HUMAN MUTATION 17:199–209 (2001)

RESEARCH ARTICLE

Molecular Analysis of Acid Ceramidase Deficiency

in Patients With Farber Disease
Julia Bär,1 Thomas Linke,1 Klaus Ferlinz,2 Ulrich Neumann,3 Edward H. Schuchman,4 and
Konrad Sandhoff1*
1
Kekulé Institut for Organic Chemistry and Biochemistry, University of Bonn, Bonn, Germany
2
Department of Physiology, University of Tübingen, Tübingen, Germany
3
Clinic for Poultry, School of Veterinary Medicine, Hannover, Germany
4
Department of Human Genetics and Institute for Gene Therapy and Molecular Medicine, Mount Sinai School of Medicine,
New York, New York

Communicated by Elizabeth Neufeld

Farber disease is a rare, autosomal recessively inherited sphingolipid storage disorder due to the
deficient activity of lysosomal acid ceramidase, leading to the accumulation of ceramide in cells
and tissues. Here we report the identification of six novel mutations in the acid ceramidase gene
causing Farber disease: three point mutations resulting in single amino acid substitutions, one
intronic splice site mutation resulting in exon skipping, and two point mutations also leading to
occasional or complete exon skipping. Of interest, these latter two mutations occurred in adja-
cent nucleotides and led to abnormal splicing of the same exon. Expression of the mutated acid
ceramidase cDNAs in COS-1 cells and subsequent determination of acid ceramidase residual
enzyme activity demonstrated that each of these mutations was the direct cause of the acid
ceramidase deficiency in the respective patients. In contrast, two known polymorphisms had no
effect on acid ceramidase activity. Metabolic labeling studies in fibroblasts of four patients showed
that even though acid ceramidase precursor protein was synthesized in these individuals, rapid
proteolysis of the mutated, mature acid ceramidase occurred within the lysosome. Hum Mutat
17:199–209, 2001. © 2001 Wiley-Liss, Inc.

KEY WORDS: acid ceramidase; AC; ASAH; Farber disease; mutation analysis; exon skipping

DATABASES:
ASAH – OMIM: 228000; GDB:6837715; GenBank: NM_004315, U70063; HGMD: ASAH

INTRODUCTION death. Depending on the subtype, hepatosple-

Farber disease, also known as Farber lipo-
granulomatosis, is a rare lysosomal sphingolipid
storage disorder with an autosomal recessive
mode of inheritance. It is caused by the defi-
ciency of acid ceramidase activity (AC; also N-
acylsphingosine amidohydrolase; ASAH; MIM#
228000; EC 3.5.1.23) [Sugita et al., 1972]. De-
ficient AC activity leads to the accumulation of
ceramide in various tissues, including liver,
spleen, and lung. Typical symptoms of Farber
disease include subcutaneous nodules, painful
and progressively deformed joints, hoarseness
due to laryngeal involvement, and premature
nomegaly and nervous system dysfunctions may
also occur [for review, see Moser, 1995].
The Farber disease phenotype has been di-
Received 12 June 2000; accepted revised manuscript 15 De-
cember 2000.
*Correspondence to: Dr. Konrad Sandhoff, Kekulé Institut for
Organic Chemistry and Biochemistry, University of Bonn,
Gerhard-Domagk-Str. 1, D-53121 Bonn, Germany. E-mail:
sandhoff@uni-bonn.de
Contract grant sponsors: Fortüne-program of University of
Tübingen; Graduiertenkolleg “Functional protein domains”; Con-
tract grant sponsor: Deutsche Forschungsgemeinschaft; Contract
grant numbers: SFB 400; DFG Fe 390/2; Contract grant sponsor:
National Institutes of Health; Contract grant number: DK-54830.

© 2001 WILEY-LISS, INC.

200 BÄ R ET AL.
vided into seven different subgroups which dif-
fer in the age of onset, the severity of symptoms,
and the tissues affected by ceramide storage. The
underlying cause of subtypes 1 to 6 is the defi-
cient activity of AC, whereas subtype 7 [accord-
ing to Moser, 1995] is caused by a deficiency of
the sphingolipid activator protein precursor
(prosaposin, pSAP).
Farber disease is commonly diagnosed by
quantification of ceramide accumulation in tis-
sues or in cultured cells [Chatelut et al., 1996;
Van Echten-Deckert et al., 1997]. The extent
of ceramide storage closely correlates with the
neurodegenerative course of the disease and the
age of death of the patients [Levade et al., 1995].
The diagnosis of Farber disease also can be made
by determination of residual AC activity in vitro,
however, the residual AC activity determined
in vitro in Farber disease patients does not cor-
relate with the severity of clinical symptoms [Van
Echten-Deckert et al., 1997].
AC catalyzes the hydrolysis of ceramide into
sphingosine and a free fatty acid [Gatt, 1963].
AC hydrolysis represents the final step in the
lysosomal degradation of all sphingolipids.
Ceramide primarily functions as a membrane
anchor for all sphingolipids. While free ceramide
is only a minor component of biological mem-
branes, ceramide and its degradation products
sphingosine and sphingosine-1-phosphate have
been recognized as important mediators in sig-
nal transduction processes [Hannun, 1994;
Spiegel et al., 1996].
Lysosomal AC is a heterodimeric protein con-
sisting of a non-glycosylated α subunit of mole-
cular weight (MW) ~13 kDa and a glycosylated
β subunit of MW ~40 kDa. Deglycosylation
with PNGase F reduces the molecular weight of
the β subunit to approximately 27 kDa [Ber-
nardo et al., 1995]. Recently, the cDNA span-
ning an open reading frame of 1185 bp was
cloned [Koch et al., 1996]. The cDNA encodes
for both subunits suggesting that the primary
translation product of 395 amino acids is cleaved
to the mature enzyme posttranslationally. The
AC gene was mapped to the chromosomal re-
gion 8p21.3-p22. It spans about 30 kb in length
and contains 14 exons [Li et al., 1999]. Until
present, four mutations in the AC gene leading
to Farber disease are described: 665C>A,
413A>T, 760A>G, and 1084C>G [Koch et al.,
1996; Li et al., 1999].
In this report we identified the mutations in
six unrelated Farber disease patients. The mu-
tations were associated with drastically reduced
AC enzyme activity when expressed in COS-1
cells. In addition, the processing of the mutated
AC polypeptides was studied in cell lines from
four patients by metabolic labeling and immuno-
precipitation.
MATERIALS AND METHODS
Patients and Cell Lines
Cultured fibroblasts from six unrelated Farber
disease patients were used in this study. Cells
were obtained from: Dr. M.T. Zabot at the Hos-
pital Debrousse, Lyon, France (patient A); Prof.
Pfeiffer, Tübingen, Germany (patient B); Dr.
A.H. Fensom, London, U.K. (patient C); Dr. E.
Vamos at the Brugmann University Hospital,
Brüssels, Belgium (patient D), and; Dr. M. Beck,
Mainz, Germany (patient E). Cell line GM 02315
was from the NIGMS Human Genetic Mutant
Cell Repository, Camden, NJ (patient F). Nor-
mal control fibroblasts were obtained from the
Johanniter Children’s Hospital, Sankt Augustin,
Germany. In all patients, the diagnosis was per-
formed by demonstrating a reduced AC enzyme
activity.
Patient A, a male infant, was the fourth child
of consanguineous Tunisian parents. He was clas-
sified as having the severe (classical) subtype of
Farber disease. He died at the age of two years
[Souillet et al., 1989].
Patient B was a female infant from consanguin-
eous Turkish parents. She was originally diagnosed
with combined Farber and Sandhoff disease (des-
ignated Farber disease subtype 6). She died at the
age of 2.5 years [Fusch et al., 1989].
Patient C, a male infant, presented with typi-
cal Farber disease symptoms. He was designated
as having Farber disease subtype 5 (neurologic-
progressive). He died at three years of age
[Colamaria et al., 1992]. No information is avail-
able regarding the consanguinity of the parents.
Patient D was a female infant of consanguin-
eous Belgian parents. She died at 22 months of
age [Toppet et al., 1978]. She was the first pa-
 MOLECULAR ANALYSIS OF FARBER DISEASE
tient in whom a mutation in the acid ceramidase
cDNA was identified [Koch et al., 1996].
Patient E presented as a female infant of non-
consanguineous parents. She has a mild Farber
disease subtype (type 3) [Klingkowski et al., 1998].
Patient F, a female child from Sicilian parents,
also presented with a mild form of Farber dis-
ease (subtype 3) [Pavone et al., 1980], charac-
terized by the late onset of the disease at 20
months of age. She died at 30 years of age. The
parents were not known to be related.
Materials
Restriction enzymes were purchased from
Amersham (Braunschweig, Germany). Culture
media, the pSVSport1 vector, oligonucleotides,
and reverse transcriptase were from Life Tech-
nologies (Eggenstein, Germany). DNA Taq poly-
merase and the Big Dye Sequencing Kit were
purchased from Perkin Elmer (Weiterstadt, Ger-
many). The RNeasy Kit, QIAamp DNA Blood
Kit, QIAquick PCR purification Kit, and QiaEx
Kit were obtained from Qiagen (Hilden, Ger-
many). The DOTAP transfection reagent was
from Boehringer Mannheim (Mannheim, Ger-
many). Radioisotopes were purchased from ICN
(Eschwege, Germany) and anti-IgY-Agarose was
from Promega (Heidelberg, Germany). All other
reagents were of the highest quality available.
Isolation of RNA and DNA, cDNA Synthesis
and PCR Amplification
Genomic DNA was obtained from cultured
skin fibroblasts or blood samples with the
QIAamp procedure according to manufacturer’s
instructions. Total RNA was isolated from the
cultured fibroblasts and purified with the RNeasy
Kit according to the manufacturer’s instructions.
Reverse transcription (200 U of SuperScript II,
42°C for 60 min) was carried out using 5 µg of
TABLE 1. Oligonucleotide Sequences
Oligonucleotide
–17s
Bam1s
614s
911i
Sal1188i
1236i
1306i
201
total RNA and 20 pmol of the internal sequence-
specific antisense primer 1306i (Table 1). cDNA
was amplified by PCR using the primer sets
–17s/911i and 614s/1236i (Table 1) under the
following conditions: 30 cycles each consisting
of 0.5 min at 94°C, 1 min at 50°C, 2 min at 72°C
and 1 cycle of 10 min at 72°C.
To confirm putative mutations identified by
cDNA sequencing (see below), amplification of
0.5 µg genomic DNA was carried out by PCR
using intronic primers surrounding the mutated
exon. Each of the 30 cycles consisted of heat
denaturation at 94°C for 0.5 min, annealing at
55°C for 1 min, and extension at 72°C for 1.5
min, and one cycle of 7 min at 72°C. The ampli-
fied genomic products were evaluated by agar-
ose gel electrophoresis and subsequently purified
from excised gel slices using the QiaEx Kit or
purified with the QIAquick PCR purification Kit.
DNA Sequencing Analysis
Cycle sequencing was performed with inter-
nal sequencing primers and the BigDye termi-
nator cycle sequencing kit according to the
manufacturer’s instructions and run on an ABI
310 sequencer (Applied Biosystems).
Cloning and Expression of the Mutant
Recombinant Proteins in COS-1 Cells
The open reading frame of the wild type AC
cDNA was amplified by PCR with the primers
Bam1s/Sal1188i and cloned into the BamHI and
SalI restriction sites of the pUC18 vector. The
insert was then subcloned into the pSVSport1
vector via KpnI and SalI restriction sites result-
ing in the construct pSVSportAC. The mutant
cDNAs were PCR amplified and cloned into the
construct pSVSportAC using suitable internal
restriction sites. All mutant constructs were veri-
fied by DNA sequencing [Sanger et al., 1977].
Sequence
5′–GGCACGAGGCTAGAGCG–3′
5′–CGGGATCCATGCCGGGCCGGAGTTGC–3′
5′–CTGGCTATGTGGGCATGTTA–3′
5′–ACATCCAATGATTCCTTTCTG–3′
5′–ACGCGTCGACTCACCAACCTATACAAGG–3′
5′–TGGTGTCTTCATGTCTCAGA–3′
5′–TCAAAGAGAACCTGCCGCGAG–3′
202 BÄ R ET AL.
Transient transfection and subsequent determi-
nation of AC activity was performed as previ-
ously described [Koch et al., 1996]. In brief, ~5
× 105 cells were plated on 21 cm2 cell culture
dishes one day before transfection. The trans-
fections were carried out with the DOTAP
lipofection reagent using 7 µg of pure plasmid
DNA for each dish in serum-free OptiMEM
medium for 12 hr. The transfected cells were
grown for an additional 48 hr, harvested, and
assayed for AC activity.
AC Activity and Protein Assays
AC activity was assayed in the transfected and
control cells with the synthetic substrate N-
lauroylsphingosine in the presence of detergents
as previously described [Bernardo et al., 1995].
Protein was determined according to the method
of Bradford using bovine serum albumin as a
standard [Bradford, 1976].
Processing Studies
One day before radioactive labeling, fibroblasts
were seeded in 21 cm2 cell culture dishes and
grown to ~70% confluency. Prior to metabolic
labeling, the fibroblasts were starved for 2 hr in
methionine and cysteine free minimum essential
medium containing 0.4% dialyzed FCS. Metabolic
labeling was initiated by the addition of 100 µCi
of Tran35S label to the culture medium for 2 hr.
Labeled cells were then washed with PBS. The
chase period was started by the addition of DMEM
containing 10% FCS and supplemented with
methionine (5 mM) and cysteine (0.5 mM). The
chase proceeded for either 6 hr or 20 hr, after
which the cells were washed twice with ice-cold
phosphate-buffered saline, harvested with a rub-
ber policeman, and lysed in 0.5 ml of Hepes buff-
ered saline containing 0.5% deoxycholate, 0.1%
SDS, 1.0% Triton X-100, 10 mM EDTA, 1% bo-
vine serum albumin, 1 mg/ml Pefabloc SC, 7 µg/
ml Pepstatin, 5 µg/ml Leupeptin, and 2 µg/ml
Aprotinin. Nucleic acids were precipitated by the
addition of 10 µl of a 3% protamine sulfate solu-
tion. After 30 min at 4°C, the suspension was cen-
trifuged for 30 min at 15000 × g at 4°C.
The supernatants (cell lysates) and the me-
dia obtained from the chase experiments were
incubated with 1 µl of chicken anti-human AC
antiserum overnight at 4°C (Ferlinz K, Bernardo
K, Linke T, Bär J, Breiden B, Neumann U, Harzer
K, Lang F, and Sandhoff K, unpublished data).
Immunoprecipitation was carried out in the pres-
ence of 20 µl anti-IgY-Agarose for 2 hr at 4°C.
For SDS-PAGE, the precipitate was heated for
3 min at 100°C in Laemmli sample buffer
containing 0.8% β-mercaptoethanol. Electro-
phoresis was performed on 12.5% tricine poly-
acrylamide gels according to the method of
Schägger and von Jagow [1987]. Radioactive
proteins were detected by autoradiography.
RESULTS
Identification of AC Mutations
The acid ceramidase cDNAs from the Farber
disease patients were amplified by RT-PCR in
two overlapping fragments (nt –17 to 911 and
614 to 1236), and the complete open reading
frames were sequenced. To confirm putative
mutations identified in these cDNAs, the exons
corresponding to each detected mutation also
was amplified by genomic PCR and sequenced.
Sequencing of the RT-PCR products of pa-
tients A, B, E, and F revealed two polymorphisms
which have been described earlier [Koch et al.,
1996]. The transition 214A>G leads to the
amino acid substitution M72V, and the transi-
tion 277G>A results in a V93I substitution (Fig.
1). In addition to these polymorphisms, several
other mutations were found (Table 2). None of
the patients’ mutations were detected in the
genomic DNA of 50 normal individuals.
Patient A. RT-PCR analysis of patient A re-
vealed products of normal size, but a transition
mutation 107A>G was found, resulting in the
substitution of Y36C in the α subunit (Fig. 1).
Analysis of the genomic DNA from this patient
confirmed that he was homoallelic for this mu-
tation.
Patient B. RT-PCR analysis of patient B re-
vealed a point mutation, 958A>G. This muta-
tion was predicted to cause an amino acid
substitution, N320D in the AC β subunit (Fig.
1). The patient was homoallelic for the muta-
tion as confirmed by sequence analysis of the
genomic DNA.
Patient C. The PCR product derived from the
3′ portion of the AC cDNA (nucleotides 614 to
 MOLECULAR ANALYSIS OF FARBER DISEASE 203

FIGURE 1. Structure of the human AC gene and protein. a: Genomic structure; b: cDNA structure; c: illustration of the AC
precursor and mature heterodimeric protein. Potential N-glycosylation sites and the location of cysteine residues are
indicated. The locations of mutations and polymorphisms are indicated by the arrows. The letters A–F refer to the patient
identity, and P1 and P2 refer to the polymorphisms.

1236) of patient C was found to be smaller com-
pared to the wild type control, indicating a de-
letion. Sequencing of the open reading frame
revealed a deletion of nucleotides 1042–1098,
which corresponds to exon 13 of the AC gene.
This deletion of 19 amino acids (348 to 366) in
the β subunit would lead to a mutated and trun-
cated protein, but no frameshift abnormality (Fig.
1). Sequencing of the genomic DNA showed
that the underlying mutation responsible for the
exon deletion was a transversion, g>t, of the
first nucleotide in intron 13 (IVS13+1G>T).
The patient was heteroallelic for this mutation,
suggesting the presence of a different mutation
on the other allele. In order to identify a second
mutation, >500 bp upstream of the ATG start
codon and the 3′ UTR of the AC gene were se-
quenced (data not shown). However, no other

TABLE 2. Patient’s Mutations

Mutation in Predicted amino Genomic
Patient AC cDNA acid substitution mutation Allelic status Literature
A 107A>G Y36C 107A>G Homoallelic This paper
B 958A>G N320D 958A>G Homoallelic This paper
C del 1042-1098 del 348-366 (exon 13) IVS13+1G>T Heteroallelic This paper
D 665C>A T222K 665C>A Homoallelic Koch et al. [1996]
E 991G>A D331N 991G>A Heteroallelic This paper
E del 383-457 del 128-152 (exon 6) 412G>T Heteroallelic This paper
F 413A>T E138V 413A>T Homoallelic Li et al. [1999]
F del 383-457 del 128-152 (exon 6) 413A>T Homoallelic This paper
204 BÄ R ET AL.
base changes were detected compared to the wild
type AC gene, perhaps indicating a mutation
located further upstream or further within the
intronic sequences.
Patient D. The mutation in patient D has
been described previously [Koch et al., 1996].
Patient E. RT-PCR analysis of patient E re-
vealed two 5′ (nt –17 to 911) AC PCR products
of different sizes, but of similar quantities. The
larger PCR product was of the same size as the
wild type PCR product and contained a transi-
tion 991G>A, leading to a D331N substitution
(Fig. 1). This mutation was confirmed at the
genomic DNA level and the patient was shown
to be heteroallelic. Examination of the genomic
DNA isolated from blood cells of the parents
demonstrated that this point mutation was trans-
mitted by the mother. The father’s sequence was
normal in this region. The smaller PCR product
carried a deletion of nucleotides 383 to 457. This
corresponded to the deletion of the entire exon
6, predicting an absence of amino acids 128 to
152 in the resulting polypeptide (Fig. 1). Unex-
pectedly, genomic sequencing of exon 6 showed
no sequence abnormalities in the adjacent in-
tron–exon boundaries or in the branch sites of
the adjacent introns. Instead, we found an ad-
ditional point mutation 412G>T in exon 6,
which could not be detected on cDNA level.
The predicted base change would result in a sub-
stitution of glutamate 138 to a stop codon
(E138X, nonsense mutation). Sequencing of the
genomic DNA of both parents showed that this
mutation was transmitted by the father.
Patient F. Analysis of patient F revealed two
AC-specific PCR products of different sizes at
the 5′ end (nt –17 to 911) (Fig. 2). The PCR
product of the larger transcript was the same size
as the wild type PCR product and contained the
point mutation 413A>T leading to a E138V
substitution (Fig. 1). As in patient E, the smaller
fragment showed a deletion of nucleotides 383
to 457, corresponding to a deletion of exon 6.
At the genomic level, however, only the
homoallelic mutation 413A>T was identified.
Effect of the Mutations on AC Enzyme Activity
In order to investigate the effects of the newly
identified mutations on the enzymatic activity
FIGURE 2. PCR products of the 5′ portion of the AC cDNA
(nt –17 to 911) from patient F and a control.
of AC, the respective mutant cDNAs were
cloned into the mammalian expression vector
pSVSport1. The wild type cDNA and a cDNA
containing both polymorphisms (214A>G and
277G>A) also were cloned into the pSVSport1
vector as positive controls.
COS-1 cells were transfected with these ex-
pression plasmids and cultured for 60 hr. Subse-
quently, the cell lysates were assayed for AC
activity. Wild type AC and AC containing the
two polymorphisms have similar specific AC ac-
tivities, indicating that both polymorphisms
M72V and V93I have no influence on AC ac-
tivity. In contrast, all other mutations lead to
the expression of mutated AC which exhibited
very low residual enzymatic activities (Fig. 3).
The residual enzymatic activity of the mutated
AC measured in vitro did not correlate with the
severity of the disease phenotype in individual
patients.
Immunoprecipitation Studies in Farber
Disease Fibroblasts
Metabolic labeling of normal and Farber dis-
ease fibroblasts with [35S]cysteine/methionine
followed by immunoprecipitation, SDS-PAGE,
and autoradiography were performed in order to
follow the intracellular maturation of AC (Fig.
4). In normal fibroblasts, AC is first synthesized
as a 53–55 kDa single chain precursor protein
which is cleaved to the mature heterodimeric
form of AC in the acidic organelles. After 6 hr
 MOLECULAR ANALYSIS OF FARBER DISEASE
of chase, some of the precursor protein and some
β subunit can be immunoprecipitated from the
cell lysate. After the 20 hr chase, the AC pre-
cursor is completely processed into the mature
α and β subunits. No AC precursor or proteolyti-
cally processed mature AC can be immunopre-
cipitated from the cell culture medium (data not
shown).
In the fibroblasts of patient A, normal
amounts of AC precursor are formed during the
2 hr pulse period. However, the β subunit is
not detectable at all in the Farber fibroblasts
and only reduced levels of the α subunit are
detected. This indicates that the point muta-
tion 107A>G, which leads to the Y36C sub-
205
FIGURE 3. Transient expression of AC
in COS-1 cells. Residual enzyme ac-
tivities are shown compared to the wild
type. Each value is the mean of 3–4
measurements. The standard devia-
tions are shown.
stitution, causes a more rapid degradation of
the mature enzyme.
In the fibroblasts of patient B, normal levels
of AC precursor protein are synthesized in com-
parison to the control fibroblasts. Both the pre-
cursor protein and β subunit were detectable six
hours after biosynthesis, as observed in the con-
trol fibroblasts. After a 20-hour chase period,
both subunits are still detectable, pointing to the
correct proteolytic processing and transport of
AC in the patient’s cells. Although this muta-
tion does not appear to have a direct effect on
the rate of biosynthesis and of maturation of the
enzyme, the COS cell transfection experiments
showed that the mutant AC exhibited a strongly
FIGURE 4. AC biosynthesis in fibro-
blasts of patients A, B, C, and D and
in controls. Immunoprecipitation of
cell homogenate of labeled normal
and Farber disease fibroblasts was
performed after a 2 hr pulse and ad-
ditional chase periods of 6 hr and 20
hr. The AC precursor and the mature
α and β subunits are marked by ar-
rows. A coprecipitating, unidentified
protein is marked by an asterisk (*).
206 BÄ R ET AL.
reduced enzyme activity, perhaps indicating a
catalytic site mutation.
In the fibroblasts of patient C, a significantly
lower amount of AC precursor protein was syn-
thesized in comparison to normal fibroblasts.
Furthermore, the molecular weight of the AC
precursor was ~2 kDa lower than in the normal
fibroblasts, presumably caused by the 19 amino
acid deletion. The precursor, which is still de-
tectable after six hours of chase, is obviously not
processed to the mature α and β subunit. After
20 hours of chase, only the α subunit, but not
the β subunit, was detected by immunoprecipi-
tation. This provides evidence for a more rapid
degradation of the β subunit which contains the
deletion. This degradation has to take place af-
ter proteolytic cleavage into the α and β sub-
units, as the α subunit is still detectable. It is
unlikely that the mutated β subunit is not rec-
ognized by the antibody because a polyclonal
antiserum was used.
In the fibroblasts of patient D, only minor
amounts of AC precursor protein were detected
after the pulse period. After the six-hour chase,
neither AC precursor protein nor the mature
subunits were detectable by immunoprecipita-
tion. After 20 hours of chase, only the α subunit
of AC was detected, whereas the β subunit was
not. This suggested that the point mutation
665C>A in patient D (T222K, β subunit) leads
to a rapid degradation of the β subunit.
DISCUSSION
In this paper we present the identification of
six novel mutations in the AC gene leading to
various subtypes of Farber disease, and confirmed
that two nucleotide substitutions that were pre-
viously classified as polymorphisms [Koch et al.,
1996] do not result in decreased AC activity. In
contrast, expression of the mutated cDNAs in
COS-1 cells led to low residual AC activity con-
sistent with the causative nature of the patient’s
mutations.
We found single nucleotide transitions in two
patients, each leading to an amino acid substi-
tution: the mutation 107A>G was localized in
the region encoding the α subunit of patient A,
and resulted in a Y36C amino acid substitution.
A possible cause of the rapid degradation as
shown by immunoprecipitation is the presence
of an additional cysteine introduced by the point
mutation. This cysteine may disrupt the forma-
tion of regular disulfide bridges in the protein
and therefore lead to a misfolded protein which
is rapidly degraded by lysosomal proteases.
The mutation 958A>G was found in the
cDNA and genomic DNA of patient B. While
metabolic labeling of the patient’s fibroblasts
revealed that normal amounts of AC precursor
was synthesized and correctly processed to the
mature heterodimeric protein in this patient, the
expression of the mutated protein decreased the
enzymatic activity by ~90% compared to the
wild type protein. Thus, this amino acid substi-
tution appears to change the catalytic proper-
ties of the enzyme without causing its premature
proteolysis.
In patient C, only one cDNA species contain-
ing a deletion of nucleotides 1042–1098 was
detected. Since the deletion covers 57 nucle-
otides, exactly 19 amino acids with a calculated
MW of 2.2 kDa are predicted to be missing from
the mutated AC protein. At the genomic level,
a heteroallelic mutation of the first nucleotide
in intron 13 was found. As a consequence, the
splice site sequence of the downstream intron is
destroyed, presumably leading to the abberant
splicing of the exon 13 sequences. Biosynthetic
studies revealed that the AC precursor derived
from this exon 13-deleted gene was ~2 kDa
smaller than normal, consistent with the pre-
dicted molecular weight loss. Since this exon also
encoded one potential N-glycosylation site and
the decrease in the MW of the AC precursor
corresponded exactly to the missing 19 amino
acids, it is likely that this N-glycosylation site is
not utilized.
Processing studies in patient D indicate that
less AC precursor protein is generated in this
patient’s cells as compared to the control fibro-
blasts. After the 20-hour chase period, only mi-
nor amounts of the α subunit were detected, and
no β subunit (which would contain the muta-
tion) was found. From this we conclude that pro-
cessing of the precursor protein must have taken
place (since the α subunit is present), followed
by rapid proteolytic degradation of the β sub-
unit. The presence of mutated and misfolded β
 MOLECULAR ANALYSIS OF FARBER DISEASE
subunit which is not recognized by the antibody
is unlikely because a polyclonal serum was used
for the immunoprecipitation.
We also have shown in the present study that
two Farber disease patients, E and F, each have
AC transcripts of abnormal size. The shortened
transcript contained a deletion of exon 6 se-
quences, predicting that the mutated AC
polypeptide would be shortened by 25 amino
acids compared to the wild type protein. The
molecular weight of the mutant polypeptide
should therefore be reduced by ~2.8 kDa. Exon
6 contains both the C-terminus of the α-sub-
unit and the N-terminus of the β-subunit of
mature, heterodimeric AC. Therefore, as a con-
sequence of this deletion, the proteolytic cleav-
age site of the AC precursor into the α- and
β-subunits would be absent.
Surprisingly, no mutation of the genomic
DNA in the exon–intron boundaries or in the
branch site of the intron adjacent to exon 6 was
detected. Instead, an additional nonsense mu-
tation 412G>T was found in the genomic DNA
of patient E within exon 6, suggested to be re-
sponsible for the exon skipping. Such an asso-
ciation between a nonsense mutation and exon
skipping has been described for other diseases
such as neurofibromatosis type 1 [Messiaen et
al., 1997], haemophilia A [Naylor et al., 1993],
cystic fibrosis [Hull et al., 1994], Hurler syn-
drome [Bach et al., 1993], and mitochondrial
acetoacetyl-CoA thiolase deficiency [Fukao et
al., 1994].
Three possible explanations for the associa-
tion of nonsense mutations with exon skipping
have been suggested [Steingrimsdottir et al.,
1992; Valentine, 1998, for review]. First, the
nonsense mutation may lead to a strongly re-
duced level of the mRNA transcript encoding
the truncated protein [Belgrader et al., 1994].
As a consequence, minor, exon-deleted RNA
species, present also in wild type cells, are vis-
ible in the patient’s cells as a result of an RT-
PCR artifact.
A second explanation for the association of
nonsense mutations with exon skipping is that
the pre-mRNA containing the point mutation
destroys a secondary structure that improves the
accessibility of splicing factors to specific splice
207
sites. Thus, certain exons may be preferentially
spliced due to secondary structure abnormalities.
Third, it has been suggested that nonsense
mutations could influence the exon definition,
which may occasionally result in exon skipping.
Exons are generally defined by the splice site
consensus sequences located at the very termini
of adjacent introns. For weakly defined exons,
internal sequences in the target exon have been
found which are important for the recognition
of such exons; such sequences are called exon
enhancers.
The induction of exon skipping by missense
mutations as found in patient F also has been
described for other human diseases: adenosine
deaminase deficiency [Ozsahin et al., 1997],
aspartylglycosaminuria [Coulter-Mackie, 1999],
Menkes disease [Das et al., 1994], and even for
silent mutations in Marfan syndrome [Liu et al.,
1997] and acute intermittent porphyria [Lle-
wellyn et al., 1996]. Therefore, it seems likely that
the mutation 413A>T causes partial exon skip-
ping and could be explained by the second and
third hypotheses described before. The mutation
413A>T already has been described [Li et al.,
1999], but no exon skipping due to this mutation
was found.
Genotype–phenotype correlations could not
be made in this study due to the limited number
of Farber disease patients examined. The
neurodegenerative course of Farber disease and
the age of death of the patients correlates with
the extent of ceramide storage in the lysosomes
[Levade et al., 1995]. In contrast, the residual
enzyme activities of Farber patients measured in
vitro does not correlate with the severity of the
clinical symptoms nor with the lipid accumula-
tion [Van Echten-Deckert et al., 1997]. The
measurement of AC activities in vitro is per-
formed in the presence of detergents which leads
to non-physiological conditions. Assays should
be done in the presence of the activator protein
SAP-D, which is necessary for ceramide catabo-
lism by AC as shown in cell culture studies [Klein
et al., 1994]. A liposomal assay in the absence
of detergents would likely mimic the in vivo situ-
ation better and therefore may give residual AC
activities which could be correlated with the
severity of Farber disease, as shown for other
208 BÄ R ET AL.
enzymes of the sphingolipid catabolism [Leine-
kugel et al., 1992; Graber et al., 1994; Meivar-
Levy et al., 1994].
In conclusion, we report the identification and
characterization of several new mutations in the
AC gene leading to Farber disease. Two of these
mutations occurred in adjacent nucleotides and
led to exon 6 splicing abnormalities. Metabolic
labeling studies also revealed that all but one of
the point mutations led to rapid proteolytic deg-
radation of the mutated AC polypeptides. These
studies provide new genetic and biochemical
insights into the pathogenesis of Farber disease
and the biology of AC.
ACKNOWLEDGMENTS
We thank Maike Schöniger and Martina
Domgörgen for excellent technical assistance,
and Ralf Klingenstein for sequencing the nor-
mal controls.
REFERENCES
Bach G, Moskowitz S, Tieu P, Matynia A, Neufeld E. 1993.
Molecular analysis of Hurler syndrome in Druze and Mus-
lim Arab patients in Israel: multiple allelic mutations of
the IDUA gene in a small geographic area. Am J Hum
Genet 53:330–338.
Belgrader P, Cheng J, Zhou X, Stephenson L, Maquat L. 1994.
Mammalian nonsense codons can be cis effectors of nuclear
mRNA half-life. Mol Cell Biol 14:8219–8228.
Bernardo K, Hurwitz R, Zenk T, Desnick RJ, Ferlinz K,
Schuchman EH, Sandhoff K. 1995. Purification, charac-
terization and biosynthesis of human acid ceramidase. J
Biol Chem 270:11098–11102.
Bradford MM. 1976. A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing
the principle of protein-dye binding. Anal Biochem
72:248–254.
Chatelut M, Feunteun J, Harzer K, Fensom AH, Basile JP,
Salvayre R, Levade T. 1996. A simple method for screen-
ing for Farber disease on cultured skin fibroblasts. Clin
Chim Acta 245:61–71.
Colamaria V, Giardina L, Simeone M, Salviati A, Fensom AH,
Dalla Bernadina B. 1992. Neurologic progressive form—
type 5—of Farber’s lipogranulomatosis (ceramidase defi-
ciency) in homozygotic twins. Meeting of the European
Neurological Society (abstract).
Coulter-Mackie MB. 1999. A novel exonic mutation in the
aspartylglucosaminidase gene causes exon skipping. J In-
herit Metab Dis 22:682–683.
Das S, Levinson B, Whitney S, Vulpe C, Packman S, Gitshier
J. 1994. Diverse mutations in patients with Menkes dis-
ease often lead to exon skipping. Am J Hum Genet
55:883–889.
Fukao T, Yamaguchi S, Wakazono A, Orii T, Hoganson G,
Hashimoto T. 1994. Identification of a novel exonic mu-
tation at –13 from 5′ splice site causing exon skipping in a
girl with mitochondrial acetoacetyl-coenzyme A thiolase
deficiency. J Clin Invest 93:1035–1041.
Fusch C, Huenges R, Moser HW, Sewell AC, Roggendorf W,
Kustermann-Kuhn B, Poulos A, Carey WF, Harzer K. 1989.
A case of combined Farber and Sandhoff disease. Eur J
Pediatr 148:558–562.
Gatt S. 1963. Enzymic hydrolysis and synthesis of ceramides.
J Biol Chem 238:3131–3133.
Graber D, Salvayre R, Levade T. 1994. Accurate differentia-
tion of neuronopathic and nonneuronopathic forms of
Niemann-Pick disease by evaluation of the effective re-
sidual lysosomal sphingomyelinase activity in intact cells.
J Neurochem 63:1060–1068.
Hannun YA. 1994. The sphingomyelin cycle and the second
messenger function of ceramide. J Biol Chem 269:3125–
3128.
Hull J, Shackleton S, Harris A. 1994. The stop mutation
R553X in the CFTR gene results in exon skipping.
Genomics 19:362–364.
Klein A, Henseler M, Klein C, Suzuki K, Harzer K, Sandhoff
K. 1994. Sphingolipid activator protein D (sap-D) stimu-
lates the lysosomal degradation of ceramide in vivo.
Biochem Biophys Res Commun 200:1440–1448.
Klingkowski U, Beck M, Zépp F. 1998. An 18-month-old girl
with hoarseness, stiff joints and subcutaneous nodules. Eur
J Pediatr 157:515–516.
Koch J, Gärtner S, Li CM, Quintern LE, Bernardo K, Levran
O, Schnabel D, Desnick RJ, Schuchman EH, Sandhoff K.
1996. Molecular cloning and characterization of a full-
length complementary DNA encoding human acid
ceramidase. J Biol Chem 271:33110–33115.
Leinekugel P, Michel E, Conzelmann E, Sandhoff K. 1992.
Quantitative correlation between the residual activity of
β-hexosaminidase A and arylsulfatase A and the severity
of the resulting lysosomal storage disease. Hum Genet
88:513–523.
Levade T, Moser HW, Fensom AH, Harzer K, Moser AB,
Salvayre R. 1995. Neurodegenerative course in ceramidase
deficiency (Farber disease) correlates with the residual ly-
sosomal ceramide turnover in cultured living patient cells.
J Neurol Sci 134:108–114.
Li CM, Park JH, He X, Levy B, Chen F, Arai K, Adler DA,
Disteche CM, Koch J, Sandhoff K, Schuchman EH. 1999.
The human acid ceramidase gene (ASAH): structure,
chromosomal location, mutation analysis and expression.
Genomics 62:223–231.
 MOLECULAR ANALYSIS OF FARBER DISEASE
Liu W, Qian C, Francke U. 1997. Silent mutation induces
exon skipping of fibrillin-1 gene in Marfan syndrome. Nat
Genet 16:328–329.
Llewellyn DH, Scobie GA, Urquhart AJ, Whatley SD,
Roberts AG, Harrison PR, Elder GH. 1996. Acute inter-
mittent porphyria caused by defective splicing of porpho-
bilinogen deaminase RNA: a synonymous codon mutation
at –22 bp from the 5′ splice site causes skipping of exon
3. J Med Genet 1996:437–438.
Meivar-Levy I, Horowitz M, Futerman AH. 1994. Analysis of
glucocerebrosidase activity using N-(1-[14C]hexanoyl)-D-
erythro-glucosylsphingosine demonstrates a correlation
between levels of residual enzyme activity and the type of
Gaucher disease. Biochem J 303:377–382.
Messiaen L, Callens T, De Paepe A, Craen M, Mortier G. 1997.
Characterisation of two different nonsense mutations,
C6792A and C6792G, causing skipping of exon 37 in the
NF1 gene. Hum Genet 101:75–80.
Moser HW. 1995. Ceramidase deficiency: Farber lipogranu-
lomatosis. In: Scriver CR, Beaudet AL, Sly WS, Valle D,
editors, Metabolic and molecular bases of Inherited dis-
ease. New York: McGraw-Hill. p 2589–2599.
Naylor J, Green P, Rizza C, Giannelli F. 1993. Analysis of fac-
tor VIII mRNA reveals defects in every one of 28 haemo-
philia A patients. Hum Mol Genet 2:11–17.
Ozsahin H, Arredondo-Vegas FX, Santisteban I, Fuhrer H,
Tuchschmidt P, Jochum W, Aguzzi A, Lederman HM,
Fleischman A, Winkelstein JA, Seger RA, Hershfield MS.
1997. Adenosine deaminase deficiency in adults. Blood
89:2849–2855.
Pavone L, Moser HW, Mollica F, Reitano C, Durand P. 1980.
Farber’s lipogranulomatosis: ceramidase deficiency and
prolonged survival in three relatives. Johns Hopkins Med
J 147:193–196.
209
Sanger F, Nicklen S, Coulsen AR. 1977. DNA sequencing
with chain terminating inhibitors. Proc Natl Acad Sci
74:5467–5483.
Schägger H, von Jagow G. 1987. Tricine-sodium dodecyl sul-
fate polyacrylamide gel electrophoresis for the separation
of proteins in the range from 1 to 100 kDa. Anal Biochem
166:368–379.
Souillet G, Guibaud P, Fensom AH, Maire I, Zabot MT. 1989.
Outcome of displacement bone marrow transplantation
in Farber’s disease: a report of a case. In: Hobbs JR, editor.
Correction of certain genetic diseases by transplantation.
London: COGENT. 1989:137–141.
Spiegel S, Foster D, Kolesnick RN. 1996. Signal transduction
through lipid second messengers. Curr Opin Cell Biol
8:159–167.
Steingrimsdottir H, Rowley G, Dorado G, Cole J, Lehmann
AR. 1992. Mutations which alter splicing in the human
hypoxanthine-guanine phosphoribosyltransferase gene.
Nucl Acids Res 20:1201–1208.
Sugita M, Dulaney JT, Moser HW. 1972. Ceramidase defi-
ciency in Farber’s disease (lipogranulomatosis). Science
178:1100–1102.
Toppet M, Vamos-Hurwitz E, Jonniaux G, Cremer N, Tondeur
M, Pelc S. 1978. Farber’s disease as a ceramidosis: clinical,
radiological and biochemical aspects. Acta Pediatr Scand
67:113–119.
Valentine CR. 1998. The association of nonsense codons with
exon skipping. Mut Res 411:87–117.
Van Echten-Deckert G, Klein A, Linke T, Heinemann T,
Weisgerber J, Sandhoff K. 1997. Turnover of endogenous
ceramide in cultured normal and Farber fibroblasts. J Lipid
Res 38:2569–2579.