Clinical/Scientific
Notes
A. Izzotti, MD, PhD
E. Fazzi, MD
S. Orcesi, MD
C. Cartiglia, PhD
M. Longobardi, PhD
V. Capra, PhD
P. Lebon, MD
A. Cama, MD
A. Pulliero, PhD
R. La Piana, MD
G. Lanzi, MD
BRAIN DAMAGE AS DETECTED BY CDNAMICROARRAY IN THE SPINAL FLUID OF PATIENTS
WITH AICARDI-GOUTIÈRES SYNDROME
Aicardi-Goutières syndrome (AGS) is a rare encephalopathy, arising during the first year of life, characterized by cerebral atrophy, leukodystrophy, basal
ganglia calcification, raised interferon (IFN)-alpha in
the CSF, CSF lymphocytosis, and negative serologic
findings for infections.1,2
Identification of AGS pathogenesis may provide
insights into the neurodegenerative mechanism resulting from exposure of the developing human brain
to IFN-alpha.
The aim of this study is to provide an insight into
global gene expression of CSF cells in AGS using
microarray technology.
Methods. CSF aliquots have been collected from
patients with AGS and controls for diagnostic purposes. The study was approved by the Ethical Committee of the IRCSS Mondino, University of Pavia,
Italy.
Patients with AGS (n ⫽ 20), 12 boys and 8 girls
(age 4.5 ⫾ 4.4 years), have been identified by the
presence of an appropriate clinical picture in presence of CSF lymphocytosis (⬎5 cells/mm3), IFNalpha in CSF ⬎2 IU/mL, in absence of any
infections.3
Controls, 11 boys and 9 girls, matched to patients
with AGS by gender, age (age 4.4 ⫾ 4.3 years), and
number (n ⫽ 20), devoid of any alteration in the
CSF, included 13 subjects with congenital hydrocephalus, 3 with endocranial tumor, 2 with myelomeningocele with hydrocephalus, 1 with choroid
plexus papilloma, and 1 with upper sella turcica cyst.
Lymphocyte was the main cell type detected
(⬎80%) in the CSF of all recruited subjects, the cell
count being 40.8 ⫾ 8.66 cells/mm3 in patients with
AGS and 2.0 ⫾ 1.13 cells/mm3 in controls (mean ⫾
SE). IFN-alpha in CSF was ⬍2 IU/mL in controls
and 52.95 ⫾ 12.15 IU/mL (mean ⫾ SE) in patients
with AGS.
RNA, as purified from CSF, underwent reverse
transcription and amplification using real-time PCR
obtaining cDNA sequences complementary to
mRNA. An aliquot (20 ␮L) of each sample was used
to determine the pre-plateau amplification cycle, the
remaining sample (80 ␮L) being amplified at this
cycle.4 Fluorescent probes complementary to cDNAs
were synthesized using amino-allyl modified uridine
nucleotides, DNA-polymerase, and Cy3/Cy5 labeling. A standardized amount of labeled oligonucleotides (20 pmol) was used for microarray
hybridization. Used microarrays (Microarray Department of the University of Amsterdam, The
Netherlands), spotted in duplicate with the expressed
sequence tags of 19,000 genes, were hybridized and
underwent laser scanning and data analysis by Genespring software (Silicongenetics, Billerica, CA).
Five genes (CathepsinD, IFRG28, BAI2,
VEGFB, and GAPDH) were analyzed for their expression also by QPCR to substantiate microarray
results.
Results. Results of patients with AGS as compared
to controls are reported in the figure, A. Differences
in gene expression accounted for a different location
of patients with AGS and controls in a threedimensional analysis-plot of the principal components of variance (figure, B). By applying the
supervised k nearest neighbor algorithm, AGS cases
and controls were correctly classified obtaining 38
correct predictions of the disease status, 0 incorrect
prediction, while two AGS cases were not classified
in any category. By using the unsupervised hierarchical cluster analysis (figure, C) all patients with AGS
were located separately from controls in the hierarchical tree. However, the two patients with AGS unclassified by the k nearest algorithm were located in
the hierarchical tree near to controls, thus indicating
that the gene expression pattern of these two samples
was similar between AGS patients and controls.
These two patients were the oldest among those
tested, being 7.6 and 11.8 years old.
One additional subject raising a diagnostic problem was included in the analysis (2-month-old girl
bearing bilateral basal-ganglia calcification, normal
neurologic examination in absence of any infection).
This subject was classified as a control when her CSF
gene-expression profile was inserted as unknown disNeurology 71
August 19, 2008
1
Figure
Results of gene expression analysis as performed in the CSF of patients with Aicardi-Goutières
syndrome (AGS) and controls by cDNA microarray
(A) Scatterplot distribution of the expression of each gene in AGS cases (vertical axis) and controls (horizontal axis). (B)
Classification of patients with AGS (red dots) and controls (yellow dots) by tridimensional principal component analysis of
variance of gene expression data. (C) Classification of patients with AGS (right part of the hierarchical tree) and controls
(left part of the hierarchical tree) by unsupervised hierarchical cluster analysis. Each column represents one subject, each
row one gene. Gene expression levels are reported on color scale (blue low, red high). (D) Functional categories of the 198
genes whose expression varied between patients with AGS and controls more than twofold and also reaching the statistical significance threshold (p ⬍ 0.05).
2
eases status in the k-nearest neighbor algorithm and
after 2-years follow-up did not develop AGS.
As far as concerns gene alterations detected in all
patients with AGS tested, 198 genes varied (p ⬍ 0.05
as tested by analysis of variance) their expression
more than twofold in patients with AGS as compared to controls (figure, D). AGS predictor genes
included upregulated genes, involved in interferondependent lymphocyte activation, and downregulated genes, involved in angiogenesis and cell cycle
suppression.
genes supports the role of interferon in AGS. Downregulation of angiogenesis-related genes in AGS is
likely to be a consequence of IFN-alpha increase in
CSF, being established that IFN-alpha hampers
angiogenesis.5
Results of the current study are consistent with
the hypothesis that AGS is an interferon-induced microangiopathy, which occurs early in life. We propose that brain injury results from insufficient vessel
development, paralleled by lymphocyte activation in
CSF.
Discussion. Obtained results provide evidence that
it is feasible to apply gene expression analysis to CSF
cells distinguishing between patients with AGS and
unaffected controls. Diagnostic performance was
good in subjects younger than 7 years, while patients
with AGS older than 7 years cannot be identified.
Functional classification of genes undergoing altered expression in AGS provides information on disease pathogenesis. Upregulation of interferon-related
e-Pub ahead of print at www.neurology.org.
Neurology 71
August 19, 2008
From the Department of Health Sciences (A.I., C.C., M.L.) and
IRCCS G. Gaslini, Department of Paediatric Sciences (V.C., A.C.),
University of Genoa; Department of Child Neurology and Psychiatry
(E.F., S.O., R.L., G.L.), IRCCS C. Mondino Institute of Neurology, University of Pavia, Italy; Hopital Cochin-St. Vincent de Paul
(P.L.), Université René Descartes, Paris, France; and International
Aicardi-Goutières Syndrome Association (A.P.), Pavia, Italy.
Supported by the International Aicardi-Goutières Syndrome Association (IAGSA) and by the CARIPLO Foundation of Milan, Italy.
Disclosure: The authors report no conflicts of interest.
Received August 28, 2007. Accepted in final form March 7, 2008.
Address correspondence and reprint requests to Prof. Alberto Izzotti,
Department of Health Sciences, University of Genoa, Via Pastore 1,
16132, Genoa, Italy; [email protected]
2.
Copyright © 2008 by AAN Enterprises, Inc.
3.
ACKNOWLEDGMENT
The authors thank Prof. Yanick J. Crow, Leeds Institute of Molecular
Medicine, University of Leeds, UK, for his contribution to the interpretation and discussion of the results obtained.
4.
5.
1.
Aicardi J, Goutieres F. A progressive familial encephalopathy in infancy with calcifications of the basal ganglia and
chronic cerebrospinal fluid lymphocytosis. Ann Neurol
1984;15:49–54.
Lebon P, Badoual J, Ponsot G, Goutieres F, HemeuryCukier F, Aicardi J. Intrathecal synthesis of interferonalpha in infants with progressive familial encephalopathy. J
Neurol Sci 1988;84:201–208.
Lanzi G, Fazzi E, Arrigo S. Aicardi-Goutieres syndrome: a
description of 21 new cases and a comparison with the
literature. Eur J Paediatr Neurol 2002;6s:1–14.
Eberwine J. Single-cell molecular biology. Nat Neurosci
2001;4s:1155–1156.
Rosewicz S, Detjen K, Scholz A, von Marschall Z. Interferon-alpha: regulatory effects on cell cycle and angiogenesis.
Neuroendocrinology 2004;80 suppl 1:85–93.
Neurology 71
August 19, 2008
3