-old man. Mean circular diameter was increased (>2 SD) in all patients and, as a consequence, the ratio of myelinated nerve fibre diameter <7 μm/>7 μm was decreased (p = 0.002) because of the selective reduction of thin myelinated fibres. A large intraindividual variation of fibre density was present between separate nerve fascicles.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Sural nerve biopsies, transverse section of a nerve fascicle stained with toluidine blue. (A) Normal fascicle from a healthy control. (B) Moderate reduction of small diameter myelinated (Aδ) fibres and normal large diameter myelinated fibres from a patient (case No 18) heterozygous for the NGFB mutation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Electron microscopy picture of the sural nerves. (A) Healthy control showing normal A and C nerve fibres (bar, 2 µm). (B) Very sparse unmyelinated C fibres and both large and small myelinated fibres from a patient (case No 22) heterozygous for the NGFB mutation (bar, 2 µm).

A comparison of symptoms, signs, neurophysiological findings and biopsy data between homozygous patients,9 heterozygous patients (old and young) and controls is shown in table 4.

DISCUSSION

The neurological signs varied widely among the heterozygous patients from those with multiple Charcot arthropathies to asymptomatic cases. Only four heterozygous patients, all >65 years old, had an Electroneurography Index suggesting polyneuropathy in large myelinated fibres also. All six biopsied patients had small fibre neuropathy with reduction of small myelinated fibres (Aδ) and unmyelinated fibres (C), important for both pain and thermal sensations and for autonomic functions. However, only 6/23 patients had impaired temperature thresholds and 6/26 had a pathological orthostatic test. MNSI history and status scores were pathological only in the older (>50 years) patients. Thus the degree of peripheral nerve changes did not seem to have a strong impact on clinical presentation, until older age. The meagre clinical findings in young asymptomatic cases may be because our diagnostic tools were not sensitive enough to diagnose the discrete alterations caused by small fibre neuropathy. Genetic background and lifestyle factors such as physical activity may also have influenced the outcome.16 In contrast with most other types of HSAN, none of our heterozygous patients was mentally retarded. Also, we found no pure insensitivity to pain or any anhidrosis, as previously reported in some of these disorders.1 6 25

Charcot arthropathies were exclusively found in the lower extremities, probably because of the fact that natural loading of these joints together with impaired deep pain perception may enhance joint destruction. Injury and excess use may play a role in the genesis of arthropathy.16 The presentation of Charcot joint stadium I may be difficult to differentiate from osteoarthritis although the deformity seems more of a problem for the patient than pain. Three of our cases were treated later on with joint replacement but we were not aware of the final diagnosis on that occasion. Looking at this treatment today, perhaps other types of procedures would have been preferable, but to date no major complications have emerged.

Almost half of our patients had clinical and neurophysiological signs indicating CTS. Furthermore, CTS was accompanied by neuropathy and/or Charcot arthropathy in about half of the cases. A child with HSAN IV has been described with CTS. His relatives also had CTS and the authors suggested that the child may be homozygous, and that HSAN IV in the heterozygous form predisposes to familial CTS.17 The mechanism is still unclear.

The reduction in both Aδ and C fibres occurs in HSAN IV and HSAN V and clinically both types demonstrate joint deformities and painless fractures. It is not unlikely that individuals heterozygous for mutations in TRKA might also have a reduction of thin sensory fibres, in common with the individuals examined in this study, but the loss might not be large enough to present clinically. The TRKA mutation in HSAN IV is likely to be a partial loss of function mutation and we believe that this is also applicable to our patients with a NGFB mutation. In our cases, unlike type IV, there were few symptoms indicating loss of sympathetic neurons and our patients also had milder autonomic dysfunction.18 19 Our patients also seemed to have no visceral pain associated with the autonomic nervous system, as judged from their case histories.

Thus the NGF mutation found in our family with a homozygous HSAN V-like clinical picture gives a different and milder neurological and orthopaedic phenotype than the NGF receptor NTRK1 mutation in HSAN type IV in its homozygous form (table 5). However, unlike what has been shown for individuals heterozygous for mutations in TRKA, individuals heterozygous for the NGFB mutation often develop symptoms as adults.