Impaired Cell Adhesion and Apoptosis in a

Novel CLN9 Batten Disease Variant

Angela Schulz, MD,1 Sumeer Dhar, PhD,1,2 Svetlana Rylova, PhD,1,3 Ghassan Dbaibo, MD,4
Joseph Alroy, DVM,5 Christian Hagel, MD,6 Isabelo Artacho, MD,7 Alfried Kohlschütter, MD,8
Simon Lin, MD,9 and Rose-Mary Boustany, MD1

We describe the ninth variant of neuronal ceroid lipofuscinosis (NCL) or Batten disease, due to defects in a putative new
gene, CLN9. We therefore refer to the new variant as CLN9-deficient. Two Serbian sisters and two German brothers are
described. Their clinical history is characteristic for juvenile NCL. They show similar gene expression patterns. The
existence of this variant is supported by the presence of curvilinear inclusions, fingerprint profiles, and granular osmi-
ophilic deposits in neurons, lymphocytes, and conjunctival cells. Enzyme screening and sequencing of the coding regions
of other NCL genes was negative. CLN9-deficient cells have a distinctive phenotype. They have rounded cell bodies, have
prominent nucleoli, attach poorly to the culture dish, and are sensitive to apoptosis but have increased growth rates.
Gene expression of proteins involved in cell adhesion and apoptosis is altered in these cells. Sphingolipid metabolism is
also perturbed. They have decreased levels of ceramide, sphingomyelin, lactosylceramide, ceramide trihexoside, and glo-
boside and increased activity of serine palmitoyl transferase.
Ann Neurol 2004;56:342–350

A novel neuronal storage disease is described in two
Serbian sisters and two German brothers. The clinical
presentation is similar to juvenile neuronal ceroid lipo-
fuscinosis (JNCL). The NCLs are a group of inherited
neurodegenerative disorders. Clinical features include
visual loss, mental and motor deterioration, seizures,
and early death.1,2 The diagnosis of NCL was based on
clinical course and appearance of inclusions in cells and
now is confirmed by enzyme and/or genetic testing.
Classic variants with known gene defects include in-
fantile NCL (INCL), late infantile NCL (LINCL),
JNCL, and a rare adult variant.3– 6 Atypical variants
include the Finnish variant, the Costa Rican or Portu-
guese variant, Northern epilepsy with mental retarda-
tion (EPMR), and the Turkish variant, with some cases
allelic to EPMR.7–11 Another atypical variant is juve-
nile variant with GRODS (granular osmiophilic depos-
its).12,13 Tissues from NCL patients contain autofluo-
rescent membrane bound inclusions with variable
ultrastructural characteristics. These inclusions are
granular, curvilinear, or fingerprint-like. The genes
CLN1 (INCL and juvenile variant with GRODS) and
CLN2 (LINCL) both code for lysosomal enzymes. The
CLN1 gene product is lysosomal palmitoyl-protein
thioesterase 1 (PPT1).14 The CLN2 gene product is
tripeptidyl-peptidase 1 (TPP1).15 The CLN3 (JNCL),
CLN8 (EPMR), and CLN6 (Portuguese/Costa Rican
variant) genes code for novel transmembrane pro-
teins.16 –19 The CLN5 protein is a soluble glycopro-
tein.20,21 The CLN8 protein possesses a Lag1 motif
like other TLC proteins (TRAM-Lag1p-CLN8).22 The
Lag1 motif imparts ceramide synthase activity to
yeast.23,24
The patients we describe had typical autofluorescent
inclusions in brain, lymphocytes, and conjunctiva. En-
zyme assays as well as molecular tests for known vari-
ants of NCL were normal. Other storage diseases with
neuronal involvement were ruled out. Fibroblasts from
patients have a distinctive phenotype. This study de-
scribes clinical characteristics of the disease and pro-
vides biological, biochemical, and gene expression clues
to its pathogenesis.

From the 1Departments of Pediatrics and Neurobiology, Duke Uni-
versity Medical Center, Durham, NC; 2Division of Clinical Phar-
macology, Department of Medical Sciences, Uppsala University
Hospital; 3Institute of Molecular Biosciences, Section of Veterinary
Medical Biochemistry, Biomedical Center, Uppsala, Sweden; 4De-
partments of Pediatrics and Biochemistry, American University of
Beirut, Beirut, Lebanon; 5Department of Pathology, Tufts Univer-
sity School of Medicine and Veterinary Medicine, New England
Medical Center, Boston, MA; 6Department of Neuropathology,
University of Hamburg, Germany; 7University Medical Center,
University of California San Francisco, Fresno, CA; 8Department of
Pediatrics, University of Hamburg, Hamburg, Germany; and 9Duke
Comprehensive Cancer Center and Duke Center for Bioinformatics
and Computational Biology, Duke University Medical Center,
Durham, NC.
Received Sep 10, 2003, and in revised form Feb 5 and Apr 27,
2004. Accepted for publication Mar 8, 2004.
Published online Jul 27, 2004 in Wiley InterScience
(www.interscience.wiley.com). DOI: 10.1002/ana.20187
Address correspondence to Dr Boustany, Duke University Medical
Center, MSRB Box 2604, Research Drive, Durham, NC, 27710.
E-mail: boust001@mc.duke.edu

342 © 2004 American Neurological Association

Published by Wiley-Liss, Inc., through Wiley Subscription Services
Materials and Methods
Cell Lines
CLN9-deficient fibroblast lines were derived from skin biop-
sies of one Serbian sister and one German brother. The
CLN1-deficient fibroblast lines were provided by Dr S. Hoff-
man (University of Texas, Southwestern Medical Center).
CLN2-, CLN3-, and CLN6-deficient fibroblast lines were
from patients at Duke University Medical Center. Use of
human cell lines in research is covered by an approved Duke
institutional review board protocol.
Tissue Culture
Fibroblasts were grown at 37°C with 5% CO2 in Dulbecco’s
modified Eagle medium (DMEM) (Gibco-Invitrogen, Grand
Island, NY) containing 10% fetal bovine serum (FBS)
(Gibco-Invitrogen) and 1% antibiotics/antimycotics.
Affymetrix GeneChip Analysis
Preparation of Total RNA/Probe Labeling
RNA was isolated using RNeasy Mini Kit (Qiagen, Valencia,
CA). Double-strand cDNA then was synthesized using Super-
script II reverse transcriptase (Gibco-Invitrogen) (Affymetrix,
Santa Clara, CA). Biotin-labeled cRNA was produced from
cDNA by an in vitro transcription reaction. Twenty micrograms
of fragmented cRNA was hybridized at 45°C for 16 hours with
the hybridization target kit, and then probes were washed and
stained in Streptavidin Phycoerythin solution (SAPE, final con-
centration 10 ␮g/ml, Molecular Probes, Eugene, OR) solution
for 30 minutes, with a final wash at 25°C at 10 cycles (four
mixes/cycle), in the Affymetrix Fluidics Station 400. Probe ar-
rays were scanned at 570nm wavelength (pixel size, 3␮m).
Data Acquisition
GeneChip analysis was run four times using the Affymetrix
HuFL 6900 GeneChip and twice using the HG-U95Av2
GeneChip. The Affymetrix Test 2 array was used to assess
quality of target RNA. After scanning probe arrays, data were
stored, converted to .txt files, imported into Microsoft Excel,
and used for data analysis and interpretation.
Data Analysis
Genechip data were analyzed with the Affymetrix software,
Microsoft Excel, and Access, Cluster, and S-plus (Seattle, WA)
software. The raw measurements of expression of each gene
were taken (AvgDiff), and a cubic root transformation was ap-
plied to the raw data. This permits inclusion of negative values
and reduces larger variation at higher expression levels. The
transformed data were normalized by a weighted linear least
squares fit. This procedure calibrates individual hybridizations
with a reference set, which was generated by averaging the ex-
pression of each gene over all hybridizations. This weighted
linear fit method takes all genes into consideration but is
weighted toward 47 housekeeping genes. After data normaliza-
tion, fold changes were computed. Genes with negative inten-
sity were reset to a minimum value of 20 to eliminate infinite-
fold changes. Only genes without absent calls (AbsCall from
Affymetrix software) from all replicates, and with greater than
twofold changes in all possible pairs of comparison of CLN9-
deficient versus normal samples, were included. Expression
levels were visualized by Cluster software with an established
color code (red for upregulation, green for downregulation).
Neuronal Ceroid Lipofuscinosis Genes
The coding sequences for CLN3, CLN5, CLN6, and CLN8
genes were amplified as previously described.25 Products were
sequenced using an automated sequencer (377XL Prism
DNA Sequencer; PE Biosystems, Foster City, CA).
Northern Blot
A Northern blot was hybridized according to the manufac-
turer’s protocol (Clontech, Palo Alto, CA). Probes were gen-
erated by polymerase chain reaction (PCR) amplification of
the CLN3 gene. PCR products were purified by QIAQuick
PCR purification kit (Qiagen) and were [32P] labeled with
the RadPrime DNA labeling system (Gibco-Invitrogen). Hy-
bridization was performed with ExpressHyb Hybridization
Solution (Clontech) at 65°C.
Immunocytochemical Analysis of Subunit c of
Mitochondrial ATP Synthase
Paraffin-embedded slides of frontal cortex tissue from one of
the CLN9-deficient and from one age-matched control pa-
tient were used for staining. Immunocytochemical staining
was performed as previously described.26 The primary anti-
body used was raised by immunizing rabbits with a synthetic
peptide corresponding to N-terminal amino acid residues
32-45 of subunit c of mitochondrial ATP synthase. The
slides were developed with 3,3⬘-diaminobenzidine (Sigma,
St. Louis, MO) at 1mg/ml in phosphate-buffered saline
(PBS) containing 0.003% H2O2 and then counterstained
with hematoxylin (Sigma). The stained tissue sections were
examined under ⫻100 and ⫻150 magnification.
Live and Dead Attached and Detached Fibroblasts
Equal numbers of normal and CLN9-deficient fibroblasts (1 ⫻
107 cells) were grown in 100 ⫻ 20mm dishes for 24 hours,
after which either adherent live and dead cells or those in the
supernatant were harvested and counted by the trypan blue dye
exclusion method. The experiment was performed in triplicate.
Counts of live and dead control fibroblasts were compared with
those of live and dead patient fibroblasts using Student’s t test.
Growth Rate
Cells (7 ⫻ 104/well) were plated in 24-well plates and then
harvested and counted at 24, 48, 72, 96, and 124 hours in
triplicate for each time point using the trypan blue dye ex-
clusion method.
Trypan Blue Dye Exclusion
Harvested cells were centrifuged at 1,200 rpm for 5 minutes
and pellets resuspended in 1 to 1 mixture of 4% trypan blue
dye (Gibco-Invitrogen) and 1 ⫻ PBS, loaded onto a hemo-
cytometer, and counted. Viable cells are white, and dead cells
are blue.
[3H]Thymidine Incorporation
Cells were plated in six-well plates at a density of 1 ⫻ 105
cells per well. After 0, 6, 12, 24, and 48 hours, cells from
Schulz et al: Adhesion and Apoptosis in CLN9 343
triplicate wells were incubated with 2␮Ci/ml [3H]thymidine sample. Reactions were incubated for 15 minutes at 37°C
(Perkin Elmer, Boston, MA) in medium for 2 hours, washed and then terminated with 1.5ml chloroform-methanol (1:2).
twice with ice-cold PBS, and DNA-precipitated with 5% tri- Organic soluble counts were extracted and quantified by liq-
chloroacetic acid. The DNA precipitate was dissolved in uid scintillation counting.27,29
0.2ml of 0.25M NaOH. Incorporated [3H]thymidine was
determined by liquid scintillation counting. Results
Identification of the CLN9-Deficient Genotype
Etoposide Treatment Affymetrix GeneChip analysis of known NCL and un-
Cells were treated with 10␮g/ml etoposide (Sigma) for 18 known cases led to the discovery of this novel CLN9-
hours. Apoptosis was determined by propidium iodide (PI)
deficient variant. RNA from the four cases had identi-
staining and expressed as a ratio of red apoptotic cells to
total cells.
cal gene expression profiles, which were distinctly
different from normal and from CLN1-, CLN2-,
Propidium Iodide Staining CLN3- and CLN6-deficient RNA. This pattern was re-
Equal numbers of cells treated/not treated with etoposide were produced six times. Figure 1 shows a partial dendro-
grown on cover slips and stained with PI (5␮g/ml) for 5 min- gram of gene expression analyses performed with the
utes. Three fields of vision were chosen randomly at ⫻100 HuFL 6900 GeneChip. Upregulated genes are labeled
magnification. The total number of cells/field of vision red; downregulated genes are green. The color black
(20 – 60 cells/field) was counted. The number of PI-positive denotes no significant change in gene expression com-
red apoptotic cells was determined under fluorescence (excita- pared with control. Expression of genes involved in cell
tion wavelength, 525nm; emission wavelength, 600nm). The adhesion and apoptosis was dysregulated in CLN9-
percentage of PI-positive cells/total cells/field of vision was cal- deficient cells. These are shown in the table posted at
culated, as well as average and standard deviation. Statistical www.dbsr.duke.edu/pub/cln9/ and referred to below in
significance was determined using the Student’s t test. the discussion.
Filipin Staining
Clinical Course
Niemann-Pick type C and normal and CLN9-deficient fibro-
The clinical course of the CLN9-deficient patients is
blasts were maintained in McCoy’s medium (Gibco-Invitrogen)
supplemented with 10% LPDS (Intracel Corporation, Freder- quite similar to that in JNCL patients. The two broth-
ick, MD) for 5 days, harvested, seeded on cover slips, and in- ers presented at age 4 years with declining vision and
cubated at 37°C for 24 hours in Dulbecco’s modified Eagle me-
dium (DMEM) (Gibco-Invitrogen) containing 10% FBS and
1% antibiotics/antimycotics. Cover slips were rinsed three times
with Dulbecco’s PBS and fixed with 10% formalin (pH 7.4;
Sigma) at room temperature for 1 hour. Monolayers were rinsed
three times with PBS and stained with 800 ␮l filipin solution
for 60 minutes. The filipin solution was prepared by dissolving
2.5mg of filipin complex (Sigma) in 1ml of dimethylformamide
(Sigma). This was added to 50ml of PBS. Cells were rinsed
three times with 2ml PBS, and the cover slips were mounted on
microscope slides with fluoromount (Southern Biotechnology,
Birmingham, AL). Slides were examined under fluorescence (ex-
citation wavelength, 372nm; emission wavelength, 446nm) at
⫻100 and ⫻400 magnification.

Sphingolipid Levels
Ceramide and sphingomyelin levels in CLN9-deficient and
normal fibroblasts were measured as previously described and
also were quantified by mass spectrometry in the Lipidomics
Core at the Medical University of South Carolina.27 Glyco-
sphingolipid levels were measured after [14C]-galactose label-
ing according to published methods.28 Results were normal-
ized to lipid phosphate.
Serine Palmitoyl Transferase Activity
Two hundred micrograms of protein was dissolved in 0.1M
Hepes, pH 8.0, 5mM DTT, 5mM EDTA, pH 7.4, and
50␮M pyridoxal-5-phosphate and reaction-initiated by add-
ing 0.2mM Palmitoyl-CoA, 1mM Serine (Sigma), and
16.75␮l [14C]-serine (179.2mCi/mmol; Sigma) per 100␮l of
Fig 1. The CLN9-deficient genotype. Partial dendrogram de-
picting gene expression patterns of CLN1-, CLN2-, CLN3-,
CLN6-, and CLN9-deficient and normal fibroblast RNA.
Upregulated genes are red, and downregulated genes are green.
No change from control is black. Note similarity of gene ex-
pression patterns in both CLN9-deficient cell lines [CLN9(1)
Serbian patient, CLN9(2) German patient] in two experi-
ments. Lanes 1, 2, and 9 and lanes 3, 4, and 10, respec-
tively, represent experiments performed on two separate days.

344 Annals of Neurology Vol 56 No 3 September 2004

Fig 2. (A–C) Electron micrographs of brain sections from CLN9-deficient patients. (A) Electron micrograph of brain tissue from
one of the German CLN9-deficient brothers demonstrating the presence of secondary lysosomes containing curvilinear bodies. Magni-
fication ⫻20,000. (B, C) Electron micrograph of right frontal lobe brain biopsy of one of the Serbian CLN9-deficient patients il-
lustrating granular osmiophilic deposits (GRODS) (arrows) and curvilinear bodies (arrow in frame). (B) Magnification ⫻54,000.
(C) (inset) Magnification ⫻100,000. (D, E) Immunohistochemical analysis for subunit c of mitochondrial ATP synthase of normal
(D) and CLN9-deficient (E) frontal cortex tissue sections. Note gray staining of neuronal cytoplasm in CLN9-deficient brain tissue
indicating increased amounts of subunit c, as opposed to the normal blue stain of control brain tissue (arrows). Magnification
⫻150; scale bar ⫽ 50␮m

Schulz et al: Adhesion and Apoptosis in CLN9 345


Fig 3. Morphology. (A) CLN9-deficient fibroblasts have a
rounded cell body and are small. (B) Normal fibroblasts are
elongated. Magnification ⫻400, scale bar ⫽ 50␮m.
seizures. Cognitive decline was apparent at age 6 years,
with ataxia and rigidity at age 9 years. They developed
dysarthria and scanning speech and were mute by age
12 years. The younger brother died at age 15 years
following a bout of pneumonia. The older brother de-
veloped hallucinations, intractable seizures, and diffi-
culty swallowing and died at age 19 years. We are un-
aware of any history of consanguinity in the German
brothers. Although there is no history of close consan-
guinity in the Serbian sisters, the great-grandmothers
came from adjacent villages. The clinical course of the
Serbian sisters is very similar to that of the German
brothers. They developed declining vision, progressive
ataxia, and seizures with onset at age 4 years. By 9
years, they could not ambulate independently. They
became mute at the age of 10 years and suffered from
frequent generalized and myoclonic seizures. Electroen-
cephalograms in all cases showed slowing, with fre-
346 Annals of Neurology Vol 56 No 3 September 2004
quent polyspike wave discharges. The older sister was
completely bedridden by age 14 years, requiring a feed-
ing tube. She is now 19 years old. The younger sister,
now 10 years old, is following an identical course.
Diagnostic Workup
Funduscopy in both brothers showed thinned vessels
and optic nerve atrophy. The older brother did not have
significant pigmentary changes at age six. The younger
brother had significant pigmentary changes in the retina
at the same age. Electroretinograms showed diminished
wave amplitudes. Electron micrographs of lymphocytes
in both brothers showed numerous membrane-bound
lysosomal vacuoles, most empty with some containing
electron-dense storage material with a fingerprint pattern
typical for JNCL (H. H. Goebel, University of Mainz,
and Dr Schwendemann, University of Hamburg). At
autopsy of the older brother, brain weight was 1,140gm,
and neurons were ballooned with fine granular material.
Dilatation of large neurons was seen in the cerebral cor-
tex, basal ganglia, thalamus, and cerebellar cortex. The
process was less marked in the red nucleus, locus cer-
uleus, and the lower olive. The substantia nigra was
atrophic with moderate astrogliosis and slight vascular
proliferation. Atrophic changes also were seen in the nu-
clei of the thalamus with moderate to high-grade astro-
gliosis. Lipopigment material was seen in neurons in the
pyramidal band of Ammon’s horn. Cerebellar Purkinje
cells were dilated by storage material. Moderate sub-
ependymal astrogliosis was seen in brain and spinal cord.
The storage material stained gray with Sudan black and
had a yellow autofluorescence (H. J. Colmant, Univer-
sity of Hamburg).
Diagnostic workup of the sisters showed progressive
cerebral and cerebellar atrophy, predominantly involv-
ing gray matter, by cranial computed tomography and
magnetic resonance imaging. Abnormal signal intensity
was seen in the periventricular white matter. These
findings were consistent with a diagnosis of NCL. A
right frontal brain biopsy from the older sister was sub-
jected to electron micrograph examination. The neu-
rons contained inclusions characteristic for NCL.
There were a combination of membrane-bound gran-
ular and curvilinear bodies (Fig 2A–C). Neurons
Table. Live and Dead Attached and Detached Fibroblasts
(patients and controls)
Live Cells Dead Cells
Type (%) (%)
Detached
Normal fibroblasts 0⫾0 100 ⫾ 0
CLN9-deficient fibroblasts 36.2 ⫾ 0.4 63.8 ⫾ 0.4
Attached
Normal fibroblasts 95.8 ⫾ 0.5 4.2 ⫾ 0.3
CLN9-deficient fibroblasts 85.9 ⫾ 1.1 14.1 ⫾ 1.0

Fig 4. (A) Growth curve of CLN9-deficient fibroblasts. Nor-
mal fibroblasts (diamond) and CLN9-deficient fibroblasts
(square: Serbian patient, triangle: German patient). Live cells
were counted at 24, 48, 72, 96, and 124 hours from tripli-
cate experiments. A significant increase in growth rate in
CLN9-deficient cells is seen after 48 hours, with continued
growth after 72 hours. Control fibroblasts became normally
confluent because of contact inhibition. (B) [3H]Thymidine
incorporation. Normal fibroblasts (diamond) and CLN9-
deficient fibroblasts (square: Serbian patient, triangle: Ger-
man patient). Proliferation was determined by [3H]thymidine
incorporation at 0, 6, 12, 24, and 48 hours in triplicate
measurements. An increase in [3H]thymidine incorporation in
CLN9-deficient cells is seen at all time points and beyond 48
hours. [3H]Thymidine incorporation in normal fibroblasts
decreases after 24 hours.
stained positively with an antibody to subunit c of mi-
tochondrial ATP synthase (see Fig 2D, E). Apoptotic
neurons with nuclei containing aggregates of chroma-
tin were present (not shown). The presence of apopto-
tic neurons was verified by terminal deoxynucleotidyl-
transferase–mediated dUTP nick end labeling or
TUNEL staining. Enzyme screening for the CLN1-
and CLN2-deficient variants was negative. The se-
quences of coding regions of CLN3, CLN5, CLN6,
and CLN8 genes were normal. A normal Northern blot
rules out the possibility of defects in the intronic or
promoter region of the CLN3 gene. The blot had a
CLN3 mRNA band of normal size and intensity com-
pared with the CLN3 band from control RNA.
Metabolic labeling and two-dimensional gel electro-
phoresis of mannose-6-phosphate glycoproteins isolated
from CLN9-deficient fibroblasts from one sister had a
pattern identical to the pattern seen in normal fibro-
blasts. Testing was performed in the laboratory of P.
Lobel. The following lysosomal storage diseases were
ruled out by normal enzyme levels: Fabry disease (␣-
galactosidase), GM1-gangliosidosis (␤-galactosidase),
Tay–Sachs disease (␤-hexosaminidase A), Niemann–Pick
disease types A and B (sphingomyelinase), Gaucher’s dis-
ease (glucosylceramidase), mannosidosis, fucosidosis,
mucopolysaccharidoses type I (␣-iduronidase), type II
(iduronate sulfatase), and type VII (␤-glucuronidase).
Niemann–Pick C disease was ruled out by negative fili-
pin staining (see Fig 5). Assays for cathepsins A, B, C,
D, H, and L were normal (laboratory of P. Lobel).
CLN9-Deficient Fibroblasts Have a Distinct
Morphology and Biology
Cells from the CLN9 patients have identical features. (1)
CLN9-deficient fibroblasts have small and rounded cell
bodies in contrast with the elongated, normal fibroblasts
(Fig 3). Filipin staining, though not indicative of excess
cholesterol storage, showed CLN9-deficient cells to have
prominent nucleoli (see Fig 5A–C). (2) CLN9-deficient
fibroblasts attached poorly and piled up in mounds.
This was seen by direct examination of the cultures with
an inverted microscope. Numbers of detached live
CLN9-deficient cells were significantly higher than de-
tached live normal fibroblasts (36.2% vs 0%, p ⬍ 0.05;
Table). Numbers of attached dead CLN9-deficient fibro-
blasts were higher than attached dead cells from controls
(14.1% vs 4.2%, p ⬍ 0.05; see Table). (3) CLN9-
deficient fibroblasts have a rapid growth rate compared
with normal, and proliferation rates of CLN9-deficient
fibroblasts compared with normal cells were increased
suggesting increased DNA synthesis (Fig 4A, B). (4)
CLN9-deficient fibroblasts have higher apoptotic rates
compared with normal cells. This was shown by PI
staining after treatment with proapoptotic etoposide. An
increased number of apoptotic red-stained nuclei was
seen in CLN9-deficient cells (Fig 5D, E: 78.9% in
CLN9-deficient fibroblasts vs 25.3% in control cells,
p ⬍ 0.005).
Perturbation of Sphingolipid Metabolism
Mass measurements of ceramide and sphingomyelin
and the glycosphingolipids lactosylceramide, ceramide
trihexoside, and globoside after metabolic labeling with
[14C]-galactose were significantly decreased in CLN9-
Schulz et al: Adhesion and Apoptosis in CLN9 347
Fig 5. (A–C) Filipin staining of (A) Niemann–Pick C fibroblasts, (B) normal fibroblasts, and (C) CLN9-deficient fibroblasts.
Slides were examined under fluorescence (excitation wavelength, 372nm; emission wavelength, 446nm) at ⫻400 magnification,
scale bar ⫽ 50␮m. (A) Niemann–Pick C fibroblasts exhibit accumulation of cholesterol within the cell body. (B) Normal fibro-
blasts. (C) CLN9-deficient fibroblasts have prominent nucleoli (white arrow), but no excess cholesterol. (D, E) Increased apoptosis
of CLN9-deficient fibroblasts. Propidium iodide staining of normal fibroblasts (D) and CLN9-deficient fibroblasts (E) after treat-
ment with etoposide. Note increase in the number of red, apoptotic CLN9-deficient fibroblasts (25.27% represents number of apo-
ptotic cells/total cells/vision field in normal fibroblasts vs 78.97% in CLN9-deficient cells, P ⬍ 0.005). Magnification ⫻100; scale
bar ⫽ 100␮m.

348 Annals of Neurology Vol 56 No 3 September 2004


Fig 6. (A) Perturbation of sphingolipid metabolism. CLN9-
deficient cells have at least 70% less ceramide (Cer), 50% less
sphingomyelin (SM), 60% less lactosylceramide, 90% less cer-
amide trihexoside (CTH), and 90% less globoside (GB4) than
normal cells. (B) SPT activity is increased above 100% or
more above the normal control.
deficient cells (Fig 6A). Activity of the key enzyme of
the ceramide de novo synthesis pathway, serine palmi-
toyl transferase (SPT), is increased by more than 100%
(see Fig 6B). These facts suggest, but do not prove,
that the CLN9 gene may have an impact on sphingo-
lipid metabolism.
Discussion
Clinical, pathological, and neuroradiological features
suggest that the four patients described represent a new
NCL variant. Genechip analysis of these unclassified
NCL cases shows a striking similarity in their gene ex-
pression patterns but sets them apart from other NCL
types. Enzyme screening and genetic testing ruled out
known NCL gene defects, and assays for known lyso-
somal enzymes, including cathepsins A, B, C, D, H,
and L, were normal.
CLN9-deficient cells also have a distinctive pheno-
type and biology. They demonstrate sensitivity to apo-
ptosis and CLN9-deficient fibroblasts are small and
rounded with prominent nucleoli (see Fig 5C). Apo-
ptotic neurons were seen in a brain biopsy from one of
the sisters. PI staining of CLN9-deficient cells con-
firmed sensitivity to apoptosis (see Fig 5D, E). This is
in agreement with previous reports of increased apo-
ptosis as a common theme in CLN1-, CLN2-, and
CLN3-deficient cells.30,31 Flupirtine, an antiapoptotic
drug, blocks apoptosis in CLN1-, CLN2-, CLN3-, and
CLN6-deficient neurons.32 Persaud-Sawin and col-
leagues have identified motifs in the CLN3 protein
that are necessary for maintaining the antiapoptotic
properties of CLN3.33
Nucleolar prominence is typically seen in tumor cells
that have both increased apoptosis and high growth
rates.34,35 CLN9-deficient cells have significantly
higher growth and proliferation rates because of in-
creased DNA synthesis (see Fig 4A, B). Gene expres-
sion of cyclins A2, B1, C, E2, G1, and T2 was also
significantly increased. Cyclin D1, a protooncogene in-
volved in malignant transformation of breast tissue,
was significantly downregulated, as was member 1A of
the tumor necrosis factor receptor superfamily probably
as a secondary response.36 Gene expressions of multiple
subunits of cytochrome c oxidase and of glutathione
S–transferase A4 also were increased.
CLN9-deficient cells may have a defect in cell adhe-
sion, consistent with upregulation of neuronal
N-cadherin and significantly decreased gene expression
of cadherin 11 and the E member of the ras homolog
gene family. This latter gene, also called Rnd1, inhibits
the formation of integrin-based focal adhesions and in-
duces loss of cell-substrate adhesion leading to cell
rounding. Rnd1 is predominantly localized to adherens
junctions.37
Ceramide and sphingomyelin levels are substantially
decreased in CLN9-deficient cells. Activity of SPT, a
key enzyme in de novo ceramide synthesis, is increased.
This may be interpreted as a compensatory response to
low ceramide levels. Glycosphingolipid levels were also
low. Glycosphingolipids play key roles in cell adhesion
and apoptosis. Ceramide trihexoside is necessary for
B-cell adhesion and CD19 interactions and can trans-
duce apoptotic signals in Burkitt’s lymphoma cells.38,39
In conclusion, we have identified a novel NCL vari-
ant with a distinctive gene expression pattern. CLN9-
deficient cells exhibit increased apoptosis, an increased
growth rate, a cell adhesion defect and perturbation of
sphingolipid metabolism. Clues to identification of the
novel gene may surface after study and characterization
of altered sphingolipid metabolism.
Schulz et al: Adhesion and Apoptosis in CLN9 349

This work was supported by the Serbian Orthodox Church (R.-M.B.)
and by the Deutsche Forschungsgemeinschaft (SCHU1597/1-1, A.S.).
We thank the affected families for participating in this study. We thank
Drs P. Lobel and D. Sleat for performing metabolic labeling and two-
dimensional gel electrophoresis of purified mannose-6-phosphate glyco-
proteins and enzyme assays for cathepsins and other lysosomal enzymes.
We thank A. Steed for performing the Northern blot for CLN3 on
CLN9-deficient fibroblasts. We are grateful for measurements of sphin-
gomyelin and ceramide levels performed by the Lipidomics Core of the
Medical University of South Carolina, under the direction of Dr A.
Bielawska. We are greatly indebted to Dr H. H. Goebel for kindly
reevaluating lymphocyte ultrastructure of the German CLN9-deficient
patients. We also thank D. Merz for help with the figures.
References
1. Haltia M. The neuronal ceroid-lipofuscinoses. J Neuropathol
Exp Neurol 2003;62:1–13.
2. Goebel HH, Mole SE, Lake BD (Eds.). The neuronal ceroid
lipofuscinoses (Batten disease). Amsterdam: IOS Press, 1999.
3. Haltia M, Rapola J, Santavuori P, Keranen A. Infantile type of
so-called neuronal ceroid-lipofuscinosis-part 2. Morphological
and biochemical studies. J Neurol Sci 1973;18:269 –285.
4. Hassin G. Amaurotic family idiocy: late infantile type
(Bielschowsky) with the clinical picture of decerebrate rigidity.
Arch Neurol Psychiat 1926;16:708 –727.
5. Harlem OK. Juvenile cerebroretinal degeneration (Spielmeyer-
Vogt). Am J Dis Child 1960;100:918 –923.
6. Goebel HH, Braak H. Adult neuronal ceroid-lipofuscinosis.
Clin Neuropathol 1989;8:109 –119.
7. Santavuori P, Rapola J, Sainio K, Raitta C. A variant of Jansky-
Bielschowsky disease. Neuropediatrics 1982;13:135–141.
8. Sharp JD, Wheeler RB, Lake BD, et al. Genetic and physical
mapping of the CLN6 gene on chromosome 15q21–23. Mol
Genet Metab 1999;66:329 –331.
9. Hirvasniemi A, Lang H, Lehesjoki AE, Leisti J. Northern epi-
lepsy syndrome: an inherited childhood onset epilepsy with as-
sociated mental deterioration. J Med Genet 1994;31:177–182.
10. Zhong N, Moroziewicz DN, Ju W, et al. Heterogeneity of late-
infantile neuronal ceroid lipofuscinosis. Genet Med 2000;2:
312–318.
11. Mitchell WA, Wheeler RB, Sharp JD, et al. Turkish variant late
infantile neuronal ceroid lipofuscinosis (CLN7) may be allelic
to CLN8. Eur J Paed Neurol 2001;5:21–27.
12. Lake BD, Bret EM, Boyd SC. A form of juvenile Batten disease
with granular osmiophilic deposits. Neuropediatrics 1996;27:
265–269.
13. Mazzei R, Conforti FL, Magariello A, et al. A novel mutation
in the CLN1 gene in a patient with juvenile neuronal ceroid
lipofuscinosis. J Neurol 2002;249:1398 –1400.
14. Vesa J, Hellsten E, Verkruyse LA, et al. Mutations in the
palmitoyl protein thioesterase gene causing infantile neuronal
ceroid lipofuscinosis. Nature 1995;376:584 –587.
15. Sleat DE, Donnelly RJ, Lackland H, et al. Association of mu-
tations in a lysosomal protein with classical late-infantile neu-
ronal ceroid lipofuscinosis. Science 1997;277:1802–1805.
16. International Batten Disease Consortium. Isolation of a novel
gene underlying Batten disease, CLN3. Cell 1995;82:949 –957.
17. Ranta S, Zhang Y, Ross B, et al. The neuronal ceroid lipofus-
cinoses in human EPMR and mnd mutant mice are associated
with mutations in CLN8. Nat Genet 1999;23:233–236.
18. Gao H, Boustany RM, Espinola JA, et al. Mutations in a novel
CLN6-encoded transmembrane protein cause variant neuronal
ceroid lipofuscinosis in man and mouse. Am J Hum Genet
2002;70:324 –335.
350 Annals of Neurology Vol 56 No 3 September 2004
19. Wheeler RB, Sharp JD, Schultz RA, et al. The gene mutated in
variant late-infantile neuronal ceroid lipofuscinosis (CLN6) and
in nclf mutant mice encodes a novel predicted transmembrane
protein. Am J Hum Genet 2002;70:537–542.
20. Savukoski M, Kestila M, Williams R, et al. Defined chromo-
somal assignment of CLN5 demonstrates that at least four ge-
netic loci are involved in the pathogenesis of human ceroid li-
pofuscinosis. Am J Hum Genet 1994;55:695–701.
21. Isosomppi J, Vesa J, Jalanko A, Peltonen L. Lysosomal localiza-
tion of the neuronal ceroid lipofuscinosis CLN5 protein. Hum
Mol Genet 2002;11:885– 891.
22. Winter E, Ponting CP. Tram, Lag1 and CLN8: members of a
novel family of lipid sensing domains? Trends Biochem Sci
2002;27:381–383.
23. Schorling S, Vallée B, Barz WP, et al. Lag1p and Lac1p are es-
sential for the Acyl-CoA-dependent ceramide synthase reaction in
Saccharomyces cerevisiae. Mol Biol Cell 2001;12:3417–3427.
24. Guillas I, Kirchman PA, Chuard R, et al. C26-CoA-dependent
ceramide synthesis of Saccharomyces cerevisiae is operated by
Lag1p and Lac1p. EMBO J 2001;11:2655–2665.
25. Teixeira CA, Espinola J, Liang H, et al. Novel mutations in the
CLN6 gene cause a variant late infantile neuronal ceroid lipo-
fuscinosis. Hum Mutat 2003;21:502–508.
26. Johnson DW, Speier S, Qian WH, et al. Role of subunit-9 of
mitochondrial ATP synthase in Batten disease. Am J Med
Genet 1995;57:350 –360.
27. Rylova SN, Amalfitano A, Persaud-Sawin DA, et al. The CLN3
gene is a novel molecular target for cancer drug discovery. Can-
cer Res 2002;62:801– 808.
28. Uemura K, Sugiyama E, Tamai C, et al. Effect of an inhibitor
of glucosylceramide synthesis on cultured rabbit skin fibro-
blasts. J Biochem 1990;4:525–530.
29. Williams RD, Wang E, Merrill AH Jr. Enzymology of long-
chain base synthesis by liver: characterization of serine palmi-
toyltransferase in rat liver microsomes. Arch Biochem Biophys
1984;228:282–291.
30. Lane S, Jolly R, Schmechel D, et al. Apoptosis as the mecha-
nism of neurodegeneration in Batten’s disease. J Neurochem
1996;67:677– 683.
31. Cho S, Dawson PE, Dawson G. Role of palmitoyl-protein thio-
esterase in cell death: implications for infantile neuronal ceroid
lipofuscinosis. Eur J Paed Neurol 2001;5:53–55.
32. Dhar S, Bitting RL, Rylova S, et al. Flupirtine blocks apoptosis
in Batten patient lymphoblasts and in human postmitotic CLN3-
and CLN2-deficient neurons. Ann Neurol 2002;51:448 – 466.
33. Persaud-Sawin DA, Van Dongen A, Boustany RM. Motifs
within the CLN3 protein: modulation of cell growth rates and
apoptosis. Hum Mol Genet 2002;11:2129 –2142.
34. Smetana K. Are nucleoli participating in programmed cell
death? J Appl Biomed 2003;1:93–97.
35. Horky M, Kotala V, Anton M, Wesierska-Gadek J. Nucleolus
and apoptosis. Ann NY Acad Sci 2002;973:258 –264.
36. Yu Q, Geng Y, Sicinski P. Specific protection against breast
cancer by cyclin D1 ablation. Nature 2001;411:1017–1021.
37. Nobes CD, Lauritzen I, Mattei MG, et al. A new member of the
rho family, Rnd1, promotes disassembly of actin filament struc-
tures and loss of cell adhesion. J Cell Biol 1998;141:187–197.
38. Jackson T, Van Exel C, Reagans K, et al. Comparison of ad-
hesion mechanisms and surface protein expression in CD77-
positive and CD77-negative Burkitt’s lymphoma cells. Cell Mol
Biol 2001;47:1195–1200.
39. Mangeney M, Lingwood CA, Taga S, et al. Apoptosis induced
in Burkitt’s lymphoma cells via GB3/CD77, a glycolipid anti-
gen. Cancer Res 1993;53:5314 –5319.