The islets of Langerhans in the pancreas are specialized
endocrine micro-organs composed of four distinct cell types:
somatostatin-producing δ-cells and pancreatic polypeptide-
producing (PP) cells. The β-cells are key metabolic regulators,
and their loss or dysfunction leads to diabetes mellitus.
During fetal development, all four endocrine cell types, as
well as the more abundant exocrine cells of the pancreas derive
from a common set of epithelial cells that originate in the early
gut endoderm (for a review of pancreatic development see
Slack, 1995). The first morphological evidence of the future
pancreas appears as a small dorsal bud at the foregut-midgut
junction at embryonic day E9.5 in the fetal mouse. A few
insulin-expressing cells appear within a day after bud
formation, and these early insulin-expressing cells often co-
express glucagon (Teitelman et al., 1993) but lack the
expression of some other β-cell markers (Oster et al., 1998;
Pang et al., 1994). Fully differentiated β-cells first appear
around E13 at the start of a massive wave of β-cell
differentiation termed the ‘secondary transition’ (Pictet and
Rutter, 1972). Although it has been proposed that the insulin-
expressing cells generated in the early pancreatic bud function
as progenitors for the β-cells that appear later during the
secondary transition (Pictet and Rutter, 1972; Teitelman et al.,
1993; Upchurch et al., 1994), this relationship has never been
proven by direct lineage tracing, leaving the possibility that the
two populations form via distinct pathways (Pang et al., 1994).
Progressive modifications in gene expression direct the
process of development from endoderm precursors to
differentiated β-cells. A number of transcription factors
have been shown to control pancreas morphogenesis or the
differentiation of the endocrine cells (Ahlgren et al., 1997;
Edlund, 1998; Gradwohl et al., 2000, Naya, 1997; Harrison et
al., 1999; Jonsson et al., 1994; Li et al., 1999; Offield et al.,
1996; Sander and German, 1997; Sander et al., 1997; Sosa-
Pineda et al., 1997; St-Onge et al., 1997). Although no
transcription factor has been identified thus far that selectively
controls β-cell formation, the NK-homeodomain factor Nkx2.2
plays an essential role in β-cell development. Nkx2.2 is
initially expressed broadly in the pancreatic bud, becomes
progressively restricted during the secondary transition, and is
eventually limited to the α, β and PP cells of the mature islet.
Disruption of the Nkx2.2 gene leads to the accumulation of
incompletely differentiated β-cells that express some β-cell
markers, but not insulin (Sussel et al., 1998).
The phenotype of Nkx2.2 mutant mice may in part result
from the loss of other transcriptional regulators, including
Nkx6.1 which is absent from the mutant pancreas after the start
of the secondary transition (Sussel et al., 1998). Like Nkx2.2,
Nkx6.1 is a homeodomain transcription factor, although it
falls into a divergent subfamily within the NK homeodomain
family (Rudnick et al., 1994). Within the pancreas, Nkx6.1 also
follows an expression pattern similar to that of Nkx2.2; the
main distinction being that it eventually becomes limited to the
β-cells alone (Jensen et al., 1996; Oster et al., 1998; Rudnick
et al., 1994).
Development 127, 5533-5540 (2000)
Printed in Great Britain © The Company of Biologists Limited 2000
Most insulin-producing β-cells in the fetal mouse pancreas
arise during the secondary transition, a wave of
differentiation starting at embryonic day 13. Here, we show
that disruption of homeobox gene Nkx6.1 in mice leads to
loss of β-cell precursors and blocks β-cell neogenesis
specifically during the secondary transition. In contrast,
islet development in Nkx6.1/Nkx2.2 double mutant embryos
is identical to Nkx2.2 single mutant islet development: β-
cell precursors survive but fail to differentiate into β-cells
throughout development. Together, these experiments
reveal two independently controlled pathways for β-cell
differentiation, and place Nkx6.1 downstream of Nkx2.2 in
the major pathway of β-cell differentiation.
Key words: Islet, β-cell, Insulin, Transcription factor, Mouse
Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of
β-cell formation in the pancreas
Maike Sander1,*, Lori Sussel1,‡, Jennifer Conners1, David Scheel1, Julie Kalamaras1, Filemon Dela Cruz1,
Valerie Schwitzgebel1,§, Andrea Hayes-Jordan1and Michael German1,2,¶
1Hormone Research Institute, University of California San Francisco, San Francisco, CA 94143-0534, USA
2Department of Medicine, University of California San Francisco, San Francisco, CA 94143-0534, USA
*Present address: Center for Molecular Neurobiology, Martinistr. 85, 20251 Hamburg, Germany
‡Present address: Barbara Davis Center for Childhood Diabetes, University of Colorado Health Sciences Center, 4200 E. 9th Ave, Denver, CO 80262, USA
§Present address: Pediatric Endocrinology, Children’s Hospital, University of Geneva, CH-1211 Geneva, Switzerland
¶Author for correspondence (e-mail: firstname.lastname@example.org)
Accepted 27 September; published on WWW 14 November 2000
In this study, we have examined the role of Nkx6.1 in the
hierarchy of transcriptional events leading to β-cell
differentiation. Consistent with its specific expression in β-
cells, homozygous mutation of the Nkx6.1 gene in mice
profoundly inhibits β-cell formation; but this defect only
becomes evident after the start of the secondary transition. The
pancreatic phenotype of mice with homozygous mutations in
both the Nkx6.1 and Nkx2.2 genes is identical to the phenotype
of Nkx2.2 homozygous single mutant mice. These studies
provide conclusive genetic evidence that Nkx6.1
downstream of Nkx2.2 in the major pathway of β-cell
MATERIALS AND METHODS
Nkx6.1 gene targeting
To generate Nkx6.1 mutant mice, a genomic clone containing exon1
and 2 of the Nkx6.1 gene was isolated from a 129J mouse genomic
library in lambda Dash (Stratagene) by Southern hybridization
screening. An 8.5 kb SpeI/EcoRV fragment was cloned into the
pBlueScript KS+ cloning vector, and an 800 bp NotI fragment
containing part of exon1 was replaced by a PGK-neo cassette.
Recombinant ES cell clones and Nkx6.1 mutant mice were generated
as described previously (Kash et al., 1997), and maintained on a
C57BL/6J mouse background. All ES cell clones and mice were
genotyped by Southern blotting with a probe external to the targeting
construct. No Nkx6.1 protein can be detected in the homozygous
mutant embryos using antisera directed at the C-terminal end of the
molecule (data not shown), showing that the mutation encodes for a
Immunohistochemistry and TUNEL assays
Immunohistochemical and immunofluorescence analyses were
performed on paraffin sections as described previously (Sander et al.,
1997). The primary antibodies used in these assays were the
following: rabbit anti-NKX6.1 diluted 1:4000; guinea pig anti-insulin
diluted 1:10000 (Linco); guinea pig anti-glucagon diluted 1:10000
(Linco); mouse anti-somatostatin diluted 1:100 (Fitzgerald); rabbit
anti-PP diluted 1:2000 (Dako); rabbit anti-IAPP diluted 1:4000
(Advanced Chemtech), rabbit anti-PC1/3 diluted 1:2000 (kindly
provided by Donald Steiner); mouse anti-BrdU diluted 1:200
(Chemicon), rabbit anti-PAX6 diluted 1:4000 (kindly provided by
Simon Saule); monoclonal anti-NKX2.2 diluted 1:50 (kindly provided
by Thomas Jessell); and monoclonal anti-ISL1 diluted 1:100
(Developmental Studies Hybridoma Bank).
Neurogenin 3 and PDX1 antigens were produced by inserting the
coding sequence for the N-terminal 95 amino acids (neurogenin3) and
the C-terminal 80 amino acids (PDX1) from the mouse genes
downstream of the glutathione-S-transferase coding sequence in the
pGEX-2T vector (Pharmacia). The resulting fusion proteins were
purified from E. coli and injected into rabbits (neurogenin 3) and
guinea pigs (PDX1). Rabbit anti-neurogenin 3 was used at a 1:5000
dilution, and guinea pig anti-PDX1 was used at a 1:4000 dilution. Pre-
immune sera give no staining at the same concentrations.
For immunohistochemistry biotinylated anti-rabbit, anti-guinea pig
or anti-mouse antibodies (Vector) were used at a 1:200 dilution, and
detected with the ABC Elite immunoperoxidase system (Vector). The
secondary antibodies used for immunofluorescence were as follows:
FITC-conjugated anti-rabbit, anti-mouse or anti-guinea pig diluted
1:100 (Jackson Laboratory); Cy3-conjugated anti-rabbit diluted 1:200
(Jackson Laboratory); and rhodamine-conjugated anti-guinea pig
diluted 1:300 (Cappel). Fluorescence was visualized with a Zeiss
axioscope and a Leica confocal microscope.
TUNEL assays on tissue sections were performed using a
commercially available kit (Oncor). For cell counting, representative
sections throughout the entire
immunohistochemistry with an anti-insulin antibody. The average
number of insulin-positive cells per section was determined by
counting stained cells in eight sections from an individual pancreas.
organ were chosen for
Quantification of insulin concentrations
Protein was extracted as described before (Sander et al., 1997). The
concentration of insulin was determined by radioimmunoassay using
a commercially available kit (Linco).
Nkx6.1 is expressed in the developing and mature pancreas and
the central nervous system (Oster et al., 1998; Qiu et al., 1998;
Rudnick et al., 1994). In the developing mouse pancreas,
Nkx6.1 protein could be detected as early as embryonic day
10.5 (E10.5) in the majority of epithelial cells (Fig. 1A). This
pattern of broad expression within the pancreatic epithelium
persisted through E12.5 (Fig. 1B). With the start of the
secondary transition, around E13, Nkx6.1 expression became
restricted; by E15.5 it was exclusively detected in insulin-
expressing cells and scattered ductal and periductal cells (Fig.
At E15.5, near the peak of new β-cell formation in the fetal
pancreas, all Nkx6.1-expressing cells co-expressed Nkx2.2,
although many Nkx2.2-positive cells did not express Nkx6.1
(Fig. 2). At the same embryonic age, Nkx6.1 expression
partially overlapped with the expression of two other
pancreatic transcription factors: the pancreatic-duodenal
homeodomain factor PDX1 and the bHLH factor neurogenin
3, a marker of islet cell progenitors (Schwitzgebel et al., 2000;
Since PDX1 and neurogenin 3 are not found in the same
cells at this stage in pancreatic development (Schwitzgebel et
al., 2000), these two factors define two populations of Nkx6.1-
expressing cells: immature progenitor cells co-expressing
neurogenin 3, and more mature cells co-expressing insulin and
PDX1. Many of the insulin-negative/Nkx6.1-positive cells also
expressed a marker of replicating cells, proliferating cell
nuclear antigen (PCNA) (Fig. 2). PCNA was not detected in
the insulin-expressing cells at this stage (data not shown).
At later developmental stages, and in the adult pancreas,
Nkx6.1 became completely restricted to insulin-expressing
cells (Fig. 1 and data not shown). In this regard Nkx6.1 is
unique: no other transcriptional regulators are known to be
restricted exclusively to the β-cells within the pancreas.
To test the role of Nkx6.1 in pancreatic development, we
generated mice with a null allele for Nkx6.1 by homologous
recombination in ES cells. A portion of exon 1 including the
translation initiation site was deleted and replaced by a PGK-
neomycin resistance cassette.
At E18.5, one day prior to birth, the pancreases of
homozygous Nkx6.1 mutant embryos were normal in size and
gross appearance. Histologically, the pancreatic islets were
smaller than normal, but the endocrine cells were organized
into islet-like clusters. Immunohistochemistry, however,
revealed a profound deficiency of insulin-expressing cells in
Nkx6.1 mutant embryos at E18.5 (Fig. 3 and Table 1). No
differences between wild type and Nkx6.1 mutant embryos
could be detected in the expression of glucagon, somatostatin
M. Sander and others
5535Homeobox gene Nkx6.1 in the pancreas
and PP (Fig. 3). When measured by radioimmunoassay in
whole pancreas extracts at E18.5, insulin content was similarly
decreased: the insulin content of Nkx6.1 mutant pancreases was
2% of the amount in wild-type pancreases (0.17±0.01 versus
7.6±0.24 µg/mg of protein).
The deficiency in β-cells first appeared at the start of the
secondary transition. In both wild-type and mutant embryos,
the first glucagon-expressing cells appeared at E9.5 in the
dorsal pancreatic bud, and insulin-producing cells appeared a
day later (data not shown). Until E12.5, the pancreatic buds
of wild type and Nkx6.1 mutant embryos contained similar
numbers of insulin-producing cells (Fig. 3 and Table 1). While
the number of insulin-expressing cells increases exponentially
after E13 in wild-type embryos, Nkx6.1 mutant embryos failed
to show this expansion of the β-cell population (Fig. 3 and
The normal initial development of insulin-expressing cells
suggests that Nkx6.1 may be exclusively required by a second
embryonic phase of β-cell neogenesis. Unlike the β-cells that
form after E13, early-generated insulin-producing cells lack
PDX1 and Nkx6.1 expression (Oster et al., 1998) (data not
shown), and may co-express glucagon (Teitelman et al., 1993).
These early insulin-expressing cells were unaffected in Nkx6.1
mutant embryos, demonstrating the presence of two distinct
pathways for generating insulin-expressing cells in the
Fig. 1. Initially broad expression of Nkx6.1
becomes restricted after the secondary
transition. Immunoperoxidase staining
(brown) shows nuclear Nkx6.1 expression
in the majority of pancreatic epithelial cells
at E10.5 (A) and E12.5 (B). By double
immunofluorescence staining at E15.5,
Nkx6.1 (green) is co-expressed with insulin
(orange) (C), but not with glucagon
(orange) (D). Nkx6.1 is also detected in
some non-hormone-expressing cells at
E15.5. In the adult pancreas, Nkx6.1 (green)
is detected exclusively in insulin-expressing
cells (orange) (E), and is not co-expressed
with glucagon (orange) (F), or somatostatin
Fig. 2. There are two populations of Nkx6.1-expressing cells at
E15.5. The expression of Nkx6.1 (green) and the expression of other
islet transcription factors (red) partially overlap (cells with yellow
appearing nuclei) at E15.5. Note that a number of Nkx6.1-expressing
cells do not express PDX1 (green appearing cells in A), whereas all
Nkx6.1 expressing cells co-express Nkx2.2 (B). Many of the Nkx6.1-
expressing cells that do not express PDX1 express neurogenin3
(C) and the proliferating cell nuclear antigen PCNA (D). At this
same time, the acinar cells are replicating rapidly, as demonstrated by
the PNA labeling of an acinus (Ac) in D.
Table 1. Number of insulin-expressing cells in wild type
and Nkx6.1 mutant embryos
Number of insulin-positive cells/section
The mean number of insulin-positive cells per section (detected by
immunostaining) was determined by counting stained cells in 12
representative sections through an individual pancreas from wild-type (+/+)
or mutant (−/−) embryos. The numbers shown represent the mean
number±standard error of the mean of insulin-positive cells per section from
four independent pancreases.
The few β-cells present in the Nkx6.1 mutant
embryos at E18.5 appeared to be fully mature:
in addition to insulin, they expressed islet
amyloid polypeptide, prohormone convertases
1/3 and 2, and the transcription factors PDX1
and Nkx2.2 (Fig. 3 and data not shown). Two
transcription factors associated with mature
islet cells, Pax6 and Isl1, were also expressed
normally in the Nkx6.1 mutant pancreas (data
not shown), demonstrating that Nkx6.1 is not
required for the expression of any of these
The presence of Nkx2.2 in the Nkx6.1
mutant pancreas suggests that Nkx6.1 functions downstream of
Nkx2.2 in β-cell development. To establish the genetic
hierarchy between Nkx6.1 and Nkx2.2, we generated embryos
homozygous for disruptions of both genes. Nkx2.2 single
mutant embryos do not form any insulin-expressing cells in
the pancreas throughout development, and instead accumulate
incompletely differentiated β-cells. Despite the absence of
any of the four classic islet hormones, insulin, glucagon,
somatostatin and PP, these incompletely differentiated cells can
be identified by their expression of the pan-islet prohormone
convertase PC2, as well as their abundant expression of PC1/3
and IAPP (Sussel et al., 1998; Fig. 4).
The pancreatic phenotype of Nkx6.1/2.2 double mutant
embryos was indistinguishable from the Nkx2.2 single mutant
phenotype (Figs 4, 5). In Nkx6.1 single mutant pancreases,
cells expressing the four classic islet hormones accounted for
all IAPP-positive and PC2-positive cells. By contrast, identical
single mutants, Nkx6.1/2.2
accumulated hormone-negative islet cells that expressed
IAPP and PC2 in their pancreases. The failure of the Nkx6.1
gene mutation to affect the development of the pancreas in
Nkx2.2 mutant embryos demonstrates that Nkx6.1 functions
genetically downstream of Nkx2.2 in pancreatic development.
Unlike the Nkx2.2 mutant embryos, the Nkx6.1 single mutant
embryos did not accumulate incompletely differentiated β-cell
precursors. Nor was there an increase in any of the other
endocrine cell types to suggest that β-cell precursors deviate to
an alternate cellular fate (Figs 3, 5), as seen in mice lacking
the paired-homeodomain transcription factor Pax4 (Sosa-
Pineda et al., 1997). At the same time, β-cell precursors formed
at a normal rate, since the mutant embryos had normal
pancreatic expression of neurogenin 3 (Fig. 3), a marker for
early stage precursors, cells at the earliest step of the islet cell
differentiation pathway (Apelqvist et al., 1999; Gradwohl et
al., 2000; Schwitzgebel et al., 2000; Fig. 3B). Therefore, the
cells are lost at a later stage in the pathway, either as late-stage
β-cell precursors (cells that no longer express neurogenin 3 but
do not yet express insulin) or as completely differentiated,
To distinguish these two possibilities, we compared the rate
of new β-cell formation after the secondary transition in wild-
type and Nkx6.1 mutant embryos. To label newly formed β-
cells, pregnant mice were injected with BrdU five times per
day between E14.5 and E16. At E16.5, the number of insulin-
expressing cells labeled with BrdU was greatly reduced in the
Nkx6.1 mutant embryos compared with their wild-type
littermates (Table 2), indicating that β-cell formation is greatly
reduced in the Nkx6.1 mutant embryos. During this period, the
replication rate of pre-existing β-cells is extremely low, as
evidenced by the 1 hour BrdU labeling of insulin-expressing
M. Sander and others
Fig. 3. The number of insulin-expressing cells is
markedly reduced in the pancreases of Nkx6.1
homozygous mutant embryos, starting with the
onset of the secondary transition. Until E12.5,
immunoperoxidase staining (brown) shows that
similar numbers of cells express insulin (A,E), and
glucagon (B,F) in the pancreatic buds of wild-type
(A,B) and Nkx6.1 mutant (E,F) pancreases. By
E14.5, the number of insulin-expressing cells
markedly increases in pancreases of wild type (C),
but not of Nkx6.1 mutant (G) embryos, while the
number of glucagon-expressing cells increases in
both (D,H). In term embryos, few insulin-
expressing cells (red) are detected in pancreases of
Nkx6.1 mutants (M), while glucagon expressing
cells (green) appear in normal numbers (I,M).
Insulin-expressing cells (orange) in both wild type
and Nkx6.1 mutant pancreases express PDX1
(green) (J,N) and Nkx2.2 (green) (K,O). The
number and pattern of cells expressing neurogenin
3 (brown) are the same in both wild-type (L) and
Nkx6.1 mutant (P) embryos, indicating that the
initial steps of the β-cell differentiation pathway are
5537Homeobox gene Nkx6.1 in the pancreas
cells at E14.5 and E15.5. Significant β-cell replication was not
detected until E18.5, when the replication rate was the same in
wild type and mutant embryos. Taken together, the data suggest
that the mutant embryos have a decrease in β-cell neogenesis,
and that this decrease is due to a defect in the late-stage
precursors that have progressed beyond the neurogenin 3 stage
Finally, to exclude increased cell death as a cause for β-cell
loss, we tested directly the rate of apoptosis in the fetal
pancreas by the TUNEL assay. Embryonic Nkx6.1 mutant
pancreases had no gross increase in apoptotic cells relative to
wild-type littermates (Table 3). This low rate of apoptosis
cannot account for a substantial loss of mature β-cells, adding
further support to a role for Nkx6.1 in β-cell neogenesis.
In this study we have determined the function of Nkx6.1 in
pancreatic development. Nkx6.1 was expressed in three
populations of cells during pancreatic development: first
broadly in the undifferentiated epithelial cells of the early
pancreatic buds; then, after the secondary transition, in a
subset of proliferating islet cell progenitors; and finally in the
differentiated β-cells. Deletion of the Nkx6.1 gene in mice
caused a marked reduction of β-cells after the secondary
transition. We excluded a role for Nkx6.1 in the formation of
islet cell progenitors or in the proliferation or survival of the
differentiated β-cells, and propose instead that the loss of
Fig. 4. Nkx2.2 single mutants and Nkx2.2
/Nkx6.1 double mutants accumulate
incompletely differentiated islet cells.
Fluorescent staining is shown for insulin,
glucagon, somatostatin and PP (IGSPP,
green) and prohormone convertase 2
(PC2, red, A-D) and IAPP (red, panels
E-H) on pancreatic sections of wild-type,
Nkx6.1 mutant, Nkx2.2 mutant, and
Nkx6.1/Nkx2.2 double mutant
homozygous embryos at E18.5. Cells co-
expressing one of the four classic islet
hormones and PC2 or IAPP appear
yellow. Cells expressing PC2 or IAPP but
lacking any of the four classic islet
hormones appear red and are seen only in
the Nkx2.2 mutant, and Nkx6.1/Nkx2.2
double mutant homozygous embryos.
Islet architecture is also similarly
disordered in Nkx2.2 mutant, and
Nkx6.1/Nkx2.2 double mutant homozygous embryos. Autofluorescent red blood cells appear as small orange cells in some panels.
Fig. 5. Nkx2.2 single mutants and Nkx2.2 /Nkx6.1 double mutants
have the same expression patterns for islet hormones in the pancreas.
Immunoperoxidase staining (brown) is shown for the markers
indicated on pancreatic sections of wild-type, Nkx6.1 mutant, Nkx2.2
mutant, and Nkx6.1/Nkx2.2 double mutant homozygous embryos at
E18.5. Islets structure is intact in the wild-type and Nkx6.1 mutant
embryos, but is disrupted in the Nkx2.2 mutant and Nkx6.1/Nkx2.2
double mutant embryos. The number of insulin-expressing cells is
reduced in Nkx6.1 mutants, but no insulin-expressing cells are
detected in Nkx2.2 and Nkx6.1/Nkx2.2 mutants. Nkx2.2 and
Nkx6.1/Nkx2.2 mutants show a similar reduction in glucagon- and
PP-expressing cells. Cells arrested in differentiation are marked by
IAPP expression in the absence of insulin or glucagon in Nkx2.2 and
Nkx6.1/Nkx2.2 mutant embryos; these cells are not seen in wild-type
or Nkx6.1 mutant embryos. The number of somatostatin-expressing
cells is similar in embryos of all four genotypes.
Nkx6.1 affects cells in transition from early neurogenin 3-
expressing precursors to differentiated β-cells. Because
Nkx2.2/Nkx6.1 double mutants display the same phenotype as
Nkx2.2 single mutants, we conclude that Nkx6.1 functions
downstream of Nkx2.2 in pancreatic development. Our
analysis defines a function for Nkx6.1 downstream of Nkx2.2
in the expansion and final differentiation of β-cell
Distinct pathways for β-cell formation
In the developing pancreas, the first insulin-expressing cells
appear as early as the bud stage, but significant β-cell numbers
cannot be detected until E13. These early and late cells
expressing insulin appear to represent two distinct cell
populations. The early insulin-expressing cells in the
pancreatic bud produce low levels of insulin (Pictet and Rutter,
1972), and often co-express glucagon (Alpert et al., 1988;
Teitelman et al., 1993), but do not express the transcription
factors PDX1 and Nkx6.1, or the glucose transporter GLUT2
(Oster et al., 1998; Pang et al., 1994). In contrast to these early
cells, high insulin-production, as well as Pdx1, Nkx6.1 and
GLUT2 expression, characterize the mature β-cells that form
in much larger numbers after E13.
It has been suggested that the early insulin/glucagon co-
expressing cells represent a transition state as cells differentiate
from the early glucagon expressing cells into mature β-cells
(Alpert et al., 1988; Teitelman et al., 1993). Our data do not
support this model. Because we do not detect significant
replication of insulin-expressing cells prior to E18, maturation
of the small early population of insulin/glucagon co-expressing
cells cannot explain the much larger number of mature β-cells
that appear after E13. Instead we propose that most of the β-
cells in the late fetal pancreas must develop from non-
hormone-expressing progenitor cells. In support of this
conclusion, Herrera and colleagues have used lineage tracing
in transgenic mouse models to show that the glucagon
promoter is not active in the progenitor cells for mature β-cells
(Herrera et al., 1994, 1998; Herrera, 2000).
In Nkx6.1 mutant embryos, the development of the early
insulin-expressing cells remains unaffected, but β-cell numbers
fail to increase with the secondary transition. Similarly,
formation of the early population requires neither Pax4 (Sosa-
Pineda et al., 1997) nor PDX1 (Offield et al., 1996), although
Pax4 is required for formation of the late population of β-cells
(Sosa-Pineda et al., 1997), and PDX1 is required for their
maintenance (Ahlgren et al., 1998). Together, these findings
suggest that distinct genetic programs control the formation of
the early and late populations of insulin-producing cells during
embryogenesis. Analogous dual pathways have been described
in the development of thyrotrope cells of the pituitary gland.
In the pituitary, an early population develops independently of
the Pou-homeodomain transcription factor Pit1, whereas
thyrotropes arising later in embryogenesis require Pit1 for their
development (Lin et al., 1994).
In late embryogenesis and early postnatal life, as the wave
of β-cell neogenesis that initiated with the secondary transition
starts to wane, the proliferation of pre-existing β-cells produces
a further increase in β-cell numbers (Finegood et al., 1995).
Supporting this notion, we observed BrdU uptake in 3% of all
β-cells at E18.5. β-cell proliferation, the third wave of β-cell
formation, does not require Nkx6.1. Those few β-cells that do
arise during the secondary transition in Nkx6.1 mutant embryos
enter the pool of replicating cells at the same rate as in wild-
M. Sander and others
Fig. 6. A proposed model for the major β-cell differentiation
pathway. Expression of neurogenin 3 commits endodermal
progenitors to an endocrine fate. Nkx2.2 then allows these
progenitors to progress to the Nkx6.1 expressing stage. In the
absence of Nkx2.2, progenitors committed to a β-cell fate fail to
differentiate into β-cells, and instead develop into incompletely
differentiated cells, but stable cells characterized by IAPP
expression. Downstream of Nkx2.2, Nkx6.1 ensures the expansion
and progression of Nkx2.2-expressing β-cell precursors to mature β-
cells. In the absence of Nkx6.1, most of the late stage precursors are
lost, and in contrast to precursors in Nkx2.2 mutants, do not adopt an
arrested or alternate fate.
Table 2. β-cell proliferation and neogenesis in wild-type and Nkx6.1 mutant embryos
Number of BrdU/insulin co-positive cells/section‡
Embryonic day (BrdU injection)*
E16.5 (E14.5- E16)¶ 5.0±0.3
E14.5 (1 hour)**0.2±0.1
E15.5 (1 hour)**0.06±0.06
E18.5 (1 hour)** 14±1
*Pregnant mice were injected intraperitoneally with 0.5 µg of BrdU per gram of body weight every 5 hours between E14.5 and E16, and then embryos were
harvested 12 hours after the last injection¶, or were injected one hour prior to harvesting the embryos**. The first experiment labels precursor cells, the second
experiment labels preexisting β-cells.
‡Pancreatic sections were co-immuno-stained with an anti-insulin and an anti-BrdU antibody. The mean number of insulin-positive and BrdU/insulin co-
positive cells per section was determined by counting stained cells in 10 representative sections through four pancreases for each data point. The mean±standard
error of the mean is shown. When precursors are labeled (labeled E14.5-E16 and harvested E16.5), the total number of BrdU/insulin co-positive cells per section
in the Nkx6.1 mutant homozygotes (−/−) is markedly reduced relative to wild type(+/+), indicating that a reduced number of precursors are progressing to mature
§The mean percentage of insulin-positive cells that are BrdU co-positive±standard error of the mean is shown. A significant decrease in the half-life of insulin-
expressing cells would cause a decrease in the average age of the cells and therefore an increase in the percentage of newly formed cells, as indicated by BrdU
labeling. The similarity in percentage of BrdU-labeled cells in wild-type and Nkx6.1 mutant homozygous embryos at E16.5 indicates that the half-life of the β-
cells is unaffected by the absence of Nkx6.1.
No. of BrdU positive/ insulin positive cells (%)§
5539Homeobox gene Nkx6.1 in the pancreas
type embryos. To date, little is known about the factors
involved in perinatal β-cell proliferation, although a recent
study by Miettinen et al. (2000) has implicated EGF signaling.
Function of Nkx6.1 in β-cell neogenesis
Defects at several points in the progression of cells from
undifferentiated epithelial precursors to mature β-cells could
account for the reduction in β-cell numbers seen in Nkx6.1
mutant embryos. Progenitor cells that have initiated the
program of endocrine differentiation can be identified by their
transient expression of the pro-endocrine bHLH transcription
factor neurogenin 3, which is expressed in a subset of epithelial
cells prior to expression of any of the hormone genes
(Apelqvist et al., 1999; Gradwohl et al., 2000; Jensen et al.,
2000; Schwitzgebel et al., 2000). The pancreases of Nkx6.1
mutant embryos contain normal numbers of neurogenin 3-
expressing cells during the secondary transition, suggesting
that the loss of Nkx6.1 does not impact the initiation of
endocrine cell differentiation and the generation of neurogenin
3-expressing progenitor cells (Fig. 6).
It is conceivable that increased cell death of differentiated β-
cells could account for the reduction in β-cell numbers in
Nkx6.1 mutant embryos. However, our data provide no
evidence that Nkx6.1 functions to maintain β-cells once they
have differentiated. Both the BrdU labeling data in Table 2 and
the apoptosis studies detailed in Table 3 demonstrate that
excess loss of differentiated β-cells cannot account for the
reduction in β-cell numbers seen in Nkx6.1 mutant embryos.
Therefore, Nkx6.1 functions at some point in the
differentiation pathway after the generation of neurogenin 3-
expressing progenitors but prior to the appearance of
But what happens in the Nkx6.1 mutants to the neurogenin
3-expressing cells that are normally destined to become β-
cells? We have no evidence that the absence of Nkx6.1 causes
β-cell precursors to deviate to an alternate endocrine cell fate,
as occurs in embryos lacking Pax4 (Sosa-Pineda et al., 1997).
Nor do incompletely differentiated β-cell precursors
accumulate in the pancreas, as occurs in embryos lacking
Nkx2.2 (Sussel et al., 1998). It appears instead that Nkx6.1 is
required for maintaining and expanding the population of β-
cell precursors as they progress from neurogenin 3-expressing
progenitors to differentiated β-cells. In the absence of Nkx6.1,
most of these β-cell precursors are lost. This loss could result
from early apoptosis, prior to the expansion of the β-cell
precursor population, or from a block in the replication of β-
cell precursors, or from a combination of both defects, thereby
preventing the normal expansion of the β-cell population
starting at E13. In the Nkx2.2 mutants, β-cell precursors
apparently escape this fate, possibly because the absence of
Nkx2.2 prevents them from reaching the Nkx6.1-dependent
Hierarchy of transcription factors
During embryonic development, a cascade of transcriptional
events controls β-cell formation in the pancreas. Different
transcription factors control distinct checkpoints along the
pathway to the differentiated β-cells. The first step of
pancreatic epithelial cells towards an endocrine fate is
controlled by neurogenin 3. Loss of neurogenin 3 function
in mice results in a complete absence of endocrine cell
differentiation (Gradwohl et al., 2000). We have shown that
neurogenin 3 expression in the pancreas does not depend on
Several transcription factors control the subsequent
differentiation of precursor cells into the four islet cell types.
Loss of function studies have shown that most transcription
factors control differentiation of more than one islet cell
type. Isl1, Pax6 and NeuroD (Neurod1 – Mouse Genome
Informatics), for example, affect the development of all
pancreatic endocrine cells (Ahlgren et al., 1997; Naya et al.,
1997; Sander et al., 1997; St-Onge et al., 1997). Other factors,
such as Pax4 and Nkx2.2 predominantly control the
differentiation of β-cells (Sosa-Pineda et al., 1997; Sussel et
In contrast to Pax4 and Nkx6.1, Nkx2.2 is required for the
development of both the early insulin-expressing cells and the
later mature β-cells (Sussel et al., 1998). In the absence of
Nkx2.2, β-cell differentiation is arrested, leading to the
accumulation of IAPP-positive, but insulin-negative islet cells.
The presence of these incompletely differentiated islet cells has
led to the hypothesis that Nkx2.2 functions in the later steps of
β-cell differentiation. In Nkx6.1 single mutant pancreases, β-
cells can escape the complete block in differentiation that
occurs in Nkx2.2 mutant embryos, apparently because
expression of Nkx2.2 is maintained.
While Nkx2.2 expression is maintained in Nkx6.1 mutant
pancreases, Nkx6.1 expression is lost after the secondary
transition in Nkx2.2 mutant pancreases, suggesting that Nkx6.1
may function downstream of Nkx2.2 in β-cell development. In
addition, the pancreatic phenotype of Nkx2.2/Nkx6.1 double
mutant embryos is indistinguishable from Nkx2.2 single mutant
embryos, demonstrating an epistatic relationship of Nkx6.1
downstream of Nkx2.2 in β-cell differentiation (Fig. 6).
In summary, our analysis establishes a requirement for
Nkx6.1 downstream of Nkx2.2 in pancreatic β-cell
development. After E13, the rate of β-cell differentiation
increases exponentially. In this wave of β-cell neogenesis,
Nkx6.1 mediates the expansion and final differentiation of β-
cell progenitors. Our experiments outline three pathways for
the formation of insulin-expressing cells under the control of
distinct molecular programs: an early minor pathway involving
Nkx2.2 but not Pax4, PDX1 or Nkx6.1; the major pathway for
neogenesis during the secondary transition that requires
Nkx6.1; and the late proliferation of pre-existing β-cells.
We thank D. Hanahan for performing the blastocyst injections. We
also thank J. Wang and Y. Zhang for excellent technical assistance,
and M. Hebrok and members of the German laboratory for helpful
Table 3. Apoptosis in wild type and Nkx6.1 mutant
Number of apoptotic cells/section*
*Pancreatic sections were assayed by TUNEL assay for apoptotic cells.
The mean number of apoptotic cells per section was determined by counting
stained cells in 10 representative sections through four pancreases for each
data point. The mean±standard error of the mean is shown.
discussion. This work is supported by grants from the Juvenile
Diabetes Foundation International (M. S. and V. S.), the Deutsche
Forschungsgemeinschaft (M. S.) and the National Institutes of Health
(DK41822 and DK21344, M. G.; NRSA NS10184, L. S.).
Ahlgren, U., Jonsson, J., Jonsson, L., Simu, K. and Edlund, H. (1998).
β-cell-specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of
the beta-cell phenotype and maturity onset diabetes. Genes Dev. 12, 1763-
Ahlgren, U., Pfaff, S. L., Jessell, T. M., Edlund, T. and Edlund, H. (1997).
Independent requirement for ISL1 in formation of pancreatic mesenchyme
and islet cells. Nature 385, 257-260.
Alpert, S., Hanahan, D. and Teitelman, G. (1988). Hybrid insulin genes
reveal a developmental lineage for pancreatic endocrine cells and imply a
relationship with neurons. Cell 53, 295-308.
Apelqvist, A., Li, H., Sommer, L., Beatus, P., Anderson, D. J., Honjo, T.,
Hrabe de Angelis, M., Lendahl, U. and Edlund, H. (1999). Notch
signalling controls pancreatic cell differentiation. Nature 400, 877-881.
Edlund, H. (1998). Transcribing pancreas. Diabetes 47, 1817-1823.
Finegood, D. T., Scaglia, L. and Bonner-Weir, S. (1995). Dynamics of beta-
cell mass in the growing rat pancreas. Estimation with a simple
mathematical model. Diabetes 44, 249-256.
Gradwohl, G., Dierich, A., LeMeur, M. and Guillemot, F. (2000).
Neurogenin3 is required for the development of the four endocrine cell
lineages of the pancreas. Proc. Natl. Acad. Sci. USA 97, 1607-1611.
Harrison, K. A., Thaler, J., Pfaff, S. L., Gu, H. and Kehrl, J. H. (1999).
Pancreas dorsal lobe agenesis and abnormal islets of Langerhans in Hlxb9-
deficient mice. Nat. Genet. 23, 71-75.
Herrera, P.-L., Huarte, J., Zufferey, R., Nichols, A., Mermillod, B.,
Philippe, J., Muniesa, P., Sanvito, F., Orci, L. and Vassalli, J.-D.
(1994). Ablation of islet endocrine cells by targeted expression of
hormone-promoter-driven toxigenes. Proc. Natl. Acad. Sci. USA 91,
Herrera, P. L. (2000). Adult insulin- and glucagon-producing cells
differentiate from two independent cell lineages. Development 127, 2317-
Herrera, P. L., Orci, L. and Vassalli, J. D. (1998). Two transgenic approaches
to define the cell lineages in endocrine pancreas development. Mol. Cell.
Endocrinol. 140, 45-50.
Jensen, J., Heller, R. S., Funder-Nielsen, T., Pedersen, E. E., Lindsell, C.,
Weinmaster, G., Madsen, O. D. and Serup, P. (2000). Independent
development of pancreatic alpha- and beta-cells from neurogenin3-
expressing precursors: a role for the notch pathway in repression of
premature differentiation. Diabetes 49, 163-176.
Jensen, J., Serup, P., Karlsen, C., Nielsen, T. F. and Madsen, O. D. (1996).
mRNA profiling of rat islet tumors reveals nkx 6.1 as a beta-cell-specific
homeodomain transcription factor. J. Biol. Chem. 271, 18749-18758.
Jonsson, J., Carlsson, L., Edlund, T. and Edlund, H. (1994). Insulin-
promoter-factor 1 is required for pancreas development in mice. Nature 371,
Kash, S. F., Johnson, R. S., Tecott, L. H., Noebels, J. L., Mayfield, R. D.,
Hanahan, D. and Baekkeskov, S. (1997). Epilepsy in mice deficient in the
65-kDa isoform of glutamic acid decarboxylase. Proc. Natl. Acad. Sci. USA
Li, H., Arber, S., Jessell, T. M. and Edlund, H. (1999). Selective agenesis
of the dorsal pancreas in mice lacking homeobox gene Hlxb9. Nat. Genet.
Lin, S. C., Li, S., Drolet, D. W. and Rosenfeld, M. G. (1994). Pituitary
ontogeny of the Snell dwarf mouse reveals Pit-1-independent and Pit-1-
dependent origins of the thyrotrope. Development 120, 515-522.
Miettinen, P. J., Huotari, M.-A., Koivisto, T., Ustinov, J., Palgi, J.,
Rasilainen, S., Lehtonen, E., Keski-Oja, J. and Otonkoski, T. (2000).
Impaired migration and delayed differentiation of pancreatic islet cells in
mice lacking EGF-receptors. Development 127, 2617-2627.
Naya, F. J., Huang, H. P., Qiu, Y., Mutoh, H., DeMayo, F. J., Leiter, A. B.
and Tsai, M. J. (1997). Diabetes, defective pancreatic morphogenesis, and
abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice.
Genes Dev. 11, 2323-2334.
Offield, M. F., Jetton, T. L., Labosky, P. A., Ray, M., Stein, R. W.,
Magnuson, M. A., Hogan, B. L. and Wright, C. V. (1996). PDX-1 is
required for pancreatic outgrowth and differentiation of the rostral
duodenum. Development 122, 983-995.
Oster, A., Jensen, J., Serup, P., Galante, P., Madsen, O. D. and Larsson,
L. I. (1998). Rat endocrine pancreatic development in relation to two
homeobox gene products (Pdx-1 and Nkx 6.1). J. Histochem. Cytochem. 46,
Pang, K., Mukonoweshuro, C. and Wong, G. G. (1994). Beta cells arise
from glucose transporter type 2 (Glut2)-expressing epithelial cells of the
developing rat pancreas. Proc. Natl. Acad. Sci. USA 91, 9559-9563.
Pictet, R. and Rutter, W. J. (1972). Development of the embryonic endocrine
pancreas. In Handbook of Physiology. Vol. 1 (eds A. P. Society, D. F. Steiner
and N. Frenkel), pp. 25-66. Washington DC: Williams and Wilkins.
Qiu, M., Shimamura, K., Sussel, L., Chen, S. and Rubenstein, J. L. (1998).
Control of anteroposterior and dorsoventral domains of Nkx-6.1 gene
expression relative to other Nkx genes during vertebrate CNS development.
Mech. Dev. 72, 77-88.
Rudnick, A., Ling, T. Y., Odagiri, H., Rutter, W. J. and German, M. S.
(1994). Pancreatic beta cells express a diverse set of homeobox genes. Proc.
Natl. Acad. Sci. USA 91, 12203-12207.
Sander, M. and German, M. S. (1997). The β-cell transcription factors and
development of the pancreas. J. Mol. Med. 15, 327-340.
Sander, M., Neubuser, A., Kalamaras, J., Ee, H. C., Martin, G. R. and
German, M. S. (1997). Genetic analysis reveals that PAX6 is required for
normal transcription of pancreatic hormone genes and islet development.
Genes Dev. 11, 1662-1673.
Schwitzgebel, V. M., Scheel, D. W., Conners, J. R., Kalamaras, J. R., Lee,
J. E., Anderson, D. J., Sussel, L., Johnson, J. D. and German, M. S.
(2000). Expression of neurogenin3 reveals an islet cell precursor population
in the pancreas. Development 127, 3533-3542.
Slack, J. M. (1995). Developmental biology of the pancreas. Development
Sosa-Pineda, B., Chowdhury, K., Torres, M., Oliver, G. and Gruss, P.
(1997). The Pax4 gene is essential for differentiation of insulin-producing
beta cells in the mammalian pancreas. Nature 386, 399-402.
St-Onge, L., Sosa-Pineda, B., Chowdhury, K., Mansouri, A. and Gruss, P.
(1997). Pax6 is required for differentiation of glucagon-producing alpha-
cells in mouse pancreas. Nature 387, 406-409.
Sussel, L., Kalamaras, J., Hartigan-O’Connor, D. J., Meneses, J. J.,
Pedersen, R. A., Rubenstein, J. L. and German, M. S. (1998). Mice
lacking the homeodomain transcription factor Nkx2.2 have diabetes due to
arrested differentiation of pancreatic beta cells. Development 125, 2213-
Teitelman, G., Alpert, S., Polak, J. M., Martinez, A. and Hanahan, D.
(1993). Precursor cells of mouse endocrine pancreas coexpress insulin,
glucagon and the neuronal proteins tyrosine hydroxylase and neuropeptide
Y, but not pancreatic polypeptide. Development 118, 1031-1039.
Upchurch, B. H., Aponte, G. W. and Leiter, A. B. (1994). Expression of
peptide YY in all four islet cell types in the developing mouse pancreas
suggests a common peptide YY-producing progenitor. Development 120,
M. Sander and others