Refinement of the 6q Chromosomal Region
Implicated in Transient Neonatal Diabetes
Hélène Cavé, Michel Polak, Séverine Drunat, Erick Denamur, and Paul Czernichow
Transient neonatal diabetes mellitus (TNDM) is estimated to occur in ~1 in 500,000 births and represents
50–60% of cases of neonatal diabetes. The pattern of
inheritance of TNDM and its association with chromosome 6 uniparental disomy is consistent with the presence on chromosome 6 of an imprinted gene involved in
pancreatic -cell development. Systematic screening
for chromosome 6 abnormalities in nine families with 13
individuals affected by TNDM revealed paternal isodisomy of chromosome 6 in one child and paternally
derived trisomy of the chromosomal region 6q in six
children from three unrelated families. To delineate
more accurately the region suspected of harboring the
gene of interest, precise mapping of the duplicated
area was performed in children exhibiting partial 6q trisomy by using microsatellite markers. The smallest
region of duplication observed in our patients was
flanked by markers D6S308 and D6S1010, which are
separated by <1 cM. These findings confirm that TNDM
may result from the overexpression of a gene located on
chromosome 6q that is exclusively expressed from the
paternal allele at least during some periods of life and
further refine the localization of this gene. Diabetes
49:108–113, 2000
T
ransient neonatal diabetes mellitus (TNDM) is a
developmental disease of insulin production that
regresses in postnatal life. It is estimated to occur
in ~1 in 500,000 births and represents 50–60% of
cases of neonatal diabetes (1,2). Patients are usually born with
intrauterine growth retardation and develop hyperglycemia,
are unable to thrive, and sometimes suffer from dehydration.
Insulin production is inadequate and exogenous insulin therapy is required to maintain euglycemia. Patients with TNDM
do not show any evidence of anti-islet antibodies or of the
type 1 diabetes–associated HLA class II susceptibility haplotypes (2). A defect in -cell maturation has been proposed to
explain this condition (3). Interestingly, exocrine pancreatic
insufficiency is present in some patients (4). However, the etiology of TNDM is still unknown. Most patients recover by
1 year of age, but some patients continue to have impaired gluFrom Laboratoire de Biochimie Génétique (H.C., S.D., E.D.); Service
d’Endocrinologie (M.P., P.C.); INSERM U457 (M.P., P.C.); and INSERM U458
(E.D.), Hôpital Robert Debré, Paris, France.
Address correspondence and reprint requests to Hélène Cavé, Laboratoire
de Biochimie Génétique, Hôpital Robert Debré, 48 Boulevard Sérurier,
75019 Paris, France. E-mail: [email protected].
Received for publication 30 June 1999 and accepted in revised form
16 September 1999.
PCR, polymerase chain reaction; STR, simple tandem repeat; TNDM,
transient neonatal diabetes mellitus.
108
cose tolerance, and diabetes has been shown to possibly
recur in late childhood or adulthood. These recurrences are
generally consistent with nonautoimmune type 1 diabetes, but
their precise mechanism (insulin deficiency and/or possibly
insulin resistance) remains to be determined (1,5,6).
Most cases are sporadic, but in about one-third of
reported cases, familial recurrences have been observed
that do not follow mendelian rules of inheritance (7). In all
reported cases, transmission was of paternal origin,
although the father might be unaffected by the disease (1,8).
Paternal isodisomy of chromosome 6 has been demonstrated in several unrelated patients with TNDM (9,10,12).
These observations are consistent with the presence on
chromosome 6 of an imprinted gene involved in the disease.
Five cases of TNDM have been shown to be associated with
partial duplications of the long arm of paternal chromosome
6 (11–14), but maternal duplications of the same region do
not result in TNDM (12). Moreover, interstitial deletions of
the same region of the maternally inherited chromosome 6
are not associated with the disease, which makes it unlikely
that TNDM is due to the absence of expression of a maternally expressed gene (15). These findings suggest that
TNDM may result from the overexpression of a gene located
on chromosome 6q and exclusively expressed from the
paternal allele at least during some periods of life.
All trisomies reported so far encompass the 6q23 region,
while a 6q24-qter paternally derived trisomy has been
reported that does not lead to TNDM (16). Using polymorphic
markers analysis and linkage analysis, Gardner et al. (17)
localized the critical region between markers D6S1699 and
D6S1010 (6q24.1–q24.3), which corresponds to an area of
3–4 cM (18).
Systematic screening for chromosome 6 abnormalities in
13 patients with TNDM enabled us to demonstrate paternal
isodisomy of chromosome 6 in one child and a paternally
derived partial trisomy of chromosome 6 in six children from
three unrelated families. Delineation of the duplicated region
permitted the chromosomal region likely to harbor the gene
involved in this pathology to be further refined.
RESEARCH DESIGN AND METHODS
Subjects. A total of nine families with 13 individuals affected by TNDM were studied (Table 1). Informed consent was obtained from the patients, their parents, or
both, as deemed appropriate. Twins D1 and D2, who are both affected by TNDM,
were subsequently shown to be monozygous by microsatellite analysis. Main
clinical features of the patients with TNDM are given in Table 1. All patients had
TNDM as defined by the following criteria: hyperglycemia, poor weight gain, and
need to introduce insulin therapy within 3 months of birth (Table 1). In all cases,
insulin therapy could be stopped between the 4th and 15th months of life; therefore, all of these children had TNDM. During the follow-up, three patients (B, E,
and H) had a weak insulin response to intravenous glucose tolerance tests. In addition, two patients (B and F) had episodes of glucose intolerance during acute sickDIABETES, VOL. 49, JANUARY 2000
H. CAVÉ AND ASSOCIATES
DIABETES, VOL. 49, JANUARY 2000
109
TRANSIENT NEONATALDIABETES
ness some time after the neonatal period. Patient F had to be treated again with
insulin from the age of 13 years, although at a very low dose. This patient had
detectable C-peptide levels that were normally reactive after glucagon injection.
In the families of patients B and F, numerous members have type 1 or type 2 diabetes. All but one child had intrauterine growth retardation. No dysmorphic features, except for macroglossia in one child, were observed in this cohort of children. Insulin levels, when measured at the time of diagnosis and before the start
of insulin therapy, were very low or undetectable (patients A, B, and E), except
in one case (patient H). In all cases, insulin secretion was clearly unadapted to
glycemia. None of the children who were tested had the immune markers or the
major histocompatibility complex class II haplotypes observed in type 1 diabetes. None of the patients had a defect of exocrine function that was clinically
relevant or detectable by fecal fat excretion determination. Ultrasonography,
when performed, did not reveal pancreatic abnormalities. Karyotypes were normal in all patients.
Markers. A total of 23 simple tandem repeat (STR) markers were used: D6S105
maps at chromosome band 6p21; D6S300 maps at chromosome band 6q21;
D6S261, D6S270, D6S292, D6S1699, D6S1569, D6S314, D6S1675, D6S471, D6S453,
D6S310, D6S308, AFM269zg5, D6S279, D6S1704, D6S1003, D6S1010, D6S1049,
D6S1703, and D6S978 maps within chromosome bands 6q22–q23; D6S311 maps
at chromosome band 6q24.1; and D6S305 maps at chromosome band 6q25.
Sequences and chromosomal assignments were obtained from the Genome
Database (http://www.gdb.org). STR markers were ordered according to the
Genethon genetic map (18), the Whitehead yeast artificial chromosome contigs
WC6–13 and WC6–15 (http://carbon.wi.mit.edu:8000/cgi-bin/contig/phys_map),
and the Sanger Center’s chromosome 6 radiation hybrid map (http://www.sanger.
ac.uk/HGP/chr6) (Fig. 1).
Genotyping. Peripheral blood of patients and their families was collected on
EDTA. DNA was extracted from leukocytes using standard methods and stored
at –20°C until analysis.
A polymerase chain reaction (PCR) was carried out with a 5 fluoresceinlabeled upstream primer (Genset, Paris) using the “hot start” procedure. DNA
(1 µl) was amplified in a final volume of 25 µl containing 10 pmol of 5 and
3 primers, 0.2mmol/l dNTP, 1.5mmol/lmgCl2, 50mmol/l KCl, 10mmol/l TRISHCl (pH 8.3), and 0.2 U Taq Polymerase (Appligene, Strasbourg, France). The
PCR cycling parameters were 10 s at 94°C and 10 s at 50°C for 35 cycles carried
out in an UNO-Thermoblock (Biometra, Göttingen, Germany). PCR products
were analyzed using a fluorescent automated laser DNA sequencer (ABI-PRISM
310, Perkin Elmer, Norwalk, CT). Areas under the peaks were determined using
the 310 Genescan (Perkin Elmer) integration software. The signal ratio for the
A1 allele to the A2 allele was calculated for patients and compared with normal
controls. Normal controls were DNA from normal subjects who display the same
allelic pattern and/or artificial reconstitutions of normal or trisomyl patterns
obtained by 1:1 and 2:1 dilutions of DNA from normal subjects exhibiting
homozygous allelic patterns. In cases where trisomy was suspected, at least three
PCR replicates were tested.
RESULTS
We studied 13 children with TNDM belonging to nine families
(Table 1). A systematic screening for uniparental disomy of
chromosome 6 was conducted by analysis of microsatellite
markers D6S105, D6S300, D6S261, D6S270, D6S292, D6S310,
D6S311, and D6S305 in patients and their parents. Paternal
isodisomy of chromosome 6 was observed in only one
patient (A). In two other patients (C and F), the allelic pattern
observed for marker D6S310, which was assigned to chromosomal band 6q23.3, showed an increased gene dosage of
the paternal allele, which is suggestive of a trisomy. The pattern observed for markers D6S292 (6q22.33) and D6S311
(6q24.1) was normal in these two children (Fig. 1).
To delineate more accurately the duplicated region, additional markers spanning the 6q22–q24 region were studied in
these two families. In both cases, the duplicated region
included markers D6S453, D6S310, D6S308, AFM269zg5,
D6S1704, and D6S1003 but excluded markers D6S471 on the
centromeric side. On the telomeric side, marker D6S1010
was duplicated in child C but not in child F (Fig. 1). The
father of child C always contributed two doses of the same
allele, indicating that the duplication arose from a single
chromosome, and trisomy was shown by an increased
dosage of paternal alleles (Fig. 2A). In child F, trisomy manifested either by an increase dosage of one allele or by the
presence of three alleles of different size, two of which were
paternally inherited (Fig. 2B). This suggests that, in the latter
case, trisomy occurred during the first meiosis. No chromosome 6 duplication was shown in the fathers of these two chil-
FIG. 1. Analysis of microsatellite
markers spread over chromosome 6
in TNDM patients with partial trisomy of chromosome 6q. , Normal
pattern; , trisomy; , noninformative. SRO, smallest region of overlap.
110
DIABETES, VOL. 49, JANUARY 2000
H. CAVÉ AND ASSOCIATES
dren, suggesting a de novo duplication. Markers spanning and
flanking the duplicated region in these two children were
analyzed in the six patients for whom no chromosome 6
abnormalities were observed. Evidence of a duplication was
obtained in a third family (children I, J, K, and L). Duplication
was detected in the four children with TNDM and in their two
fathers (Fig. 3). The two fathers were unaffected by the disease, but subsequent oral glucose tolerance testing revealed
impaired tolerance in both. It is likely that the genetic defect
was transmitted via their respective mother and grandmother. Unfortunately, no genetic material was available for
testing. In this family, trisomy was restricted to markers
AFM269zg5, D6S1704, and D6S1003, while markers D6S308
and D6S1010 were normal (Table 2 and Fig. 1).
Numerous members of the paternal family of child F had diabetes. To investigate any relationship with the occurrence of a
case of TNDM in this family, we analyzed the allelic pattern of
chromosome 6 markers in the paternal aunt of child F, who has
type 1 diabetes, and in the paternal grandmother and great-aunt,
who have type 2 diabetes. Trisomy was not found in these
three people, and they did not display the paternally inherited
haplotypes of child F in the 6q22–24 region, which thereby
excludes a direct implication of this locus in familial diabetes.
DISCUSSION
FIG. 2. Typical results of microsatellite analysis obtained in patients
with duplication. After electrophoresis of fluorescent PCR products,
each allele gave rise to a major peak, the size of which depends on the
number of nucleotide repeats, associated to two or three smaller, less
intense products or “slutter bands.” A: Example of a duplication with
increased dosage of one paternal allele (marker D6S1704). Duplications
were scored as increases in intensity of one allele relative to the other.
The ratio of signal for the two alleles of the child with TNDM (R = 2.60)
is twice the ratio calculated for his father (R = 1.34) and normal control subjects (data not shown). B: Example of a duplication evidenced
by the presence of three different alleles, two of which are paternally
inherited, of marker D6S308 in a child with TNDM (child F).
This study confirms that TNDM, either familial or sporadic,
is linked to the chromosomal region 6q23–24. The alterations we found are paternal uniparental disomy of chromosome 6 or 6q23–24 duplications, which are consistent
with the hypothesis that TNDM is due to an increased gene
dosage of a paternally expressed gene. Pooling our results
with those of Gardner et al. (19) indicates that of a total of
33 unrelated children with TNDM, an abnormality of chromosome 6 was observed in 15 cases (45%) with 9 cases of uniparental disomy (27%) and 6 cases of dup(6)(q23–q24)
(18%). The hypothesis that some cases of TNDM could be
associated with another locus cannot be formally excluded.
However, it is likely that at least some of the children for
whom no defect was found in our and other studies harbor
a genetic defect at the same gene(s), but the defect is the
result of molecular mechanisms that cannot be detected
using microsatellite typing and cytogenetics.
FIG. 3. Family tree of patients I, J, K, and
L. Only patients III2, III3, III4, II6, IV E, IVF,
IVG, and IV H have been tested for chromosome 6 abnormalities. , 6q23–24 trisomy
without TNDM; , 6q23–24 trisomy with
TNDM.
DIABETES, VOL. 49, JANUARY 2000
111
TRANSIENT NEONATALDIABETES
Gardner et al. (17) recently localized the region of interest
in TNDM between markers D6S1699 and D6S1010. Precise
mapping of the duplicated area in three families enabled us to
further restrict the region believed to harbor the gene of interest. The smallest region of overlap between the duplicated
areas observed in our patients is contained within the region
described by Gardner et al (17). It is flanked by markers
D6S308 and D6S1010, which are separated by <1 cM. PC-1 has
been proposed as a candidate gene because of its overexpression in the fibroblasts of patients with type 2 diabetes (20).
Results from chromosome 6 radiation hybrid map consistently show that the PC-1 gene maps are centromeric to our
region and thus cannot be the gene responsible for TNDM. Several expressed sequence tags map in the region that we have
delineated. In the absence of information concerning their
function, a systematic search for those that are imprinted
could help to identify the gene(s) responsible for TNDM.
A familial history for diabetes has been reported in 31% of
TNDM and 40% of permanent neonatal diabetes, but only
few reports state the type of diabetes (8). It is worth noting
that the child in whom diabetes recurred (patient F) and the
child who had glucose intolerance during acute sickness
(patient B) have numerous family members with type 2 diabetes, in whom insulin resistance is an important causal phenomenon. Study of chromosome 6 markers in the members
of the family of child F with diabetes suggests that the locus
responsible for TNDM is independent of the one responsible
for familial diabetes. In these two children, an additional
genetic cause of insulin resistance could have played a role
in the recurrence of diabetes.
ACKNOWLEDGMENTS
We wish to thank Drs. P. Garandeau, M. Beregszaszi, M.
Nicolino, A.M. Bertrand, E. Khallouf, and C. Metz for allowing us to study their patients. We thank Drs. G. Tachdjian and
F. Fellmann for karyotypes. We thank R. Sharfmann and B.
Grandchamp for helpful discussion.
REFERENCES
1. Von Mühlendahl KE, Herkenhoff H: Long-term course of neonatal diabetes.
N Engl J Med 333:704–708, 1995
2. Shield JP, Gardner RJ, Wadsworth EJ, Whiteford ML, James RS, Robinson DO,
Baum JD, Temple IK: Aetiopathology and genetic basis of neonatal diabetes.
Arch Dis Child 76:39–42, 1997
3. Ferguson AW, Milner RDG: Transient neonatal diabetes mellitus in sibs. Arch
Dis Child 45:80–81, 1970
4. Fösel S: Transient and permanent neonatal diabetes. Eur J Pediatr 154:944–
948, 1995
5. Shield JPH, Baum JD: Transient neonatal diabetes and later onset diabetes:
a case of inherited insulin resistance. Arch Dis Child 72:56–57, 1995
6. Geffner ME, Clare-Salzler M, Kaufman DL, Ting GS, Gottschalk ME, Schatz DA:
Permanent diabetes developing after transient neonatal diabetes (Letter).
Lancet 341:1095, 1993
7. Temple IK, James RS, Crolla JA, Sitch FL, Jacobs PA, Howell WN, Betts P,
Baum JD, Shield JP: An imprinted gene(s) for diabetes? Nat Genet 9:110–112,
1995
8. Coffey JD, Killelea DE: Transient neonatal diabetes in half sisters: a sequel.
Am J Dis Child 136:626–627, 1982
9.Whiteford ML, Narenda A, White MP, Cooke A, Wilkinson AG, Robertson KJ,
Tolmie JL: Paternal uniparental disomy for chromosome 6 causes transient
neonatal diabetes. J Med Genet 34:167–168, 1997
10. Christian SL, Rich BH, Loebl C, Israel J, Vasa R, Kittikamron K, Spito R,
Rosenfield R, Ledbetter DH: Significance of genetic testing for paternal uniparental disomy of chromosome 6 in neonatal diabetes mellitus. J Pediatr
134:42–46, 1999
11. Pivnik EK, Qumsiyeh MB, Tharapel AT, Summit JB, Wilroy RS: Partial duplication of the long arm of chromosome 6: a clinically recognisable syndrome.
112
DIABETES, VOL. 49, JANUARY 2000
H. CAVÉ AND ASSOCIATES
J Med Genet 27:523–526, 1990
12. Ziel B, Zneimer SM, Bachman R: Partial trisomy of chromosome 6q, and
interstitial duplication of the long arm. Am J Hum Genet 53:131–141, 1995
13. Temple IK, Gardner RJ, Robinson DO, Kibirige MS, Ferguson AW, Baum JD,
Barber JC, James RS, Shield JP: Further evidence from imprinted gene for
neonatal diabetes localized to chromosome 6q22–6q23. Hum Mol Genet
5:1117–1121, 1996
14. Arthur EI, Zlotogora J, Lerer I, Dagan J, Marks K, Abeliovich D: Transient
neonatal diabetes mellitus in a child with invdup(6)(q22q23) of paternal origin. Eur J Hum Genet 5:417–419, 1997
15. Pandya A, Braverman N, Pyeritz RE, Ying KL, Kline AD, Falk RE: Interstitial
deletion of the long arm of chromosome 6 associated with unusual limb
anomalies: report of two new patients and review of the literature. Am J Med
Genet 59:38–43, 1995
16. Chase TR, Jalal SM, Martsolf JT, Wasdahl WA: Duplication 6q24–6qter in an
infant from a balanced paternal translocation. Am J Med Genet 14:347–351,
DIABETES, VOL. 49, JANUARY 2000
1983
17. Gardner RJ, Mungall AJ, Dunham I, Barber JCK, Shield JPH, Temple IK,
Robinson DO: Localization of a gene for transient neonatal diabetes mellitus
to an 18.72 cR3000 (~5.4 Mb) interval on chromosome 6q. J Med Genet
36:192–196, 1999
18. Dib C, Faure S, Fizames C, Samson D, Drouot N, Vignal A, Millasseau P, Marc
S, Hazan J, Seboun E, Lathrop M, Gyapay G, Morisette J, Weissenbach J: A comprehensive genetic map of the human genome based on 5,264 microsatellites.
Nature 380:152–154, 1996
19. Gardner RJ, Mungall AJ, Shield JPH, Robinson DO, Barber JCK, Temple IK:
Further localisation of a gene for transient neonatal diabetes mellitus to chromosomal band 6q24 (Abstract). Am J Hum Genet 63 (Suppl.):A55, 1998
20. Maddux BA, Sbraccia P, Kumakura S, Sasson S, Youngren J, Fisher A, Spencer
S, Grupe A, Henzel W, Stewart TA, Reaven GM, Goldfine ID: Membrane glycoprotein PC-1 and insulin resistance in non-insulin-dependent diabetes mellitus. Nature 373:448–451, 1995
113