Note: Descriptions are shown in the official language in which they were submitted.
CA 022160~7 1997-09-19
POLYMORPHIC CAG REPEAT-CONTAINING GENE, DIAGNOSIS OF
PSYCHIATRIC DT~~.S AND THERAPEUTIC USES T~R~OF
R~CK~OUND OF THE INV~NTION
(a) Field of the Invention
The invention relates to hGTl gene, a polymor-
phic CAG repeat-containing gene and its uses thereof
for the diagnosis, prognosis and treatment of psychiat-
ric diseases, such as schizophrenia.
(b) DescriPtion of Prior Art
Schizophrenia is a chronic brain disorder char-
acterized by a behavioral syndrome combining in various
degrees hallucinations, delusions, social withdrawal,
affective flattening, disorganized behavior and formal
thought disorders. It affects up to 1% of the general
population and results in a lower level of social and
occupational functioning. Many recent studies indicate
that schizophrenia may originate from neural cell dis-
turbances occurring in the developing/maturing brain.
Genetic factors are known to play a major role in the
etiology of this disorder as demonstrated by extensive
family, twin and adoption studies. However, the quest
for genes conferring susceptibility to schizophrenia
has been difficult and has not yielded consistent find-
ings using both association and linkage studies. It isthought that these difficulties are in part due to het-
erogeneity in etiology, both of genetic and non-genetic
origins, resulting in a highly variable phenotype with
respect to age at onset, symptom profile, course of
illness, response to medication, long term outcome and
performance on neuropsychological tests.
One promising avenue to guide research in this
search for genes increasing susceptibility to schizo-
phrenia may be to distinguish patients on the basis of
therapeutic response to neuroleptics. Indeed, while
most schizophrenic patients are improved by neuroleptic
CA 022160~7 1997-09-19
medication, a substantial number of subjects (15 to
25%) remains severely symptomatic despite multiple and
adequate neuroleptic therapeutic attempts. In contrast
to this between subject variability, within subject
(from one episode to the other) consistency of neu-
roleptic response have been reported. Clinical pre-
treatment characteristics that correlate with good neu-
roleptic response include spontaneous high blink-rate
and blink-rate decrease under Haloperidol challenge,
absence of spontaneous movement disorders, and absence
of dysphoric reaction within 24-48 hours of neuroleptic
initiation. On a long term basis, it has been demon-
strated that good response to neuroleptics (but not the
severity of the symptoms prior to neuroleptic medica-
tion) in the early stages of the disease, predicts abetter outcome. Neurophysiological characteristics
that correlate with good neuroleptic response include
high-frequency waves and few alpha and slow waves in
computerized EEG prior to the treatment with neurolep-
tics, a specific profile of changes in quantified EEGspectrum under neuroleptics and high degree of elec-
trodermal activity prior to neuroleptic treatment. An
important number of studies indicate that dopamine neu-
rotransmission is disturbed predominantly in the
responsive schizophrenic patients. High pre-treatment
plasma levels of HVA have been shown to predict good
response to neuroleptics in most of the studies. Pre-
liminary genetic epidemiological data indicate that
poor or delayed response to neuroleptic treatment is
associated with an increased prevalence of schizophre-
nia spectrum disorders in relatives of schizophrenic
probands. These convergent lines of evidence suggest
that long term response to neuroleptic medication may
be considered as a bioclinical dimension with an
etiologically relevant significance; the two extremes
-
CA 022160~7 1997-09-19
of this dimension being occupied by two groups of
schizophrenic patients, at least partially, distinct
with respect to the pathogeny of their illness.
It would be highly desirable to be provided
with a tool for the diagnosis, prognosis and treatment
of psychiatric diseases, such as schizophrenia.
SUMMARY OF THE INVENTION
One aim of the present invention is to provide
a tool for the diagnosis, prognosis and treatment of
psychiatric diseases, such as schizophrenia.
Another aim of the present invention is to
detect association between allelic variants of CAG
repeat-containing genes and schizophrenia or its pheno-
typic variability with respect to long term response toneuroleptic medication.
In accordance with the present invention, we
compared the allelic frequencies of various polymorphic
candidate genes between two groups of schizophrenic
patients carefully screened on the basis of their long
term response to typical neuroleptics (excellent
responders, Rs; non-responders, NRs) and controls.
This report summarizes our finding while considering
CAG containing genes as candidates for schizophrenia.
This family of candidate genes was deemed attractive
for the following reasons: (1) CAG repeat instability
was associated with several neurodegenerative brain
diseases that display genetic anticipation, a feature
believed to be present in schizophrenia, (Z) some iso-
lated, though promising, reports indicate that expandedCAG repeats are more prevalent in schizophrenic
patients compared to normal controls, (3) CAG repeats
are often very polymorphic and have been found to be
over represented in coding sequences of the human
genome particularly those coding for DNA-binding pro-
teins/transcription factors. These factors are impor-
CA 022160~7 1997-09-19
tant actors in the regulation of the genetic program
and neurodevelopmental processes and have been impli-
cated in several human neurodevelopmental diseases
including one that may present with schizophrenia-like
symptoms, and; (4) CAG repeats (or the polyglutamine
stretches for which they encode) might modulate the
function of the genes (or protein) they are part of
suggesting that they might be functional polymorphisms
and not silent ones.
In accordance with the present invention there
is provided a hGTl gene containing transcribed polymor-
phic CAG repeat, which comprises a sequence as set
forth in Fig. 3.
The allelic variants of CAG repeat of hGTl gene
may be associated with schizophrenia, affective dis-
eases such as manic depression, neurodevelopmental
brain diseases or with phenotypic variability with
respect to long term response to neuroleptic medica-
tion.
More precisely, there are 5 allelic variants of
CAG repeat which are identified as follows:
Size of PCR amplifiedPredicted No. of CAG repeats SIIOll~al to longest
r~ay~"~lS (bp)
171 11 -3
174 12 -2
177 13 -1
180 14 0
183 15
In accordance with the present invention there
is provided a method for the prognosis of severity of
schizophrenia of a patient, which comprises the steps
of:
a) obtaining a nucleic acid sample of the patient;
and
CA 022160~7 1997-09-19
b) determining allelic variants of CAG repeat of
the hGTl gene, and wherein long allelic vari-
ants are indicative of severe schizophrenia.
The preferred nucleic acid sample used in
accordance with the present invention is DNA. For RNA
sample, an additional step is carried out, which con-
sists in using a reverse transcriptase to transcribe
the RNA into DNA.
More precisely, the allelic variants identified
as short or as having between about 171 and 177 bp
(referred to as -3, -2 and -1) are associated with mild
schizophrenia and long or as having between about 180
and 183 bp (referred to as 0 and 1) are associated with
severe schizophrenia.
In accordance with the present invention there
is provided a method for the identification of patient
responding to neuroleptic medication, which comprises
the steps of:
a) obtaining a nucleic acid sample of the patient;
and
b) determining allelic variants of CAG repeat of
the hGTl gene, and wherein short allelic vari-
ants are indicative of neuroleptic response.
More precisely, the allelic variants identified
as short or as having between about 171 and 177 bp
(referred to as -3, -2 and -1) are associated with
patient capable of neuroleptic response and long or as
having between about 180 and 183 bp (referred to as 0
and 1) are associated with non-response to neuroleptic
medication.
In accordance with the present invention there
is provided a non-human mammal model for the hGTl gene,
whose germ cells and somatic cells are modified to
express at least one allelic variant of the hGTl gene
and wherein the allelic variant of the hGTl being
CA 022160~7 1997-09-19
introduced into the mammal, or an ancestor of the mam-
mal, at an embryonic stage.
In accordance with the present invention there
is provided a method for the identification of patient
responding to neuroleptic medication, which comprises
the steps of:
a) obtaining a nucleic acid sample of the patient;
and
b) determining allelic variants of CAG repeat of
the hGTl gene, and wherein short allelic vari-
ants (from about 171 to about 177 bp) are
indicative of neuroleptic response.
In accordance with the present invention there
is provided a method for the screening of therapeutic
agents for the prevention and/or treatment of schizo-
phrenia, which comprises the steps of:
a) administering said therapeutic agents to the
non-human mammal of the present invention or
schizophrenia patients; and
b) evaluating the prevention and/or treatment of
development of schizophrenia in said mammal or
said patients.
In accordance with the present invention there
is provided a method to identify genes part of or
interacting with a biochemical pathway affected by hGTl
gene, which comprises the steps of:
a) designing probes and/or primers using the hGTl
gene of the present invention and screening
psychiatric patients samples with said probes
and/or primers; and
b) evaluating the identified gene role in psychi-
atric patients.
In accordance with the present invention there
is provided a method of stratifying psychiatric
CA 022160~7 1997-09-19
patients based on the allelic variants of the hGTl gene
for clinical trials purposes, which comprises:
a) obtaining a nucleic acid sample of the
patients; and
b) determining allelic variants of CAG repeat of
the hGTl gene, wherein patients are stratified
with respect to their allelic variants and
wherein short allelic variants are indicative
of neuroleptic response.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 illustrates the average allelic lengths
of the GCTlOD04 EST CAG repeat in controls, responsive
(R) and non-responsive (NR) patients, showing the
shorter (S) allele only, longer (L) allele only and the
sum (L+S) of the two alleles in the three groups of
subjects;
Fig. 2 illustrates the correlation between the
average length of the (CAG)n polymer of the short (a),
the long (b) alleles and the sum of 2 alleles (C) and
seventy of schizophrenia in the different classes of
severity of the disease; and
Fig. 3 illustrates the sequence homology
between the human GCTlOD04 sequence and the mouse GTl
gene.
DET~TT~n DESCRIPTION OF THE INVENTION
In accordance with the present invention, the
main objective was to detect allelic variants of CAG
repeat containing genes associated with schizophrenia
or its phenotypic variability with respect to the pres-
ence or absence of schizophrenia and long term response
to neuroleptic medication.
Accordingly, CAG repeat allelic variants were
compared between three groups of subjects: two groups
CA 022160~7 1997-09-19
of schizophrenic patients, one neuroleptic-responsive
(Rs; n=43) and one -non-responsive (NRs; n=63), and; a
group of controls screened out for DMS-IV axis I psy-
chiatric disorders (Ci n=87). Assessment of response
to conventional neuroleptics was based on a comprehen-
sive review of medical files according to a priori
defined criteria and blind to genotyping. Genes con-
taining polymorphic CAG repeats were identified by
means of genetic sequences data base searches.
The results in accordance with the present
invention shows that short CAG repeat allelic variants
of the hGTl gene were associated with schizophrenia
irrespective of neuroleptic response (% short alleles
SCZ=45%; C=31%, p=0.005). This association was highly
significant in Rs (52%, p=0.0009) and marginal in NRs
(40%, p=0.12) groups. A statistically significant cor-
relation (Gamma=0.37, p=0.0024) between the CAG repeat
length and the overall pattern of severity of schizo-
phrenia was also observed.
Surprisingly and in accordance with the present
invention, CAG repeat allelic variants of the hGTl gene
show strong association with neuroleptic responsive
schizophrenia and length correlation with the overall
pattern of severity of the disease.
Clinical
The study was conducted between 1994 and summer
1997. Patients have been recruited in the Douglas hos-
pital (n=82), the Clinique Jeunes Adultes of L.H.
Lafontaine Hospital (n=15) and the Schizophrenia Clinic
of the Royal Ottawa Hospital (n=9). 333 schizophrenic
patients were identified as potential subjects for this
study. 123 patients did not meet the criteria for
schizophrenia or Rs/NRs (undifferentiated response)
diagnoses. 125 and 85 patients met respectively the
criteria for NRs-schizophrenia and Rs-schizophrenia. 62
CA 022160~7 1997-09-19
NRs and 42 Rs subjects were not included in the study
because refusal or other exclusion criteria.
NRs schizophrenic patients were recruited
according to the following criteria: (1) they all met
axis I diagnosis of schizophrenia, according to the
Diagnostic and Statistical Manual of Mental Disorders,
version IV (American Psychiatric association, Diagnos-
tic and Statistical Manual of Mental Disorders, APA;
1994) (DSM-IV), (2) they did not experience remission
of psychotic symptoms within the past 2 years, (3) in
the preceding 5 years, all patients underwent at least
3 periods of treatment with typical neuroleptics, from
at least two distinct families of drugs, at therapeutic
dosage (equal to or greater than 750 mg Chlorpromazine
equivalent in monotherapy or 1000 mg chlorpromazine
equivalent, when a combination of neuroleptics is
used), for a continuous period of at least 6 weeks at a
time, with no significant relief of symptoms, and; (4)
Unable to function without supervision in all or nearly
all domains of social and vocational activities with a
Global Assessment Score (GAS) < 40 within the last 12
months.
Criteria for the selection of neuroleptic Rs
patients were as follows: (1) all patients met the cri-
teria for schizophrenia according to DSM-IV, (2) all
were admitted at least once to a psychiatric care
facility because of acute psychotic episode, (3) during
all hospitalizations, patients experienced full or par-
tial remission in response to treatment with typical
neuroleptics, at recommended dosage, within six-eight
weeks of continuous treatment; remission being defined
as a rapid reduction of schizophrenic symptoms with
limited residual symptoms, (4) all patients were able
to function with only occasional supervision in all or
nearly all domains of social and vocational activities
CA 022160~7 1997-09-19
-- 10 --
with a GAS score > 60 within the last 12 months, (5) no
patients had to be admitted to hospitals because of
psychotic exacerbation, if and when compliant to treat-
ment and treated continuously with typical neurolep-
tics, and; (6) at least one psychotic relapse when neu-
roleptic medication is reduced or interrupted. Exclu-
sion criteria for schizophrenic patients were brain
trauma, any neurological condition, drug or alcohol
abuse in the last two years.
All schizophrenic patients were directly inter-
viewed by the PI, a research psychiatrist trained in
the use of the Diagnostic Interview for Genetic Stud-
ies (DIGS) (Nurnberger JI et al., A~chives of General
Psychiatry. 1994jSl:849-59) and their medical records
were comprehensively reviewed. Complementary informa-
tion from the treating physician and nurses in charge
of the patients and their close relatives was obtained,
whenever possible. A best estimate diagnosis was
established on the basis of all the available data.
Responsiveness to typical neuroleptic medication was
evaluated according to a 7 point's scale. The severity
of symptoms and overall psychosocial functioning were
assessed using the following instruments: (1) Brief
Psychiatric Rating Scale (BPRS) (Woerner MG et al.,
Psychopharmacology Bulletin. 1988j24:112-117), (2) the
Scale for the Assessment of Negative Symptoms (SANS),
the (3) Scale for the Assessment of Positive Symptoms,
(4) the GAS, (5) the Pattern of Severity Scale, a 5
point's scale assessing overall course and outcome of
the disease (American Psychiatric association, Diagnos-
tic and Statistical Manual of Mental Disorders, Fourth
Edition, American Psychiatric association, Washington
D.C.; 1994), and; (7) the Pattern of Symptoms subtypes,
a categorical classification of patients according to
the combination and changes over the course of the dis-
CA 022160~7 1997-09-19
ease of positive and negative symptoms (Nurnberger JI
et al., Archives of General Psychiatry. 1994;51:849-
59). All these evaluation tools, except the BPRS are
part of the DIGS.
The control group (C) was made by healthy vol-
unteers recruited through advertisement in local papers
(n=49) and married-in individuals from a linkage study
(n=38). All subjects in this group underwent a struc-
tured psychiatric interview in order to exclude those
who meet DSM-IV axis I disorders. Subjects recruited
through advertisement have also been screened for
schizophrenia spectrum disorders and have been tightly
matched for ethnic background (mother and father eth-
nicity) with schizophrenic patients. All, (except one
responsive~, patients and controls were Caucasians.
All of them gave informed and written consent. The
research protocol has been approved by the three hospi-
tals ethic committee where the research took place.
Genetic methods
To identify sequences potentially encoding
polymorphic polyglutamine tracts, we conducted a number
of Basic Local Alignment Search Tool (BLAST) (Altschul
SF et al., Journal of Molecular Biology. 1990;215:403-
410) searches using the following sequences: (1)
(CAG)30 or (CAA)30 (BLASTn, unfiltered against the non-
redundant nucleic acid and the expressed sequence Tag
(dbEST) databases) and, (2) Q30 (BLASTp, unfiltered,
against the non-redundant protein database or tBLASTn
against dbEST). Sequences containing homopolymer
tracts of >7 CAG or CAA repeats or potentially encoding
a tract of >12 glutamine residues were used to design
PCR primers able to amplify the CAG or CAA repeats.
PCR primers were designed using DNASTAR Inc. (Madison,
Wisconsin) software.
CA 022160~7 1997-09-19
- 12 -
Genomic DNA was isolated from peripheral lym-
phocytes using standard methods. CAG repeat-containing
fragments were amplified by PCR using specific primers
for each repeat. PCR was performed in a total volume
of 13 ~1 containing 30 ng of human genomic DNA, lOmM
Tris-HCl (pH 8.8), 1.5 mM MgC12, 50mM KCl, 1% Dimethyl-
sulfoxide, 250 mM each of dCTP, dGTP, and dTTP, 25 mM
dATP, 1.5 uCi alpha 35S-dATP, 100 ng of each primer,
and 3 units of Taq polymerase (Perkin-Elmer). DNA was
denatured at 94~C for 5 min., then subjected to 30
cycles of a 1 min. denaturation at 94~C, a 1 min.
annealing at the optimized annealing temperature for
each primer pair and a 1 min. elongation at 72~C. This
was followed by a final extension at 72~C for 5 min.
PCR products were electrophoresed on denaturing
6% Polyacrylamide gels and visualized by autoradiogra-
phy. Absolute allele sizes were estimated according to
an M13 sequence ladder. Since differences in absolute
allele sizes were in all cases multiples of 3 base
pairs, we assumed that variations in allele sizes were
due to differences in the number of trinucleotide
repeat units in the amplified sequences. By conven-
tion, we designated the most common allele as 0, with
less common alleles as positive or negative integers
according to their number of trinucleotide repeats
(e.g. if allele 0 had 20 repeats, allele +2 and -2
would have respectively 22 and 18 repeats).
Analysis:
Each subject was assigned two numeric values
which represent respectively the lengths of his short
(S) and long (L) alleles. Under the assumption of a
quantitative effect of the CAG tract length, data were
initially analyzed using a non-parametric analysis of
variance (Kruskal-Wallis median statistic) where the
independent variable is the diagnostic status (Rs, NRs
CA 022160~7 1997-09-19
and C) and the dependent variable is the length of the
CAG repeat of S, L or the sum of the two alleles. In
the case of a significant overall group effect in the
ANOVA, pair-wise contrasts between the different groups
were performed using the Mann and Whitney U-statistic.
This approach allows to control for the inflation of
type I error secondary to multiple testing.
We also analyzed data by contrasting allelic
frequencies in different pairs of groups using the %Z
statistic. Alleles were grouped in different classes
in accordance to the pattern of results found in the
analysis of variance. Since both patient and control
groups include an important number of subjects with a
French Canadian ethnic origin, we reanalyzed any asso-
ciation finding after stratifying subjects according tothe ethnic origin of their parents (both parents from
French Canadian origin vs. at least one parent with
non-French Canadian origin). This analysis allows to
control for associations resulting from ethnically
based differences in allelic frequencies (population
stratification) as opposed to those attributable to the
pathological condition under study (true association).
When a particular EST showed allelic or size
association with schizophrenia and/or responsiveness to
medication, further analyses were performed to investi-
gate the putative relation linking various clinical
dimensions (age at onset, pattern of severity) to the
length of the CAG repeat. For this purpose, we used the
Gamma correlation statistic, a non parametric statistic
recommended when there are many ties in the data set.
Clinical dimension that were used as criteria to define
the two groups of patients (GAS, severity of current
symptoms, neuroleptic responsiveness scores) were not
included in this analysis. Relations between categori-
cal variables (schizophrenia subtypes of illness, pat-
CA 022160~7 1997-09-19
- 14 -
tern of symptoms) and the CAG alleles were explored by
a % statistic with the appropriate degree of freedom.
Logistic regression was used to determine the attribut-
able risk conferred by any EST allelic variants which
showed a positive association with schizophrenia or
neuroleptic responsiveness. All analyses were made
using the Statistica software (Statsoft).
Table 1 Demo~raphic and clinical characteristics of patients and controls
Non-respongive(62) Responsive (43) Controls (C) (87)
Mean age in y~ars +SD (n~ 38J 7 (62) 40i lo (43) 44* 13 (87)
Education inyears+ SD (n) 11+ 2 0 (59) 11+ 2 ~ (43) 14+ 3 3 (49)
SES of HH+SD (n)- 54+ 24 (53) 51+ 24(41) 59+ 20(49)
Sex, % M 74% 67% 45%
Ethnic origin FCIOB 27/35 26117 40147
Subtype, UIPIDIC 271301411 61361110
Mean age at C, in years +SD (n) 181 3 9(55) 24:t 4.8 (43)
Illness duration in years +SD (n) 20+ 7 0 (55) 16i 8 8 (43)
~/O time as in-patient~(n) 62% (61)-- 8 2% (43)
BPRS total score i SD (n) 49+ 8 9 (53) ~- 241 3 9 (53)
NLP response score 1 83 + 0 74 (58) ~ 6 3 ~ 0 67 (43)
Patternofseverity 40+00(55)~- 19+05(43)
SES of HH indicates socioeconomic status head of house hold; FC/OB, French Canadian/other ethnic background; U,
undifferentiated; P, paranoid; D, disorganized; C, catatonic SrLi ~j L .,..ia; C~, first consultation; BPRS. Brief Psychiatric
Rating Scale; and, NLP neuroleptic. ~-, p<0 001
Table 1 shows the demographic and clinical
characteristics of the three groups of subjects Rs,
NRs, and C. The two groups of patients were comparable
with respect to age, level of education and socio-
economic status of the head of household. As expected,they differed significantly according to the severity
of psychosis ~BPRS scores, F=280, p<0.000), the percent
of time spent as inpatient since their first contact
with the psychiatric institution (F=81, p<0.000) and
CA 02216057 1997-09-19
the age at first contact with psychiatric care facili-
ties (F=47, p<O.OOO).
Table 2 shows the sixteen different candidate
expressed CAG repeats identified and analyzed and
includes mapping, homology, and polymorphism informa-
tion.
Table 2~ it of diffelent sLufied ESTs
Seq~elD PCR RT-PCR PNQ 1~ ,potentialf~lion PolynK~phic M~pdata
T08930 + na 15 1 ',~ ." ab~RNA!' ' _, ' - -
CUG~BPmNabS0 and xc~us E~R-I gene
R98242 - + 27 1 ' ._,~." acAMP~pQlsivei '.''
~ ,! geneexE~ssion
L37868 + na 21 N-Oct-3TF,P~Ud~mainTF H.
U23868 + na 26 No known l~mology. + HCHI
U23862OS + na 7 pc6sible l ' ~, I . ' - -
p~tein(S. cercisiae)
N55395 + na 15 b.~n _~ _ ~ ' TF +
L10379 + na 28 noknownlKmology +++
Z78314 + na 20 nol
X85326 - notdale 11 noknownb.~logy na
T90581 - + 10 nokn~wnbunology
L10375 - + 16 no l I ',, +
X82209 + na 27 27Q, l~nnnl gene disnQted byabalanced ++
1~26155 + na 23 SW12/SN~2,awidelangci '.' factor, ++
idaacls wi~ ERand RAreceptors
GCT5EII + na 22 noknownb.~nology +++ HCH3
TATABF + na 39 TFllDTATAbo~bir~lingp~teir~gen~
factor
GCIIOD04 + na 14 b,amology to a rn~3se retinoic acid ind~ceble gene +++ HCH17
1~ ~ ' PDGFTF,
ER, . ~ R~ ~c~n~ic acid ra~r.
Seven of the candidate sequences showed homol-
ogy or identity with DNA binding domains or transcrip-
tion factors. Most of the candidates (12/16) gave a
PCR product with the predicted size. Candidates that
amplified a larger than expected fragment or no prod-
ucts at all were further analyzed by RT-PCR to control
for possible intronic interruptions in the genomic DNA.
Three candidates gave an RT-PCR product of the pre-
CA 022160~7 1997-09-19
- 16 -
dicted size; only one was polymorphic using a small
sample of chromosomes. Overall, 10/16 candidate
sequences contained a polymorphic CAG repeat. Allelic
frequencies of these polymorphic CAG repeats were com-
pared in the four groups of subjects.
Only allelic variants of the GCTlOD04 locus
(primers; SCZ15:GGGGCAGCGGGTCCAGAATCTTC,
SCZ16: TGGCCTTGCTGCCCGTAGTGCT; annealing temperature
62~C) showed an overall significant group effect for
the L allele (Kruskal-Wallis H (2, N = 194) = 12.18,
p = .002)i the CAG repeat average length being the
shortest in the neuroleptic-responders (Rs),
intermediate in the non-responders (NRs) and longest in
the control group (C) (Fig. 1).
The reference point to measure the CAG repeat
length is the most common allele (180 bp fragment or 14
predicted repeats), which is taken as 0. Alleles with
n repeats above or below the 0 allele are scored +n or
-n. C indicates the control groupi Rs, neuroleptic
responsive schizophrenic patients group; and NRs, neu-
roleptic non-responsive schizophrenic patients group.
A similar trend was observed for the S allele
(Kruskal-Wallis H (2, N= 194) = 5.32, p =0.06). Post-
hoc analysis using the U-statistic showed that this
global effect was mainly due to the difference between
neuroleptic-responders and normal controls (C) (L
allele: adjusted-Z=-3.52, p=0.0004; S allele: adjusted-
Z= -2.28, p=0.02). Resistant schizophrenic patients
showed also a trend toward smaller CAG repeat average
size of the L allele compared to controls (C)
(adjusted-Z=-1,68, p=0.09). When we analyzed the sum
of the two alleles, the three groups were statistically
different (p=0.01) and the difference between controls
and Rs was significant at the level of p=0.004
(adjusted-z=-2.8). Further analysis were carried out,
CA 02216057 1997-09-19
testing the hypothesis that short alleles of the
GCTlOD04 were more frequent in schizophrenic patients.
For that purpose, two distinct classes of alleles, long
(O,l) and short (-3, -2, -1), were defined and allelic
frequencies between the four groups were reexamined
(Table 3).
Table 3: frequencies of the CAG Allele short variants of hGT 1 gene
p~ ~ ~ patientS
Controls SCZ Rs NRs
Number(2n) 174 212 86 126
% of short alleles 31% 45% 52% 40%
%2 p=0.005 p=0.0009 1'=0.12
Both parents are French C'
Nurnber (2n) 80 106 52 54
% of short alleles 35% 47% 54% 41%
%2 p=O O9 P=0.03 P=0.5
at least one par0t is non French C - )
Number (2n) 94 106 34 72
% of short alleles 28% 42% 50% 39%
%2 1'=0.03 P=0.018 1~0.12
Allelic r~ ;;C;, are given as percent of alleles shorter than 0 (<0) . Fre~pl~.n~ i~s are analyzed according
to different diagnosis groups and ethnic l~c~uulld of parents. All rl~u~ were c~ ast~Al with the
frequencies of the alleles shorter than 0 in the control group. SCZ indicates ~l~h~ c patients; Rs,
neuroleptic~ ol~ive schizophrenic patients; NRs, n~ ~ non-l~ol~iv~ GO~JIII~ patients,
and; x2, Chi 2 statistic with I degree of freedom
Schizophrenic patients, irrespective of their
neuroleptic response status were more likely to carry
one of the short alleles compared to controls (%2=7.6,
df=l, p=0.005). This difference was mainly due to Rs
schizophrenic patients who were significantly more
CA 022160~7 1997-09-19
- 18 -
likely to have small alleles compared to controls
(% =11.0, df=l; p=0.0009) and to NRs patients (% =3.30,
df=l, p=0.07). Neuroleptic-non responders were margin-
ally different from controls (X =2.41, df=l, p=0.12).
When subjects with both parents of French Canadian ori-
gin or those with at least one parent from non French
Canadian origin were analyzed separately, the same pat-
tern emerged (Rs vs. C: %2=4.6, df=l, p=0.03i schizo-
phrenics vs. C: % =2.7, df=l, p=0.09).
Finally, a correlation analysis indicated that
the size of the CAG repeat tract is linearly related to
the pattern of severity of schizophrenia (measured
blindly to genotype, using a 1-5 scoring system defined
as follows: (1, episodic shift) episodes of illness
interspersed between periods of health or near normal-
ity, (2, mild deterioration) periods of illness occur,
but there are periods of return to near normality, with
some ability to work at a job and near normal or normal
social functioning, (3, moderate deterioration), the
subject may occasionally experience some resolution of
symptoms, but overall the course is downhill culminat-
ing in a relatively sever degree of social and occupa-
tional incapacitation, (4, severe deterioration), the
subject illness has become chronic resulting in inabil-
ity to maintain employment (outside of a shelteredworkshop) and social impairment, and; (5, relatively
stable), the subject illness has not changed signifi-
cantly (since it started at a severe level of impair-
ment); the longer the size, the worse and poorer is the
outcome (Gamma statistic for S, L and L+S alleles
respectively: 0.25, p=0.01; 0.37, p= 0.002; 0.29,
p=0.002) (Fig. 2).
To evaluate the proportion of variance attrib-
utable to the CAG polymorphism in the phenotype respon-
sive schizophrenia (as contrasted to the phenotype nor-
CA 022160~7 1997-09-19
-- 19 --
mal controls), we performed a logistic regression where
the S and L alleles were the independent variables.
This analysis indicate that the length of the two
alleles contribute 10 % to the variance of this pheno-
type.
A sequence homology search was performed using
the GCTlOD04 nucleic acid sequence (GenBank acc. no.
G09710) against the non-redundant nucleic acid database
(BLASTn, GenBank). The GCTlOD04 sequence was 84%
homologous to a mouse gene (GTl, GenBank D29801, see
figure 3) from which is transcribed a 7.2 kb cDNA
encoding a 196 kDa protein of unknown function, sug-
gesting that GCTlOD04 represents a portion of the human
homologue, which we term hGTl. The murine GTl gene is
inducible with retinoic acid in the mouse embryonic
carcinoma cell line Pl9 and is expressed at highest
levels in neurons but not in glial cells. A sequence
homology search using the mGTl protein sequence identi-
fied several conserved domains in another mouse gene
(stromelysin PDGF responsive element binding protein
transcription factor, GenBank U20282) and in its human
homologue (ARl, GenBank U19345), suggesting that the
hGTl protein may also function as a transcription fac-
tor.
Common allelic variants, rather than rare muta-
tions, may be responsible for the familial aggregation
observed in complex diseases such as schizophrenia.
Allelic variants that are neither necessary nor suffi-
cient to cause a disease may not be identified by link-
age analysis, particularly when the attributable risk
is less than 10%. In contrast, association studies are
sensitive to detect such variants.
To identify genes that may confer susceptibil-
ity to schizophrenia and/or its phenotypic variability
with respect to neuroleptic responsiveness, we
CA 022160~7 1997-09-19
- 20 -
recruited patient according to their long term respon-
siveness to neuroleptic medication, a strategy that
might reduce the putative genetic heterogeneity of
schizophrenia. Control and patient groups were strati-
fied according to the ethnic background of parents;thus reducing the risk of population stratification
bias.
In accordance with the present invention, neu-
roleptic-responsive-schizophrenic patients were sig-
nificantly more likely to have hGTl gene alleles withshort CAG repeats as compared to patients who are char-
acterize by long term poor response to neuroleptics and
outcome. Furthermore, a significant correlation
between the size of the hGTl CAG repeat and the pattern
of severity of the disease (the longer is the CAG
repeat the more severe is the outcome) was identified
in the group of schizophrenic patients regardless of
the quality of their response to neuroleptic medica-
tion.
One major limitation of association studies
with a relatively small number of subjects and a poten-
tially high number of genes to be tested is an
increased risk of false positive findings (type
error). In this study, we focused on candidate genes
containing expressed and polymorphic CAG repeats, thus
markedly reducing the number of genes to be tested;
the number of CAG repeats is thought to be around 700
in the total human genome. Polymorphic CAG repeats
containing transcripts might be much less represented.
Based on these numbers, the Bonferroni corrected p-
value for our tested hypothesis ought to be between 2X
10 4 and 7x10 5. In our study, and in spite of the
small sample sizes, short alleles were likely to be
more frequent in responsive schizophrenia compared to
controls at a p-value of 9x10 , which is suggestive of
CA 022160~7 1997-09-19
a true association in the case of a complex disease
such as schizophrenia. Moreover, the fact that the
association is detected in an ethnically very homoge-
nous subgroup (both parents French Canadians) as well
as in a mixed subgroup (at least one parent is non
French Canadian~, suggests that this allelic associa-
tion is very likely not to be due to stratified popula-
tion bias. Furthermore, the fact that hGTl gene has a
high homology with a mouse gene involved in neural cell
differentiation induced by retinoic acid is consistent
with both the neurodevelopmental and retinoic acid
hypotheses of schizophrenia.
Patients who presented with episodic shifts and
good between-episode recovery were more likely to have
shorter CAG repeats in both of their hGTl alleles. This
finding could be interpreted in various different ways:
(1) it could indicate that hGTl short alleles have a
causative effect in the disease of patients with favor-
able outcome (good between episodes recovery, slow pro-
gression of functional deficits) whereas resistantpatients with sever pattern of severity (continuos psy-
chosis, no psychotic free episodes, rapid decline of
psychosocial functioning) have other genetic or envi-
ronmental factors involved in their disease. Patients
falling between these two levels of severity may be a
more mixed group difficult to relate to either one of
the two extremes using clinical criteria (heterogeneity
hypothesis), (2) it could indicate that the hGTl poly-
morphism modulates the pattern of severity of the
schizophrenia phenotype but not the susceptibility to
schizophrenia per se (modifier gene hypothesis), and
(3); hGTl gene could influence susceptibility to
schizophrenia irrespective of the pattern of severity
and responsiveness to neuroleptics; the weak associa-
tion in the group of resistant schizophrenic patients
CA 022160~7 1997-09-19
- 22 -
being the result of a selection bias. Indeed, should
another gene with a higher attributable risk than the
hGTl be acting in the resistant form, the enrichment of
hGTl short alleles in non-responsive patients with
severe pattern of the disease would be relaxed and the
association would be more difficult to identify in this
group. In accordance with this hypothesis, family
studies have suggested that neuroleptic-delayed
response and marked deterioration in the psychosocial
functioning are associated with a higher degree of
familial aggregation of the disease; suggesting the
presence of gene(s) with relatively high penetrance.
Transcription factors are major actors in all
neurodevelopmental phases, and might be very important
when developmental activity is intensive such as brain
fetal development or synaptic pruning occurring in the
adolescence phase of human development. They have been
implicated in animal complex behavioral traits and have
also a major role in the transduction pathways involved
in the biological adaptation of the central nervous
system to environmental changes (ranging from physical
conditions such as viral infections to psychological
conditions such as nurturing behavior in mice). It is
also of interest to note that all antipsychotic drugs
2 5 modulate DNA transcription in specific areas of the
brain and ultimately results in modifications of neu-
ronal interconnectivity. Variable number of tandem
repeats, including trinucleotide repeats, have been
found to be over represented in genes coding for DNA-
binding proteins/transcription factors. Such repeatsmay be the basis of a fine modulation of gene activity.
We speculate that one or multiple transcription factors
might be involved in the etiology of schizophrenia or
its phenotypic variability (including the quality of
the response to different drugs). It is therefore of
CA 022160~7 1997-09-19
- 23 -
interest to consider transcription factors containing
polymorphic CAG repeats as a putative candidate "family
of genes" for schizophrenia and other psychiatric
disorders thought to be of a neurodevelopmental origin.
The present invention will be more readily un-
derstood by referring to the following examples which
are given to illustrate the invention rather than to
limit its scope.
While the invention has been described in con-
nection with specific embodiments thereof, it will be
understood that it is capable of further modifications
and this application is intended to cover any varia-
tions, uses, or adaptations of the invention following,
in general, the principles of the invention and
including such departures from the present disclosure
as come within known or customary practice within the
art to which the invention pertains and as may be
applied to the essential features hereinbefore set
forth, and as follows in the scope of the appended
claims.
