Phosphoglucomutase Genetic Polymorphism

and Body Mass

FULVIA GLORIA-BOTTINI, MD; ANDREA MAGRINI, MD; ELENA ANTONACCI, MD;
MAURO LA TORRE, MD; LAURA DI RENZO, MD; ANTONINO DE LORENZO, MD;
ANTONIO BERGAMASCHI, MD; EGIDIO BOTTINI, MD
ABSTRACT: Background: We have searched for a pos-
sible association of the genetic polymorphism of Phos-
phoglucomutase locus 1 (PGM
1
), a key enzyme in car-
bohydrate metabolism, with body mass. Methods:
Adults (n 257) with type 2 diabetes, 74 children
referred for obesity, and 740 consecutive healthy new-
born infants were studied. Body mass index, body
weight, birth weight, and PGM
1
phenotype were deter-
mined. Sexes were analyzed separately. Results: In type
2 diabetes, females carrying the PGM
1
*2 allele are less
represented among subjects with extreme body mass
index deviation as compared with other classes of sub-
jects. Among children referred for obesity, females
carrying the PGM
1
*2 allele are less represented among
children with extreme body weight deviation. Among
consecutive infants, in both sexes the proportion of
those showing a birth weight higher than the 3rd quartile
is lower in homozygous PGM
1
2/2 subjects than in other
PGM
1
phenotypes. Conclusions: The data suggest that
during extrauterine life, females carrying the PGM
1
*2
allele are relatively protected from extreme body mass
increase. During intrauterine life, PGM
1
2/2 homozy-
gotes show a tendency to low body mass increase.
Because PGM
1
enzymatic activity depends on its phos-
phorylation status by the kinase Pak1, both structural
differences of the PGM
1
allelic product and different
rates of activation by Pak between sexes might be respon-
sible for the pattern observed. At present, the effect of other
genes near PGM
1
and in linkage disequilibrium with it
cannot be ruled out. KEY INDEXING TERMS: PGM
1
; Ge-
netic polymorphism; Body mass index; Birth weight; Over-
weight. [Am J Med Sci 2007;334(6):421425.]
O
besity is a heterogeneous group of disorders
having in common energy balance disturbances
and fat cell increase. Besides excess caloric intake
and a sedentary lifestyle, genetic factors are also of
importance in the development of obesity. Moreover
variation in energy intake and participation in phys-
ical activity are not independent from genetic fac-
tors
1,2
. At present, more than 400 chromosomal re-
gions have been identified as candidate genes.
37
Moreover, obesity shows different patterns in males
and females, suggesting that sex hormones may
modify the effect of genes involved in the develop-
ment of obesity disorders.
Phosphoglucomutase (PGM) is a highly polymor-
phic enzyme that plays a key role in glycide metab-
olism. Associations with birth weight have been ob-
served both in normal and diabetic pregnancy,
8,9
suggesting that its genetic variability may be impor-
tant for body mass development. Differences be-
tween males and females concerning the effect of
PGM
1
on birth weight have been also reported.
10
Phosphoglucomutase Polymorphism
Phosphoglucomutase is an enzyme widely distrib-
uted in nature that catalyzes the reversible reaction:
glucose-1-phosphate glucose-6-phosphate, an
essential step in carbohydrate metabolism. Four sep-
arate loci determine distinct sets of PGM isozymes:
PGM
1
, PGM
2
, PGM
3
, and PGM
4.
1114
About 85% to
95% of total PGM activity is determined by the PGM
1
locus, that shows an electrophoretic polymorphism de-
termined by the occurrence of 2 codominant alleles:
PGM
1*
1 and PGM
1
*2 at a locus on the short arm of
chromosome No. 1). In vitro the activity associated
with PGM
1
*2 is higher than that associated with
PGM
1
*1.
15
Recent studies have shown that Pak1 (p-21
activated kinase) binds to, phosphorylates, and en-
hances the enzymatic activity of PGM
1
.
16
Differences
in activation between the 2 allelic products of PGM
1
From the Department of Biopathology and Imaging Diagnostics
(FG-B, AM, EB) and the Department of Neurosciences (LDR, ADL),
University of Rome Tor Vergata, School of Medicine, Rome, Italy;
the Center of Diabetology (EA) and the Department of Pediatrics
(MLT), S. Massimo Hospital, Penne, Italy; and the Institute of
Occupational Health Medicine (AB), Catholic University of Holy
Hearth, Rome, Italy.
Submitted February 16, 2007; accepted in revised form April 11,
2007.
Correspondence: Dr. Fulvia Gloria-Bottini, Department of Bio-
pathology and Imaging Diagnostics, University of Rome, Tor
Vergata, Via Montpellier, 1, 00133 Rome, Italy (E-mail: gloria@med.
uniroma2.it).
421 THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
are likely but have not yet been investigated. Thus,
in vivo, differences in enzymatic activity between
PGM
1
phenotypes could be more marked relative to
those observed in vitro.
Isoelectrofocusing
1719
and heat denaturation
20
studies have revealed a greater variability within
the PGM
1
locus leading to the identification of 8
different PGM
1
alleles. Recent evidence indicates
that intragenic recombination has an important role
in generating PGM
1
variability.
2125
The central role inglycide metabolism, the organspec-
ificity of some sets of PGMisozymes (the PGM
4
locus is
active only during lactation and its isozymes are
present only in milk
14
), the variable proportion of dif-
ferent sets of isozymes between different tissues and
within a given tissue,
13
and the presence of genetic poly-
morphism in all human populations strongly suggest
that genetic variability of PGM is adapted to specific
functions of tissue and organs and may have impor-
tance in energy expenditure and physical activity.
In the present study, we searched for a possible asso-
ciation of the PGM
1
phenotype with body mass in adults
and children and with birth weight in newborns from
normal pregnancy. Sexes have beenanalyzedseparately.
Materials and Methods
Samples Studied
Adults With Type 2 Diabetes. Adults (n 257) with type 2
diabetes from the Caucasian population of Penne, a small rural
town in southeastern Italy, were studied. The sample was chosen
randomly froma population of about 2000 subjects under care at the
Center of Diabetology of the local hospital. Samples were collected
over a period of about 18 months from patients scheduled for met-
abolic control on a previously fixed day of the week. The sample
includes male and female patients (mean age, 66.3 years; SD, 9.8).
Children Referred for Obesity. Caucasian children (n
74; ages between 3 and 14 years), referred for obesity to the
out-patient Department of our Pediatric Clinic, were studied.
Children were considered obese if their weight exceeded more
than 20% the mean weight for their age, height, and sex. All
children were above median height. The cutoff point between
moderate and severe deviations of body mass was fixed on 4
standard deviations above the mean. Cases considered for our
analysis were not associated with known disease and did not
show abnormalities of the following parameters: glycemia, tri-
glyceridemia, cholesterolemia, and urinalysis with determination
of free cortisol in the urine.
Newborns FromHealthy Women. Consecutive healthy new-
born infants (n 740) collected from the Caucasian populations
of Rome and Penne were studied. No significant differences in
PGM
1
distribution, birth weight, and gestational age were ob-
served between the sample of Rome and that of Penne. Birth
weight percentile was evaluated for each class of gestational age
according to tables for the population of Florence.
25
Gestational
length was estimated from the date of the last menstrual period
and checked with Dubowitz score as an additional index of neo-
natal maturity.
Blood samples in adults and children were obtained by vasopunc-
ture. Newborn blood samples were collected from the placental side
of umbilical cord after its section. The investigation was approved by
the Department of Biopathology and Imaging Diagnostics, and pa-
tients or parents of children gave informed consent to be included or
include their children in this investigation.
Laboratory and Statistical Methods
PGM
1
phenotype was determined in all subjects by starch gel
electrophoresis according to the method of Spencer et al.
11
The
2
test of independence was performed, using SPSS programs.
26
Three-way contingency tables were analyzed, using a log-linear
model according to Sokal and Rohlf.
27
Combined probabilities
were evaluated according to Sokal and Rohlf.
27
Results
Table 1 shows the proportion of subjects with
extreme body mass distribution (body mass index
Table 1. Percent Proportion of Subjects With Severe Body Mass in Adults With Type 2 Diabetes and of Children Referred for
Obesity in Relation to PGM1 Phenotype and Gender
Adult Subjects With Type 2 Diabetes
Males Females
PGM
1
1/1 Phenotype PGM
1
*2 Carriers PGM
1
1/1 Phenotype PGM
1
*2 Carriers
Proportion of subjects with
BMI 35
9.2% 16.1% 15.5% 6.1%
No. of subjects 65 56 71 65
Three-way contingency table analysis by log linear model x PGM
1
; y BMI; z sex; xyz interaction P 0.045
Independence between BMI and gender: in PGM
1
1/1 P 0.300: in carriers of PGM
1
*2 allele P 0.070
Children Referred for Obesity
Males Females
PGM
1
1/1 Phenotype PGM
1
*2 Carriers PGM
1
1/1 Phenotype PGM
1
*2 Carriers
Proportion of children with
weight 4 SD
33.3% 27.8% 34.6% 6.7%
N of children 15 18 26 15
Three way contingency table analysis by log linear model x PGM
1
; y BMI; z sex; xyz interaction P 0.150
Independence between BMI and gender: in PGM
1
1/1 P 0.980: in carriers of PGM
1
*2 allele P 0.024
Combining probabilities: for interaction P 0.040; for independence between body mass and gender in carriers of PGM
1
*2 allele P 0.013
PGM
1
and Body Mass
422 December 2007 Volume 334 Number 6
[BMI] 35) in subjects with type 2 diabetes and in
children referred for obesity (weight 4 SD) in
relation to PGM
1
phenotype and gender. In subjects
with type 2 diabetes, there is a significant interac-
tion between PGM
1
phenotype, BMI, and gender. In
carriers of the PGM
1
*2 allele (PGM
1
2/1 and 2/2),
severe obesity (BMI35) is less frequent in females
than in males. In children referred for obesity
(weight 4 SD), the interaction between PGM
1
phe-
notype, body weight, and gender does not reach a
level of statistical significance, but the pattern is
similar to that observed in subjects with type 2
diabetes, showing that in carriers of the PGM
1
*2
allele, severe body mass increase is much less fre-
quent in females than in males.
In Table 2, we have considered the proportion of
PGM
1
*2 alleles in subjects with type 2 diabetes and
in children referred for obesity in relation to sex
and body mass deviation. Control subjects (healthy
newborns from the same population) are reported
but have not been considered in statistical analysis.
In subjects with type 2 diabetes, there is a signifi-
cant interaction between the PGM
1
allele, BMI, and
gender. In severely obese subjects (BMI 35), the
proportion of the PGM
1
*2 allele is much lower in
females than in males. In children referred for obe-
sity the interaction between the PGM
1
allele, weight
and gender do not reach a level of statistical signifi-
cance, but the pattern of association is similar to that
observed in type 2 diabetes, showing that in severely
obese subjects (weight 4 SD) the proportion of
PGM
1
*2 alleles is lower in females than in males.
Table 3 shows the proportion of newborn infant with
a birth weight higher than 3rd quartile in relation to
Table 2. Percent Proportion of PGM1*2 Allele in Relation to Sex and Severe Increase of Body Mass in Adults With Type 2 Diabetes
and in Children Referred for Obesity
Type 2 Diabetes
BMI 35 BMI 35 Controls
%PGM
1
*2 No. Alleles %PGM
1
*2 No. Alleles %PGM
1
*2 No. Alleles
Males 28.3% 212 43.3% 30 28.1% 766
Females 28.5% 242 13.3% 30 29.4% 714
Three way contingency table analysis by a log linear model x PGM
1
; y BMI; z sex; xyz interaction P 0.012
Independence between PGM
1
allele and gender: BMI 35 P 0.980, BMI 35 P 0.010
Children Referred for Obesity
Weight 4 SD Weight 4 SD Controls
%PGM
1
*2 No. Alleles %PGM
1
*2 No. Alleles %PGM
1
*2 No. Alleles
Males 32.6% 45 25.0% 20 28.1% 766
Females 27.4% 62 5.0% 20 29.4% 714
Three way contingency table analysis by a log linear model x PGM
1
; y BMI; z sex; xyz interaction P 0.140
Independence between PGM
1
allele and gender Body weight 4 SD P 0.910 Body weight 4 SD P 0.080
Combining probabilities: for interaction P 0.014; for independence between PGM
1
allele and gender in severely obese subjects P 0.003
Controls are shown for comparison only and have not been considered in statistical analysis.
Table 3. Percent Proportion of Newborn Infants in Relation to Birth Weight Quartile, PGM1 phenotype, and Gender
Males Females
1st Q 2nd Q 3rd Q 4th Q No. 1st Q 2nd Q 3rd Q 4th Q No.
PGM
1
1/1 and PGM
1
2/1
phenotypes
22.9% 23.2% 28.3% 25.6% 336 21.4% 24.5% 22.6% 31.4% 318
PGM
1
2/2 phenotype 20.6% 32.3% 35.3% 11.8% 34 43.5% 8.7% 39.1% 8.7% 23
Three-way contingency table analysis by a log linear model x birth weight; y PGM1; z gender
1st Q vs 2nd Q vs 3rd Q vs 4th QP (1st Q 2nd Q 3rd Q) vs 4th QP
xyz interaction 0.070 0.570
xy independence 0.007 0.006
xz independence 0.120 0.230
yz independence 0.090 0.480
Independence of x from y and z 0.008 0.005
Q quartile.
Gloria-Bottini et al
423 THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
PGM
1
phenotype and gender. In both males and fe-
males carrying the homozygous PGM
1
2/2 phenotype,
the proportion of those with a birth weight higher than
the 3rd quartile is much lower than in the other PGM
1
phenotypes (PGM
1
1/1 and PGM
1
1/2).
Discussion
The data suggest that during extrauterine life,
females carrying the PGM
1
*2 allele are relatively
protected from extreme body mass increase. During
intrauterine life PGM
1
2/2 subjects of either sex
show a tendency to reduced body mass increase. The
association between PGM
1
and body mass is similar
in 3 independent conditions arguing against the
possibility of sampling chance artifact.
Considering the key role of PGM
1
in glycide metab-
olism, we are disposed to consider a causal mechanism
for the statistical associations reported. However, the
effect of other genes near PGM
1
and in linkage dis-
equilibrium with it cannot be excluded. In addition to
PGM
1
, other genes important in the glycolytic path-
way are clustered on the short arm of chromosome 1
including glucose dehydrogenase (GDH), 6-phosphoglu-
conate dehydrogenase (PGD), UDP-galactose-4-epimer-
ase (GALE), and glucose transport 1 (GLUT1).
28
PGM
1
is located at 1p31, and in the same area is
also located the leptin gene promoter (LEPR), which
has been implicated in obesity risk.
2931
Since PGM
1
and LEPR are nearly 2 millions of bp apart, a strong
linkage disequilibrium appears unlikely
32
but can-
not ruled out. Therefore, the possibility of a causal
role of LEPR cannot be excluded.
At present, there is great interest in elucidating
the molecular mechanisms, especially those con-
cerning the signal transduction pathways, involved
in gender differences concerning development and
body composition.
33
PGM
1
and Pak1 seem promising
candidate genes for studies in this area.
From a practical point of view, the identification of
factors involved in metabolism that show a genetic
variability associated to predisposition/resistance to
an increase of BMI may have clinical relevance for
prevention and cure of obesity disorders.
Acknowledgments
We thank Prof. James MacMurray for helpful
advice.
References
1. Webber J. Energy balance in obesity. Proc Nutr Soc 2003;
62:53943.
2. Loktionov A. Common gene polymorphisms and nutrition:
emerging links with pathogenesis of multifactorial chronic
diseases (review). J Nutr Biochem 2003;14:42651.
3. Speakman JR. Obesity, the integral roles of environment
and genetics. J Nutr 2004;134:2090S105S.
4. Cancello R, Tounian A, Poitou CH, et al. Adiposity sig-
nals, genetic and body weight regulation in humans. Diabetes
Metab 2004;30:21527.
5. Tremblay A, Doucet E. Obesity: a disease or a biological
adaptation? Obes Rev 2000;1:2735.
6. Ochoa Mdel C, Marti A, Martinez JA. Obesity studies in
candidate genes. Med Clin 2004;122:54251.
7. Snyder EE, Walts B, Perusse L, et al. The human obesity
gene map: the 2003 update. Obes Res 2004;12:369439.
8. Bottini E, Gerlini G, Fallucca F, et al. Diabetic pregnancy:
fetal development and maternal phosphoglucomutase pheno-
type. Dis Markers 1985;3:839.
9. Gloria-Bottini F, Lucarini N, Bonci E, et al. Genetic
polymorphism and intrauterine development: the role of ma-
ternal PGM
1
and MNSs genotypes. Am J Dis Child 1986;140:
3246.
10. Gloria-Bottini F, Lucarini N, La Torre M, et al. Birth
weight and parental PGM
1
alleles. Am J Hum Biol 2001;13:
41720.
11. Spencer N, Hopkinson DA, Harris H. Phosphoglu-
comutase polymorphism in man. Nature 1964;21:7425.
12. Hopkinson DA, Harris H. A third phosphoglucomutase
locus in man. Ann Hum Genet 1968;31:35967.
13. McAlpine PJ, Hopkinson DA, Harris H. The relative activ-
ities attributable in three phosphoglucomutase loci (PGM1,
PGM2, PGM3) in human tissues. Ann HumGenet 1970;34:169
75.
14. Cantu` JN, Ibarra B. Phosphoglucomutase evidence for a new
locus expressed in human milk. Science 1982;216:63940.
15. Scacchi R, Corbo RM, Palmarino R, et al. Human phos-
phoglucomutase locus 1: red cell enzymatic activities associ-
ated with common isoelectric focusing phenotypes. Hum
Hered 1983;33:21822.
16. Guguraj A, Barnes CJ, Vadlamudi RK, et al. Regulation
of phosphoglucomutase 1 phosphorylation and activity by a
signalling kinase. Oncogene 2004;23:811827.
17. Bark JE, Harris MJ, Firth M. Typing of the common
phosphoglucomutase variants using isoelectric focusing- a
new interpretation of the phosphoglucomutase system. J Fo-
rensic Sci Soc 1976;16:11520.
18. Kuhnl P, Schmidtmann U, Spielmann W. Evidence for
two additional common alleles at the PGM1 locus (phospho-
glucomutase, E.C.2.7.5.1): a comparison by three different
techniques. Hum Genet 1977;35:21923.
19. Sutton JG, Burgess R. Genetic evidence for four common
alleles at the phosphoglucomutase 1 locus (PGM1) detectable
by isoelectric focusing. Vox Sang 1978;34:97103.
20. Scozzari R, Trippa G, Santachiara-Benerecetti AS, et al.
Further genetic heterogeneity of human red cell phosphoglu-
comutase 1: a nonelectrophoretic polymorphism. Ann Hum
Genet 1981;45:31322.
21. Whitehouse DB, Putt W, Lovegrove JU, et al. Phospho-
glucomutase 1: complete human and rabbit mRNA sequences
and direct mapping on this highly polymorphic marker on
human chromosome 1. Proc Natl Acad Sci U S A 1992;89:
4115.
22. March RE, Putt W, Hollyoake M, et al. The classical
human phosphoglucomutase (PGM1) isozyme polymorphism
is generated by intragenic recombination. Proc Natl Acad Sci
U S A 1993;90:107303.
23. Takahashi N, Neel JV. Intragenic recombination at the
human phosphoglucomutase 1 locus: predictions fulfilled.
Proc Natl Acad Sci U S A 1993;90:107259.
24. Yip SP, Lovegrove JU, Rana NA, et al. Mapping recom-
bination hotspots in human phosphoglucomutase (PGM1).
Hum Mol Genet 1999;8:1699706.
PGM
1
and Body Mass
424 December 2007 Volume 334 Number 6
25. Panero C, Romano S, Cianciulli D, et al. Auxometric
parameters and gestational age in 9751 newborns in Flo-
rence. J Foet Med 1983;3:95103.
26. SPSS/PC Version 5.0. Chicago, SPSS Inc, 1992.
27. Sokal RR, Rohlf FJ. Biometry. New York, Freeman,
1981.
28. McKusick VA. Mendelian inheritance in man. 11th edition.
Baltimore, MD, Johns Hopkins University Press, 1994.
29. Portoles O, Sorli JV, Frances F, et al. Effect of genetic
variation in the leptin gene promoter and the leptin receptor
gene in obesity risk in a population-based case-control study
in Spain. Eur J Epidemiol 2006;21:60512.
30. Park KS, Shin HD, Park BL, et al. Polymorphisms in the
leptin receptor (LEPR): putative association with obesity and
T2DM. J Hum Genet 2006;51:8591.
31. Duarte SF, Francischetti EA, Genelhu-Abreu V, et al.
p.Q223R leptin receptor polymorphism associated with obe-
sity in Brazilian multiethnic subjects. Am J Hum Biol 2006;
18:44853.
32. Dawson E, Abecasis GR, Bumpstead S, et al. A first-
generation linkage disequilibrium map of human chromo-
some 22. Nature 2002;418:5448.
33. Whitacre CC, Reingold SC, OLooney PA. A gender map
in autoimmunity. Science 1999;283:12778.
Gloria-Bottini et al
425 THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES