MafA Is a Key Regulator of Glucose-Stimulated Insulin Secretion

Abstract

MafA is a transcription factor that binds to the promoter in the insulin gene and has been postulated to regulate insulin
transcription in response to serum glucose levels, but there is no current in vivo evidence to support this hypothesis. To
analyze the role of MafA in insulin transcription and glucose homeostasis in vivo, we generated MafA-deficient mice. Here
we report that MafA mutant mice display intolerance to glucose and develop diabetes mellitus. Detailed analyses revealed that glucose-, arginine-,
or KCl-stimulated insulin secretion from pancreatic β cells is severely impaired, although insulin content per se is not significantly
affected. MafA-deficient mice also display age-dependent pancreatic islet abnormalities. Further analysis revealed that insulin
1, insulin 2, Pdx1, Beta2, and Glut-2 transcripts are diminished in MafA-deficient mice. These results show that MafA is a
key regulator of glucose-stimulated insulin secretion in vivo.

Full text

MOLECULAR AND CELLULAR BIOLOGY, June 2005, p. 4969–4976 Vol. 25, No. 12
0270-7306/05/$08.00⫹0 doi:10.1128/MCB.25.12.4969–4976.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
MafA Is a Key Regulator of Glucose-Stimulated Insulin Secretion
Chuan Zhang,
1
Takashi Moriguchi,
1,2
Miwako Kajihara,
1
Ritsuko Esaki,
1
Ayako Harada,
1
Homare Shimohata,
1
Hisashi Oishi,
1
Michito Hamada,
1
Naoki Morito,
1
Kazuteru Hasegawa,
1
Takashi Kudo,
1
James Douglas Engel,
2
Masayuki Yamamoto,
3
and Satoru Takahashi
1
*
Institute of Basic Medical Sciences and Laboratory Animal Resource Center, University of Tsukuba,
1-1-1 Tennodai, Tsukuba 305-8575, Japan
1
; University of Michigan Medical School, Ann Arbor,
Michigan 48109-0616
2
; and Center for Tsukuba Advanced Research Alliance,
University of Tsukuba, 1-1-1 Tennodai, Tsukuba 305-8577, Japan
3
Received 1 October 2004/Returned for modification 9 December 2004/Accepted 22 March 2005
MafA is a transcription factor that binds to the promoter in the insulin gene and has been postulated to
regulate insulin transcription in response to serum glucose levels, but there is no current in vivo evidence to
support this hypothesis. To analyze the role of MafA in insulin transcription and glucose homeostasis in vivo,
we generated MafA-deficient mice. Here we report that MafA mutant mice display intolerance to glucose and
develop diabetes mellitus. Detailed analyses revealed that glucose-, arginine-, or KCl-stimulated insulin
secretion from pancreatic ␤ cells is severely impaired, although insulin content per se is not significantly
affected. MafA-deficient mice also display age-dependent pancreatic islet abnormalities. Further analysis
revealed that insulin 1, insulin 2, Pdx1, Beta2, and Glut-2 transcripts are diminished in MafA-deficient mice.
These results show that MafA is a key regulator of glucose-stimulated insulin secretion in vivo.
Insulin is the only polypeptide hormone that is essential for
the regulation of blood glucose levels and is synthesized exclu-
sively in ␤ cells of the islets of Langerhans in the pancreas. The
molecular mechanisms that control ␤-cell-specific insulin gene
transcription are well characterized. Three conserved cis-reg-
ulatory elements within the promoter, E1, A3, and RIPE3b/
C1, respectively, appear to be indispensable for proper insulin
gene regulation (22, 25). Islet-restricted transcription factors
Beta2/NeuroD and Pdx1 bind to the E1 and A3 elements in
vitro. Gene disruption experiments in mice have revealed that
both Beta2 and Pdx1 play critical roles in insulin gene regula-
tion as well as in islet development and function (1, 8, 21).
Furthermore, mutations in both the Beta2 and Pdx1 genes
have been identified within populations of patients with type II
diabetes (18, 29, 30).
The third regulatory element, RIPE3b/C1, has also been
shown to play a critical role in ␤-cell-specific insulin gene
transcription as well as in glucose-regulated expression. Previ-
ous studies identified a pancreatic ␤-cell-restricted factor,
called the RIPE3b1 activator, that is enriched in response to
glucose in pancreatic ␤-cell nuclear extracts. Very recently,
four groups reported that the RIPE3b1 activator is a member
of the Maf family of transcription factors, MafA (10, 12, 20,
26). The large Maf proteins, MafA/L-Maf/SMaf1 (2, 9, 24),
MafB (11), c-Maf (23), and Nrl (31), each contain a basic motif
followed by a leucine zipper, and all four family members
harbor acidic domains that act as transcriptional activation
domains. Although a role for MafA in insulin gene regulation
was hypothesized, in vivo tests of the hypothesis have not been
reported. To elucidate MafA function in insulin gene regula-
tion, we generated MafA-deficient mice.
MATERIALS AND METHODS
Targeted disruption of the mafA gene. mafA genomic clones were isolated
from a 129/SvJ genomic library (Stratagene) using a partial mouse MafA cDNA
as a probe. The targeting vector was constructed with the bacterial lacZ gene
containing a nuclear localization signal (NLS), and a neomycin resistance (neo
R
)
cassette. NLS-lacZ-neo
R
genes were inserted into the mafA open reading frame
between the Eco47III/EcoRI and EcoRV/Eco47III sites in order to delete both
the transactivation and basic motif-leucine zipper domains (Fig. 1A). The diph-
theria toxin A gene was inserted at the 3⬘ end of short arm of the targeting vector
for negative selection. Chimeras generated from two correctly targeted embry-
onic stem (ES) cell clones were bred to ICR mice with resultant germ line
transmission.
Mouse genotyping by PCR/Southern blot analysis. Genotyping was performed
on genomic DNA isolated from ES cells or tails by PCR or Southern blotting.
The 5⬘ and 3⬘ primers for the mutant allele (⬃700 bp amplified) were 5⬘-ATG
CGAAGTGGACCTGGGACCGCGCCGC-3⬘ and 5⬘-CTGCGCTGGCGAGG
GCTCCCGAGGGAAG-3⬘ under the following conditions: 30 cycles of 98°C for
10 s, 71°C for 30 s, and 72°C for 30 s, followed by 1 cycle of 72°C for 7 min. The
5⬘ and 3⬘ primers for the mafA gene were 5⬘-GAGGCCTTCCGGGGTCAGA
GCTTCGCGG-3⬘ and 5⬘-TCTGTTTCAGTCGGATGACCTCCTCCTTGC-3⬘
under the following conditions: 1 cycle of 94°C for 3 min and 30 cycles of 98°C
for 10 s, 71°C for 30 s, and 72°C for 30 s, followed by 1 cycle of 72°C for 7 min,
which results in an ⬃400-bp-amplified fragment for the wild-type mafA allele.
For mafA Southern blot analysis, DNA was digested with NheI and hybridized
with probe.
Quantitative transcript analysis in pancreatic islets by competitive RT-PCR.
Competitive reverse transcription-PCR (RT-PCR) analysis was performed on
total RNA prepared from isolated adult pancreatic islets essentially as previously
described (15). Competitor DNA plasmids carrying a small deletion within the
respective cDNAs were constructed by appropriate restriction endonuclease
digestion, as shown in Table 1. Each PCR product was electrophoresed in a 2%
agarose gel and visualized by ethidium bromide staining. The intensity of the
amplified fragment was quantified using an NIH image system. To ascertain the
efficiency of cDNA preparation from total RNAs, the competitive RT-PCR
analysis of hypoxanthine phosphoribosyltransferase (HPRT) transcript was per-
formed in each sample as the internal control. Reactions were plotted on indi-
vidual standard curves to derive the actual quantity of individual transcripts.
* Corresponding author. Mailing address: Institute of Basic Medical
Sciences and Laboratory Animal Resource Center, University of
Tsukuba, 1-1-1 Tennodai, Tsukuba 305-8575, Japan. Phone: 81-298-
53-7516. Fax: 81-298-53-6965. E-mail: satoruta@md.tsukuba.ac.jp.
4969

Histological analysis. Immunohistochemistry and light microscopy were per-
formed as follows. Pancreata were dissected, weighed, fixed overnight in 4%
paraformaldehyde, and then incubated overnight in 30% sucrose. Frozen sec-
tions were mounted on slides. For the quantification of the endocrine mass and
determination of the ␤-cell/␣-cell ratio, sections were immunostained with both
guinea pig anti-insulin (Linco) and rabbit antiglucagon (DAKO) antibodies.
Detection was performed using fluorescein secondary antibodies (Cortex Bio-
chem and ZYMED). Sections were incubated for 5 min with 0.01% Hoechst
stain to reveal nuclei. All islets in the sections were photographed, and analyzed
with Adobe Photoshop software (Adobe System Inc.). X-Gal (5-bromo-4-chloro-
3-indolyl-␤-
D-galactopyranoside) staining was performed as previously described
(32).
Glucose tolerance test and insulin release. Mice were fasted for 12 h and then
injected intraperitoneally (i.p.) with glucose (2 g/kg of body weight). Venous
blood was obtained from the retro-orbital plexus at 0, 15, 30, 60, and 120 min
after the injection. Plasma glucose levels were measured using a Fuji Drichem
3500 (Fuji-Film, Tokyo, Japan). For insulin release, glucose (3 g/kg of body
weight) or
L-arginine (1 g/kg of body weight) was injected i.p. and venous blood
was collected at 0, 2, 5, and 15 min in heparinized tubes. Pancreatic insulin was
extracted by the acid-ethanol method as described previously (7). Serum insulin
levels and insulin contents of the pancreata were measured with an Ultra-
sensitive insulin enzyme-linked immunosorbent assay kit (Morinaga Bioscience,
Yokohama, Japan).
Islet isolation and insulin release. To obtain pancreatic islets, pancreata were
removed and islets were isolated by collagenase digestion using the protocol
described in reference 16. The islets were individually dissected under a ste-
reomicroscope. Batches of 10 islets of similar size were collected and incubated
in RPMI 1640–10% fetal calf serum at 37°C in 5% CO
2
for 2 h. These islets were
washed and preincubated in 0.5% (wt/vol) bovine serum albumin–Krebs-Ringer
HEPES-buffered saline in 2.8 mM glucose at 37°C in 5% CO
2
for 30 min and
then transferred to 0.5% (wt/vol) bovine serum albumin–Krebs-Ringer HEPES-
buffered saline in 2.8 mM glucose, stimulatory 20 mM glucose alone, or 30 mM
KCl at 2.8 mM glucose. After incubation at 37°C in 5% CO
2
for 30 min, the
supernatants were measured for insulin release as described above.
RESULTS
MafA is expressed in pancreatic ␤ cells. A positive or neg-
ative targeting vector was constructed in which the bacterial
FIG. 1. Targeting strategy for mafA mutagenesis and LacZ expression in pancreatic islets in MafA mutant mice. (A) Schematic representation
of the wild-type allele, the targeting construct, and the expected product of homologous recombination between them. (B) Southern blot of tail
DNA showing NheI-cleaved DNA fragments corresponding to the wild-type (11 kbp) and targeted (16 kbp) mafA alleles. (C) ␤-Galactosidase
expression (blue) in MafA
⫹/⫺
pancreatic ␤ cells. Immunostaining was performed with antiglucagon (green) and anti-insulin (red) antibodies.
Nuclear ␤-galactosidase staining overlaps that of insulin-expressing cells (merged).
4970 ZHANG ET AL. M
OL.CELL.BIOL.

␤-galactosidase gene containing a nuclear localization signal
and a phosphoglycerate kinase promoter-neomycin resistance
(PGK-neo
R
) cassette was used to replace MafA coding se
-
quences (Fig. 1A and B). Chimeras obtained from two cor-
rectly targeted ES cell clones were bred to ICR mice and
resulted in germ line transmission. ␤-Galactosidase expression
in these heterozygous mutant mice was detected in the lens,
somites, and olfactory bulb (data not shown). In addition,
strong X-Gal staining was observed in pancreatic islets. ␤-Ga-
lactosidase expression overlapped anti-insulin immunostain-
ing, but not antiglucagon staining, indicating that MafA is
expressed exclusively in pancreatic ␤ cells (Fig. 1C).
We intercrossed MafA
⫹/⫺
mice to determine the viability of
homozygous MafA mutant mice. Analysis of offspring from the
first line of heterozygous intercrosses revealed that MafA
⫺/⫺
mice were recovered, but the frequency was low (data not
shown). To confirm the results, we analyzed the offspring from
a second independent line of MafA
⫹/⫺
mutant intercrosses.
Analysis of 223 offspring from 26 litters demonstrated that 59
mice (26.5%) were homozygous mutant. Since the phenotypes
of MafA-deficient mice from the two lines were indistinguish-
able, we concluded that MafA deficiency does not confer em-
bryonic lethality. MafA
⫺/⫺
mice survived until adulthood.
MafA-deficient mice develop diabetes mellitus. Since it was
suggested that MafA is a primary candidate regulator for in-
sulin gene transcription, we screened the MafA
⫺/⫺
mutant
mice for blood glucose levels. Fasting blood glucose levels of
MafA
⫺/⫺
female mice were significantly higher than those of
wild-type female littermates at 4 weeks of age (wild-type glu-
cose level was 121 ⫾ 10.9 mg/dl, MafA
⫹/⫺
glucose level was 150
⫾ 5.1 mg/dl, and MafA
⫺/⫺
glucose level was 152 ⫾ 8.0 mg/dl,
respectively) (Fig. 2A), and at 8 weeks of age, both male and
female MafA
⫺/⫺
mice had high blood glucose levels (P ⬍
0.05). The differences in blood glucose became more signifi-
cant at 12 weeks postnatally. Although there was no significant
difference in the mean body weight of MafA
⫺/⫺
and wild-type
male mice, the mutant female animals displayed significant
growth retardation postnatally from 8 weeks onward (Fig. 2B).
We followed MafA
⫺/⫺
mice to determine whether or not
they developed overt diabetes mellitus. Under random feeding
conditions, the blood glucose levels of 5/11 male and 1/9 fe-
male MafA
⫺/⫺
mice were over 300 mg/dl at 12 weeks of age.
Five of 19 adult MafA
⫺/⫺
male mice became urinary glucose
positive by 20 weeks of age, and the mean blood glucose levels
of these mice reached ca. 600 mg/dl. By 50 weeks, 1/3 male and
1/5 female MafA
⫺/⫺
mice had over 500 mg/dl blood glucose.
Our preliminary results suggest that there was no significant
difference in the rate of survival between MafA
⫺/⫺
and wild-
type mice until 50 weeks. These data show that the MafA-
deficient mouse is a new model for overt diabetes mellitus.
Glucose-stimulated insulin secretion (GSIS) is impaired in
MafA-deficient mice. To elucidate developmental mechanisms
that might contribute to diabetes mellitus, we tested the glu-
cose tolerance of 8-week-old MafA
⫺/⫺
mice by an i.p. glucose
tolerance test (ipGTT). Blood glucose levels were rapidly in-
duced in MafA
⫺/⫺
mutants after i.p. glucose injection and were
sustained at high levels for at least 2 h following glucose ad-
ministration. Plasma glucose levels of MafA-deficient mice
were 521 ⫾ 31.7 mg/dl, while the level in wild-type mice was
243 ⫾ 16.9 mg/dl 2 h after glucose injection (P ⬍ 0.01). Fur-
thermore, MafA heterozygous mutant mice also had signifi-
cantly higher blood glucose levels (417 ⫾ 29.8 mg/dl) than
wild-type littermates, demonstrating a MafA haploinsufficiency
in pancreatic function (Fig. 3A).
To elucidate the mechanism of this impaired glucose toler-
ance, we analyzed plasma insulin levels in the ipGTT mice.
TABLE 1. Oligonucleotide primers and sizes of PCR products for competitive RT-PCR
Gene product Primer
a
Size of PCR products (bp)
Restriction enzyme(s) for construction
of competitor DNA
Target Competitor
Insulin 1 5⬘-CCAGCTATAATCAGAGACCA-3⬘ 377 328 SmaI-BsgI
5⬘-GGGCCTTAGTTGCAGTAGTT-3⬘
Insulin 2 5⬘-AGGAAGCCTATCTTCCAGGT-3⬘ 401 293 EcoNI-SmaI
5⬘-ATTCATTGCAGAGGGGTAGG-3⬘
Glucagon 5⬘-CTGGACAATCTTGCCACCAGGGAC-3⬘ 413 358 SphI-BstXI
5⬘-CCACTACGGTTACCAGGTGGTCATGTC-3⬘
Pdx1 5⬘-TCGCTGGGATCACTGGAGCA-3⬘ 365 247 BlpI-MluI
5⬘-CGGGAGATGTATTTGTTAAATAAGAATTC-3⬘
Beta2 5⬘-CTCCAGGGTTATGAGATCGTCAC-3⬘ 525 329 PpuMI
5⬘-GATCTCTGACAGAGCCCA-3⬘
Glut-2 5⬘-TGGGATGAAGAGGAGACTGAA-3⬘ 652 442 XcmI-BsrGI
5⬘-CATCCGTGAAGAGCTGGATCA-3⬘
Glucokinase 5⬘-GATGTATTCCATCCCCGAGGACG-3⬘ 410 291 StyI-MscI
5⬘-GCTCCACATTCTGCATCTCCTCC-3⬘
HPRT 5⬘-GGCTTCCTCCTCAGACCGCTTT-3⬘ 702 523 BclI-PpuMI
5⬘-AGGCTTTGTATTTGGCTTTTCC-3⬘
a
The sequence of the forward primer is given on the top, and that of the reverse primer is given on the bottom.
VOL. 25, 2005 MafA DEFICIENCY INDUCES DIABETES MELLITUS 4971

Plasma insulin levels in the MafA
⫺/⫺
mutants were significantly
lower than in wild-type littermates. Very interestingly, there
was almost no acute insulin secretion in response to the ele-
vated blood glucose levels in MafA
⫺/⫺
mice (Fig. 3B). Plasma
insulin levels of wild-type, MafA
⫹/⫺
, and MafA
⫺/⫺
at 2 min
after glucose injection were 1.51 ⫾ 0.25 ng/ml, 0.84 ⫾ 0.12
ng/ml, 0.51 ⫾ 0.08 ng/ml, respectively. We also measured se-
rum insulin levels after arginine stimulation to ascertain
whether or not insulin secretion from ␤ cells was affected.
Serum insulin levels in the MafA
⫺/⫺
mutants were significantly
lower than in the wild-type mice, and the animals displayed
almost no response to arginine administration (Fig. 3C). These
results indicated two possibilities: either dysfunction of the
insulin secretory machinery or decreased insulin content in ␤
cells. To clarify the reason for the impaired response, we mea-
sured the amounts of insulin in the pancreata of these animals.
The absolute insulin content in MafA
⫺/⫺
mutant pancreata was
comparable to that of wild-type mice (insulin contents in wild-
type, MafA
⫹/⫺
, and MafA
⫺/⫺
mice were 148 ⫾ 23.5 ␮g/g, 140
⫾ 7.8 ␮g/g, and 128 ⫾ 13.0 ␮g/g, respectively; Fig. 3D). These
results summarily demonstrate that MafA is a vital regulator of
glucose-stimulated insulin secretion but not of insulin produc-
tion.
Isolated MafA-deficient islets display abnormalities of
GSIS. To analyze the underlying defect in glucose-stimulated
insulin secretion in the MafA mutants, pancreatic islets from
each mouse were isolated and measured for insulin secretion
in response to glucose or KCl stimulation in vitro. Islets from
wild-type mice secreted insulin in response to glucose admin-
istration (2.66 ⫾ 0.46 ng/islet/h), but islets from the MafA-
deficient mice did not (1.12 ⫾ 0.19 ng/islet/h; P ⬍ 0.05). Islets
recovered from MafA heterozygous mice also displayed a im-
paired response to glucose (1.40 ⫾ 0.68 ng/islet/h; Fig. 4),
although it was not significant. In addition, KCl stimulation
had very little effect on insulin secretion in MafA-deficient
islets (P ⬍ 0.01 versus wild type; Fig. 4). The results clearly
indicate that MafA-deficient islets have a ␤-cell autonomous
defect in GSIS.
Abnormal architecture of pancreatic islets in adult MafA-
deficient mice. Since MafA
⫺/⫺
mutant mice displayed impaired
GSIS and developed diabetes mellitus, and since targeted dis-
ruption of other ␤-cell-affiliated transcription factors such as
Beta2, Pdx1, and Pax6 (28) cause disruption of islet cell devel-
opment, we next analyzed histologically the pancreatic islets of
MafA mutant mice. Using anti-insulin and antiglucagon anti-
bodies, the number and size of islets were measured in wild-
type, MafA
⫹/⫺
, and MafA
⫺/⫺
littermates at P1 and at 12 weeks
of age. There was no significant difference in the architecture
of pancreatic islets observed in equivalent cross-sections of the
islets at P1 (Fig. 5A). The ratio of ␤ to ␣ cells, as determined by
histologically assessing simultaneous expression of insulin and
glucagon, in the MafA
⫺/⫺
mouse pancreas was comparable to
that of wild-type mice at P1 (Fig. 5B). In contrast, a significant
difference was observed in the structure of islets and also in the
ratio of ␤ to ␣ cells in the MafA mutant and wild-type mice at 12
weeks of age (Fig. 5A and C), although there was no obvious
difference in the diameters of islets among these three genotyped
mice (Fig. 5D). These results indicate that MafA is dispensable
for embryonic pancreatic development but is indispensable for
the maintenance of adult pancreatic architecture and function.
Gene expression in MafA-deficient islets. In order to iden-
tify the underlying molecular mechanisms for impaired GSIS
and the abnormal architecture of adult pancreatic islets, we
next analyzed the levels of several other well-characterized
␤-cell effectors in the MafA
⫺/⫺
mutants (Fig. 6).
The quanti-
tative RT-PCR data using RNA recovered from isolated pan-
creatic islets of each genotype show that MafA mRNA is not
detected in MafA
⫺/⫺
mutant islets, and about half the amount
of MafA mRNA was detected in MafA
⫹/⫺
mutant islets, as
expected (data not shown). The amounts of insulin 1, insulin 2,
or Glut-2 mRNAs were reduced in the MafA
⫺/⫺
islets in com
-
parison to wild-type islets, while no difference was observed in
glucagon or glucokinase mRNA levels. Moreover, there was a
tendency toward reduced expression of Pdx1 or Beta2 mRNAs
in the MafA
⫺/⫺
islets as compared to wild type.
DISCUSSION
Here we report the generation and initial characterization of
MafA
⫺/⫺
mutant mice and show that they display abnormal
FIG. 2. Development of diabetes mellitus in MafA
⫺/⫺
mice.
(A) Fasting blood glucose levels of offspring derived from intercrosses
of MafA
⫹/⫺
mice were determined using a semiautomated analyzer.
Results represent the mean ⫾ standard error of the mean. (B) Mean
body weight ⫾ standard error of the mean of offspring derived from
intercrosses of MafA
⫹/⫺
mice at the indicated ages (weeks). Both sets
of data are from 7 to 22 animals of each genotype. * indicates P ⬍ 0.05,
while ** represents P ⬍ 0.01.
4972 ZHANG ET AL. M
OL.CELL.BIOL.

GSIS and adult islet structure, thus leading to pathological
development of diabetes mellitus. The data demonstrate that
transcription factor MafA is a crucial regulator of insulin se-
cretion and of islet structural maintenance but is not required
for ␤-cell development per se.
Three major observations on pancreatic function are re-
ported here. First, MafA-deficient mice display almost normal
insulin content in pancreata, although transcription of insulin 1
and insulin 2 is markedly reduced in MafA-deficient mice.
Since MafA has been identified as a transcription factor that
binds to a promoter element of the insulin gene and is thought
to regulate insulin transcription in response to serum glucose
levels (10, 12, 13, 19, 20, 26), we expected that insulin tran-
scription and insulin content would be diminished in MafA-
deficient mice. As expected, insulin 1 and insulin 2 transcrip-
tion is markedly reduced in MafA-deficient mice. These results
indicate that MafA is an important regulator of insulin tran-
scription in vivo, as well as in vitro. The reduction of Pdx1 and
Beta2 may also have a synergistic effect on diminished insulin
transcription. Alternatively, the insulin content of a MafA
⫺/⫺
pancreas is not significantly diminished in comparison to that
FIG. 3. Glucose tolerance and arginine tolerance tests and their effects on insulin production. (A) Glucose tolerance tests (ipGTT) after
intraperitoneal loading with 2 g D-glucose/kg were performed on 8-week-old male animals of the indicated genotypes following a 12-h fast. Each
symbol represents the following: *, P ⬍ 0.05, MafA
⫺/⫺
versus MafA
⫹/⫹
; **, P ⬍ 0.01, MafA
⫺/⫺
versus MafA
⫹/⫹
; ***, P ⬍ 0.01, MafA
⫺/⫺
versus
MafA
⫹/⫺
;#,P ⬍ 0.05, MafA
⫹/⫺
versus MafA
⫹/⫹
; ##, P ⬍ 0.01, MafA
⫹/⫺
versus MafA
⫹/⫹
. (B) Level of plasma insulin of each MafA genotype
during ipGTT. **, P ⬍ 0.01, MafA
⫺/⫺
versus MafA
⫹/⫹
; ***, P ⬍ 0.05, MafA
⫺/⫺
versus MafA
⫹/⫺
;#,P ⬍ 0.05, MafA
⫹/⫺
versus MafA
⫹/⫹
. (C) Level
of plasma insulin after intraperitoneal arginine administration of each MafA genotype. *, P ⬍ 0.05, MafA
⫺/⫺
versus MafA
⫹/⫹
. (D) Insulin content
of wild-type, MafA
⫹/⫺
, and MafA
⫺/⫺
mice. All data represent the mean values ⫾ standard error for at least five male mice (8 to 14 weeks of age)
of each genotype.
FIG. 4. Insulin secretion from isolated pancreatic islets in vitro.
Insulin secretion in response to the indicated secretagogues. Values
are expressed in nanograms of insulin islet
⫺1
h
⫺1
, as the mean ⫾
standard error of the mean of at least three male mice (8 to 12 weeks
of age) per genotype. * indicates P ⬍ 0.05, while ** represents P ⬍
0.01.
V
OL. 25, 2005 MafA DEFICIENCY INDUCES DIABETES MELLITUS 4973

in the wild-type sibling pancreas. This paradoxical observation
could be explained by either of two hypotheses. An abnormal-
ity of GSIS may be one possibility. Since GSIS of MafA-
deficient mice is impaired as we demonstrated here, the secre-
tion of insulin might be diminished, and thus the steady-state
insulin content in the MafA
⫺/⫺
pancreas is not significantly
affected. Another possibility is posttranscriptional regulation
of insulin synthesis. Leroux et al. reported that the amount of
insulin 2 protein in insulin 1-deficient mice is augmented as
compared with that of wild-type mice, even though the amount
of insulin 2 transcript is unchanged (17). These results indicate
the existence of posttranscriptional regulation of insulin syn-
thesis and suggest that the insulin content in MafA mutant
mice may be regulated by posttranscriptional mechanism.
The second major observation is that MafA deficiency had
no effect on embryonic development of pancreatic islets. This
is in striking contrast to the consequences of Pdx1 or Beta2
mutation, which are also known to be important for insulin
gene expression in vitro, and loss of these factors led to severe
islet development abnormalities in vivo. Pdx1 deficiency leads
to pancreatic agenesis, while Beta2-deficient mice display de-
velopmental arrest between embryonic day 14.5 (E14.5) and
FIG. 5. Histological analysis of pancreatic islets. (A) Insulin (red) and glucagon (green) immunoreactivity in wild-type (⫹/⫹) and MafA
homozygous mutant (⫺/⫺) mice at P1 or 12 weeks of age. Scale bar, 20 ␮m. (B and C) ␤-Cell/␣-cell ratio of the pancreatic islets from mice of
each genotype at P1 (B) and 12 weeks of age (C) (male mice). Pancreatic sections were double stained with anti-insulin and antiglucagon
antibodies. Data are the mean ␤-cell/␣-cell ratios ⫾ standard error of the mean for at least three mice of each genotype. * indicates P ⬍ 0.05, while
** represents P ⬍ 0.01. (D) Morphometric analysis of islet diameter in pancreata from wild-type (⫹/⫹), heterozygous (⫹/⫺), and MafA
homozygous mutant (⫺/⫺) 12-week-old male mice. Data represent the mean ratios ⫾ standard error of at least three male mice of each genotype.
4974 ZHANG ET AL. MOL.CELL.BIOL.

E17.5, a period characterized by a major expansion of the
␤-cell population (8, 21). Matsuoka et al. recently reported
that MafA is expressed initially in insulin-expressing cells at
E13.5 but is not detected in Nkx6.1-null mutant pancreata (19).
These results may indicate that MafA is hypostatic to Nkx6.1
during pancreatic islet development.
In MafA mutant adult mice, the ␣ cells are located inside of
the pancreatic islets and the islet structure becomes abnormal.
While we have not identified the specific molecular mechanism
leading to this aberrant structural anomaly, diminished Pdx1
expression is implicated, since Pdx1
⫹/⫺
mice display similar
structural defects in adult islets but not in newborn mice (1).
As abnormal islet structure is often accompanied by impaired
glucose-stimulated insulin secretion, as seen in Glut-2-null or
glucokinase-null mutant mice (5, 33), aberrant islet architec-
ture itself may cause moderate impairment of insulin secretion.
The third important observation reported here is that MafA
is a key regulator of GSIS in vivo. The data demonstrate that
MafA-deficient mice and islets are unable to respond to glu-
cose, arginine, or KCl administration. GSIS consists of two
stimulatory pathways, ionic and nonionic. Whereas the glu-
cose-induced ionic pathway (i.e., closure of K
⫹
ATP
channels,
membrane depolarization, activation of L-type voltage-depen-
dent Ca
2⫹
channels, Ca
2⫹
influx, elevation of cytosol-free
Ca
2⫹
) is the major signaling pathway in ␤-cell insulin secretion,
the nonionic glucose activity (termed K
⫹
ATP
channel-indepen
-
dent action of glucose) has significant physiological relevance.
The activation of a cyclic AMP-protein kinase A pathway in ␤
cells by GLP-1 augments Ca
2⫹
-stimulated insulin release but
also appears to enhance insulin secretion of a distal event,
beyond the elevation of Ca
2⫹
influx (14). Arginine or KCl, like
GLP-1, potentiates insulin secretion in the presence, but not in
the absence, of glucose. Arginine or KCl directly depolarizes
the ␤-cell membrane and thereby elicits Ca
2⫹
-dependent elec
-
trical activity, Ca
2⫹
entry, and insulin secretion. Thus, the
unresponsiveness of MafA-deficient mice or isolated islets to
glucose, arginine, or KCl stimulation indicates that both the
ionic and nonionic pathways are affected by MafA deficiency.
Since Pdx1 mRNA is diminished in MafA-deficient mice, Pdx1
could be the one of the causes of the observed unresponsive-
ness to glucose, since Pdx1
⫹/⫺
islets display abnormal response
to glucose and KCl accompanied with decreased protein levels
of Glut-2 and glucokinase (3). Samaras et al. also reported that
Pdx1 expression is regulated by MafA in ␤ cells (27). Accord-
ingly, we hypothesize that abnormal GSIS observed in MafA-
deficient mice may be partially explained by this down-regula-
tion of Pdx1. Further studies, especially the comparison of
gene expression profiles using a DNA microarray, must per-
formed to elucidate the target genes of MafA in the ␤ cells of
MafA-deficient mice.
Since other regulators of insulin gene expression (Pdx1 and
Beta2) are associated in some populations of patients with type
2 diabetes and mature onset diabetes of the young (4, 6, 18,
29), it will be of significant interest to determine whether
mutations in the human MafA gene are associated with disease
susceptibility. Finally, we suspect that these MafA
⫺/⫺
mutant
mice will serve as a very useful new model to develop novel
therapies for treating human diabetes mellitus.
ACKNOWLEDGMENTS
We would like to thank N. Minegishi, T. Yokomizo, M. Ema, H.
Shimano, M Ishikawa, and H. Sone (Tsukuba University) for helpful
discussions. We are grateful to A. Godo for excellent assistance.
This work was supported in part by the NIH (R01 CA80088), a
Grant-in-Aid from the Ministry of Education, Science, Sports and
Culture, and the Environmental Response Project of JST-ERATO.
REFERENCES
1. Ahlgren, U., J. Jonsson, L. Jonsson, K. Simu, and H. Edlund. 1998. ␤-Cell-
specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of the ␤-cell
phenotype and maturity onset diabetes. Genes Dev. 12:1763–1768.
2. Benkhelifa, S., S. Provot, O. Lecoq, C. Pouponnot, G. Calothy, and M. P.
Felder-Schmittbuhl. 1998. mafA, a novel member of the maf proto-onco-
gene family, displays developmental regulation and mitogenic capacity in
avian neuroretina cells. Oncogene 17:247–254.
3. Brissova, M., M. Shiota, W. E. Nicholson, M. Gannon, S. M. Knobel, D. W.
Piston, C. V. Wright, and A. C. Powers. 2002. Reduction in pancreatic
transcription factor PDX-1 impairs glucose-stimulated insulin secretion.
J. Biol. Chem. 277:11225–11232.
4. Furuta, H., Y. Horikawa, N. Iwasaki, M. Hara, L. Sussel, M. M. Le Beau,
E. M. Davis, M. Ogata, Y. Iwamoto, M. S. German, and G. I. Bell. 1998.
␤-Cell transcription factors and diabetes: mutations in the coding region of
the BETA2/NeuroD1 (NEUROD1) and Nkx2.2 (NKX2B) genes are not
associated with maturity-onset diabetes of the young in Japanese. Diabetes
47:1356–1358.
FIG. 6. Comparison of gene expression in pancreatic islets. For
quantitative analysis using competitive RT-PCR, pancreatic islets from
8-week-old male mice were used. The amount of each transcript was
normalized to the amount of HPRT transcript. Data represent the
mean ratios ⫾ standard error of three mice of each genotype.
V
OL. 25, 2005 MafA DEFICIENCY INDUCES DIABETES MELLITUS 4975

5. Guillam, M. T., E. Hummler, E. Schaerer, J. I. Yeh, M. J. Birnbaum, F.
Beermann, A. Schmidt, N. Deriaz, and B. Thorens. 1997. Early diabetes and
abnormal postnatal pancreatic islet development in mice lacking Glut-2. Nat.
Genet. 17:327–330.
6. Hansen, L., J. N. Jensen, S. Urioste, H. V. Petersen, F. Pociot, H. Eiberg,
O. P. Kristiansen, T. Hansen, P. Serup, J. Nerup, O. Pedersen et al. 2000.
NeuroD/BETA2 gene variability and diabetes: no associations to late-onset
type 2 diabetes but an A45 allele may represent a susceptibility marker for
type 1 diabetes among Danes. Diabetes 49:876–878.
7. im Walde, S. S., C. Dohle, P. Schott-Ohly, and H. Gleichmann. 2002. Mo-
lecular target structures in alloxan-induced diabetes in mice. Life Sci. 71:
1681–1694.
8. Jonsson, J., L. Carlsson, T. Edlund, and H. Edlund. 1994. Insulin-promoter-
factor 1 is required for pancreas development in mice. Nature 371:606–609.
9. Kajihara, M., S. Kawauchi, M. Kobayashi, H. Ogino, S. Takahashi, and K.
Yasuda. 2001. Isolation, characterization, and expression analysis of ze-
brafish large Mafs. J. Biochem. (Tokyo) 129:139–146.
10. Kajihara, M., H. Sone, M. Amemiya, Y. Katoh, M. Isogai, H. Shimano, N.
Yamada, and S. Takahashi. 2003. Mouse MafA, homologue of zebrafish
somite Maf 1, contributes to the specific transcriptional activity through the
insulin promoter. Biochem. Biophys. Res. Commun. 312:831–842.
11. Kataoka, K., K. T. Fujiwara, M. Noda, and M. Nishizawa. 1994. MafB, a new
Maf family transcription activator that can associate with Maf and Fos but
not with Jun. Mol. Cell. Biol. 14:7581–7591.
12. Kataoka, K., S. I. Han, S. Shioda, M. Hirai, M. Nishizawa, and H. Handa.
2002. MafA is a glucose-regulated and pancreatic ␤-cell-specific transcrip-
tional activator for the insulin gene. J. Biol. Chem. 277:49903–49910.
13. Kataoka, K., S. Shioda, K. Ando, K. Sakagami, H. Handa, and K. Yasuda.
2004. Differentially expressed Maf family transcription factors, c-Maf and
MafA, activate glucagon and insulin gene expression in pancreatic islet
alpha- and ␤-cells. J. Mol. Endocrinol. 32:9–20.
14. Komatsu, M., T. Schermerhorn, T. Aizawa, and G. W. Sharp. 1995. Glucose
stimulation of insulin release in the absence of extracellular Ca
2⫹
and in the
absence of any increase in intracellular Ca
2⫹
in rat pancreatic islets. Proc.
Natl. Acad. Sci. USA 92:10728–10732.
15. Kudo, T., Y. Ikehara, A. Togayachi, K. Morozumi, M. Watanabe, M. Naka-
mura, S. Nishihara, and H. Narimatsu. 1998. Up-regulation of a set of
glycosyltransferase genes in human colorectal cancer. Lab. Investig. 78:797–
811.
16. Lacy, P. E., and M. Kostianovsky. 1967. Method for the isolation of intact
islets of Langerhans from the rat pancreas. Diabetes 16:35–39.
17. Leroux, L., P. Desbois, L. Lamotte, B. Duvillie, N. Cordonnier, M. Jackerott,
J. Jami, D. Bucchini, and R. L. Joshi. 2001. Compensatory responses in mice
carrying a null mutation for Ins1 or Ins2. Diabetes 50(Suppl. 1):S150–S153.
18. Malecki, M. T., U. S. Jhala, A. Antonellis, L. Fields, A. Doria, T. Orban, M.
Saad, J. H. Warram, M. Montminy, and A. S. Krolewski. 1999. Mutations in
NEUROD1 are associated with the development of type 2 diabetes mellitus.
Nat. Genet. 23:323–328.
19. Matsuoka, T. A., I. Artner, E. Henderson, A. Means, M. Sander, and R.
Stein. 2004. The MafA transcription factor appears to be responsible for
tissue-specific expression of insulin. Proc. Natl. Acad. Sci. USA 101:2930–
2933.
20. Matsuoka, T.-A., L. Zhao, I. Artner, H. W. Jarrett, D. Friedman, A. Means,
and R. Stein. 2003. Members of the large Maf transcription family regulate
insulin gene transcription in islet ␤ cells. Mol. Cell. Biol. 23:6049–6062.
21. Naya, F. J., H. P. Huang, Y. Qiu, H. Mutoh, F. J. DeMayo, A. B. Leiter, and
M. J. Tsai. 1997. Diabetes, defective pancreatic morphogenesis, and abnor-
mal enteroendocrine differentiation in BETA2/neuroD-deficient mice.
Genes Dev. 11:2323–2334.
22. Naya, F. J., C. M. Stellrecht, and M. J. Tsai. 1995. Tissue-specific regulation
of the insulin gene by a novel basic helix-loop-helix transcription factor.
Genes Dev. 9:1009–1019.
23. Nishizawa, M., K. Kataoka, N. Goto, K. T. Fujiwara, and S. Kawai. 1989.
v-maf, a viral oncogene that encodes a “leucine zipper” motif. Proc. Natl.
Acad. Sci. USA 86:7711–7715.
24. Ogino, H., and K. Yasuda. 1998. Induction of lens differentiation by activa-
tion of a bZIP transcription factor, L-Maf. Science 280:115–118.
25. Ohlsson, H., K. Karlsson, and T. Edlund. 1993. IPF1, a homeodomain-
containing transactivator of the insulin gene. EMBO J. 12:4251–4259.
26. Olbrot, M., J. Rud, L. G. Moss, and A. Sharma. 2002. Identification of
␤-cell-specific insulin gene transcription factor RIPE3b1 as mammalian
MafA. Proc. Natl. Acad. Sci. USA 99:6737–6742.
27. Samaras, S. E., L. Zhao, A. Means, E. Henderson, T. A. Matsuoka, and R.
Stein. 2003. The islet ␤ cell-enriched RIPE3b1/Maf transcription factor
regulates pdx-1 expression. J. Biol. Chem. 278:12263–12270.
28. Sander, M., and M. S. German. 1997. The ␤ cell transcription factors and
development of the pancreas. J. Mol. Med. 75:327–340.
29. Stoffers, D. A., J. Ferrer, W. L. Clarke, and J. F. Habener. 1997. Early-onset
type-II diabetes mellitus (MODY4) linked to IPF1. Nat. Genet. 17:138–139.
30. Stoffers, D. A., N. T. Zinkin, V. Stanojevic, W. L. Clarke, and J. F. Habener.
1997. Pancreatic agenesis attributable to a single nucleotide deletion in the
human IPF1 gene coding sequence. Nat. Genet. 15:106–110.
31. Swaroop, A., J. Z. Xu, H. Pawar, A. Jackson, C. Skolnick, and N. Agarwal.
1992. A conserved retina-specific gene encodes a basic motif/leucine zipper
domain. Proc. Natl. Acad. Sci. USA 89:266–270.
32. Takahashi, S., K. Onodera, H. Motohashi, N. Suwabe, N. Hayashi, N. Yanai,
Y. Nabesima, and M. Yamamoto. 1997. Arrest in primitive erythroid cell
development caused by promoter-specific disruption of the GATA-1 gene.
J. Biol. Chem. 272:12611–12615.
33. Terauchi, Y., H. Sakura, K. Yasuda, K. Iwamoto, N. Takahashi, K. Ito, H.
Kasai, H. Suzuki, O. Ueda, N. Kamada, K. Jishage, K. Komeda, M. Noda, Y.
Kanazawa, S. Taniguchi, I. Miwa, Y. Akanuma, T. Kodama, Y. Yazaki, T.
Kadowaki. 1995. Pancreatic ␤-cell-specific targeted disruption of glucokinase
gene. Diabetes mellitus due to defective insulin secretion to glucose. J. Biol.
Chem. 270:30253–30256.
4976 ZHANG ET AL. MOL.CELL.BIOL.