OUP user menu

Highly variable neural involvement in sphingomyelinase-deficient Niemann–Pick disease caused by an ancestral Gypsy mutation

Violeta Mihaylova , Janina Hantke , Ivanka Sinigerska , Silvia Cherninkova , Margarita Raicheva , Sonja Bouwer , Radka Tincheva , Djako Khuyomdziev , Jaume Bertranpetit , David Chandler , Dora Angelicheva , Ivo Kremensky , Pavel Seeman , Ivailo Tournev , Luba Kalaydjieva
DOI: http://dx.doi.org/10.1093/brain/awm026 1050-1061 First published online: 14 March 2007


Niemann–Pick disease (NPD), an autosomal recessive disorder resulting from mutations in the sphingomyelin phosphodiesterase 1 (SMPD1) gene, is subdivided into the acute, lethal neuronopathic type A, and the chronic visceral type B, explained by the different residual activity levels of acid sphingomyelinase (ASMase). An increasing number of reports on intermediate forms, challenging this traditional clinical classification, have described a broad range of neurological manifestations; however genotype–phenotype correlations have been compromised by relatively small sample sizes and/or allelic heterogeneity. Here we present a genetically homogeneous group of 20 Gypsy patients with intermediate NPD, where we observed a surprising diversity of neurological features. All affected subjects were homozygous for the same ancestral mutation, W391G in SMPD1, yet displayed the entire spectrum of phenotypic variation observed previously in unrelated affected subjects of diverse ethnicity and disease-causing mutations, ranging from subclinical retinal involvement to severe ataxia, cognitive deficits and psychiatric disorders. The clinical heterogeneity of W391G homozygotes points to additional factors, beyond SMPD1 and residual ASMase, which determine the localization, extent and severity of neural involvement. The phenotype similarity of affected relatives suggests a possible role of genetic modifying factors. In practical terms, W391 is common in the Gypsy population and the diagnosis of NPD should be borne in mind despite the atypical course of the disease. Generally, our findings indicate that mutation analysis is of limited value in predicting brain damage, and the option of enzyme replacement therapy should be considered in intermediate NPD.

  • intermediate Niemann–Pick disease
  • neurological manifestations
  • Gypsy founder mutation


Niemann–Pick disease (NPD) is an autosomal recessive disorder, caused by acid sphingomyelinase (ASMase, EC deficiency and an ensuing accumulation of sphingomyelin and other lipids in cells of the monocyte–macrophage lineage and in ganglion cells in the central nervous system (Schuchman and Desnick, 2001). ASMase, encoded by the sphingomyelin phosphodiesterase 1 (SMPD1) gene on chromosome 11, is a lysosomal hydrolase responsible for the breakdown of sphingomyelin into ceramide and phosphocholine (Kanfer et al., 1966; Schuchman and Desnick, 2001). ASMase is also involved in the ceramide signalling pathway and thus implicated in the cellular stress response and the induction of the apoptotic pathway (reviewed in Kolesnick, 2002). Traditional classifications divide NPD phenotypes into type A (OMIM 257200), a fatal neurodegenerative condition with onset in infancy and death within the first 2–3 years of life, and a slowly progressing non-neuronopathic type B (OMIM 607616) with visceral involvement of variable severity, compatible with survival into adulthood (Schuchman and Desnick, 2001). Occasional early reports on individual patients with intermediate, protracted neurovisceral forms have questioned the NPD type A–type B dichotomy (Saidi et al., 1970; Sogawa et al., 1978; Elleder and Cihula, 1983; Shah et al., 1983; Elleder et al., 1986; Matthews et al., 1986; Dubois et al., 1990), however generally such cases have been considered to be rare exceptions. Recently, a high prevalence of the intermediate type of NPD was demonstrated in German, Czech and Slovak NPD patients, whose clinical features suggested a continuum rather than two well-defined extreme clusters (Harzer et al., 2003; Pavlú-Pereira et al., 2005). In another recent study of the prevalence and phenotype characteristics of intermediate forms among NPD type B patients, performed at the Mount Sinai School of Medicine in New York, Wasserstein et al. (2006) observed global progressive neurological abnormalities in 8% and minor non-progressive abnormalities in a further 22% of patients classified as NPD type B. While the extent of variation of nervous system involvement and its correlation with SMPD1 mutations and ASMase activity in intermediate cases is of particular interest, assessment is confounded by allelic heterogeneity and a limited number of patients with identical genotypes. The most common SMPD1 mutation associated with an intermediate NPD phenotype to date, Q292K, has been detected in a total of 6 homozygotes and 13 compound heterozygotes of diverse ethnic background and a variety of other mutant alleles in the compound heterozygotes (Harzer et al., 2003; Pavlú-Pereira et al., 2005; Wasserstein et al., 2006).

Here we describe for the first time the characteristics of intermediate Niemann–Pick disease in patients of Gypsy ancestry. Similar to other autosomal recessive disorders in this genetically isolated founder population of limited diversity (Kalaydjieva et al., 2005), NPD is caused by a single ancestral mutation, W391G in SMPD1. The genetic homogeneity of our patients provided a rare opportunity to examine the phenotypic spectrum of intermediate NPD, revealing a surprisingly broad range of nervous system involvement.

Material and methods


The study included 20 Gypsy patients (15 males) from Bulgaria, suffering from Niemann–Pick disease. They were recruited among the subjects diagnosed as NPD at the National Genetic Laboratory, where 25 out of 27 known patients with this diagnosis had declared Gypsy ethnicity. The initial diagnosis was based on clinical and biochemical findings. Enzyme assays included ASMase activity in leucocytes or cultured skin fibroblasts, measured with the artificial substrate hexadecanoylamino 4-nitrophenyl phosphocholine (Gal et al., 1975) and/or chitotriosidase in plasma, using 4-methylumbelliferyl β-d-N,N′,N″-triacetylchitotrioside 3 hydrate as substrate (Guo et al., 1995).

At the time of the most recent clinical examination, the patients were aged between 7 months and 35 years. They belonged to 16 unrelated families from diverse Gypsy sub-isolates, residing in different parts of Bulgaria. Subjects 5 (deceased) and 6 were siblings, as were Subjects 1 to 4 (all deceased), who were the first atypical NPD cases recognized in Bulgaria in the 1970s by one of the co-authors (Khuyomdziev et al., 1979).

The molecular epidemiology of NPD in the Gypsy population was investigated by mutation analysis of 642 de-identified control samples from various sub-isolates in Bulgaria (Kalaydjieva et al., 2005), Romania, Hungary, Serbia, the Czech Republic and Spain.

Written informed consent has been obtained from all subjects, parents or guardians. The study complies with the ethical guidelines of the institutions involved.

Genetic analysis

For the SMPD1 sequencing in DNA samples from affected subjects, the full-length genomic sequence (Schuchman et al., 1992) was amplified using PCR primers SMPD1-1F ACCGAGAGATCAGCTGTCAG and SMPD1-6R CTGTCTGCACTGCTGTGTGC and an eLONGase PCR reaction mix (Gibco BRL Life Technologies). Sequencing of the 4.734 kb PCR product was done with internal primers (available upon request) and ETT chemistry (Amersham). The reactions were run on an ABI 377 DNA Analyzer. The data were analysed with Sequence Navigator, in comparison to the reference mRNA sequence NM_000543 and the chromosome 11 genomic sequence accessed via Genome Browser at http://genome.cse.ucsc.edu.

High-throughput carrier testing was performed with the Amplifluor SNPs HT Genotyping system (Chemicon). For the detection of the wild-type allele we used primer GAAGGTCGGAGTCAACGGATTTTGTTCCCGTGAGAACTTCT labelled with 6-carboxy-4′,5′-dichloro-2′,7′-dimethoxyfluorescein (JOE); fluorescein (FAM)-labelled primer GAAGGTGACCAAGTTCATGCTTTGTTCCCGTGAGAACTTCG detected the mutant allele; and ACCAGCCACTGGAGCTGT served as the common reverse primer. The assay was run in a 384-well plate format on an ABI PRISM 7900HT Sequence Detection System.

Alternatively, small-scale mutation analysis of individual DNA samples can be performed using a PCR-based RFLP assay. The PCR primers GAGGACCAGGATTGGAACAA and CAGTGACCATGAGCTGAATC generate a 365 bp product. The mutation creates a BanII restriction site such that, in mutant alleles, the 365 bp amplicon is cut into two fragments, 141 and 224 bp in length.

Phenotype characterization

Affected families were contacted and visited to discuss the study, examine patients and collect current information and medical history. In the course of the study, nine patients were hospitalized for further investigations. Information on the remaining subjects (including the deceased patients) was compiled from medical records and the recent data collected during field trips.

Detailed physical, neurological and psychiatric examinations of each patient were performed by two independent investigators.

Neuropsychological testing measured general cognitive functioning (Mini-Mental State Examination), basic and higher cognitive functions, namely orientation, memory (verbal learning, short-term memory), visual perception (object, colour, number and letter recognition), praxis (ideomotor and constructive), language (expressive speech, naming, comprehension, reading, writing) and arithmetic. Psychometric assessment was based on Standard Raven's Progressive Matrices for adult patients and Coloured Raven's Progressive Matrices and HAWIK-R, Bulgarian standardization (1996) for children <18 years of age.

Electrophysiological studies included nerve conduction velocity (NCV), elecroencephalography (EEG) and brainstem auditory evoked potentials (BAEPs).

Radiological investigations included X-rays of the chest, hand, wrist and forearm; lung and abdominal CT scans, ultrasonography of the heart and abdomen and brain imaging (CT and MRI scans).

Venous blood samples were drawn for DNA extraction, blood counts and biochemical analyses, including total serum protein, albumin, aspartate and alanine aminotransferases, gamma-glutamyl transpeptidase, alkaline phosphatase, blood glucose, total and conjugated bilirubin, urea, creatinine, electrolytes, fibrinogen, prothrombin time international normalized ratio and lipid profiles.

Bone marrow aspirates were obtained for histological analysis.

Statistical analysis of the variation in neurological manifestations, comparing phenotype differences between sibships and the group of non-familial cases, was performed with ANOVA, including variance component analysis, using SPSS.


Genetic analysis

Our sequencing data, in comparison to the reference NM_000543 sequence, identified a T → G substitution at position 1351, corresponding to c.1177 for numbering starting from the first methionine codon. We observed an additional difference to NM_000543, a known polymorphism (rs3838786) in the number of Leu-Ala repeats in the signal peptide: six repeats in the reference sequence and five in the patients’ samples. Taking into account that variant, the position of the mutated nucleotide is c.1171 T → G, leading to the substitution of glycine for the conserved tryptophane residue at amino acid position 391 (p.Trp391Gly or W391G). The same mutation has been described previously as W391G by Ferlinz et al. (1995) and we will adhere to this nomenclature, as it corresponds to the structure of the gene in Gypsy NPD chromosomes, as well to the original SMPD1 sequence reported by Schuchman et al. (1991).

All our living NPD patients were homozygous for the W391 mutation. It was also confirmed in a stored DNA sample from Patient 5 (deceased). In the ‘historical’ case of the first diagnosed Gypsy NPD family (siblings 1 to 4), the genotype was inferred based on the detection of W391G carriers among the living unaffected first-degree relatives.

A common origin of the molecular defect was supported by the identical intragenic haplotype shared by mutant chromosomes: Val36 (rs12417689)—5 Leu-Ala repeat units in the signal peptide (rs3838786)—C (rs7951904)—T (rs11601088)—A (rs8164) and a novel private variant IVS2 + 731 G > A.

The screening of 642 unrelated control subjects from different Gypsy populations detected seven heterozygotes, equivalent to an overall carrier rate of ∼1.1%. By Gypsy group identity, the highest frequency of 4.6% was found among the Rudari (4 carriers identified among 86 samples from Bulgaria and Serbia). One carrier was also detected among the 36 samples from Romania, where group identity is unknown but is also very likely Rudari. The heterozygote frequency in this sub-isolate can thus be estimated at ∼4%, with the expected incidence of new cases amounting to 1 in 2500 live births.

Clinical phenotypes

Visceral involvement

The signs and symptoms of visceral involvement, as well as their variability, were similar to those described previously in studies of NPD type B patients (Simonaro et al., 2002; Wasserstein et al., 2004). The individual data are summarized in Table 1.

View this table:
Table 1

Visceral involvement in NPD patients homozygous for the W391G mutation

Patient no.SexAge (years) at onsetAge (years) at deathPresenting signs at time of diagnosisDelayed early developmentSphingomyelinase activity in leucocytes or fibroblastsaChitotriosidase in plasmabFindings at most recent examination
Age (years) at examinationLow staturecHepatomegalydSerum AST/ALTeSplenomegalydHypersplenismFoam cells in bone marrowPulmonary infiltrationsHeart valve fibrosisAbnormal blood lipid profile/ HDL-CfOther
 1M1Early adulthood Intellectual deficit, ear infections, hepatosplenomegaly+NDND12++ND+ND+NDNDNDCoarse facial features
 2FNDEarly adulthoodProfound intellectual deficit, hepatosplenomegaly+NDND20ND+ND+ND+NDNDNDJoint ankylosis
 3MNDEarly adulthoodIntellectual deficit, hepatosplenomegaly+NDND16ND+ND+NDNDNDNDNDCoarse facial features
 4FND23Intellectual deficit, hepatosplenomegaly+NDND10ND+ND+ND+NDNDNDCoarse facial features
 5M234Intellectual deficit, ataxia, hepatosplenomegaly+NDND23ND515/84±NDNDIchthyosis vulgaris
 7M<1NARecurrent infections, hepatosplenomegaly61F4901456123/977++++/0.96Kidney ptosis
 8M<1NAHepatosplenomegaly, anaemia14F20871610545/2611++++/0.5Osteoporosis
 9M<1NARecurrent infections23L22552510151/1318++++/0.44Ascending aorta wall thickening, delayed bone age, osteoporosis
10M<1NAHepatosplenomegaly, anaemia+ND67117<36194/1308++++/0.51Ichthyosis vulgaris, delayed bone age, nephrolithiasis
11M<1NAAcute bronchitis, hepatosplenomegaly2L44115<3667/897ND++/0.59
12M<1NAChest infections, hepatosplenomegaly19F1463515793/108*ND+++/0.18Ascending aorta wall thickening Gall bladder polyposis
13M∼3NAAcute bronchitis, hepatosplenomegaly14L 26F18216<3633/387++++/0.52
14M7NAHepatosplenomegaly+23L 32F246312<313ND9±NDNDNDNDDelayed bone age, osteoporosis
15M<1NARecurrent infectionsND36L188021<36ND9NDNDNDND
16F<1NARecurrent infections, hepatosplenomegaly20L 36F155175782/1179ND+++/0.43
17M1NARecurrent infections, hepatosplenomegaly+ND6171.5++273/128++NDNDNDND
18F2NARecurrent infections, hepatosplenomegaly+20L 29F14442.5++299/250+NDNDNDND
19M<1NARecurrent infections, hepatosplenomegaly+17F12410.6++179/111++NDNDNDND
  • NA, not applicable; ND, no data; +, yes; −, no.

  • aSphingomyelinase activity measured with the artificial substrate 2-(N-hexadecanoylamino)-4-nitrophenyl phosphocholine in leucocytes (L), reference range 24–135 nmol/17 h/mg protein (N = 36) and in fibroblasts (F), reference range 102–590 nmol/h/mg protein (N = 86). bReference range 34 ± 49 nmol/h/ml (N = 180). Chitotriosidase in plasma derives from activated macrophages and is used as a marker for the formation of pathological lipid–laden macrophages in NPD [Brinkman et al., 2005]. cIn percentiles of age- and sex-adjusted reference values for the Bulgarian population; (+) denotes patients where growth retardation and/or low stature are mentioned in medical records without specific details. dShown as centimetres below costal margin or as present (+) when no specific data were available. *Patient 12 has been splenectomized. eReference values ALT < 41 U/l, AST < 37U/l. fHDL-C (mmol/l): children desirable >1.2 mmol/l, low <0.9 mmol/l; adults desirable ≥1.6 mmol/l, low <1 mmol/l.

Onset of the disease was within the first 2 years of life in nearly all cases, however the diagnosis was established in early childhood in only 7/20 patients and delayed until age 9 to 27 years in the remainder (Subject 5 was diagnosed as NPD retrospectively, after his death). Initial manifestations included hepatosplenomegaly (invariably present at the first examination), recurrent respiratory infections and delayed developmental milestones.

At the most recent examination, low stature was common: all patients were below the 25th and five were below the 3rd percentile of the reference values. Hepatosplenomegaly was present in all patients with changes in hepatic echogenicity, pointing to diffuse parenchymal involvement, detected by abdominal ultrasound in 6/10. Serum aminotransferase activity (ALT and AST) was increased in 11/14 cases. The increase was not age-related: patients with values within the reference range were 12 to 26 years old, while markedly abnormal values were observed in some paediatric cases aged <2 years. No patient showed current evidence of impaired liver synthetic function, however decreased prothrombin time had been documented previously in three cases. Abnormal haematological indices (white blood cell count and/or platelet count), consistent with hypersplenism, were observed in 9/16 patients with no evidence of age-related progression.

The majority of patients reported recurrent chest infections. Lung infiltrations were detected by recent chest X-rays in 7 out of 10 patients (symptom-free at the time of the examination), and mild fibrotic changes were found in one additional case by a high-resolution CT scan. Pulmonary signs, such as cyanosis or clubbing, were not seen in any of the subjects.

Mild fibrotic changes of the mitral and aortic valves, which have not been described in previous studies, were detected by echocardiography in 6/9 subjects.

The blood lipid profiles were consistently abnormal. High-density lipoprotein cholesterol values were grossly reduced in 9/9 patients (including seven children). The children also presented with abnormally high total cholesterol, low-density lipoprotein cholesterol and triglycerides.

The characteristic Niemann–Pick foam cells were seen in the bone marrow aspirates of seven patients, but were not detected in three further cases (aged 7 months, 2 and 13 years).

ASMase activity in leucocytes, measured with the artificial substrate 2-(N-hexadecanoylamino)-4-nitrophenyl phosphocholine, did not provide definitive diagnostic information. Reduced activity was observed in only three out of eight cases investigated, with a significant decrease found in only one of these (Patient 11). Enzyme analysis in cultured skin fibroblasts was more informative: decreased ASMase activity was found in all the eight patients tested, ranging between 15 and 60% of the lowest value in the normal range. In contrast, plasma chitotriosidase activity was a reliable indicator, showing the expected increase in 15/15 subjects investigated.

Nervous system involvement

Evidence of nervous system involvement was present in 19 of our 20 patients. The exception was a 7-month-old child (Patient 19), where normal development and lack of neurological symptoms and of macular halos were documented during hospitalization in a paediatric gastroenterology clinic.

Both the localization and the extent of neural involvement varied widely. In some patients, it was limited to instrumentally detectable lipid accumulation in retinal cells without any functional consequences, whereas in others it resulted in various neurological abnormalities, psychotic behaviour and cognitive deficits (Table 2 for individual data).

View this table:
Table 2

Nervous system involvement

Patient no.Age (years) at last examinationMacular haloCNS involvementPNS involvement
Clinical examinationIntellectual deficitPsychiatric symptomsEEG abnormalitiesBrain imaging abnormalitiesClinical examinationReduced MNCVReduced CMAP/SNAP amplitude
 523++Ataxia, pyramidal and extrapyramidal signs++++ND
 626+++Pyramidal signs+++++++
1115++Pyramidal signs+++++
1521YesMild ataxiaNDND+NDND
  • ND, no data; +, yes; −, no. Macular halo: + <0.5 DD (disc diameter); ++ 0.5–0.75 DD; +++ >0.75; ‘yes’ indicates the presence of halo without quantitative evaluation. Rating of intellectual deficit (IQ values): – normal/borderline (85–70); + mild (<70); ++ moderate (<50); +++ profound (<30); +* denotes delayed psychomotor development reported by referring paediatricians, no formal assessment of cognitive function. Psychiatric symptoms: Patient 8 (±) has developed psychotic episodes related to hepatic encephalopathy. EEG abnormalities: ± denotes mild changes with diffuse slow theta waves and sharp waves; + generalized slow waves or focal sharp waves; ++ paroxysmal activity or grossly disorganized background activity. Brain imaging: Patient 6—generalized cerebral atrophy; Patient 8—pontocerebellar atrophy; Patient 12—cerebellar atrophy.

Ocular features

Macular halos were the most consistent finding, present in 16 of the 18 patients who had a documented ophthalmological examination. The two cases (Patients 18 and 19) where fundoscopy was unremarkable were aged 2.5 years and 7 months, respectively.

Recent re-examination of nine patients by a single ophthalmologist (S.C.) showed that, while the characteristic silver–grey rings around the brown–reddish foveola were invariably present, they differed in appearance, possibly reflecting lipid accumulation in different types of retinal ganglion cells. In Subjects 6, 7, 8 and 16 (age 14 to 26 years) the rings were homogeneous in density and composed of discrete granules (Fig. 1A). In contrast, in Subjects 11, 12 and 13 (age 15 to 35), the halo appeared stippled with fine white deposits of variable confluence (Fig. 1B). In the youngest patient examined (Patient 9, age 5 years), a small, cloudy, greyish–white halo was seen, which was poorly demarcated from the adjacent retina (Fig. 1C).

Fig. 1

(A) Fundus photograph of Subject 6 showing macular halo with discrete granular appearance. (B) Fundus photograph of Subject 11. Note the scattered white deposits around the foveola which do not form a homogeneous ring. (C). Fundus photograph of Subject 9 demonstrating an opaque poorly demarcated macular halo without granules.

We used the ratio between the outer diameter of the halo and the diameter of the optic disc as a quantitative measure to assess the evolution of the ocular changes. The mean ratio ± SD for the overall group was 0.75 ± 0.28. An age-related progression in the size of the halo was evident, with the ratio increasing from 0.33 in the youngest patient (aged 5 years) to 1.1 and 1.2 in the two oldest (aged 26 and 35 years).

All patients had normal corrected visual acuity. The anterior segment was intact and extraocular movements were preserved.

Involvement of the central nervous system

Neurological abnormalities

Cerebellar involvement was evident in four affected subjects. Ataxia was the most prominent and disabling symptom in Subject 5 from early childhood. Less severe abnormalities of volitional movements (dysmetria and bilateral intention tremor on finger-to-nose testing) were detected in Subjects 8 (age 16 years) and 12 (35 years). The latter also had bilateral intention tremor on knee-to-heel testing together with dysdiadochokinesis at pronation–supination of the forearm. Subject 15 had intention tremor when running the heel down the front of the shin. Brain MRI showed cerebellar atrophy and a small pons in Patient 8, and cerebellar atrophy and enlargement of the cisterna magna in Patient 12 (Fig. 2). No neuroimaging studies were possible in Patients 5 and 15.

Fig. 2

Brain MRI of Patient 12 demonstrating cerebellar atrophy.

A lesion of the corticospinal tracts was evident in three patients. Subjects 5 and 6 presented with absence or asymmetry of the superficial abdominal cutaneous reflexes. Subject 11 had a positive Babinski sign on the right, together with upper motoneuron paresis of the right lower limb and pes equinovarus.

Extrapyramidal signs were observed only in Subject 5, who had athetosis affecting all four limbs.

Psychiatric manifestations

Impulsive aggressive behaviour and sexual disinhibition were observed in Subjects 5 (age 23 years) and 11 (age 15 years), both also classified as having a mild intellectual deficit.

Patient 12 was diagnosed with seasonal affective disorder (periodic depression) at age 30 years. He has had a total of three episodes of depression accompanied by persecutory delusions. Risperidone treatment was initiated during the third episode, resulting in complete remission.

Patient 6 developed hallucinations and paranoid delusions at age 23 years and has been treated with a variety of neuroleptic drugs (clozapine, haloperidol, chlorpromazine, fluphenazine). During the recent hospitalization, at age 26, she had psychotic behaviour, with ideas of influence, persecutory delusions and auditory and somatic hallucinations leading to the diagnosis of schizophrenia-like psychosis. Three months of risperidone treatment resulted in only minor improvement. In this patient, a brain CT scan at age 10 years failed to detect any abnormalities, however repeated neuroimaging (CT and MRI) at age 26 years documented generalized cortical atrophy.

Cognitive function

Out of the 16 individuals where detailed data could be obtained, cognitive and social functioning was preserved in six (Subjects 7, 9, 12, 13, 15 and 16), aged 5 to 35 years. All (except the youngest, Subject 9) had completed primary education in a general school, and had IQ values within the low normal to borderline range (70–85) with developed reading, writing and simple arithmetic skills and abstract reasoning at different levels of complexity. Subject 9 had appropriate receptive and expressive language, object-, colour-, body parts- and form-recognition and ideomotor and fine motor praxis.

Ten patients were classified as cognitively deficient following ICD-10 criteria. Four (Subjects 3, 4, 5 and 11), aged 10–23 years, had mild and mild-to-moderate deficit with simple conversational speech (some with articulation defects), difficulties in dealing with money and symbols (letters, numbers), lack of abstract reasoning, fine motor and constructive apraxia and problematic or not fully independent social functioning. One (Subject 5) could read letter-by-letter and do simple arithmetic (adding and subtracting to 10) and coped with daily living activities. Four patients (Subjects 1, 6, 8 and 10), aged 12–26 years, were classified as moderately deficient; they presented with simplified, stereotypic, sometimes agrammatic or inappropriate speech, inability to deal with money, poor time and space orientation, immature reasoning and behaviour, and limited, dependent social functioning. Two patients (Subjects 2 and 14), aged 12 and 20 years, had profound intellectual deficit with severely impaired verbal and non-verbal communication, motor stereotypes and fully dependent behaviour.

Electroencephalographic data

Standard EEG recordings and brain mapping were done in 10 patients (Subjects 6 and 12 were on risperidone medication at the time of the study). Spontaneous electrical activity was normal in all except Subject 5. In eight patients, the recordings revealed diffuse slow θ waves and sharp waves. The EEG was manifestly abnormal in five cases: Subject 5 had grossly disorganized background activity with multiple θ and δ waves, particularly over the frontal regions. Subject 16 showed generalized slow wave (4 Hz) activity predominantly in the frontal and occipital regions. Focal sharp-wave activity in the right temporal and parietal regions was detected in two patients—Subjects 8 and 11. Two individuals had paroxysmal activity: Subject 6 displayed generalized sharp waves with a frontal maximum, while Subject 11 had parietal sharp wave discharges tending to generalize.

Brainstem auditory evoked potentials (BAEPs)

BAEP recordings (click stimuli at 80 dB hearing pressure level with amplifier band frequency at 200–2000 Hz) were normal in 7/7 patients. The mean latency values ± SD (reference values in brackets), combining right and left side, were as follows: wave I 1.57 ± 0.11 (1.6 ± 0.2); wave III 3.74 ± 0.13 (3.7 ± 0.2); wave V 5.69 ± 0.1 (5.8 ± 0.2); I–III interpeak latency 2.2 ± 0.1(2.2 ± 1.48); III–V interpeak latency 1.96 ± 0.17 (1.95 ± 1.49); I–V interpeak latency 4.06 ± 0.11 (4.15 ± 0.22). The wave V/I amplitude ratio was also normal in all individuals examined.

Involvement of the peripheral nervous system

Clinical evidence of mild peripheral neuropathy, with more pronounced motor involvement, was obtained in 2 out of 12 patients examined recently. Subject 7 presented with lower limb areflexia and symmetrically reduced tendon reflexes in the upper limbs. Sensory testing revealed a slightly diminished sense of vibration at the ankle and the tibial tuberosity. Patient 8 had bilaterally depressed tendon reflexes in the upper limbs. Distally accentuated lower limb hyperaesthesia in response to pinprick stimuli, suggesting involvement of the small myelinated sensory fibres, was recorded in Subjects 13 and 15.

Nerve conduction studies showed abnormalities in 8/10 patients. A measurable, albeit small, reduction in the motor nerve conduction velocity (MNCV) was seen in six patients (excluding measurements in Subjects 12 and 13, who had co-existing vertebral L4-L5 and L5-S1 disc herniation). The mean MNCV ± SD was 43.68 ± 0.75 m/s for the median (reference value >46 m/s), 49.5 ± 2.54 m/s (reference >54 m/s) for the ulnar and 37.48 ± 4.57 m/s (reference >46.2 m/s) for the peroneal nerve. Distal latencies were normal. Low amplitudes of compound muscle action potentials were observed in seven subjects, with mean values ± SD of 1.21 ± 0.99 mV (reference >4 mV) at median nerve, 2.73 ± 0.61 mV (reference >5 mV) at peroneal nerve and 2.53 ± 1.26 mV (reference >5 mV) at tibial nerve stimulation. Abnormal amplitudes of the sensory nerve action potentials (SNAPs) were seen in eight patients. Low SNAP values in the median nerve, found in three individuals, were 12.1 to 17.2 µV, the mean ulnar nerve values were 13.86 ± 1.97 µV and those for the sural nerve were 10.36 ± 4.23 µV (reference values for all three are >20 µV). Sural SNAPs were unobtainable in Subject 16.

Analysis of variance of neurological phenotypes in W391G homozygotes

To examine the contribution of genetic modifying factors to phenotype variation, we compared the manifestations of nervous system involvement between four groups of patients: non-familial cases, family 1 (the sibship of Patients 1 to 4), family 2 (patients 5 and 6), as well as the original family from Serbia with six affected subjects, where the W391G mutation was first identified (Sperl et al., 1994; Ferlinz et al., 1995). We included six variables, namely macular halos, intellectual deficit, clinically detectable neurological manifestations indicating CNS involvement, EEG abnormalities, psychiatric symptoms, and manifestations of PNS involvement (clinically detectable abnormalities and reduced MNCV). All children before prepubertal age were considered unknown for psychiatric manifestations. The rating used was as shown in Table 2, except for macular halos, where only presence or absence was considered, ignoring age-related progression.

The ANOVA results showed no significant differences between the four groups as regards the presence of macular halos (P = 0.454) and intellectual deficit (P = 0.098). In the case of CNS involvement, between-group variation was significantly higher than within-group variability (P = 0.002 for clinical manifestations and P = 0.003 for EEG abnormalities), reflecting phenotype similarity between affected members of the same family. The differences were also significant for psychiatric symptoms (0.040) and PNS involvement (0.013). The results reflect phenotypic homogeneity within families and differentiation among groups (between the individual families and also in comparison to non-familial cases).

Estimates of the partition of variance showed values ranging from 43 to 72% for the among-group variation and from 28 to 57% for the within-group variation.


This study was triggered by the observation of a disproportionately high representation of patients of Gypsy descent among NPD cases in Bulgaria, suggesting that the disorder is common in this ethnic group. We hypothesized that, similar to other autosomal recessive disorders in the Gypsy population (reviewed in Kalaydjieva et al., 2005), NPD is caused by an ancestral mutation, which has spread by founder effect and genetic drift. We identified the mutation as W391G in SMPD1, present in the homozygous form in all NPD cases of Gypsy ethnicity. Our screening data point to a high frequency of W391G in the general Gypsy population, with the observed 1.1% carrier rate identical to that of the joint frequency of the R496L, L302P and fsP330 mutations, causing NPD type A among Ashkenazi Jews (Li et al., 1997). In some sub-isolates, heterozygosity for the W391G mutation is ∼4%, making Niemann–Pick disease an important health problem in such communities.

The W391G mutation was first identified in a previous study of an NPD family with six affected individuals, described as ‘members of the same clan with roots in a small village in Serbia’ (Sperl et al., 1994; Ferlinz et al., 1995). Based on the structure of the affected family with lack of close consanguinity and the introduction of the same mutation by several unrelated individuals marrying in, Ferlinz et al. (1995) concluded that W391G has a high prevalence in the Serbian population. This conclusion appears unlikely in view of the outbred nature and genetic diversity of Balkan populations, and the rarity of genetic isolates and of consanguineous marriages. Our findings in Gypsy patients from neighbouring Bulgaria suggest that the previous studies in fact described the founder NDP mutation in the Gypsy population. Interestingly, W391G was also detected in 3 out of 64 patients (all three were compound heterozygotes) in the Mount Sinai NPD type B series (Wasserstein et al., 2006). Information on the ethnic origins of the current and future patients identified outside Bulgaria, and extended analysis of W391G-associated chromosomal haplotypes will allow a distinction between the possible scenarios leading to the current observations: (i) the W391G mutation originated in the Gypsy population and subsequently spread with the Gypsy diaspora, similar to the founder mutation causing GM1-gangliosidosis (Sinigerska et al., 2006); (ii) W391G is common in other ethnic groups and has been ‘imported’ into the Gypsy population, as is the case with cystic fibrosis and delF508 in Gypsies (Angelicheva et al., 1997); (iii) W391G has multiple origins in different populations. Whatever the history of this mutation, its detection in patients and carriers from Bulgaria, Serbia, Romania and the United States makes it an important contributor to the molecular epidemiology of Niemann–Pick disease.

The pathogenic effects of W391G have been documented by biochemical and cellular analyses (Sperl et al., 1994; Ferlinz et al., 1995). In vitro measurements of ASMase activity with the radioactively labelled natural substrate showed severe enzyme deficiency, in the range typical of NPD type A patients. Lack of the catalytically active mature 70 kDa form, in the presence of normal levels of the ASMase precursor forms, cell trafficking and uptake from the medium, suggested that the enzyme deficiency resulted from decreased stability and rapid intralysosomal degradation of the mutant protein (Ferlinz et al., 1995).

Our sample of W391G homozygotes displayed an intermediate type of Niemann–Pick disease. The visceral component can be summarized as generally severe, with variation similar to that reported in NPD type B patients with a diversity of underlying molecular defects in SMPD1 (Simonaro et al., 2002; Wasserstein et al., 2004). Our patients also showed a surprisingly high degree of variation in nervous system involvement despite identical SMPD1 genotypes and a genetic background of restricted diversity (Kalaydjieva et al., 2005). Previously published descriptions of intermediate Niemann–Pick disease outline a broad range of clinical and subclinical neural manifestations, varying in localization and severity, such as peripheral neuropathy, extrapyramidal signs, EEG abnormalities and epileptic seizures, cognitive impairment ranging from learning difficulties to dementia and psychotic episodes, etc. (Saidi et al., 1970; Sogawa et al., 1978; Elleder and Cihula, 1983; Shah et al., 1983; Elleder et al., 1986; Matthews et al., 1986; Dubois et al., 1990; Pavlu-Pereira et al., 2005). The interpretation of these findings has been complicated by the diversity of SMPD1 mutations or the lack of molecular data. W391G homozygotes presented with a mosaic of signs and symptoms spanning the entire spectrum observed in unrelated affected subjects of diverse ethnic backgrounds.

At one end of the spectrum, W391G homozygotes were indistinguishable from the classical NPD type B phenotype. Such patients had no neurological manifestations or only subclinical retinal involvement (macular halos) of no functional significance, which also appear to be common in NPD type B (McGovern et al., 2004). This mildly affected group included Subject 19 in our study and Subjects 1, 2, 5 and 6 in the original W391G family from Serbia (Sperl et al., 1994).

PNS involvement, affecting both Schwann cells and axons and leading to subclinical or mild peripheral neuropathy, was detected in most patients where electrophysiological studies could be performed. In Patient 9 it was the only abnormality, in addition to the macular halo, that distinguished the phenotype from the pure visceral form.

Clinical evidence of CNS involvement was found in 6/18 cases in our sample. Cerebellar ataxia was observed in four and was the leading manifestation in one. Cerebellar atrophy was detected in the two ataxic patients where neuroimaging could be performed. Cerebellar symptoms and MRI-documented atrophy have also been described by Pavlu-Pereira et al. (2005) and loss of Purkinje cells and general atrophy of the cerebellum were among the prominent pathological features of the mouse model of ASMase deficiency (Horinouchi et al., 1995). The data in human patients and the mouse phenotype both point to the vulnerability of the cerebellum to the effects of ASMase deficiency. Pyramidal signs were the next most common manifestation in our sample, found in three patients.

EEG changes were observed in 8/10 patients, ranging from diffuse mild abnormalities to focal and paroxysmal activity but no epileptic seizures.

Serious psychiatric disorders, requiring hospital admissions and treatment with neuroleptics, developed in two patients: Subject 6 with schizophrenia-like psychosis and Subject 12 with recurrent depression.

Cognitive impairment appeared to be common in W391G homozygotes, with formal testing showing deficit in 10/16 individuals in our study and in 1/6 affected members of the family described by Sperl et al. (1994). These data should be interpreted with caution, taking into account the problematic use of measures which have not been calibrated in the specific population, the lack of schooling and often insufficient command of the language (some of the patients speak mostly Romanes at home) and the social isolation, all of which can affect the results. Nonetheless, the findings were unequivocal in the two subjects, falling in the range of profound deficit, and in four others, classified as moderately deficient.

Our current data do not allow an assessment of the evolution (and hence the predictability) of nervous system involvement in W391G homozygotes. The progression of retinal lipid storage was supported by the observed age-related increase in halo size in our sample, and has been documented previously by patient follow-up in the McGovern et al. (2004) and Sperl et al. (1994) studies. Some of our patients showed additional evidence of progressive neurodegeneration. For example, in Subject 5 cerebellar ataxia was present from childhood, at age 13 years EEG abnormalities were detected and the ataxia had deteriorated, and pyramidal and extrapyramidal signs and behavioural changes developed subsequently. His sister (Subject 6) had no clinically detectable neurological abnormalities at age 10 years, developed psychotic illness at age 23 years, and presented with psychosis, pyramidal signs, EEG abnormalities and cerebral atrophy at age 27 years. At the same time, older affected individuals in the family from Serbia (Patients 1 and 2, aged 28 and 31 years) (Sperl et al., 1994) had a stabilized clinical course with no clinically manifest neurological involvement. Prospective follow-up studies of a larger sample of W391G homozygotes should provide conclusive data.

Similar to other inborn errors of metabolism, diverse SMPD1 mutations and the resulting different levels of residual enzyme activity have been invoked to explain the differences in clinical phenotypes (Schuchman and Desnick, 2001; Wasserstein et al., 2004). Imprinting at the SMPD1 locus may be another important contributor to the phenotypic diversity of NPD, where the nature of the mutation on the preferentially expressed maternal allele is likely to define phenotype severity in compound heterozygotes (Réthy, 2000; Simonaro et al., 2006). A parent-of-origin effect appears irrelevant in our group of W391G homozygotes, whose clinical diversity suggests further complexity of genotype–phenotype correlations. In Mendelian disorders, well-documented clinical differences between patients carrying the same mutation(s) have led to the concept of phenotypes as complex traits, resulting from the compound effects of modifiers, thresholds and systems dynamics (Dipple and McCabe, 2000a, b). In this context, it should come as no surprise that ASMase deficiency leads to a continuum of phenotypes (Pavlu-Pereira et al., 2005), and that the clinical manifestations in W391G homozygotes cover almost the entire clinical spectrum. Studies of patients with phenylketonuria, caused by missense mutations, have suggested that in vivo phenylalanine hydroxylase activity is largely dependent on the rate of degradation of the mutant protein and these differences in chaperone and proteolytic activity may contribute to phenotypic differences between individuals sharing the same mutation (Waters et al., 1998; Scriver and Waters, 1999). Such variation may also play a role in shaping the NPD phenotype in W391G patients, while the involvement of ASMase in cell signalling and apoptosis (Kolesnick, 2002) introduces a host of potential additional modifiers.

The involvement of genetic factors modifying the neurological NPD phenotype is suggested by the tendency towards intrafamilial clinical concordance among W391G homozygotes. While our analysis of phenotypic variance was performed on a small group of patients and is thus only preliminary in nature, the results point to similarities between affected members of the same family, with differences between families and in comparison to non-familial cases. Given the frequency of the W391G mutation and the large size of traditional families, the Gypsy population is well placed to address the issues of variability of nervous system involvement and the nature of the modifying effects, among which genetic factors could play an important role.

Our findings also have practical implications for the diagnosis and management of NPD patients. The atypical course of the disease, as well as the prevalent textbook concept of NPD as a rare condition presenting as either infantile neuronopathic type A or chronic visceral type B, may give rise to considerable diagnostic difficulties in W391G homozygotes. The inconsistent results of ASMase activity measurements with the artificial substrate, observed in our series and reported previously in other studies (Wenger and Williams, 1991; Harzer et al., 2003) may complement these difficulties. W391G mutation analysis is a simple test of high sensitivity and specificity, which should facilitate the process if applied to every Gypsy patient presenting with hepatosplenomegaly and elevated plasma chitotriosidase activity, regardless of age and nervous system involvement. The poor correlation between SMPD1 genotype and CNS involvement, observed in our sample, is of general relevance to patients with the chronic neurovisceral form of Niemann–Pick disease. In view of our findings, predictions of brain damage and hence a poor therapeutic effect of exogenous ASMase appear of limited validity, suggesting that the option of enzyme replacement therapy should be considered in intermediate NPD.


We thank affected individuals and family members for their participation in the study and Dr P. Janeva, Dr R. Kumanova, Dr E. Lazarova, Dr B. Radeva, Dr D. Konstantinova and Dr H. Jelev for patient referral. Wellcome Trust support of our research into genetic disorders of the Gypsies is gratefully acknowledged.

acid sphingomyelinase
Niemann–Pick disease
SMPD1 gene
sphingomyelin phosphodiesterase 1 gene


View Abstract