Molecular Genetics and Metabolism 97 (2009) 53–59
Contents lists available at ScienceDirect
Molecular Genetics and Metabolism
journal homepage: www.elsevier.com/locate/ymgme
Neurodegenerative lysosomal storage disease in European Burmese cats
with hexosaminidase b-subunit deficiency
Allison M. Bradbury a, Nancy E. Morrison a, Misako Hwang a, Nancy R. Cox a,b, Henry J. Baker a,b,
Douglas R. Martin a,b,*
a
b
Scott-Ritchey Research Center, College of Veterinary Medicine, Auburn University, Auburn, AL 36849-5525, USA
Department of Pathobiology, College of Veterinary Medicine, Auburn University, Auburn, AL 36849-5525, USA
a r t i c l e
i n f o
Article history:
Received 15 November 2008
Received in revised form 13 January 2009
Accepted 13 January 2009
Available online 23 February 2009
Keywords:
Gangliosidosis
Lysosomal storage disease
Hexosaminidase
RNA splicing
Models, animal
Feline
a b s t r a c t
GM2 gangliosidosis is a fatal, progressive neuronopathic lysosomal storage disease resulting from a deficiency of b-N-acetylhexosaminidase (EC 3.2.1.52) activity. GM2 gangliosidosis occurs with varying
degrees of severity in humans and in a variety of animals, including cats. In the current research, European Burmese cats presented with clinical neurological signs and histopathological features typical of a
lysosomal storage disease. Thin layer chromatography revealed substantial storage of GM2 ganglioside in
brain tissue of affected cats, and assays with a synthetic fluorogenic substrate confirmed the absence of
hexosaminidase activity. When the hexosaminidase b-subunit cDNA was sequenced from affected cats, a
91 base pair deletion constituting the entirety of exon 12 was documented. Subsequent sequencing of
introns 11 and 12 revealed a 15 base pair deletion at the 30 end of intron 11 that included the preferred
splice acceptor site, generating two minor transcripts from cryptic splice acceptor sites in affected Burmese cats. In the cerebral cortex of affected cats, hexosaminidase b-subunit mRNA levels were approximately 1.5 times higher than normal (P < 0.001), while b-subunit protein levels were substantially
reduced on Western blots.
Ó 2009 Elsevier Inc. All rights reserved.
b-N-acetylhexosaminidase (EC 3.2.1.52) is a lysosomal enzyme
responsible for removal of the terminal N-acetyl-galactosamine
residue from GM2 ganglioside. Functional hexosaminidase activity
requires the coordinated action of three distinct proteins, a nonhydrolytic GM2 activator protein and the two hydrolytic subunits,
a and b. Subunits may combine to form the two major hexosaminidase isozymes, each with different substrate specificities: HexB
(bb) or HexA (ab), the isozyme responsible for degradation of
GM2 ganglioside in humans. Thus, a defect in the a or b-subunit
or in the GM2 activator protein causes GM2 gangliosidosis, classified as follows based on the remaining isozyme(s): variant AB
(GM2 activator protein deficiency), variant B (a-subunit deficiency,
Tay-Sachs Disease), or variant 0 (b-subunit deficiency, Sandhoff
disease).
In addition to the human disease, GM2 gangliosidosis variant 0
has been reported in animals, including cats [1–3], pigs [4], dogs
[5,6] and knockout mice [7,8]. Feline GM2 gangliosidosis is unique
among large animal models of disease in the number and variety of
mutations that have been reported. Pathogenic mutations include
* Corresponding author. Address: Scott-Ritchey Research Center, College of
Veterinary Medicine, Auburn University, Auburn, AL 36849-5525, USA. Fax: +1
334 844 5850.
E-mail address: [email protected] (D.R. Martin).
1096-7192/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.ymgme.2009.01.003
a 4 base pair deletion in the GM2 activator protein gene (GM2A)
[9] and three distinct mutations in the b-subunit gene (HEXB), all
of which produce a premature termination codon: a 25 base pair
inversion at the 30 terminus of the coding region [10], a single base
deletion in exon 1 [11], or a nonsense mutation in exon 7 [12]. The
current research reports a fourth distinct mutation of feline HEXB, a
15 base pair deletion that includes the splice acceptor site of intron
11, resulting in improper RNA splicing, removal of exon 12 from the
mRNA and a premature termination codon. In addition, the deletion
activates cryptic RNA splice sites in the vicinity of the mutation,
generating novel transcripts not detected in normal cat tissue.
Materials and methods
Materials
Unless otherwise stated, standard laboratory reagents and fluorogenic enzyme substrates were purchased from Sigma–Aldrich
(St. Louis, Missouri, USA). 4-methylumbelliferyl-6-Sulfo-2-acetamido-2-deoxy-b-D-glucopyranoside (MUGS) was purchased from
Toronto Research Chemicals (Toronto, Ontario, Canada). Ganglioside standards for thin layer chromatography were purchased from
Matreya (Pleasant Gap, Pennsylvania, USA). SuperScript II reverse
transcriptase was purchased from Invitrogen (Carlsbad, California,
54
A.M. Bradbury et al. / Molecular Genetics and Metabolism 97 (2009) 53–59
USA), and SYBR green PCR core reagents were from Applied Biosystems (Foster City, California, USA). Taq DNA Polymerase was purchased from Fisher Scientific (Waltham, Massachusetts, USA).
Secondary antibody and chemiluminescent substrate (Supersignal
West Dura) were purchased from Thermo-Pierce (Rockford,
Illinois, USA).
10 cm; Whatman, Florham Park, New Jersey, USA) which had been
heat-activated for 30 min at 110 °C. Ganglioside standards and extracted samples were loaded onto the plate and developed with a
chloroform–methanol-0.4% calcium chloride in water solution of
11:9:2. Plates were dried and sprayed with fresh resorcinol reagent, covered with clean glass plates, and heated for 20–30 min
at 100 °C for color development.
Enzyme assay
Sequencing
The enzyme activity of b-galactosidase, hexosaminidase, and amannosidase was measured with the corresponding 4-methylumbelliferone (4MU) substrate in cerebral cortex homogenates from
two European Burmese cats showing severe neurological disease
signs and normal cats not of European Burmese descent. b-galactosidase activity was measured by incubating 20 ll of homogenate
with 0.5 mM 4MU-N-acetyl-b-D-glucosaminide, pH 3.8, for
60 min. For hexosaminidase A activity, 20 ll of sample was added
to 1 mM MUGS, pH 4.2, for 60 min. Total hexosaminidase activity
was tested by combining 5 ll of homogenate and 1 mM 4MU-Nacetyl-b-D-glucosaminide (MUG), pH 4.3, for 30 min. To assay amannosidase activity, 10 ll of sample was incubated with 2 mM
4MU-mannopyranoside, pH 4.2, for 60 min. After the allotted time,
the enzyme activity was terminated with 3 ml of cold glycine carbonate buffer. Fluorescence of the 4MU compound was then measured on a Perkin-Elmer LS-5B luminescence spectrophotometer
with excitation at 360 nm and emission at 450 nm. Protein concentrations were measured by the method of Lowry, and specific activity was expressed as nmol 4MU cleaved/mg protein/h.
Thin layer chromatography
High-performance thin layer chromatography was performed
on brain tissue samples from affected European Burmese cats
taken at necropsy and stored at 80 °C. The cortex samples (0.2–
0.6 g) were homogenized in 2:1 chloroform:methanol and allowed
to extract for 1 h at room temperature. The samples were centrifuged and the pellet extracted with 1:1 chloroform:methanol and
again with 1:2 chloroform:methanol. Supernatants from each sample were then combined and dried under an air stream. The residues were treated with 3 ml of 0.1 N NaOH in methanol and
incubated at 37 °C for 2 h and again dried under an air stream.
The sample residues were then dissolved in 5 ml of ice cold water
and neutralized, pH 4.5, with drop-wise addition of HCl. Lipids
were further purified by reverse-phase column chromatography
with a SEP-PAK C18 cartridge (Waters Associates Inc., Milford,
Massachusetts, USA).
Extracted samples were applied to pre-coated silica gel 60 highperformance thin layer chromatography (HPTLC) plates (10 Total RNA from frozen liver tissue of affected European Burmese
cats was reverse transcribed using oligo-dT and SuperScript II at
50 °C for 1 h. The resultant cDNA was used as template for amplification of the 1700 base pair HEXB cDNA, sequenced in three overlapping segments as described previously for domestic shorthair
cats [10]. Once the putative mutation was identified in the cDNA
(deletion of exon 12), introns 11 and 12 were sequenced from
genomic DNA of two known mutants (that exhibited clinical disease signs), one obligate heterozygote (confirmed by mating),
and one normal cat from an unrelated feline breeding colony.
Primers for intron amplification and sequencing were designed
based on conserved exon/intron boundaries in humans and mice
[13–15] and may be found in Table 1: Intron 11, 1338c and
1406n; Intron 12, 1387c and 1532n. To detect transcripts derived
from cryptic splice acceptor sites in affected Burmese cats, nested
PCR of the 50 end of exon 12 was performed as follows. In round 1,
primers 3U and 1378n [10] were used to amplify exons 9 (30 end),
10, 11 and 12 (50 end), followed by a second-round of nested PCR
with primers 992c and 1378n to amplify minor transcripts containing exon 12 (Table 1). The PCR products were gel purified (Qiagen, Valencia, California, USA) and sequenced with amplification
primers on an Applied Biosystems model 3100 Genetic Analyzer
by the Auburn University Genomics and Sequencing Laboratory.
Diagnostic assay
After the putative mutation was identified in intron 11 of the
feline HEXB gene, a diagnostic assay was developed using primers
1406n and Int11c (within intron 11, Table 1) to confirm the mutation and to genotype feline blood samples submitted by European
Burmese breeders worldwide. Anti-coagulated blood samples
(1 ml) in EDTA were submitted by overnight courier at ambient
temperature for genomic DNA extraction and genotype analysis.
Amplification was performed on a BioRad iCycler using Taq DNA
polymerase with the following cycling parameters: initial denaturation at 95° C for 5 min followed by 38 cycles of 95° C for 45 s,
58° C for 45 s, and 72° C for 2 min, followed by a final elongation
at 72° C for 6 min.
Table 1
Primers for investigation of European Burmese HEXB mutation.
Category
Fragment
Primer
Sequence (50 ? 30 )
Locationa
Sequencing
Intron 11
1338c
1406n
1387c
1532n
3U
1378n
992c
AGT GGA CCC TCT TCA TTT TG
AGA CAA GCT TCT CCA CCA ATG
ATT GGT GGA GAA GCT TGT CTG T
CTG TCA GTC TGT TGT AGG CAT T
TTT CCA GAT CAC TTT GTT CAC TTG
GAC AAG CTT CTC CAC CAA TGA CA
GGA GGA GAT GAA GTG GAA T
1333–1352
1401–1381
1382–1403
1529–1508
968–991
1400–1378
992–1010
Int11c
1406n
GCT GCC TCT TTT TGT GCC AA
AGA CAA GCT TCT CCA CCA ATG
Intron 11
1401–1381
3U
1040n
1280c
1378n
TTT CCA GAT CAC TTT GTT CAC TTG
TCT GCT TCA TGA AAC CCT G
GCT TCC CTG TGA TCC TTT CTG CTC
GAC AAG CTT CTC CAC CAA TGA CA
968–991
1058–1040
1257–1280
1400–1378
Intron 12
Exon 12 (Nested)
Diagnostic
qRT-PCR
Exon 12
+Exon 12
a
Primers based on GenBank sequence S70340 (normal feline HEXB). Additional primers for sequencing feline HEXB can be found in the previous Ref. [10].
55
A.M. Bradbury et al. / Molecular Genetics and Metabolism 97 (2009) 53–59
Quantitative RT-PCR
Table 2
Specific activitya of lysosomal enzymes in cerebral cortex of European Burmese cats.
Quantitative reverse transcription-polymerase chain reaction
(RT-PCR) assays were performed on frozen cerebral cortex tissue
from affected Burmese and normal cats using primers in Table 1.
Total RNA was reverse transcribed using oligo-dT and SuperScript
II reverse transcriptase. Using 0.5 ll of the cDNA template per reaction, primers 3U and 1040n [10], Taq polymerase, and a SYBR
Green core reagents kit, quadruplicate 25 ll quantitative PCR were
performed. An Applied Biosystems model 7700 Sequence Detection
System was used for the assay with the following cycling conditions: 50 °C for 2 min, 95° C for 10 min, and then 45 cycles of
95° C for 30 s, 55° C for 30 s and 72° C for 45 s. As a control, quantitative RT-PCR assays also were performed with primers 1280c
and 1378n [10], which anneals in HEXB exon 12 (the exon deleted
in affected Burmese cats). Because primers 1280c and 1378n detected low but reproducible levels of HEXB transcript (see Results),
it was hypothesized that the Burmese mutation resulted in alternative transcripts containing at least a portion of exon 12. Therefore,
nested PCR was performed to amplify minor transcripts not
detected upon initial sequencing of the HEXB cDNA in affected
Burmese cats (see Sequencing section above for details). Data from
HEXB assays was normalized to glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.12, GAPD) by amplification with primers 101c
(CCT TCA TTG ACC TCA ACT ACA T) and 225n (GAA GAT GGT GAT
GGG CTT T), as described previously [16]. Data is expressed as
mean ± standard deviation (SD).
Sample
bGal
HexA
Hex total
a-Mann
Affected #1
Affected #2
GM2
Normal
135.4 ± 1.9
143.8 ± 11.0
138.9 ± 5.1
42.1 ± 4.5
0.0
0.0
0.0
10.8 ± 0.8
0.0
0.0
2.4 ± 3.4
566.6 ± 65.7
682.5 ± 5.1
823.2 ± 31.9
880.6 ± 193.9
148.3 ± 0.4
Western blot
Western blots were performed with 75 lg protein on a 10%
SDS–polyacrylamide gel. Protein was transferred to nitrocellulose
membranes (BioRad, Hercules, California, USA), which were
blocked in 7.5% nonfat dry milk with 0.05% Tween 20 and probed
with hybridoma supernatant from serum-free cultures. Secondary
antibody (horseradish peroxidase-conjugated goat anti-mouse
IgG/M) was diluted 1:190,000 and visualized with Supersignal
West Dura chemiluminescent substrate. Three independent experiments were performed.
Statistical significance
A one-sample, two-tailed t test was used to determine statistical significance. The null hypothesis was defined as Ho:l = 1,
where l is the quotient defined by GM2 mutant HEXB normal
HEXB after both values were normalized to GAPD levels. Therefore,
if there is no difference in HEXB levels between the GM2 mutant
and normal samples, then l = 1.
Abbreviations used: bGal, b-Galactosidase; HexA, A isozyme of hexosaminidase
measured by MUGS; Hex total, total hexosaminidase measured by MUG; a-Mann,
a-Mannosidase; GM2, cat genotyped as homozygous for the 25 base pair inversion
known to cause GM2 gangliosidosis variant 0 in domestic shorthair cats [10];
Normal, normal cat from unrelated breeding colony.
a
Specific activity expressed as nmol 4MU/mg/h ± SD.
very similar to those described previously for feline GM2 gangliosidosis [1,3,9,17]. Grossly swollen neurons were evident in the
spinal cord and in every region of the brain, including cerebral
cortex, cerebellum, hippocampus and brainstem. Grey and white
matter contained variable increases in both microglia and macroglia, including prominent reactive astrocytes. Degenerative lesions
were evident in white matter. In addition, numerous swollen hepatocytes were apparent in liver sections, which contained large, single vacuoles that stained faintly with PAS, possibly due to leaching
of water-soluble glycoproteins from lysosomes during fixation
(data not shown). Cryopreserved tissue was not available for PAS
staining. Similar hepatocyte pathology has been reported in other
animal models of GM2 gangliosidosis, variant 0 [1,3,5,8]. The presence of distended Kupffer cells also was noted throughout the liver.
After histopathology consistent with a gangliosidosis was observed, lysosomal enzyme assays were performed. Although
b-galactosidase and a-mannosidase activities were elevated above
normal, activity toward MUG and MUGS was severely reduced, suggesting a deficiency of the b-subunit of hexosaminidase (Table 2).
To corroborate the findings of the enzyme assay, high-performance thin layer chromatography was performed on brain lipids.
Storage of GM2 ganglioside in the cerebral cortex of affected Burmese cats was observed when compared to normal cats (in which
GM2 is undetectable) or cats known to be affected with either GM1
[16,18] or GM2 gangliosidosis [1,10] (Fig. 1). Also of interest was
the apparent reduction of GM1 ganglioside in HPTLC samples from
affected Burmese cats or a cat homozygous for a separate GM2
gangliosidosis mutation [10], compared to samples from normal
cat brain.
Because two other HEXB mutations are screened routinely in
our laboratory by DNA-based diagnostic assays, genomic DNA from
the affected Burmese cats was tested for these mutations [10,11],
which were not present (data not shown). Since these results suggested the presence of a novel mutation, HEXB cDNA from Burmese
cats was sequenced in three overlapping segments, as described
Results
Multiple litters of kittens were obtained by breeding a male
European Burmese cat from Norway to a female European Burmese
cat in Illinois (USA). Three of the offspring presented with clinical
disease traits very similar to those reported previously for HEXB
mutations in domestic shorthair and Korat cats [1,3]. Disease onset
appeared at 6–8 weeks of age, characterized by mild intention
tremors progressing to ataxia, hypermetria and occasional falling.
Severe ambulation difficulties and compromise of motor control
led to difficulty eating and using the litter box. Affected cats were
smaller than normal in body size and weight, and euthanasia was
performed at approximately 4 months of age due to neurological
disease progression. Parents and siblings of affected animals
showed no clinical disease signs.
Brain and liver tissue from affected cats exhibited histopathological abnormalities typical of a lysosomal storage disease and
Fig. 1. Thin layer chromatography of brain lipids from affected and normal cats.
Total lipids were extracted from the cerebral cortex and separated as detailed in
Materials and methods. Samples from affected Burmese cats (A1 and A2) demonstrated storage of GM2 ganglioside, as did the positive control tissue (+) from a cat
genotyped to have the previously reported 25 base pair inversion causing GM2
gangliosidosis in domestic shorthair cats [10]. In comparison, no GM2 ganglioside
was observed in two normal cats (N1 and N2) or in a cat with GM1 gangliosidosis
[16] (G1). Gangliosides GM1, GM2 and GM3 were run as standards on the left-hand
side of the plate.
56
A.M. Bradbury et al. / Molecular Genetics and Metabolism 97 (2009) 53–59
Fig. 2. Mutation analysis of hexosaminidase b-subunit in affected Burmese cats. (A) Partial sequence of intron 11 and exon 12 from Normal cats. The intron 11–exon 12
junction is designated by a vertical line, and the 15 base pair deletion in affected Burmese cats is underlined. Splice sites (AG) are designated by horizontal lines atop the text,
and numbers 1–3 correlate with protein products in B derived from the use of each splice site. Site 1 is the normal splice site, while sites 2 and 3 are active only in affected
cats. (B) Schematic representation of the hexosaminidase b-subunits in Normal and Affected Burmese cats. In normal cats, the preferred splice acceptor site (1) is used,
producing a 538 amino acid (AA) protein. Shown are the first 19 AAs derived from exon 12. In affected Burmese cats, three putative mutant proteins are derived from the
three splice sites shown in (A): (1*) The preferred splice acceptor is deleted resulting in splicing of exon 11 directly to exon 13, which produces an 18 amino acid substitution
in the mutant protein (LGQVPLVKDSGVQKTSRV) beginning at codon 454. In addition, a premature termination codon truncates the mutant protein to 472 AAs. (2) The use of a
cryptic splice site at position 24 of intron 11 introduces a 3 AA substitution (AIQ) followed by a premature termination codon, truncating the protein to 456 AAs. (3) A
cryptic splice site at base 1363 (position +8 of exon 12) results in deletion of the first 3 AAs derived from exon 12, yielding a 535 AA protein that is otherwise identical to the
normal feline b-subunit.
previously [10]. Comparison of the Burmese HEXB sequences with
the normal feline cDNA identified a 91 base pair deletion at the 30
end of the coding region. Upon examination of conserved intron–
exon boundaries for HEXB in humans and mice, it was determined
that the 91 base pair deletion resulted from the complete excision
of exon 12. Using genomic DNA from affected cats, the introns
flanking exon 12 were sequenced with primers 1338c and 1406n
(for intron 11) and 1387c and 1532n (for intron 12, see Materials
and methods). A 15 base pair deletion was found at the 30 end of
intron 11 and included the AG splice acceptor site. Deletion of
the AG splice acceptor site resulted in the removal of intron 11,
exon 12, and intron 12 from the mRNA. Splicing of exon 11 directly
to exon 13 caused an 18 amino acid substitution followed by premature termination at codon 472, compared to the normal protein
of 538 amino acids (Fig. 2). The mutation was further verified by
sequencing the genomic DNA of obligate heterozygote (confirmed
by mating) and presumed normal Burmese cats. Beginning at the
mutation site, overlapping peaks were evident in the fluorescent
trace files of obligate heterozygotes, demonstrating the presence
of two separate alleles in cats known to have produced affected offspring (data not shown).
Because neurological diseases similar to that of the probands
were reported in other European Burmese families, blood samples
were submitted from international sources to be screened for the
15 base pair deletion. To date, 145 cats have been tested from five
countries, including the USA (nine states). Twenty-one heterozygotes have been identified, for a carrier frequency rate of 14.5%.
Heterozygotes were detected in 11 of the 26 total catteries that
submitted samples (i.e., 42.3% of all catteries had at least one heterozygote). No difference in carrier frequency was noted between
North American and non-North American catteries. All non-hetero-
zygote animals identified through the international screening program have been normal, since the only known affected European
Burmese cats were the probands described above.
To determine the level of HEXB mRNA expressed in affected
cats, quantitative RT-PCR was performed with primers 3U and
1040n [10] and data was normalized to GAPD. In cerebral cortical
tissue from two clinically affected cats, HEXB mRNA levels were
1.53 ± 0.31 (SD) times higher than normal (P < 0.001) (Table 3).
Melting curve analysis of the PCR product demonstrated a single
peak suggestive of a single amplicon, which subsequently was
gel purified and sequenced for confirmation of identity and purity
(data not shown). To further confirm the mutation, quantitative
RT-PCR was performed with a primer set (1280c–1378n) in which
one of the primers annealed in exon 12 (1378n). Primers 1280c and
1378n generated low but reproducible levels of HEXB amplification
in affected Burmese brain tissue (2.5 ± 0.47% of normal, P < 0.001,
Table 3), suggesting that a small percentage of mRNA from affected
cats contained at least a portion of exon 12.
Table 3
HEXB expressiona in Burmese cats with GM2 gangliosidosis compared to normal cats.
Primer pair
Affected #1
Affected #2
Combined
3U–1040n
1280c–1378n
1.59 ± 0.34*
0.026 ± 0.006**
1.48 ± 0.31*
0.024 ± 0.004**
1.53 ± 0.31**
0.025 ± 0.005**
a
HEXB mRNA levels in the cerebral cortex of two affected Burmese cats were
normalized to GAPD expression. Data is expressed as the fold change in HEXB levels
compared to normal. Primers 3U and 1040n anneal upstream of the mutation in
exons 8 and 9, respectively; primers 1280c and 1378n anneal in exons 11 and 12
(mutation site), respectively.
*
P < 0.05.
**
P < 0.001.
A.M. Bradbury et al. / Molecular Genetics and Metabolism 97 (2009) 53–59
Examination of the intron 11–exon 12 DNA sequence of feline
HEXB revealed that two additional AG sequences exist in the vicinity of the preferred splice acceptor: position 24 of intron 11 and
position 1363 of the normal feline HEXB cDNA (eight bases downstream of the beginning of exon 12; numbering based on GenBank
Accession No. S70340). Nested PCR was performed with a non-coding strand primer that annealed in exon 12 to amplify potential
minor HEXB transcripts from affected Burmese cat brain tissue.
Second-round PCR product from the nested reactions was sequenced, revealing the expected exon 12 sequence in normal cats
and two alternative sequences in affected Burmese cats derived
from the cryptic splice acceptor sites at base 24 of intron 11
and base 1363 of exon 12 (Fig. 2).
To analyze hexosaminidase b-subunit protein levels in affected
Burmese cats, Western blots were performed with an antibody that
recognizes the b-subunit from normal cats and cats possessing the
previously described 25 base pair inversion in the HEXB gene [10].
As shown in Fig. 3, profoundly reduced levels of the Hex b-subunit
were detected in the cerebral cortex of affected Burmese cats. Because the faint band from the affected Burmese cerebral cortex is of
similar molecular weight to the normal band, it represents the 535
amino acid protein generated by cryptic splice site #3 (see Fig. 2).
Although the two other mutant transcripts detected by RT-PCR and
automated fluorescent DNA sequencing were expected to generate
truncated b-subunit proteins of 472 and 456 amino acids, no bands
of lower molecular weight were detected on Western blots.
Discussion
As in humans, GM2 gangliosidosis in cats is caused by several
distinct mutations, the fourth of which is reported herein to be a
15 base pair deletion at the 30 end of intron 11 that includes the
splice acceptor site and leads to aberrant RNA splicing. In humans,
at least eight splicing mutations have been defined, four of which
affect the splice acceptor site (see Human Gene Mutation Database,
http://www.hgmd.cf.ac.uk/ac/index.php) [19–22]. Like the muta-
Fig. 3. Western analysis of mutant hexosaminidase b-subunit in Burmese cats.
Protein from affected Burmese (lane 1) and normal (lane 2) cats was probed with a
previously described antibody (#240) to the feline Hex b-subunit [10]. Hex bsubunit levels were profoundly reduced in the affected Burmese brain. Since the
antibody was generated with a peptide mapping approximately 200 AAs upstream
from the mutation, it was expected to recognize all of the mutant b-subunits
generated from cryptic splice sites in Burmese cats (see Fig. 2). However, neither the
472 or 456 AA variant was detected on Western blots, suggesting that the mutant
proteins were not translated or were degraded after synthesis. Probing the blots
with an antibody to glyceraldehyde-3-phosphate dehydrogenase (GAPD) verified
that equal amounts of protein were loaded. Molecular weights in kilodaltons are
depicted on the right-hand side of the blot. Shown is one representative example of
three experiments.
57
tion described herein, at least two of the human mutations result
in the use of cryptic splice sites, thereby generating abnormal transcripts [19,21,23]. Interestingly, reports from human Sandhoff disease patients have shown that mutations causing aberrant splicing
need not directly involve the splice acceptor site and may include
the following: (1) exon sequences, such as a C to T substitution in
HEXB exon 11 that prevents normal splicing of exon 10 to exon 11
[21,23]; (2) intron mutations that generate novel splice sites preferred by the cell’s splicing machinery, such as a G to A substitution
at position 26 of intron 12 [20], and (3) intron mutations remote
from the splice acceptor site resulting in selection of a cryptic
splice site that is distinct from both the normal and mutated
sequence. For example, an A to G substitution at position 17 of
intron 10 leads to the almost exclusive use of a cryptic splice site
at 37 of the same intron [19]. As the above reports demonstrate,
numerous signals and mechanisms exist for selection and use of
splice sites in mammalian transcripts (reviewed in [24,25]).
In addition to loss of exon 12 from the mRNA and activation of
cryptic splice sites as discussed above, the 15 base pair deletion in
affected Burmese cats led to HEXB mRNA levels 1.5 times normal.
Presumably, this is the result of a molecular feedback mechanism
in which low enzymatic activity or increased levels of GM2 ganglioside activate compensatory increases in HEXB transcription.
Although rigorous measurement of the proportional contribution
of each transcript to overall HEXB mRNA levels was not performed,
it is clear that the transcript pool consisted primarily of exon
12-deficient mRNA, based on the following information: (a) automated fluorescent DNA sequencing of the affected HEXB cDNA
showed no evidence of background peaks that would have been
evident if transcripts generated from cryptic splice sites were present in substantial quantities; (b) when quantitative RT-PCR was
performed with a primer pair specific for exon 12-containing transcripts, only 2.5% of the normal level was detected, compared to
153% of normal levels when a primer pair upstream of exon 12
was used (Table 3); (c) successful amplification of exon 12-containing transcripts was possible only after two rounds of nested
PCR, probably due to low levels of such transcripts in the starting
template pool.
Although transcripts containing exon 12 were relatively rare,
transcript 3 derived from a cryptic splice site within exon 12
(Fig. 2) was sufficiently abundant to generate a faint signal upon
Western blotting. Transcript 3, containing an in-frame deletion of
the first nine nucleotides of exon 12, generated a Hex b-subunit
of 535 amino acids identical to the normal b-subunit but for a three
amino acid deletion. Because the size difference was minor and the
antibody employed recognized an epitope remote from the mutation, the affected (transcript 3) and normal b-subunits were indistinguishable on Western blots. Therefore, it is clear that only a
small amount of transcript is needed to generate detectable quantities of the stable b-subunit protein.
Conversely, no protein bands on Western blots could be attributed to the other splice variants identified in affected brain tissue,
including the major, exon 12-deleted transcript. Differing substantially from the normal b-subunit, the lower molecular weight
mutant b-subunits of 472 and 456 amino acids (Fig. 2) may be targeted for endoplasmic reticulum-associated degradation (ERAD),
as reported for other lysosomal enzyme mutants [26–28] or abnormal proteins resulting from enzyme mutants [29–31]. In this scenario, components of the cell’s unfolded protein response
machinery could be responsible for degrading mutant b-subunits
that are marked as abnormal.
The affected European Burmese cats analyzed in this report
demonstrated deficient activity of both HexA and HexB accompanied by elevated levels of other lysosomal enzymes such as
b-galactosidase and a-mannosidase. While GM2 ganglioside was
stored in affected cerebral cortex, GM1 ganglioside appeared to
58
A.M. Bradbury et al. / Molecular Genetics and Metabolism 97 (2009) 53–59
be below normal. However, HPTLC plates are loaded based on the
sialic acid content of the samples. In affected GM2 cats, a higher
percentage of the total sialic acid content in brain is contributed
by GM2 ganglioside. Therefore, comparing a fixed amount of total
sialic acid from normal and affected brain samples causes an
apparent reduction in the level of other gangliosides (such as
GM1) in affected tissue. Similar results have been reported in Sandhoff dogs, mice, cats and humans [5,6,32]. In most instances,
ganglioside content was measured as the percentage of total material on the HPTLC plate by comparison of band intensities. When
actual ganglioside concentration was measured by scraping bands
from the HPTLC plate and measuring sialic acid levels, a much less
pronounced reduction of GM1 ganglioside was documented [32].
Thus, the actual concentration (versus percentage) of GM1 may
be reduced in affected GM2 cats, but the reduction is likely to be
exaggerated on HPTLC plates loaded based on total sialic acid content of the samples. If real, small reductions in GM1 levels may be
due to overexpression of b-galactosidase, the lysosomal enzyme
responsible for GM1 catabolism.
Cats traditionally have been the favored species for experimental neuroscience and continue to serve as exceptional models for
translational therapeutic studies of human diseases. Domestic cats
are maintained easily in breeding colonies, produce 4–6 progeny
per breeding, have multiple breeding cycles per year, have a short
gestation of 63 ± 1 days, and respond well to surgery and experimental procedures [33]. There is substantial knowledge about
feline prenatal and postnatal hematopoiesis, fetal, neonatal and
adult immunology and bone marrow transplantation [34,35].
Unlike rodents, cats can be evaluated readily by routine clinical
procedures and are large enough to permit frequent sampling of
tissues and body fluids. Even though inbred mice provide some
experimental advantages, cats more closely mimic the therapeutic
challenges and limitations that are likely to be encountered in human patients, such as immunologic heterogeneity, brain complexity and size. While the mouse brain is 1000 times smaller than
that of a human infant [36], the cat brain is only 16 times smaller
than an infant’s brain and presents a more realistic model to study
the challenges of distributing therapeutic agents globally throughout the central nervous system. Additionally, a number of naturally
occurring neurodegenerative diseases have been documented in
cats, with pathology very similar to the corresponding human conditions [37].
Feline GM2 gangliosidosis and mucopolysaccharidosis VI
[38,39] are unique among large animal models of lysosomal storage disease in that multiple pathogenic mutations have been reported. This suggests that cats effectively model the molecular
genetics of human lysosomal storage diseases as well. Previously
described feline HEXB mutations include a 25 base pair inversion
at the 30 terminus of the coding region [10], a single base deletion
in exon 1 [11], and a nonsense mutation in exon 7 [12]. Although
the mutations are of different types, all generate premature termination codons and normal to elevated levels of HEXB mRNA. It is
not known whether the exon 1 or exon 7 mutations produce
cross-reactive material, but the 25 base pair inversion generates
mutant b-subunit protein levels <20% of normal [10]. Aberrant
splicing from cryptic splice sites has not been reported with the
other feline mutations, but alternative splicing and removal of
exon 7 has been documented as a normal physiological process
in both Korat and Japanese domestic shorthair cats [11,12]. The
mutation reported herein further establishes the similarity between molecular genetic mechanisms of Sandhoff disease in cats
and humans, providing a more comprehensive understanding of
pathogenesis which may be utilized for development of therapeutic strategies such as enzyme enhancement therapy [40], stem cell
transplantation [41] and/or gene therapy. To this end, studies are
underway to evaluate adeno-associated viral vectors for treatment
of the feline gangliosidoses, testing in cats a therapeutic strategy
proven enormously successful in knockout mice [42,43]. Translational therapeutic studies in cats are one component of a multi-faceted project designed to prepare gene therapy for eventual
treatment of humans with GM2 gangliosidosis (see www.tsgtconsortium.com). The continued development and characterization of
large animal models of GM2 gangliosidosis is a critical step in
translational research designed to initiate human clinical trials.
Acknowledgments
We express our sincere appreciation to Ms. Robin Bryan of the
Chamsey European Burmese Cattery, Urbana, IL, for her commitment to discover the cause of the illness that befell her kittens
and for organizing the international screening program for detection of carriers in the European Burmese breed. We also thank
Dr. Gaurav Tyagi, Veterinary Pathology Resident, Department of
Veterinary Pathobiology, College of Veterinary Medicine, University of Illinois, Urbana, IL who necropsied the proband kittens
and provided us with tissues for the studies reported herein.
References
[1] L.C. Cork, J.F. Munnell, M.D. Lorenz, J.V. Murphy, H.J. Baker, M.C. Rattazzi, GM2
ganglioside lysosomal storage disease in cats with beta-hexosaminidase
deficiency, Science 196 (1977) 1014–1017.
[2] D. Hasegawa, O. Yamato, M. Kobayashi, M. Fujita, S. Nakamura, K. Takahashi, H.
Satoh, T. Shoda, D. Hayashi, M. Yamasaki, Y. Maede, T. Arai, H. Orima, Clinical
and molecular analysis of GM2 gangliosidosis in two apparent littermate
kittens of the Japanese domestic cat, J. Fel. Med. Surg. 9 (2007) 232–237.
[3] E.A. Neuwelt, W.G. Johnson, N.K. Blank, M.A. Pagel, C. Maslen-McClure, M.J.
McClure, P.M. Wu, Characterization of a new model of GM2-gangliosidosis
(Sandhoff’s disease) in Korat cats, J. Clin. Invest. 76 (1985) 482–490.
[4] K.R. Pierce, S.D. Kosanke, W.W. Bay, C.H. Bridges, Animal model of human
disease: GM2 gangliosidosis, Am. J. Pathol. 83 (1976) 419–422.
[5] J.F. Cummings, P.A. Wood, S.U. Walkley, A. de Lahunta, M.E. DeForest, GM2
gangliosidosis in a Japanese spaniel, Acta Neuropathol. 67 (1985) 247–253.
[6] Y. Eto, L. Autilio-Gambetti, J.T. McGrath, Canine GM2-gangliosidosis: chemical
and enzymatic features, Adv. Exp. Med. Biol. 174 (1984) 431–440.
[7] K. Sango, S. Yamanaka, A. Hoffmann, Y. Okuda, A. Grinberg, H. Westphal, M.P.
McDonald, J.N. Crawley, K. Sandhoff, K. Suzuki, Mouse models of Tay-Sachs and
Sandhoff diseases differ in neurologic phenotype and ganglioside metabolism,
Nat. Genet. 11 (1995) 170–176.
[8] D. Phaneuf, N. Wakamatsu, J.Q. Huang, A. Borowski, A.C. Peterson, S.R.
Fortunato, G. Ritter, S.A. Igdoura, C.R. Morales, G. Benoit, B.R. Akerman, D.
Leclerc, N. Hanai, J.D. Marth, J.M. Trasler, R.A. Gravel, Dramatically different
phenotypes in mouse models of human Tay-Sachs and Sandhoff diseases,
Hum. Mol. Genet. 5 (1996) 1–14.
[9] D.R. Martin, N.R. Cox, N.E. Morrison, D.M. Kennamer, S.L. Peck, A.N. Dodson,
A.S. Gentry, B. Griffin, M.D. Rolsma, H.J. Baker, Mutation of the GM2 activator
protein in a feline model of GM2 gangliosidosis, Acta Neuropathol. 110 (2005)
443–450.
[10] D.R. Martin, B.K. Krum, G.S. Varadarajan, T.L. Hathcock, B.F. Smith, H.J. Baker,
An inversion of 25 base pairs causes feline GM2 gangliosidosis variant 0, Exp.
Neurol. 187 (2004) 30–37.
[11] L.L. Muldoon, E.A. Neuwelt, M.A. Pagel, D.L. Weiss, Characterization of the
molecular defect in a feline model for type II GM2-gangliosidosis (Sandhoff
disease), Am. J. Pathol. 144 (1994) 1109–1118.
[12] Y. Kanae, D. Endoh, O. Yamato, D. Hayashi, S. Matsunaga, H. Ogawa, Y. Maede,
M. Hayashi, Nonsense mutation of feline beta-hexosaminidase beta-subunit
(HEXB) gene causing Sandhoff disease in a family of Japanese domestic cats,
Res. Vet. Sci. 82 (2007) 54–60.
[13] R.L. Proia, Gene encoding the human beta-hexosaminidase beta chain:
extensive homology of intron placement in the alpha- and beta-chain genes,
Proc. Natl. Acad. Sci. USA 85 (1988) 1883–1887.
[14] B.L. Triggs-Raine, G. Benoit, T.J. Salo, J.M. Trasler, R.A. Gravel, Characterization
of the murine beta-hexosaminidase (HEXB) gene, Biochim. Biophys. Acta 1227
(1994) 79–86.
[15] S. Yamanaka, O.N. Johnson, F. Norflus, D.J. Boles, R.L. Proia, Structure and
expression of the mouse beta-hexosaminidase genes, Hexa and Hexb,
Genomics 21 (1994) 588–596.
[16] D.R. Martin, B.A. Rigat, P. Foureman, G.S. Varadarajan, M. Hwang, B.K. Krum,
B.F. Smith, J.W. Callahan, D.J. Mahuran, H.J. Baker, Molecular consequences of
the pathogenic mutation in feline GM1 gangliosidosis, Mol. Genet. Metab. 94
(2008) 212–221.
[17] O. Yamato, S. Matsunaga, K. Takata, K. Uetsuka, H. Satoh, T. Shoda, Y. Baba, A.
Yasoshima, K. Kato, K. Takahashi, M. Yamasaki, H. Nakayama, K. Doi, Y. Maede,
H. Ogawa, GM2-gangliosidosis variant 0 (Sandhoff-like disease) in a family of
Japanese domestic cats, Vet. Rec. 155 (2004) 739–744.
A.M. Bradbury et al. / Molecular Genetics and Metabolism 97 (2009) 53–59
[18] H.J. Baker Jr., J.R. Lindsey, G.M. McKhann, D.F. Farrell, Neuronal GM1
gangliosidosis in a Siamese cat with beta-galactosidase deficiency, Science
174 (1971) 838–839.
[19] M. Fujimaru, A. Tanaka, K. Choeh, N. Wakamatsu, H. Sakuraba, G. Isshiki, Two
mutations remote from an exon/intron junction in the beta-hexosaminidase
beta-subunit gene affect 30 -splice site selection and cause Sandhoff disease,
Hum. Genet. 103 (1998) 462–469.
[20] T. Nakano, K. Suzuki, Genetic cause of a juvenile form of Sandhoff disease.
Abnormal splicing of beta-hexosaminidase beta chain gene transcript due to a
point mutation within intron 12, J. Biol. Chem. 264 (1989) 5155–5158.
[21] N. Wakamatsu, H. Kobayashi, T. Miyatake, S. Tsuji, A novel exon mutation in
the human beta-hexosaminidase beta-subunit gene affects 30 splice site
selection, J. Biol. Chem. 267 (1992) 2406–2413.
[22] T. Yoshizawa, Y. Kohno, S. Nissato, S. Shoji, Compound heterozygosity with
two novel mutations in the HEXB gene produces adult Sandhoff disease
presenting as a motor neuron disease phenotype, J. Neurol. Sci. 195 (2002)
129–138.
[23] B. McInnes, M. Potier, N. Wakamatsu, S.B. Melancon, M.H. Klavins, S. Tsuji, D.J.
Mahuran, An unusual splicing mutation in the HEXB gene is associated with
dramatically different phenotypes in patients from different racial
backgrounds, J. Clin. Invest. 90 (1992) 306–314.
[24] A.E. House, K.W. Lynch, Regulation of alternative splicing: more than just the
ABCs, J. Biol. Chem. 283 (2008) 1217–1221.
[25] Z. Wang, C.B. Burge, Splicing regulation: from a parts list of regulatory
elements to an integrated splicing code, RNA 14 (2008) 802–813.
[26] J.Q. Fan, S. Ishii, N. Asano, Y. Suzuki, Accelerated transport and maturation of
lysosomal alpha-galactosidase A in Fabry lymphoblasts by an enzyme
inhibitor, Nat. Med. 5 (1999) 112–115.
[27] S. Ishii, H.H. Chang, K. Kawasaki, K. Yasuda, H.L. Wu, S.C. Garman, J.Q. Fan,
Mutant alpha-galactosidase A enzymes identified in Fabry disease patients
with residual enzyme activity: biochemical characterization and restoration of
normal intracellular processing by 1-deoxygalactonojirimycin, Biochem. J. 406
(2007) 285–295.
[28] I. Ron, M. Horowitz, R. Department of Cell, T.A.U.R.A.I. Immunology, ER
retention and degradation as the molecular basis underlying Gaucher disease
heterogeneity, Hum. Mol. Genet. 14 (2005) 2387–2398.
[29] S.J. Kim, Z. Zhang, E. Hitomi, Y.C. Lee, A.B. Mukherjee, Endoplasmic reticulum
stress-induced caspase-4 activation mediates apoptosis and neurodegeneration
in INCL, Hum. Mol. Genet. 15 (2006) 1826–1834.
[30] S.J. Kim, Z. Zhang, Y.C. Lee, A.B. Mukherjee, Palmitoyl-protein thioesterase-1
deficiency leads to the activation of caspase-9 and contributes to rapid
neurodegeneration in INCL, Hum. Mol. Genet. 15 (2006) 1580–1586.
59
[31] Z. Zhang, Y.C. Lee, S.J. Kim, M.S. Choi, P.C. Tsai, Y. Xu, Y.J. Xiao, P. Zhang, A.
Heffer, A.B. Mukherjee, Palmitoyl-protein thioesterase-1 deficiency mediates
the activation of the unfolded protein response and neuronal apoptosis in
INCL, Hum. Mol. Genet. 15 (2006) 337–346.
[32] R.C. Baek, D.R. Martin, N.R. Cox, T.N. Seyfried, Comparative analysis of brain
lipids in mice, cats, and humans with Sandhoff Disease, Lipids, in press,
doi:10.1007/s11745.008.3268.
[33] B. Griffin, H.J. Baker, Domestic cats as laboratory animals, in: J.G. Fox, L.C.
Anderson, F.M. Lowe, F.W. Quimby (Eds.), Laboratory Animal Medicine,
Academic Press, New York, NY, 2002, pp. 460–482.
[34] M. Haskins, H.J. Baker, E. Birkenmeier, P.M. Hoogerbrugge, B.J.H.M. Poorthuis,
T. Sakiyama, R.M. Shull, R.M. Taylor, M.A. Thrall, S.U. Walkley, Transplantation
in animal model systems, in: R.J. Desnick (Ed.), Treatment of Genetic Diseases,
Churchill Livingstone, New York, 1991, pp. 183–201.
[35] S.U. Walkley, K. Dobrenis, Bone marrow transplantation for lysosomal
diseases, Lancet 345 (1995) 1382–1383.
[36] C.H. Vite, M.A. Passini, M.E. Haskins, J.H. Wolfe, Adeno-associated virus vectormediated transduction in the cat brain, Gene Ther. 10 (2003) 1874–1881.
[37] R.D. Jolly, S.U. Walkley, Lysosomal storage diseases of animals: an essay in
comparative pathology, Vet. Pathol. 34 (1997) 527–548.
[38] A.C. Crawley, F.H. Muntz, M.E. Haskins, B.R. Jones, J.J. Hopwood, Prevalence of
mucopolysaccharidosis type VI mutations in Siamese cats, J. Vet. Int. Med. 17
(2003) 495–498.
[39] A.C. Crawley, G. Yogalingam, V.J. Muller, J.J. Hopwood, Two mutations within a
feline mucopolysaccharidosis type VI colony cause three different clinical
phenotypes, J. Clin. Invest. 101 (1998) 109–119.
[40] G.H. Maegawa, M. Tropak, J. Buttner, T. Stockley, F. Kok, J.T. Clarke, D.J.
Mahuran, Pyrimethamine as a potential pharmacological chaperone for lateonset forms of GM2 gangliosidosis, J. Biol. Chem. 282 (2007) 9150–9161.
[41] J.P. Lee, M. Jeyakumar, R. Gonzalez, H. Takahashi, P.J. Lee, R.C. Baek, D. Clark,
H. Rose, G. Fu, J. Clarke, S. McKercher, J. Meerloo, F.J. Muller, K.I. Park, T.D.
Butters, R.A. Dwek, P. Schwartz, G. Tong, D. Wenger, S.A. Lipton, T.N. Seyfried,
F.M. Platt, E.Y. Snyder, Stem cells act through multiple mechanisms to benefit
mice with neurodegenerative metabolic disease, Nat. Med. 13 (2007) 439–
447.
[42] M.B. Cachon-Gonzalez, S.Z. Wang, A. Lynch, R. Ziegler, S.H. Cheng, T.M. Cox,
Effective gene therapy in an authentic model of Tay-Sachs related diseases,
Proc. Natl. Acad. Sci. USA 103 (2006) 10373–10378.
[43] M.L. Broekman, R.C. Baek, L.A. Comer, J.L. Fernandez, T.N. Seyfried, M. SenaEsteves, Complete correction of enzymatic deficiency and neurochemistry in
the GM1 gangliosidosis mouse brain by neonatal adeno-associated virusmediated gene delivery, Mol. Ther. 15 (2007) 30–37.