Content uploaded by Kevin Gerrish
Author content
All content in this area was uploaded by Kevin Gerrish
Content may be subject to copyright.
Content uploaded by Kevin Gerrish
Author content
All content in this area was uploaded by Kevin Gerrish
Content may be subject to copyright.
MOLECULAR AND CELLULAR BIOLOGY, Dec. 2009, p. 6366–6379 Vol. 29, No. 24
0270-7306/09/$12.00 doi:10.1128/MCB.01259-09
Copyright © 2009, American Society for Microbiology. All Rights Reserved.
Transcription Factor Glis3, a Novel Critical Player in the Regulation of
Pancreatic -Cell Development and Insulin Gene Expression
䌤
Hong Soon Kang,
1
† Yong-Sik Kim,
1,2
† Gary ZeRuth,
1
Ju Youn Beak,
1
Kevin Gerrish,
3
Gamze Kilic,
4
Beatriz Sosa-Pineda,
4
Jan Jensen,
2
Julie Foley,
5
and Anton M. Jetten
1
*
Cell Biology Section, Division of Intramural Research,
1
and Microarray Lab Core,
3
National Institute of Environmental Health Sciences,
National Institutes of Health, Research Triangle Park, North Carolina 27709; Department of Stem Cell Biology and Regenerative
Medicine, Lerner Research Institute, Cleveland Clinic, Cleveland, Ohio 44195
2
; Department of Genetics and Tumor Cell Biology,
St. Jude Children’s Research Hospital, Memphis, Tennessee 38105
4
; and Laboratory of Experimental Pathology,
National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park,
North Carolina 27709
5
Received 17 September 2009/Returned for modification 22 September 2009/Accepted 25 September 2009
In this study, we report that the Kru¨ppel-like zinc finger transcription factor Gli-similar 3 (Glis3) is induced
during the secondary transition of pancreatic development, a stage of cell lineage specification and extensive
patterning, and that Glis3
zf/zf
mutant mice develop neonatal diabetes, evidenced by hyperglycemia and hypo-
insulinemia. The Glis3
zf/zf
mutant mouse pancreas shows a dramatic loss of  and ␦ cells, contrasting a smaller
relative loss of ␣, PP, and cells. In addition, Glis3
zf/zf
mutant mice develop ductal cysts, while no significant
changes were observed in acini. Gene expression profiling and immunofluorescent staining demonstrated that
the expression of pancreatic hormones and several transcription factors important in endocrine cell develop-
ment, including Ngn3, MafA, and Pdx1, were significantly decreased in the developing pancreata of Glis3
zf/zf
mutant mice. The population of pancreatic progenitors appears not to be greatly affected in Glis3
zf/zf
mutant
mice; however, the number of neurogenin 3 (Ngn3)-positive endocrine cell progenitors is significantly reduced.
Our study indicates that Glis3 plays a key role in cell lineage specification, particularly in the development of
mature pancreatic  cells. In addition, we provide evidence that Glis3 regulates insulin gene expression
through two Glis-binding sites in its proximal promoter, indicating that Glis3 also regulates -cell function.
Proteins Glis1 to -3 constitute a subfamily of Kru¨ppel-like
zinc finger transcriptional regulators that share a highly con-
served five-C
2
H
2
-type zinc finger domain with members of the
Gli and Zic subfamilies (6, 25, 27, 30–32, 35, 38, 39, 43, 56).
Glis1 to -3 regulate gene transcription by binding specific DNA
sequences referred to as Glis-binding sites (Glis-BS) in pro-
moter regulatory regions of target genes (10, 30, 31). Although
their precise physiological functions are still poorly under-
stood, genetic studies have implicated Glis1 to -3 in several
pathologies (7, 8, 24, 29, 33, 39, 45).
Glis3 is abundantly expressed in the adult kidney, pituitary,
pancreas, uterus, and thyroid gland (31, 45). During mouse
embryonic development, Glis3 is expressed in a spatiotemporal
pattern suggesting that Glis3 regulates gene expression at spe-
cific stages during development (31). Genetic alterations in the
human GLIS3 gene have been linked to a rare syndrome char-
acterized by neonatal diabetes and congenital hypothyroidism
(NDH) (45, 50). Depending on the nature of the GLIS3 mu-
tation, NDH patients can also display facial abnormalities,
glaucoma, liver fibrosis, and polycystic kidney disease. Re-
cently, a genome-wide association study identified the GLIS3
gene as a susceptibility locus for type 1 diabetes (8). These
studies, together with evidence that Glis3 is expressed in pan-
creatic  cells, suggest that Glis3 has an important regulatory
role in the pancreas.
Although major advances have been made in understanding
pancreatic development, many of the molecular mechanisms
that regulate progenitor cell dynamics and cell differentiation
are still not precisely understood (1, 18, 19, 21, 26, 28, 37). At
approximately embryonic day 9 (E9) of mouse embryogenesis,
the pancreas first appears from distinct ventral and dorsal
anlagen as evaginations of the distal foregut endoderm (21,
36). The buds grow and initiate branching morphogenesis at
about E11.5. Early multipotent pancreatic progenitors, marked
by Pdx1, Ptf1a, Nkx2.2, and Cpa1 expression (12, 14, 41, 57),
are the source of all differentiated cells of the exocrine, ductal,
and endocrine cell lineages. Lineage determination is a com-
plex process that involves many transcription factors and sig-
naling pathways. Induction of Ngn3 marks the differentiation
of pancreatic progenitors into proendocrine progenitors (15,
17, 22, 36, 51). Differentiation into the different endocrine cell
lineages involves the induction of a combination of additional
transcription factors, including Myt1, NeuroD, Isl1, Pax4,
Pax6, and Arx (2, 13, 22, 34, 36, 37, 48, 49). Defects in the
expression or activity of these transcription factors in mice and
humans often result in abnormal pancreatic development and
function that can lead to diabetes.
To obtain greater insights into the physiological and molec-
ular functions of Glis3, we recently generated Glis3
zf/zf
mutant
mice that are deficient in Glis3 transactivating activity (24). In
this study, we characterize the pancreatic phenotype of these
mice and analyze the role of Glis3 in pancreatic development.
We demonstrate that, in addition to cyst formation in the
* Corresponding author. Mailing address: National Institute of En-
vironmental Health Sciences, National Institutes of Health, 111 T. W.
Alexander Drive, Research Triangle Park, NC 27709. Phone: (919)
541-2768. Fax: (919) 541-4133. E-mail: jetten@niehs.nih.gov.
† These authors contributed equally to this work.
䌤
Published ahead of print on 5 October 2009.
6366
pancreatic ducts, Glis3
zf/zf
mutant mice develop neonatal dia
-
betes that is associated with an almost total loss of  cells. We
provide evidence which indicates that Glis3 plays a key role in
cell lineage specification, particularly in the development of
mature pancreatic  cells. We further identify Glis3 as a reg-
ulator of insulin 2 gene expression. Our study shows that Glis3
has multiple functions in the pancreas and suggests that Glis3
might provide a new therapeutic target to intervene in diabetes.
MATERIALS AND METHODS
Glis3
zf/zf
mutant mice. Glis3
zf/zf
mutant mice, in which a 3.5-kb region that
includes exon 4 and parts of introns 3 and 4 was deleted, were described previ-
ously (24). This deletion results in the removal of the fifth zinc finger motif (ZF5)
of Glis3, disruption of its DNA-binding capacity, and impairment of Glis3 tran-
scriptional activation function (9, 24). All of our animal studies followed the
guidelines outlined in the NIH Guide for the Care and Use of Laboratory
Animals, and the protocols were approved by the Institutional Animal Care and
Use Committee at the National Institute of Environmental Health Sciences
(NIEHS).
Histopathological evaluation. Histopathological evaluation of male and fe-
male wild-type (WT) and Glis3
zf/⫹
and Glis3
zf/zf
mutant mice was performed on
postnatal day 1 (PND1) and PND3. Tissues were collected and fixed by immer-
sion in 10% neutral buffered formalin or 4% paraformaldehyde–phosphate-
buffered saline for 24 h, subsequently embedded in paraffin, and sectioned at 5
m. The sections were stained with hematoxylin and eosin for histopathological
evaluation.
Immunohistochemical analysis. Formalin-fixed/paraffin-embedded or cryo-
genic sections were used for immunohistochemistry or immunofluorescence
staining. For cryosections, embryos were fixed overnight at 4°C in 4% parafor-
maldehyde, transferred to 30% sucrose for 30 h, and subsequently embedded in
OCT (Tissue-Tek, Hatfield, PA). Antibodies against insulin, cadherin 1 (E-
cadherin), ␣-acetylated tubulin, and amylase were purchased from Sigma (St.
Louis, MO). Fluorescein isothiocyanate-labeled Dolichos biflorus agglutinin
(DBA) was obtained from Vector Laboratories (Burlingame, CA). Antibodies
against glucagon, somatostatin, and pancreatic polypeptide were from Dako
(Carpinteria, CA). Rabbit anti-Hnf6 was from Santa Cruz Biotech (Santa Cruz,
CA), rabbit anti-osteopontin was from R&D Systems (Minneapolis, MN), rabbit
anti-Sox9 and anti-Glut2 were from Chemicon (Billerica, MA), and rabbit anti-
MafA was from Bethyl Laboratories (Montgomery, TX). Mouse anti-Nkx6.1
(AB2022) and rabbit anti-Ngn3 (44) antibodies were kindly provided by Chris-
topher Newgard, Duke University, and Michael German, University of Califor-
nia, San Francisco, respectively. Rabbit and guinea pig anti-Pdx1 antibodies were
from Abcam (Cambridge, MA), and rabbit anti-Cpa1 was from Serotec (Raleigh,
NC). Alexa Fluor-conjugated secondary antibodies were purchased from Molec-
ular Probes (Carlsbad, CA). Fluorescence was observed in a Leica DMRBE
microscope (Leica, Wetzlar, Germany).
In situ hybridization. In situ hybridization was performed with cryosections of
pancreas and embryos as described previously (31). The Glis3 probe, encoding
the region from Q204 to S507, was generated by PCR and subsequently cloned
into pGEM Teasy (Promega, Madison, WI). Plasmid DNA from forward and
reverse clones was then linearized with SpeI to generate T7-generated sense and
antisense transcripts, respectively, and labeled riboprobe produced with digoxi-
genin-substituted UTP (Roche, Indianapolis, IN). Sense probes did not show any
signal (data not shown).
Glucose and insulin assays. Blood and urine were collected from PND3 pups.
Glucose levels were measured with Multistix 10 SG from Bayer. Blood insulin
levels were measured with an enzyme-linked immunosorbent assay kit from
LINCO Research (St. Charles, MO) according to the manufacturer’s instruc-
tions.
Microarray analysis. Microarray analyses were carried out by the NIEHS
Microarray Group with Agilent whole-genome mouse oligonucleotide arrays
(Agilent Technologies, Palo Alto, CA) as described previously (23). Total RNA
was isolated from the pancreata of PND3 and E15.5 WT and Glis3
zf/zf
mutant
mice with a Qiagen RNeasy Mini kit and subsequently amplified in accordance
with the Agilent Low RNA Input Fluorescent Linear Amplification kit protocol.
RNAs from three or four individual mice of each strain were analyzed in dupli-
cate. Hybridizations were performed as described previously. Data were ob-
tained with the Agilent Feature Extraction software (v7.1) by using defaults for
all parameters. Images and GEML files, including error and P values, were
exported from the Agilent Feature Extraction software and deposited into Ro-
setta Resolver (version 3.2, build 3.2.2.0.33; Rosetta Biosoftware, Kirkland,
WA). The resultant ratio profiles were combined into ratio experiments as
described previously (23). Intensity plots were generated for each ratio experi-
ment, and genes were considered “signature genes” if the P value was less than
0.001.
Real-time QRT-PCR analysis. Quantitative reverse transcription-PCRs
(QRT-PCRs) were carried out in triplicate in a 7300 Real Time PCR system
(Applied Biosystems, Foster City, CA) as previously described (23). All results
were normalized to an internal control, either the 18S or the glyceraldehyde-3-
phosphate dehydrogenase transcript. Primers were designed with Primer Express
2.0 software and synthesized at Sigma/Genosys (St. Louis, MO). The primers and
probes used in this study are shown in Table 1.
Generation of reporter and expression constructs. To generate mIns2 and
hINS reporter constructs, the indicated regions of either the mouse insulin 2 or
human insulin promoter were amplified by PCR from genomic DNA (Promega)
and ligated into the pGL4.10 luciferase reporter vector (Promega) to generate
mIns2(⫺696)-Luc or hINS(⫺700)-Luc. Site-directed mutagenesis of the two
putative Glis-binding sites within the mouse Ins2 promoter was carried out with
the QuikChange site-directed mutagenesis kit (Stratagene) in accordance with
the manufacturer’s protocol. In the following sequences, the mutated bases are
underlined: Glis-BS(⫺263), 5⬘-GGAACAATGTCTTCTGCTGTGAAC; Glis-
BS(⫺99), 5⬘-CTGC-TGACCTACTTCACCTGGAGCCC. The Glis3 expression
vector p3xFlag-CMV-Glis3 was described previously (9). Plasmids p3xFlag-
CMV-Glis3⌬N302 and p3xFlag-CMV-Glis3⌬C748 were generated by PCR am-
plification of the respective fragments and insertion into the p3xFlag-CMV10
expression vector (Sigma). All constructs were verified by restriction enzyme
analysis and DNA sequencing.
Electrophoretic mobility shift assay (EMSA). Binding of Glis3 to
32
P-labeled
Glis-BS oligonucleotides was performed as described previously (9). In addition
to the Glis-BS consensus sequence, binding of Glis3 to the putative Glis-BS in
the mouse Ins2 proximal promoter region at ⫺84 to ⫺109 (Glis-BS1, 5⬘-CTGC
TGACCTACCCCACCTGGAGCCC), ⫺253 to ⫺276 (Glis-BS2, 5⬘-GGAACA
ATGTCCCCTGCTGTGAAC), and their mutant oligonucleotides (shown
above) was examined.
Cells. Rat insulinoma INS-1 (832/13) cells were a generous gift from H.
Hohmeier (Duke University) and were maintained in RPMI 1640 medium sup-
plemented with 10% fetal calf serum, 10 mM HEPES, 2 mM glutamine, 1 mM
sodium pyruvate, 100 U/ml penicillin, 100 g/ml streptomycin, and 50 M
-mercaptoethanol.
Reporter assays. Cells were plated in 12-well dishes at 1 ⫻ 10
5
/well, incubated
for 24 h at 37°C, and subsequently cotransfected with 1 g of the indicated
reporter, 0.3 g pCMV--galactosidase, and 0.5 g of the indicated expression
vector in Opti-MEM (Gibco) with Lipofectamine 2000 (Invitrogen) in accor-
dance with the manufacturer’s instructions. Each transfection cocktail was used
to transfect triplicate sets of wells. Cells were harvested 24 h later, and reporter
activity was measured with a luciferase assay kit (Promega). -Galactosidase
levels were measured with a luminometric -galactosidase detection kit (Clon-
tech) by following the manufacturer’s protocol. Each data point was assayed in
triplicate, and each experiment was performed at least twice. Luciferase data
were normalized to -galactosidase activity and are presented as the mean ⫾ the
standard error of the mean. Statistical significance was determined by analysis of
variance and Tukey-Kramer comparison tests with InStat software (GraphPad
Software Inc.).
Microarray data accession number. The microarray data discussed in this
study have been deposited in the NCBI Gene Expression Omnibus as GSE18172
(GEO, http://www.ncbi.nlm.nih.gov/geo/).
RESULTS
Glis3 is expressed during pancreatic development. Recent
human genetic studies implicated GLIS3 in a syndrome with
NDH and suggested a regulatory function for GLIS3 in the
pancreas (8, 45, 50). To gain greater insights into the role of
Glis3 in the pancreas and neonatal diabetes, we analyzed Glis3
expression during early mouse pancreatic development. QRT-
PCR analysis of Glis3 expression during E11.5 to E18.5 of
pancreatic development showed that Glis3 expression was rel-
atively low at E11.5 but significantly increased at E12.5, which
is at the beginning of the second transition of pancreatic de-
velopment, and remained high through E18.5 (Fig. 1A). In situ
VOL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6367
hybridization analysis detected expression of Glis3 mRNA in
E11.5 pancreata, as well as at later stages (E12.5, E13.5, E16.5,
and E18.5) of pancreatic development (Fig. 1C to G). At
E18.5, which is at the beginning of the third transition of
pancreatic development, Glis3 expression was restricted to is-
lets and ducts (Fig. 1G). These results suggest that Glis3 may
have a critical role early during the second transition of pan-
creatic development and that its function may be restricted to
pancreatic islets and ducts in the maturing pancreas. This con-
clusion is in agreement with data showing that Glis3 mRNA
was highly expressed in TC (-cell-like) and PANC1 (pan-
creatic duct-like) cells, but at low levels in ␣TC (␣-cell-like)
cells (Fig. 1B) and is consistent with a previous study showing
high expression of Glis3 in  cells (45).
Glis3
zf/zf
mutant mice display overt postnatal diabetes. To
study the physiological functions of Glis3 in the pancreas, we
recently generated Glis3 mutant (Glis3
zf/zf
) mice in which ZF5
of Glis3 is deleted, yielding a Glis3 defective in its transcrip-
tional activity (9, 24). Glis3
zf/zf
mutant mice have a normal
general appearance but die within several days after birth (24).
Biochemical analysis showed that blood and urine glucose lev-
els were significantly elevated in Glis3
zf/zf
mutant mice com
-
pared to those in WT and heterozygous mice (Fig. 2A and B),
while urine pH and ketone levels remained unchanged and the
mice did not develop signs of proteinuria (data not shown).
Normally, elevation of blood glucose levels induces an increase
in circulating insulin; however, serum insulin levels in Glis3
zf/zf
mutant mice were significantly lower than those in WT and
Glis3
⫹/zf
mutant mice (Fig. 2C). The development of hypergly
-
cemia and hypoinsulinemia is consistent with the development
of neonatal diabetes in Glis3
zf/zf
mutant mice. This phenotype,
together with the greatly shortened life span and the develop-
ment of polycystic kidneys displayed by Glis3
zf/zf
mutant mice
(24), shows a great resemblance to the abnormalities observed
in human NDH patients with mutations in the GLIS3 gene (45,
50). These similarities suggest that the Glis3
zf/zf
mutant mice
can serve as a suitable model to study this syndrome.
Glis3
zf/zf
mutant mice display a reduced endocrine compart
-
ment. At PND3, Glis3
zf/zf
mutant mice had developed both
ventral and dorsal pancreatic derivatives, acini appeared to
have formed normally, and little difference in the size of the
pancreas was observed between WT and Glis3
zf/zf
mutant mice
(data not shown). The development of hyperglycemia and hy-
poinsulinemia in Glis3
zf/zf
mutant mice suggested that loss of
Glis3 function might result in abnormal -cell development/
function. Although islet structures were present, they were
greatly reduced in size (Fig. 2H), morphologically distinct, and
less tightly organized than those of WT mice (Fig. 2D to G).
These observations suggested that Glis3
zf/zf
mutant mice might
exhibit a specific defect in endocrine cell development and
maturation, alongside a more subtle effect on pancreatic duct
morphology. No morphological differences were noticeable be-
tween the pancreata from WT and heterozygous mice (data
not shown). These results suggest that the Glis3
zf/zf
mutation is
recessive, in agreement with GLIS3 mutations associated with
NDH patients (45).
To determine the influence of the Glis3 mutation on specific
endocrine cell types, sections of pancreas from PND3 mice
were examined by immunofluorescence with antibodies against
several islet hormones (Fig. 3). This analysis showed that the
TABLE 1. QRT-PCR primers and probes used in this study
Gene
5⬘33⬘ sequence
Forward primer Reverse primer Probe
a
Ins1 ACCATCAGCAAGCAGGTCAT CACTTGTGGGTCCTCCACTT
Ins2 CAGCAAGCAGGAAGCCTATC TTGTGCCACTTGTGGGTCCT
Gcg AGGCCGAGGAAGGCGA TGCCTGCGGCCGAGT TCCCAGAAGAAGTCGCCATTGCTGA
Sst CCACCGGGAAACAGGAACTG GGGCCAGGAGTTAAGGAAGA
Ppy TAGCTCAGCACACAGGATGG GCCTGGTCAGTGTGTTGATG
Ghrl GGCAGGCTCCAGCTTCCT GGCTTCTTGGATTCCTTTCTCTG AGCCCAGAGCACCAGAAAGCCCA
Isl1 ACATGGGCGATCCACCAA TCGTGAATTTGATTGCCGC ACCAACACACAGGGAAATCAGACGTT
TTTTT
Ngn3 TTCTTTTGAGTCGGGAGAAC
TAGG
GGGACACTTGGATGGTGAGC TGGCGCCTCATCCCTTGGATG
Pax4 CGGGACAAGCCGAGGC CGGCCACTGAATCTGGATA TGGAGAAAGAGTTTCAGCGTGGGCA
Pax6 CCACCACACCTGTCTCCTCCT TTGGTGAGGGCGGTGTCT CATCAGGTTCCATGTTGGGCCGAA
Nkx2.2 TCGCTGACCAACACAAAGAC GCTTTGGAGAAGAGCACTCG
Nkx6.1 CCCGGAGTGATGCAGAGT AGAGTTCGGGTCCAGAGGTT
MafA CAGCAAGGAGGAGGTCATCC GCGTAGCCGCGGTTCTT CTGAAACAGAAGCGGCGCACGC
NeuroD1 TCCGGTGCCGCTGC GCGAATGGCTATCGAAAGACA TCGCTGCGAGATCCCCATAGACAACAT
Pdx1 AAATCCACCAAAGCTCACGC CTCGGGTTCCGCTGTGTAAG CTCCTGCCCACTGGCCTTTCCA
Abcc8 TTGCTGAAACCGTGGAAGG GGAGCTTCTGCTGGAACCG CCATCCGTGCCTTCAGGTACGAGG
Amylase GCTTATCAGGTCAGAAAT
TTCG
CATTCCACTTGCGGATAACTG
Ptf1a ATCGAGGCACCCGTTCAC CGATGTGAGCTGTCTCAGGA
Rat Glis3 GTGAAGGCACATTCTTCCA
AAGA
GGAGATCTGGATGGAGCTCAGT CAACAAGCAAGGAAAAAGCTACGG
TCCA
Rat Ins1 CCTGCTCGTCCTCTGGGAGC
CCAAG
CTCCAGTGCCAAGGTCTGAAG
ATCC
Rat Ins2 CCTGCTCATCCTCTGGGAGCC
CCGC
CTCCAGTGCCAAGGTCTGAAG
GTCA
a
Each probe includes 6-carboxyfluorescein at the 5⬘ end and 6-carboxytetramethylrhodamine at the 3⬘ end.
6368 KANG ET AL. MOL.CELL.BIOL.
numbers of insulin-positive  cells and somatostatin-positive
␦ cells were dramatically decreased in pancreatic islets of
Glis3
zf/zf
mutant mice compared to those of WT mice, while the
numbers of glucagon (Gcg)-positive and pancreatic polypep-
tide (Ppy)-positive cells were moderately reduced (Fig. 3A to C
and G to I). The reduction in insulin-positive  cells was the
most striking and probably responsible for the development of
hypoinsulinemia and hyperglycemia observed in Glis3
zf/zf
mu
-
tant mice. Immunofluorescence staining indicated that the
level of hormone expression was not significantly different be-
tween the pancreata of heterozygous and WT mice (Fig. 3D to
F). These data indicate that Glis3 plays a critical role in the
regulation of pancreatic endocrine functions and of  and ␦
cells in particular.
Pancreatic ducts are dilated in PND3 Glis3
zf/zf
mutant mice.
Staining of PND3 pancreatic sections with the lectin DBA,
a ductal marker, identified an additional phenotypic change in
the pancreata of Glis3
zf/zf
mutant mice. As in the WT pancreas,
DBA
⫹
cells were prominent in the pancreata of PND3
Glis3
zf/zf
mutant mice, indicating that ductal cell differentiation
had proceeded (Fig. 4A to D). However, pancreatic ducts in
Glis3
zf/zf
mutant mice were regularly dilated or cystic (Fig. 4B
and E). The latter matches the development of a polycystic
renal phenotype observed in Glis3
zf/zf
mutant mice (24), as well
as in some NDH patients (45). Formation of pancreatic and
renal cysts often coincides with and has been linked to defects
in the primary cilium structure or in primary cilium-associated
signaling pathways (16). Analysis of ␣-acetylated tubulin, a
component of cilia, indicated a significant reduction in the
percentage of cells with primary cilia in pancreatic cysts com-
pared to normal ducts (Fig. 4C to F) similar to what we ob-
served in renal tubules (24). These observations are consistent
with the hypothesis that the development of pancreatic and
renal cysts in Glis3
zf/zf
mutant mice involves a common mech
-
anism.
Loss of Glis3 function affects the expression of endocrine
cell-related genes. Because Glis3 functions as a transcriptional
regulator (9, 10, 24, 31), one might expect that loss of Glis3
would result in alterations in gene expression profiles in tissues
where it is expressed. Therefore, we compared the gene ex-
pression profiles in the pancreata of WT and Glis3
zf/zf
mutant
mice at E15.5 and PND3 by microarray analysis. These anal-
yses showed that the expression of genes associated with en-
docrine cell differentiation and function, including several pan-
creatic hormones and transcription factors, was downregulated
in the pancreata of Glis3
zf/zf
mutant mice (Table 2).
Very few
genes were found to be upregulated in the Glis3
zf/zf
pancreas
(data not shown). Although the reduction in the expression of
Ins1 and Ins2 mRNA was the most dramatic, the expression of
Sst, Ppy, Gcg, and Ghrl was also significantly diminished in the
FIG. 1. Expression of Glis3 during pancreatic development. (A) The expression of Glis3 mRNA (n ⫽ 4) during pancreatic development was
examined by QRT-PCR analysis. The expression of Glis3 was normalized to glyceraldehyde-3-phosphate dehydrogenase. (B) Comparison of the
relative levels of Glis3 mRNA expression in ␣TC (pancreatic ␣-cell-like line), TC (-cell-like line), and PANC1 (pancreatic epithelial cell-like
line) cell lines and pancreatic islets. The relative expression of Glis3 was analyzed from the Genespeed database (http://genespeed.ccf.org/). (C to
G) The expression of Glis3 mRNA in E11.5, E12.5, E13.5, E16.5, and E18.5 pancreata was examined by in situ hybridization. Sto, stomach; vp,
ventral pancreas; dp, dorsal pancreas; Int, intestine.
V
OL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6369
E15.5, as well as in the PND3, Glis3
zf/zf
mutant mouse pan
-
creas. The downregulation of the expression of all major pan-
creatic hormones, including insulin 1 and 2 (Ins1 and Ins2),
somatostatin (Sst), Gcg, Ppy, and ghrelin (Ghrl), in the PND3
Glis3
zf/zf
mutant pancreas was confirmed by QRT-PCR analysis
(Fig. 3J). These observations are consistent with the immuno-
fluorescence analyses shown in Fig. 3. In addition to Ins1 and
Ins2, the expression of several other -cell-associated genes,
including islet amyloid polypeptide (Iapp), ATP-binding cas-
sette subfamily C8 (Abcc8), and Glut2, was decreased in
Glis3
zf/zf
mutant mice (Table 2). Moreover, several genes re
-
ported to be associated with both neuroendocrine and  cells,
including those for chromogranin A (Chga), synaptotagmin
(Syt), and secretagogin (Scgn), were expressed at reduced lev-
els in the Glis3
zf/zf
mutant pancreas.
In contrast to endocrine markers, the expression of exocrine
cell-related genes was generally unchanged in Glis3
zf/zf
mutant
mice. For example, no significant alterations were observed in
FIG. 2. Development of hyperglycemia and hypoinsulinemia in Glis3
zf/zf
mutant mice. (A) Elevated blood glucose levels in WT (n ⫽ 11),
Glis3
⫹/zf
mutant (n ⫽ 12), and Glis3
zf/zf
mutant (n ⫽ 16) mice. (B) Analysis of urine glucose levels in WT (n ⫽ 11), Glis3
⫹/zf
mutant (n ⫽ 9), and
Glis3
zf/zf
mutant (n ⫽ 11) mice. (C) Comparison of blood insulin levels in WT (n ⫽ 10), Glis3
⫹/zf
mutant (n ⫽ 11), and Glis3
zf/zf
mutant (n ⫽ 14)
mice. (D to G) The size of pancreatic islets is significantly reduced in Glis3
zf/zf
mutant mice. Shown are hematoxylin-and-eosin-stained sections of
pancreata from PND3 WT and Glis3
zf/zf
mutant mice. Arrows indicate pancreatic islets. (H) Islet volume was reduced in the PND3 Glis3
zf/zf
mutant
pancreas (n ⫽ 6) compared to that in the WT pancreas (n ⫽ 6). By a systematic random sampling approach, the islet load was assessed in every
10th section throughout the pancreas. An asterisk indicates a P value of ⬍0.02.
6370 KANG ET AL. M
OL.CELL.BIOL.
the expression of the transcription factor Ptf1a, which plays a
key role in exocrine cell differentiation, or the exocrine marker
amylase (Fig. 3J to L). These findings corroborate the notion
that the major effect of the Glis3 mutation is on the endocrine
and not the exocrine compartment of the pancreas.
In addition to the expression of several pancreatic hor-
mones, gene expression profile analysis showed that the ex-
FIG. 3. Loss of Glis3 function causes a dramatic reduction in en-
docrine cells, particularly  cells. (A to I) Sections of PND3 WT and
Glis3
⫹/zf
and Glis3
zf/zf
mutant pancreata were examined by immuno
-
fluorescence with antibodies against insulin (Ins), glucagon (Gcg),
pancreatic polypeptide (Ppy), and somatostatin (Sst) as indicated. Nu-
clei were visualized by 4⬘,6-diamidino-2-phenylindole (DAPI) staining.
(J) The expression of several endocrine cell-specific, but not exocrine
cell-specific, genes is greatly reduced in Glis3
zf/zf
mutant mice. RNAs
isolated from pancreata of PND3 WT and Glis3
⫹/zf
and Glis3
zf/zf
mu
-
tant mice were analyzed by QRT-PCR for the expression of insulin 1
and 2 (Ins1 and Ins2), somatostatin, glucagon, ghrelin (Ghrl), and
pancreatic polypeptide mRNAs and the exocrine cell-related genes for
amylase and Ptf1a. An asterisk indicates a P value of ⬍0.02. (K, L)
Expression of amylase is not significantly altered in the Glis3
zf/zf
mutant
mouse pancreas. Sections of E16.5 WT and Glis3
zf/zf
mutant mouse
pancreata were stained with antibodies against the exocrine enzyme
amylase (red) and the ductal marker osteopontin (Opn; green).
FIG. 4. Development of dilated and cystic pancreatic ducts in
Glis3
zf/zf
mutant mice. (A, B) Pancreata from PND3 WT and Glis3
zf/zf
mutant mice were stained with the ductal marker DBA and DAPI. (C
to E) Pancreatic ducts and cilia were stained by DBA and anti-␣-
acetylated tubulin, respectively. Representative images of a pancreatic
duct from a PND3 WT mouse (C) and a normal (D) and a cystic
(E) pancreatic duct from a PND3 Glis3
zf/zf
mutant mouse are shown. In
WT mice, about 50% of the ductal cells were stained positively with
anti-␣-acetylated tubulin antibody. The number of ciliated cells was
greatly reduced in cysts of the Glis3
zf/zf
mutant pancreas. Arrowheads
indicate primary cilia. (F) The percentage of ciliated cells is reduced in
cystic pancreatic ducts of Glis3
zf/zf
mutant mice. Sections from three
different mice in each group were selected randomly and stained with
DBA and anti-␣-acetylated tubulin antibody. The percentage of DBA-
positive cells (n ⫽ 50 to 70) containing a primary cilium was calculated.
V
OL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6371
pression of several transcription factors with established roles
in the regulation of differentiation of proendocrine progenitors
into different endocrine cell lineages, including Ngn3, Nkx6.1,
Pax4, Pax6, Isl1, NeuroD1, and MafB, was downregulated in
the pancreata of E15.5 and/or PND3 Glis3
zf/zf
mutant mice
(Table 2) (1, 14, 18, 19, 26, 28, 37, 57). The reduced expression
of several transcription factors was supported by QRT-PCR
analysis (Fig. 5A). No significant differences in the expression
of these genes were observed between heterozygous and WT
mice. Immunofluorescence analysis showed that the number of
cells positive for Nkx6.1, Pdx1, and MafA, all of which are
-cell markers, was significantly reduced in pancreatic islets of
PND3 Glis3
zf/zf
mutant mice compared to those of WT mice
(Fig. 5B to G). Although Pdx1 and Nkx6.1 are expressed in
common pancreatic progenitors, they eventually become re-
stricted to mature  cells (40). The mature -cell marker Glut2
was also dramatically downregulated in the pancreata of
Glis3
zf/zf
mutant mice, even within the few remaining Pdx1
⫹
cells (Fig. 5H to M). Staining for the apoptotic marker caspase
3 (data not shown) did not indicate increased apoptosis in
endocrine cells from Glis3
zf/zf
mutant mice, suggesting that the
reduced expression of these genes is related to impairment in
the generation rather than degeneration of  cells. These re-
sults are in agreement with the concept that Glis3 plays a
critical role in the development of mature  cells.
Endocrine cell progenitors are reduced in Glis3
zf/zf
mutant
mice. To obtain further insight into the role of Glis3 in the
regulation of -cell development, we examined the expression
pattern of several transcription factors during pancreatic de-
velopment in Glis3
zf/zf
mutant mice. Considering that Glis3 is
induced during E11.5 to E13.5 and is not expressed in mature
exocrine cells, we were interested in analyzing the role of Glis3
during this stage of pancreatic development. At E11.5, pancre-
atic progenitor cells (Pxd1
⫹
Ptf1a
⫹
Cpa1
⫹
Sox9
⫹
) rapidly pro
-
liferate (36, 57), while at E13.5 they become patterned in a
distal/proximal manner, a process that can be monitored by the
expression of Pdx1, Sox9, and Cpa1. Comparison of the ex-
pression patterns of these proteins in WT and Glis3
zf/zf
mutant
TABLE 2. Genes downregulated in the pancreata of PND3 Glis3
zf/zf
mutant mice
a
GenBank
accession no.
Sequence description Designation
Fold change at:
E15.5 PND3
NM_008386 Insulin I Ins1 ⫺11.8 ⫺17.8
NM_008387 Insulin II Ins2 ⫺38.4 ⫺10.6
NM_021488 Ghrelin Ghrl ⫺2.2 ⫺5.3
NM_008918 Pancreatic polypeptide Ppy ⫺3.3 ⫺5.1
NM_021331 Glucose-6-phosphatase, catalytic, 2 G6pc2 ⫺4.5 ⫺4.9
NM_009215 Somatostatin Sst ⫺4.7 ⫺4
AF213386 ATP-binding cassette, subfamily C Abcc8 ⫺3.7 ⫺3.8
NM_030725 Synaptotagmin XIII Syt13 ⫺5 ⫺3.7
X63963 Paired box gene 6 Pax6 ⫺3.8 ⫺3.5
NM_007693 Chromogranin A Chga ⫺10.8 ⫺3.4
NM_013640 Proteasome subunit, beta type 10 Psmb10 NC
b
⫺3.3
NM_145399 Secretagogin Scgn ⫺3.2 ⫺3.2
NM_019741 Solute carrier family 2, member 5 Slc2a5 (Glut5) ⫺1.8 ⫺3.1
NM_011255 Retinol binding protein 4 Rbp4 ⫺2.1 ⫺3
NM_010290 Gap junction membrane channel protein alpha 9 Gja9 ⫺3.2 ⫺3
NM_011356 Frizzled-related protein Frzb ND
c
⫺3
NM_008240 Forkhead box J1 Foxj1 NC ⫺2.9
NM_013757 Synaptotagmin-like 4 Sytl4 ⫺2.1 ⫺2.9
NM_021391 Protein phosphatase 1, r1A Ppp1r1a ⫺4.1 ⫺2.8
NM_020626 Transmembrane protein 27 (collectrin) Tmem27 ⫺2.9 ⫺2.5
NM_008100 Glucagon Gcg ⫺2.3 ⫺2.3
NM_009504 Vitamin D receptor Vdr ⫺2.8 ⫺2.2
AK122226 Regulating synaptic membrane exocytosis 3 Rims3 ND ⫺2.2
NM_023182 Chymotrypsin-like Ctrl 2 ⫺2.2
NM_021459 ISL1 transcription factor Isl1 ⫺1.9 ⫺2.2
NM_009162 Secretogranin V Scg5 ⫺2.2 ⫺2.2
NM_008792 Proprotein convertase subtilisin/kexin type 2 Pcsk2 ⫺3.8 ⫺2.1
AK003606 Signal recognition particle 9 Srp9 NC ⫺2.1
AK035727 Slit homolog 3 Slit3 ⫺1.4 ⫺2.1
NM_010658 Musculoaponeurotic fibrosarcoma oncogene B MafB ⫺2.7 ⫺2.1
AK037670 Low-density lipoprotein receptor Ldlr ND ⫺2.1
NM_145435 Peptide YY Pyy ⫺3.1 ⫺2
NM_144955 NK6 transcription factor, locus 1 Nkx6.1 ⫺3.5 ⫺1.9
NM_031197 Solute carrier family 2, member 2 Slc2a2 (Glut2) ⫺2.5 ⫺1.8
NM_010894 Neurogenic differentiation 1 NeuroD1 ⫺5.1 ⫺1.6
NM_010491 Islet amyloid polypeptide Iapp ⫺1.6 ⫺1.4
NM_009719 Neurogenin 3 Ngn3 ⫺3.6 ND
NM_011038 Paired box gene 4 Pax4 ⫺2.9 NC
a
Gene expression profiles of pancreata from E15.5 and PND3 WT and Glis3
zf/zf
mutant mice were examined by microarray analysis. Shown is the decrease in gene
expression in Glis3
zf/zf
mutant mice (partial list).
b
NC, no change.
c
ND, not detectable.
6372 KANG ET AL. MOL.CELL.BIOL.
pancreata showed no significant difference in the staining pat-
tern for Pdx1 at E11.5 (Fig. 6). Likewise, little difference was
observed in the staining patterns for Pdx1 and Sox9 between
E13.5 WT and Glis3
zf/zf
mutant pancreata (Fig. 7A to D).
In
addition, little difference in glucagon staining was observed
between the pancreata of WT and Glis3
zf/zf
mutant E13.5 em
-
bryos (Fig. 7E and F). At E14.5, Cpa1
⫹
cells formed in a
distalized fashion within the branching tips of both WT and
Glis3
zf/zf
mutant pancreata (Fig. 8A to F).
These observations
indicate that Glis3 is not required for pancreatic progenitor
maintenance, its patterning, and early endocrine cell develop-
ment but acts at a later stage of mouse pancreagenesis. We
noted, however, that following the secondary transition at
E16.5, a stage at which endocrine, exocrine, and ductal cells
are organized into islets, acini, and ducts and the expression of
Pdx1 becomes confined largely to  cells, the number of Pdx1
⫹
cells was greatly diminished in the Glis3
zf/zf
mutant pancreas
(Fig. 7I and J). This decrease correlated with the reduction in
Ins
⫹
cells (Fig. 7K and L) and is consistent with the hypothesis
that the endocrine (-cell) compartment is greatly reduced in
Glis3
zf/zf
mutant mice.
Because Ngn3 plays a key role in the induction of endocrine
cell progenitors and promotes the expression of several more
terminal endocrine transcription factors, including NeuroD1,
Pax4, Pax6, Nkx6.1, and MafB, we compared the patterns of
Ngn3 expression in the pancreata of WT and Glis3
zf/zf
mutant
embryos (15, 17, 51, 53). In contrast to E11.5, the number of
Ngn3
⫹
cells was greatly reduced in the pancreata of E13.5
Glis3
zf/zf
mutant embryos compared to that in WT embryos
(Fig. 7G and H). The difference in the number of Ngn3
⫹
cells
was even more dramatic at E16.5 (Fig. 7O and P). Consistent
with these data, QRT-PCR analysis showed that the level of
Ngn3 mRNA expression was significantly reduced in E13.5 and
E15.5 Glis3
zf/zf
mutant pancreata (Fig. 7Q and R). These ob
-
servations suggest that the Glis3 mutation may affect the gen-
eration of endocrine cell progenitors and implicate Glis3 in the
formation of the pool of endocrine cell progenitors.
Regulation of Ngn3 is complex and controlled by multiple
transcription factors, including positive regulation by Myt1 and
Hnf6 and negative regulation by a Notch/Hes feedback loop (4,
20, 51). At E15.5, no significant difference in Notch expression
was observed between the pancreata of WT and Glis3
zf/zf
mutant
embryos, while Hnf6 mRNA was slightly elevated (Fig. 7R). Im-
munohistochemistry showed no significant difference in Hnf6
staining between pancreata from E16.5 WT and Glis3
zf/zf
mutant
FIG. 5. The expression of several endocrine cell-related transcrip-
tion factors is significantly reduced in the Glis3
zf/zf
mutant pancreas.
(A) The expression of Nkx6.1, Pdx1, Pax4, Ngn3, and MafA, genes
associated with endocrine development, was examined by QRT-PCR
in pancreas samples from PND3 WT and Glis3
⫹/zf
and Glis3
zf/zf
mutant
mice (n ⫽ 3). An asterisk indicates a P value of ⬍0.02. (B to M)
Pancreas sections from PND3 WT and Glis3
zf/zf
mutant mice were
examined by immunofluorescence with antibodies for Nkx6.1, Pdx1,
insulin, MafA, and/or Glut2, as indicated. D, G, J, and M, merged.
DAPI, blue staining.
FIG. 6. Analysis of the expression of Pdx1, Ngn3, and glucagon at
E11.5 of pancreatic development. (A to H) Expression of Ngn3, Pdx1,
and glucagon in pancreas sections from E11.5 Glis3
zf/zf
mutant and WT
embryos was examined by immunofluorescent staining. Representative
images are shown. No significant differences were observed. Ngn3,
green in panels A and E; Pdx1, red in panels B and F; merged in panels
C and G. (D, H) Merged glucagon (Gcg; green) and Pdx1 (red)
staining of E11.5 pancreas.
V
OL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6373
embryos (Fig. 7M and N). Moreover, Glis3 expression appeared
unaffected in the Hnf6 null pancreas (data not shown).
Regulation of Ins2 expression by Glis3. Our study provides
strong evidence of a role for Glis3 in the regulation of -cell
development. Many transcription factors with roles in pancreas
development have been reported to exhibit multiple functions.
For example, in addition to their role in lineage determination,
Pdx1, Ngn3, NeuroD1, and MafA have also been implicated in
the regulation of -cell maturation and maintenance of -cell
function (3, 5, 51). Because loss of Glis3 function affected the
expression of Ins2 the most dramatically, we examined whether
Glis3 regulates this gene. Comparison of the human INS and
mouse Ins2 proximal promoter region identified two putative
Glis-binding sites (Glis-BS1 and Glis-BS2) at ⫺94 and ⫺266 and
at ⫺99 and ⫺263, respectively (Fig. 9A and B). EMSA analysis
showed that Glis3 was able to bind both of these sites effectively,
although not as efficiently as the consensus Glis-BS (Fig. 9C to E).
Mutations at these sites abolished Glis3 binding. In addition,
FIG. 7. Comparison of Sox9, Pdx1, Ngn3, and HNF6 expression during the development of WT and Glis3
zf/zf
mutant pancreata. (A to H) The Sox9,
Pdx1, glucagon, Ngn3, and E-cadherin proteins were analyzed by immunofluorescence in sections of E13.5 WT and Glis3
zf/zf
mutant pancreata. (I to N)
The Pdx1, Ins, and Hnf6 proteins were analyzed in sections of E16.5 WT and Glis3
zf/zf
mutant pancreata by immunofluorescence staining as indicated.
(I, J) The numbers of cells expressing high levels of Pdx1 were dramatically decreased in the Glis3
zf/zf
mutant pancreas, as indicated by arrows. Arrowheads
indicate cells expressing low levels of Pdx1. (K, L). The number of cells expressing insulin was dramatically decreased in the Glis3
zf/zf
mutant pancreas,
as indicated by arrows. (M, N) The expression of Hnf6 is high in the forming ducts (indicated by arrows) and low in the exocrine tissue (indicated by
arrowheads); the asterisks indicate nonspecific fluorescence in vessels. (O, P) Ngn3 mRNA expression was examined by in situ hybridization in sections
of E16.5 WT and Glis3
zf/zf
mutant pancreata. (Q) The expression of Ngn3, Gcg, and Pdx1 mRNAs was analyzed in samples from WT and Glis3
⫹/zf
and
Glis3
zf/zf
mutant E13.5 embryos (n ⫽ 3) by QRT-PCR. (R) Expression of Ngn3, Notch1 to -3, Hnf6, and HNF6 mRNAs was compared between E15.5
WT (n ⫽ 3) and Glis3
zf/zf
mutant (n ⫽ 3) pancreata. Expression was analyzed by QRT-PCR. An asterisk indicates a P value of ⬍0.05.
6374 KANG ET AL. M
OL.CELL.BIOL.
Glis3 was able to induce Ins2 and INS promoter activity in rat
insulinoma INS-1 832/13 cells (Fig. 9F). A C-terminal deletion
mutant of Glis3 (Glis3⌬C748) that lacks transactivation function
exhibited a greatly diminished ability to activate the Ins2 and INS
promoter (Fig. 9F), while the N-truncated mutant Glis3⌬N302
was more effective in activating the Ins2 promoter (Fig. 9G). The
latter is in agreement with previous observations showing in-
creased transactivating activity of N-terminal deletion mutants of
Glis3 (31), which may suggest the presence of a repressor domain
at the N terminus of Glis3. Point mutations in only one of the two
Glis-BS reduced Glis3-mediated activation only partially (Fig.
9G, right panel); however, mutations in both sites almost totally
abolished this activity and also reduced basal activation of the Ins2
promoter by endogenous Glis3 (Fig. 9G, left panel). Again,
Glis3⌬N302 was more active. These results strongly support the
conclusion that Glis3 functions as a transcriptional regulator of
Ins2 gene expression and suggest that both Glis-BS are important
in this regulation. This was further supported by data showing
that exogenous expression of Glis3 in INS-1 832/13 cells enhanced
the expression of endogenous Ins2 expression (Fig. 9H). Al-
though another potential Glis-BS was found at ⫺342 to ⫺348,
Glis3 did not bind this sequence and mutation of this site did not
influence Glis3-mediated activation of the Ins2 promoter (data
not shown), suggesting that this site is not a functional Glis3-
binding site.
DISCUSSION
We recently reported that Glis3
zf/zf
mutant mice, which are
defective in their transcriptional activation function, have a
very short life span and develop polycystic kidney disease (24).
In this study, we demonstrate that these mice also display
hyperglycemia and hypoinsulinemia, consistent with the devel-
opment of neonatal diabetes. During revision of the manu-
script, a similar phenotype was reported in an alternate Glis3
mutant mouse model (52). These phenotypic changes resemble
the abnormalities observed in NDH patients with mutations in
GLIS3 (45, 50) and suggest that GLIS3 plays a critical role in
the regulation of pancreatic endocrine functions and the de-
velopment of type 1 diabetes. The latter is supported by a
recent report identifying the GLIS3 gene as a susceptibility
locus for type 1 diabetes (8). In this study, we further demon-
strate that Glis3 plays a critical role in cell lineage specifica-
tion, particularly the development of  cells, and in the regu-
lation of mature -cell function. In addition, we show that
Glis3 is important in maintaining normal functions in ductal
epithelial cells but does not appear to have a major role in the
development or function of acini.
Analysis of Glis3 expression during pancreatic development
showed that Glis3 mRNA was detectable at E11.5 and ex-
pressed at increased levels at E12.5 and later stages of pancre-
atic development. At E18.5, Glis3 is particularly highly ex-
pressed in ductal cells and islets but was not detectable in acini,
suggesting that Glis3 functions as a transcriptional regulator
particularly in endocrine cell progenitors,  cells, and ductal
cells. Immunohistological analysis of the pancreata of PND3
Glis3
zf/zf
mutant mice showed that although the size of the
pancreas is not significantly different, the islets are consider-
ably smaller than those of WT mice. Pancreatic acini develop
normally and exocrine markers are normally expressed in
Glis3
zf/zf
mutant mice. The reduced islet size was largely due to
a dramatic decline in the number of fully differentiated  cells,
as indicated by the reduction in insulin-, Pdx1-, MafA-, and
Glut2-positive cells, as well in the expression of several other
-cell selective genes. This decrease in  cells explains the
development of diabetes observed in Glis3
zf/zf
mutant mice.
These observations suggest that Glis3
zf/zf
mutant mice exhibit a
specific defect in endocrine, rather than exocrine, cell devel-
opment and indicate that Glis3 is a key factor in regulating the
generation and/or maintenance of  cells. Analysis of pancre-
atic sections for active caspase 3 did not indicate increased
apoptosis in the pancreata of E13.5 or E15.5 Glis3
zf/zf
mutant
mouse embryos or in the islets of PND3 Glis3
zf/zf
mutant mice.
This, together with the observations that Glis3 is expressed
early in pancreatic development and that the reduction in the
proendocrine compartment in Glis3
zf/zf
mutant mice was de
-
tectable as early as E13.5, suggests that Glis3 plays a major role
in regulating the generation rather than the survival of the
-cell population.
Early multipotent pancreatic progenitors, marked by Pdx1,
Ptf1a, Nkx2.2, Sox9, and Cpa1 expression (12, 57), are the
source of all differentiated cells of the exocrine, ductal, and
endocrine cell lineages. Sox9 expression is restricted to a mi-
totically active, Notch-responsive subset of Pdx1 multipotent
progenitor cells and is not expressed in committed endocrine
cell precursors or differentiated cells (46). Sox9 regulates the
maintenance and promotes the proliferation and survival of
these multipotent progenitors. Our results show that at E13.5,
the expression pattern of Sox9 in the Glis3
zf/zf
pancreas is not
significantly changed (Fig. 7A and B). Moreover, the patterns
of Pdx1 staining in the pancreata of WT and Glis3
zf/zf
mutant
FIG. 8. The expression of Cpa1 was not significantly changed in pan-
creata of WT and Glis3
zf/zf
mutant E14.5 embryos. Cpa1, green in panels
A and B; Pdx1, red in panels C and D; merged in panels E and F.
V
OL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6375
E11.5 and E13.5 embryos were not significantly different (Fig.
6B and F and 7C and D). However, the number of Pdx1
⫹
cells
was dramatically reduced at E16.5 and PND3 (Fig. 7I and J
and 5C and F). After E15.5, expression of Pdx1 is known to
become largely associated with  cells. Thus, this decrease in
Pdx1
⫹
cells may largely reflect the reduction of -cell devel
-
opment in Glis3
zf/zf
mutant mice. These observations suggest
that Glis3 does not regulate pancreatic progenitor mainte-
nance or its patterning but acts at a later stage of mouse
pancreagenesis (Fig. 10). This is consistent with our observa-
tion that the size of the PND3 pancreas was not significantly
changed in Glis3
zf/zf
mutant mice; a reduction in the pancreatic
progenitor pool would likely result in hypoplasia of all differ-
entiated cell compartments.
A large number of transcription factors affect the generation
of  cells (1, 18, 19, 21, 26, 28, 36). Ngn3 has been considered
to be one of the key factors promoting the differentiation of
pancreatic multipotent progenitors into endocrine cell progen-
itors (15, 17, 21, 36, 53). Ngn3 is dramatically induced at the
beginning of the second transition, peaks at E15.5, and then
rapidly decreases. Ngn3 expression is under complex control,
which includes positive regulation by Hnf6 (20) and a Notch1/
FIG. 9. Glis3 functions as a transcriptional regulator of insulin gene expression. (A) Schematic representation of potential Glis-BS and other
enhancers in the murine and rat Ins2 and human INS 5⬘ regulatory regions. The scale at the top is in base pairs relative to the transcriptional start
site (represented as ⫹1). (B) Comparison of the nucleotide sequences of the consensus Glis-BS with the two putative Glis-BS in the mouse Ins2
and human INS proximal promoter regions. (C) Two putative Glis-BS-containing regions in the mouse insulin 2 promoter (Glis-BS1 at ⫺84 to
⫺109 and Glis-BS2 at ⫺253 to ⫺276) and their respective mutant nucleotides were labeled and used as probes in an EMSA. Competition was
carried out with 5-fold (⫹) and 25-fold (⫹⫹) excesses of unlabeled oligonucleotides as indicated. Arrow indicates the Glis3-Glis-BS complex. (D,
E) Binding of Glis3 to
32
P-labeled consensus Glis-BS was examined in the absence or presence of unlabeled consensus Glis-BS, mouse Glis-BS1,
and mGlis-BS2 sites, their mutants (mut), or the corresponding human INS Glis-BS. (F) INS-1 (832/13) cells were cotransfected with the pGL4.10
empty vector, mIns2-696-Luc, or hINS-700-Luc and either the p3xFlag-CMV-10 empty vector, p3xFlag-CMV-Glis3, or p3xFlag-CMV-Glis3⌬C748
as indicated. After 24 h, cells were assayed for luciferase and -galactosidase activities and the relative Luc activity was calculated and plotted. Each
bar represents the mean ⫾ the standard error of the mean. (G) INS-1 (832/13) cells were cotransfected with the pGL4.10 empty vector,
mIns2(⫺696)-Luc, or the specified mIns2(⫺696)-Luc mutant and either p3xFlag-CMV-10, p3xFlag-CMV-Glis3, or p3xFlag-CMV-Glis3⌬N302 as
indicated. Relative Luc activity was calculated and plotted. In the insert, shaded boxes indicate Glis-BS and X indicates mutated Glis-BS. (H) INS-1
(832/13) cells were transiently transfected with p3xFlag-CMV-Glis3 or p3xFlag-CMV-Glis3⌬N302 and 48 h later analyzed by QRT-PCR for the
expression of rat Ins1 and Ins2 and mouse Glis3.
6376 KANG ET AL. M
OL.CELL.BIOL.
Hes negative feedback loop (4, 18). Ngn3 regulates the tran-
scription of Myt1, Insm1, Pax4, Nkx2.2, MafA, and NeuroD1,
all of which play important roles in endocrine development.
Analysis of Ngn3 expression shows that the number of Ngn3
⫹
cells is greatly diminished in the Glis3
zf/zf
mutant mouse pan
-
creas at E13.5 and E16.5. Given the important role of Ngn3 in
the generation of proendocrine progenitors, our results suggest
that Glis3 either regulates the transition of pancreatic progen-
itors to endocrine cell progenitors and/or the maintenance of
endocrine cell progenitors (Fig. 10). Further studies are
needed to determine the precise role of Glis3 in endocrine
development, whether Glis3 directly regulates Ngn3 transcrip-
tion, and whether Glis3 and Ngn3 act cooperatively in endo-
crine development.
Observations showing that Glis3 is highly expressed in 
cells rather than other endocrine cells (45) and that the gen-
eration of  cells is affected rather than that of ␣ cells indicate
that Glis3 may have an additional role in selectively promoting
the differentiation of endocrine progenitors along one or more
different lineages and/or in regulating the function of mature 
cells. The latter is supported by evidence showing that Glis3
also plays a role in the regulation of insulin gene expression.
We identified two putative Glis-BS, Glis-BS1 and Glis-BS2, in
the proximal region of the mouse Ins2 and human INS genes
that are able to bind Glis3. The Glis-BS2 site was recently
reported to be functional in Glis3-mediated activation of the
rat Ins2 promoter (54). In this study, we demonstrate that
optimal activation of the mouse Ins2 and human INS promot-
ers by Glis3 requires both Glis-BS1 and Glis-BS2. Our results
suggest that, in addition to being a critical factor in -cell
development during pancreagenesis, Glis3, along with other
transcription factors, including Pdx1, Nkx6.1, Ngn3, and MafA
(3, 5, 51), is implicated in maintaining -cell function (Fig. 10).
In addition to the impairment of -cell generation, PND3
Glis3
zf/zf
mutant mice develop pancreatic cysts, suggesting that
Glis3 plays a role in regulating the function and maintenance
of ductal epithelial cells (Fig. 10). This is consistent with the
observed Glis3 expression in pancreatic ducts. We previously
showed that Glis3 mutant mice also develop renal cysts (24), as
was also reported for NDH patients (45). Development of
pancreatic and renal cysts, together with hepatic fibrosis, often
coincides, suggesting a common contributing mechanism. Al-
though the precise molecular mechanisms responsible for cyst
development have yet to be established, it has become evident
that dysfunction of the primary cilium and defects in cilium-
associated signal transduction pathways are key factors in the
etiology of cystic diseases (47, 55). We have proposed that
Glis3 is part of a primary cilium-linked signaling pathway; after
its activation, Glis3 translocates to the nucleus, where it regu-
lates gene transcription (24). Thus, the development of renal
and pancreatic cysts and of hepatic fibrosis in NDH patients
and Glis3
zf/zf
mutant mice appears also to be linked to a defect
in a primary cilium-associated (Glis3-dependent) pathway, as
has been demonstrated for other cystic diseases. However,
diabetes is not commonly associated with primary cilium dys-
function, suggesting that the regulation of -cell development
by Glis3 is mediated by a different mechanism. Interestingly,
several other transcription factors with important functions in
pancreatic development, including Hnf6, have been implicated
in cystic disease (42). Hnf6 plays a critical role in the regulation
of Ngn3 and formation of endocrine progenitors (20, 42). In
addition, Hnf6, by regulating Hnf1 expression, is a key factor
in ductal cell specification and maturation. Loss of Hnf6 or
Hnf1 results in -cell deficiency, cyst formation within pan-
creatic ducts, and the development of polycystic kidney disease
in mice (11, 42). However, no immediate connection was found
between the expression of Hnf6 and that of Glis3. Hnf6 was
not significantly altered in the pancreata of E16.5 Glis3
zf/zf
mutant embryos (Fig. 8M and N).
In summary, our study demonstrates that mice deficient in
Glis3 function develop neonatal diabetes due to impairment in
the generation of  cells and insulin expression. This is asso-
ciated with a significant decrease in the Ngn3
⫹
endocrine pro
-
genitor population, suggesting that Glis3 is involved in the
lineage determination of endocrine progenitors into endocrine
cells and  cells in particular (Fig. 10). Glis3
zf/zf
mutant mice
will provide an excellent model in which to study the molecular
mechanism underlying the development of  cells and the role
of Glis3 in neonatal diabetes.
ACKNOWLEDGMENTS
We thank Kristin Lichti-Kaiser and Xiaoling Li for their comments
on the manuscript, F. P. Lemaigre for providing pancreas RNA sam-
ples from developing Hnf6 null embryos, Jennifer Collins from the
Microarray Group for her assistance with the microarray analyses, and
Laura Miller for her assistance with animal breeding.
This research was supported by the Intramural Research Program of
the NIEHS, NIH (Z01-ES-100485).
We have no conflicts of interest to declare.
REFERENCES
1. Ackermann, A. M., and M. Gannon. 2007. Molecular regulation of pancre-
atic beta-cell mass development, maintenance, and expansion. J. Mol. En-
docrinol. 38:193–206.
2. Ahlgren, U., S. L. Pfaff, T. M. Jessell, T. Edlund, and H. Edlund. 1997.
Independent requirement for ISL1 in formation of pancreatic mesenchyme
and islet cells. Nature 385:257–260.
FIG. 10. Schematic representation of the critical roles of Glis3 in
pancreatic development and insulin gene expression. Pdx1
⫹
Ptf1a
⫹
Cpa1
⫹
progenitor cells can differentiate into proendocrine, ductal, and
exocrine lineages. Glis3 expression is induced during the second tran-
sition of pancreatic development, and Glis3 remains highly expressed
in proendocrine and ductal cells and in  cells at later stages of
development. Loss of Glis3 transactivation activity results in the for-
mation of pancreatic cysts and greatly impedes the generation of 
cells, which leads to the development of diabetes. In addition to reg-
ulating the generation of  cells, Glis3 also regulates the transcription
of the insulin gene. These observations demonstrate that Glis3 has
multiple functions in the pancreas and is a critical regulator of -cell
development and insulin expression.
V
OL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6377
3. Andrali, S. S., M. L. Sampley, N. L. Vanderford, and S. Ozcan. 2008.
Glucose regulation of insulin gene expression in pancreatic beta-cells. Bio-
chem. J. 415:1–10.
4. Apelqvist, A., H. Li, L. Sommer, P. Beatus, D. J. Anderson, T. Honjo, M.
Hrabe de Angelis, U. Lendahl, and H. Edlund. 1999. Notch signalling con-
trols pancreatic cell differentiation. Nature 400:877–881.
5. Aramata, S., S. I. Han, and K. Kataoka. 2007. Roles and regulation of
transcription factor MafA in islet beta-cells. Endocr. J. 54:659–666.
6. Aruga, J. 2004. The role of Zic genes in neural development. Mol. Cell.
Neurosci. 26:205–221.
7. Attanasio, M., N. H. Uhlenhaut, V. H. Sousa, J. F. O’Toole, E. Otto, K.
Anlag, C. Klugmann, A. C. Treier, J. Helou, J. A. Sayer, D. Seelow, G.
Nurnberg, C. Becker, A. E. Chudley, P. Nurnberg, F. Hildebrandt, and M.
Treier. 2007. Loss of GLIS2 causes nephronophthisis in humans and mice by
increased apoptosis and fibrosis. Nat. Genet. 39:1018–1024.
8. Barrett, J. C., D. G. Clayton, P. Concannon, B. Akolkar, J. D. Cooper, H. A.
Erlich, C. Julier, G. Morahan, J. Nerup, C. Nierras, V. Plagnol, F. Pociot, H.
Schuilenburg, D. J. Smyth, H. Stevens, J. A. Todd, N. M. Walker, and S. S.
Rich. 2009. Genome-wide association study and meta-analysis find that over
40 loci affect risk of type 1 diabetes. Nat. Genet. 41:703–707.
9. Beak, J. Y., H. S. Kang, Y.-S. Kim, and A. M. Jetten. 2008. Functional
analysis of the zinc finger and activation domains of Glis3 and mutant
Glis3(NDH1). Nucleic Acids Res. 36:1690–1702.
10. Beak, J. Y., H. S. Kang, Y.-S. Kim, and A. M. Jetten. 2007. Kru¨ppel-like zinc
finger protein Glis3 promotes osteoblast differentiation by regulating FGF18
expression. J. Bone Miner. Res. 22:1234–1244.
11. Bingham, C., M. P. Bulman, S. Ellard, L. I. Allen, G. W. Lipkin, W. G. Hoff,
A. S. Woolf, G. Rizzoni, G. Novelli, A. J. Nicholls, and A. T. Hattersley. 2001.
Mutations in the hepatocyte nuclear factor-1 gene are associated with
familial hypoplastic glomerulocystic kidney disease. Am. J. Hum. Genet.
68:219–224.
12. Burlison, J. S., Q. Long, Y. Fujitani, C. V. Wright, and M. A. Magnuson.
2008. Pdx-1 and Ptf1a concurrently determine fate specification of pancreatic
multipotent progenitor cells. Dev. Biol. 316:74–86.
13. Collombat, P., A. Mansouri, J. Hecksher-Sorensen, P. Serup, J. Krull, G.
Gradwohl, and P. Gruss. 2003. Opposing actions of Arx and Pax4 in endo-
crine pancreas development. Genes Dev. 17:2591–2603.
14. Gittes, G. K. 2009. Developmental biology of the pancreas: a comprehensive
review. Dev. Biol. 326:4–35.
15. Gradwohl, G., A. Dierich, M. LeMeur, and F. Guillemot. 2000. neurogenin3
is required for the development of the four endocrine cell lineages of the
pancreas. Proc. Natl. Acad. Sci. USA 97:1607–1611.
16. Gresh, L., E. Fischer, A. Reimann, M. Tanguy, S. Garbay, X. Shao, T.
Hiesberger, L. Fiette, P. Igarashi, M. Yaniv, and M. Pontoglio. 2004. A
transcriptional network in polycystic kidney disease. EMBO J. 23:1657–1668.
17. Gu, G., J. Dubauskaite, and D. A. Melton. 2002. Direct evidence for the
pancreatic lineage: NGN3
⫹
cells are islet progenitors and are distinct from
duct progenitors. Development 129:2447–2457.
18. Habener, J. F., D. M. Kemp, and M. K. Thomas. 2005. Minireview: tran-
scriptional regulation in pancreatic development. Endocrinology 146:1025–
1034.
19. Herrera, P. L., V. Nepote, and A. Delacour. 2002. Pancreatic cell lineage
analyses in mice. Endocrine 19:267–278.
20. Jacquemin, P., S. M. Durviaux, J. Jensen, C. Godfraind, G. Gradwohl, F.
Guillemot, O. D. Madsen, P. Carmeliet, M. Dewerchin, D. Collen, G. G.
Rousseau, and F. P. Lemaigre. 2000. Transcription factor hepatocyte nuclear
factor 6 regulates pancreatic endocrine cell differentiation and controls ex-
pression of the proendocrine gene ngn3. Mol. Cell. Biol. 20:4445–4454.
21. Jensen, J. 2004. Gene regulatory factors in pancreatic development. Dev.
Dyn. 229:176–200.
22. Jørgensen, M. C., J. Ahnfelt-Ronne, J. Hald, O. D. Madsen, P. Serup, and
J. Hecksher-Sorensen. 2007. An illustrated review of early pancreas devel-
opment in the mouse. Endocr. Rev. 28:685–705.
23. Kang, H. S., M. Angers, J. Y. Beak, X. Wu, J. M. Gimble, T. Wada, W. Xie,
J. B. Collins, S. F. Grissom, and A. M. Jetten. 2007. Gene expression
profiling reveals a regulatory role for ROR␣ and ROR␥ in phase I and phase
II metabolism. Physiol. Genomics 31:281–294.
24. Kang, H. S., J. Y. Beak, Y.-S. Kim, R. Herbert, and A. M. Jetten. 2009. Glis3
is associated with primary cilia and Wwtr1/TAZ and implicated in polycystic
kidney disease. Mol. Cell. Biol. 29:2556–2569.
25. Kasper, M., G. Regl, A. M. Frischauf, and F. Aberger. 2006. GLI transcrip-
tion factors: mediators of oncogenic Hedgehog signalling. Eur. J. Cancer
42:437–445.
26. Kemp, D. M., M. K. Thomas, and J. F. Habener. 2003. Developmental
aspects of the endocrine pancreas. Rev. Endocr. Metab. Disord. 4:5–17.
27. Kim, S. C., Y.-S. Kim, and A. M. Jetten. 2005. Kru¨ppel-like zinc finger
protein Gli-similar 2 (Glis2) represses transcription through interaction with
C-terminal binding protein 1 (CtBP1). Nucleic Acids Res. 33:6805–6815.
28. Kim, S. K., and R. J. MacDonald. 2002. Signaling and transcriptional control
of pancreatic organogenesis. Curr. Opin. Genet. Dev. 12:540–547.
29. Kim, Y. S., H. S. Kang, R. Herbert, J. Y. Beak, J. B. Collins, S. F. Grissom,
and A. M. Jetten. 2008. Kru¨ppel-like zinc finger protein Glis2 is essential for
the maintenance of normal renal functions. Mol. Cell. Biol. 28:2358–2367.
30. Kim, Y. S., M. Lewandoski, A. O. Perantoni, S. Kurebayashi, G. Nakanishi,
and A. M. Jetten. 2002. Identification of Glis1, a novel Gli-related, Kru¨ppel-
like zinc finger protein containing transactivation and repressor functions.
J. Biol. Chem. 277:30901–30913.
31. Kim, Y. S., G. Nakanishi, M. Lewandoski, and A. M. Jetten. 2003. GLIS3, a
novel member of the GLIS subfamily of Kru¨ppel-like zinc finger proteins
with repressor and activation functions. Nucleic Acids Res. 31:5513–5525.
32. Lamar, E., C. Kintner, and M. Goulding. 2001. Identification of NKL, a
novel Gli-Kru¨ppel zinc-finger protein that promotes neuronal differentia-
tion. Development 128:1335–1346.
33. Lukashova-v Zangen, I., S. Kneitz, C. M. Monoranu, S. Rutkowski, B.
Hinkes, G. H. Vince, B. Huang, and W. Roggendorf. 2007. Ependymoma
gene expression profiles associated with histological subtype, proliferation,
and patient survival. Acta Neuropathol. 113:325–337.
34. 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.
35. Merzdorf, C. S. 2007. Emerging roles for zic genes in early development.
Dev. Dyn. 236:922–940.
36. Murtaugh, L. C. 2007. Pancreas and beta-cell development: from the actual
to the possible. Development 134:427–438.
37. Murtaugh, L. C., and D. A. Melton. 2003. Genes, signals, and lineages in
pancreas development. Annu. Rev. Cell Dev. Biol. 19:71–89.
38. Nakanishi, G., Y.-S. Kim, T. Nakajima, and A. M. Jetten. 2006. Regulatory
role for Kru¨ppel-like zinc-finger protein Gli-similar 1 (Glis1) in PMA-
treated and psoriatic epidermis. J. Investig. Dermatol. 126:49–60.
39. Nakashima, M., N. Tanese, M. Ito, W. Auerbach, C. Bai, T. Furukawa, T.
Toyono, A. Akamine, and A. L. Joyner. 2002. A novel gene, GliH1, with
homology to the Gli zinc finger domain not required for mouse development.
Mech. Dev. 119:21.
40. Nelson, S. B., A. E. Schaffer, and M. Sander. 2007. The transcription factors
Nkx6.1 and Nkx6.2 possess equivalent activities in promoting beta-cell fate
specification in Pdx1
⫹
pancreatic progenitor cells. Development 134:2491–
2500.
41. Oliver-Krasinski, J. M., and D. A. Stoffers. 2008. On the origin of the beta
cell. Genes Dev. 22:1998–2021.
42. Pierreux, C. E., A. V. Poll, C. R. Kemp, F. Clotman, M. A. Maestro, S. Cordi,
J. Ferrer, L. Leyns, G. G. Rousseau, and F. P. Lemaigre. 2006. The tran-
scription factor hepatocyte nuclear factor-6 controls the development of
pancreatic ducts in the mouse. Gastroenterology 130:532–541.
43. Ruiz i Altaba, A., C. Mas, and B. Stecca. 2007. The Gli code: an information
nexus regulating cell fate, stemness and cancer. Trends Cell Biol. 17:438–
447.
44. Sander, M., L. Sussel, J. Conners, D. Scheel, J. Kalamaras, F. Dela Cruz, V.
Schwitzgebel, A. Hayes-Jordan, and M. German. 2000. Homeobox gene
Nkx6.1 lies downstream of Nkx2.2 in the major pathway of beta-cell forma-
tion in the pancreas. Development 127:5533–5540.
45. Sene´e, V., C. Chelala, S. Duchatelet, D. Feng, H. Blanc, J. C. Cossec, C.
Charon, M. Nicolino, P. Boileau, D. R. Cavener, P. Bougneres, D. Taha, and
C. Julier. 2006. Mutations in GLIS3 are responsible for a rare syndrome with
neonatal diabetes mellitus and congenital hypothyroidism. Nat. Genet. 38:
682–687.
46. Seymour, P. A., K. K. Freude, M. N. Tran, E. E. Mayes, J. Jensen, R. Kist,
G. Scherer, and M. Sander. 2007. SOX9 is required for maintenance of the
pancreatic progenitor cell pool. Proc. Natl. Acad. Sci. USA 104:1865–1870.
47. Singla, V., and J. F. Reiter. 2006. The primary cilium as the cell’s antenna:
signaling at a sensory organelle. Science 313:629–633.
48. Sosa-Pineda, B., K. Chowdhury, M. Torres, G. Oliver, and P. Gruss. 1997.
The Pax4 gene is essential for differentiation of insulin-producing beta cells
in the mammalian pancreas. Nature 386:399–402.
49. St-Onge, L., B. Sosa-Pineda, K. Chowdhury, A. Mansouri, and P. Gruss.
1997. Pax6 is required for differentiation of glucagon-producing alpha-cells
in mouse pancreas. Nature 387:406–409.
50. Taha, D., M. Barbar, H. Kanaan, and J. Williamson Balfe. 2003. Neonatal
diabetes mellitus, congenital hypothyroidism, hepatic fibrosis, polycystic kid-
neys, and congenital glaucoma: a new autosomal recessive syndrome? Am. J.
Med. Genet. A 122A:269–273.
51. Wang, S., J. Hecksher-Sorensen, Y. Xu, A. Zhao, Y. Dor, L. Rosenberg, P.
Serup, and G. Gu. 2008. Myt1 and Ngn3 form a feed-forward expression loop
to promote endocrine islet cell differentiation. Dev. Biol. 317:531–540.
52. Watanabe, N., K. Hiramatsu, R. Miyamoto, K. Yasuda, N. Suzuki, N. Os-
hima, H. Kiyonari, D. Shiba, S. Nishio, T. Mochizuki, T. Yokoyama, S.
Maruyama, S. Matsuo, Y. Wakamatsu, and H. Hashimoto. 2009. A murine
model of neonatal diabetes mellitus in Glis3-deficient mice. FEBS Lett.
583:2108–2113.
53. White, P., C. L. May, R. N. Lamounier, J. E. Brestelli, and K. H. Kaestner.
2008. Defining pancreatic endocrine precursors and their descendants. Di-
abetes 57:654–668.
6378 KANG ET AL. MOL.CELL.BIOL.
54. Yang, Y., B. H. Chang, S. L. Samson, M. V. Li, and L. Chan. 2009. The
Kru¨ppel-like zinc finger protein Glis3 directly and indirectly activates insulin
gene transcription. Nucleic Acids Res. 37:2529–2538.
55. Yoder, B. K., S. Mulroy, H. Eustace, C. Boucher, and R. Sandford. 2006.
Molecular pathogenesis of autosomal dominant polycystic kidney disease.
Expert Rev. Mol. Med. 8:1–22.
56. Zhang, F., G. Nakanishi, S. Kurebayashi, K. Yoshino, A. Perantoni, Y.-S.
Kim, and A. M. Jetten. 2001. Characterization of Glis2, a novel gene encod-
ing a Gli-related, Kru¨ppel-like transcriptional factor with transactivation and
repressor functions. Roles in kidney development and neurogenesis. J. Biol.
Chem. 12:10139–10149.
57. Zhou, Q., A. C. Law, J. Rajagopal, W. J. Anderson, P. A. Gray, and D. A.
Melton. 2007. A multipotent progenitor domain guides pancreatic organo-
genesis. Dev. Cell 13:103–114.
VOL. 29, 2009 PANCREATIC -CELL REGULATION BY Glis3 6379
