270
Xenopus
pancreatic hypoplasia/agenesis has also been thoroughly
reviewed (Salanga and Horb 2015; Kofent and Spagnoli
2016). We therefore highlight studies in Xenopus that have
defned mechanisms governing a pancreas versus liver fate
choice that occurs in multipotent ventral foregut endoderm
cells.
During neurula patterning stages, canonical Wnt/βCatenin signaling must be suppressed in foregut endoderm to allow for subsequent liver and pancreas induction
(McLin et al. 2007; Li et al. 2008). Studies in Xenopus,
mouse, zebrafsh, and human PSC-derived endoderm have
all demonstrated that ventral pancreas and liver progenitors arise from multipotent ventral foregut endoderm cells,
and prolonged FGF and BMP signaling promotes liver over
pancreatic fate (Deutsch et al. 2001; Wandzioch and Zaret
2009 ; Shifey et al. 2012; Kenny et al. 2012; Loh et al. 2014;
Twaroski et al. 2015). BMP signaling must be suppressed in
a subset of these foregut progenitors to allow for pancreatic
induction. This involves the TALE-family homeodomain
protein Tgif2, which is initially expressed in the multipotent foregut progenitors and then selectively maintained in
the pancreatic lineage. Tgif2, identifed in Xenopus as direct
target of the TF Gata5 in foregut endoderm, is required
for pdx1 expression, can physically interact with SMAD1,
and functionally suppresses BMP signaling to permit pancreas fate (Spagnoli and Brivanlou 2008; Cerdá-Esteban
et al. 2017). As RNA-seq and ChIP-seq studies identif ed
Gata5 as a direct BMP/pSMAD1 target (Stevens et al.
2017), a BMP>pSMAD1>Gata5>Tgif2—| BMP negative
feedback cascade thus regulates in pancreas induction. It
is unclear if Tgif2 is actively repressed from the Gata5expressing hepatic lineage or if additional signals, such as
non-canonical Wnt and RA signaling, both of which also
promote pancreatic fate (Stafford et al. 2004; RodríguezSeguel et al. 2013), selectively maintain tgif2 expression in
pancreatic progenitors.
Non-canonical Wnt signaling promotes pancreatic fate.
Xenopus studies demonstrated that Wnt5a/Atf2 signaling
can expand the expression domains of the master pancreatic TFs Pdx1 and Ptf1a at the expense of hhex-positive/
nr1h5-positive liver progenitors (Rodríguez-Seguel et al.
2013). Similar effects of WNT5A and the expression of noncanonical WNT pathway components have been observed
during the directed differentiation of both mouse and human
PSCs into pancreatic progenitors (Rodríguez -Seguel et al.
2013; Cebola et al. 2015).
RA is an essential, conserved signal necessary for patterning the posterior foregut and for induction of Pdx1expressing pancreas progenitors (Stafford et al. 2004; Chen
et al. 2004). Transcriptional profling in Xenopus has identifed direct RA targets, which include the TF hnf1b and the
Wnt-receptor frizzled 4 (fzd4) (Gere-Becker et al. 2018).
Hnf1b is a conserved regulator of liver and pancreatic development in vertebrates (Gere-Becker et al. 2018; Lokmane
et al. 2008; Poll et al. 2006; Kotalova et al. 2015) and heterozygous mutations in human HNF1B resulting in MODY
(OMIM #606391). RNA-seq studies in mouse also identif ed
Fzd4 expression to be enriched in pancreatic progenitors
(Rodríguez-Seguel et al. 2013), but it remains to be tested if
the non-canonical Wnt signal that promotes pancreatic fate
utilizes Fzd4. Finally, it is still unclear how the FGF, BMP,
RA, and non-canonical Wnt pathways interact in an epistatic
and combinatorial manner to regulate a pancreatic versus
hepatic GRN.
Repression of alternative fate by the master pancreatic
TFs Pdx1 and Ptf1a is also important during pancreas induction. Experimental overexpression of pdx1/ptf1a messenger
RNAs is suffcient to change the fate of Xenopus foregut
endoderm into pancreatic tissue (Afelik et al. 2006; Jarikji
et al. 2007); consistent with these observations, ChIP-seq
studies in human pancreatic progenitors revealed that PDX1
binds and represses hepatic genes to ensure pancreatic
lineage commitment (Teo et al. 2015; Wang et al. 2018).
Additional Xenopus studies have shown that the histone
methyltransferase Setd7 is necessary for pdx1 expression
and for establishment of active histone marks at pancreatic promoters (Kofent et al. 2016). How Setd7 is specif -
cally recruited to pdx1 and other pancreatic loci is currently
unclear, but an intriguing hypothesis is that FOX TF activity and/or RA signaling, both known to regulate enhancer
status, could be involved (Wang et al. 2015; Vinckier et al.
2020 ).
18.5.4. STOMACH
RA and FGF signaling during NF25–NF35 is required
for the formation of sox2-expressing stomach progenitors,
and these signals act in part via inducing expression of the
zinc-fnger transcriptional repressors Osr1 and Osr2, which
restrain BMP signaling (Rankin et al. 2012). Experimental
BMP/Smad1 gain-of-function during this developmental
window suppresses stomach progenitor induction (Rankin
et al. 2012). Conserved mechanisms that drive the curvature
of the stomach have also been identifed in Xenopus ( Davis
et al. 2017). During stomach morphogenesis from NF34–
NF39, the left stomach epithelium becomes polarized and
undergoes radial rearrangement; these asymmetries are
observed in both mouse and Xenopus and are dependent
on the left-right patterning genes foxj1, nodal1, and pitx2
(Davis et al. 2017).
18.5.5. INTESTINE
Intestinal fate is promoted by posteriorizing Wnt/BMP/FGF
signals, which cooperate to induce and maintain expression
of the master intestinal TF Cdx2 (Zorn and Wells 2009;
Stevens et al. 2017; Rankin et al. 2018). RA signaling during
NF14-NF25 acts as a molecular toggle between cdx2- positive/gata4-positive midgut and cdx2 -positive/satb2 -positive
hindgut fate (Figure 18.5B; Rankin et al. 2018; Múnera et al.
2017). Although Xenopus has a cloaca and not a colon, the
existence of gata4-positive midgut and satb2- positive distal hindgut progenitors demonstrates conserved molecular
pattern of the intestine amongst frog, mouse, and human;
Xenopus
pancreatic hypoplasia/agenesis has also been thoroughly
reviewed (Salanga and Horb 2015; Kofent and Spagnoli
2016). We therefore highlight studies in Xenopus that have
defned mechanisms governing a pancreas versus liver fate
choice that occurs in multipotent ventral foregut endoderm
cells.
During neurula patterning stages, canonical Wnt/βCatenin signaling must be suppressed in foregut endoderm to allow for subsequent liver and pancreas induction
(McLin et al. 2007; Li et al. 2008). Studies in Xenopus,
mouse, zebrafsh, and human PSC-derived endoderm have
all demonstrated that ventral pancreas and liver progenitors arise from multipotent ventral foregut endoderm cells,
and prolonged FGF and BMP signaling promotes liver over
pancreatic fate (Deutsch et al. 2001; Wandzioch and Zaret
2009 ; Shifey et al. 2012; Kenny et al. 2012; Loh et al. 2014;
Twaroski et al. 2015). BMP signaling must be suppressed in
a subset of these foregut progenitors to allow for pancreatic
induction. This involves the TALE-family homeodomain
protein Tgif2, which is initially expressed in the multipotent foregut progenitors and then selectively maintained in
the pancreatic lineage. Tgif2, identifed in Xenopus as direct
target of the TF Gata5 in foregut endoderm, is required
for pdx1 expression, can physically interact with SMAD1,
and functionally suppresses BMP signaling to permit pancreas fate (Spagnoli and Brivanlou 2008; Cerdá-Esteban
et al. 2017). As RNA-seq and ChIP-seq studies identif ed
Gata5 as a direct BMP/pSMAD1 target (Stevens et al.
2017), a BMP>pSMAD1>Gata5>Tgif2—| BMP negative
feedback cascade thus regulates in pancreas induction. It
is unclear if Tgif2 is actively repressed from the Gata5expressing hepatic lineage or if additional signals, such as
non-canonical Wnt and RA signaling, both of which also
promote pancreatic fate (Stafford et al. 2004; RodríguezSeguel et al. 2013), selectively maintain tgif2 expression in
pancreatic progenitors.
Non-canonical Wnt signaling promotes pancreatic fate.
Xenopus studies demonstrated that Wnt5a/Atf2 signaling
can expand the expression domains of the master pancreatic TFs Pdx1 and Ptf1a at the expense of hhex-positive/
nr1h5-positive liver progenitors (Rodríguez-Seguel et al.
2013). Similar effects of WNT5A and the expression of noncanonical WNT pathway components have been observed
during the directed differentiation of both mouse and human
PSCs into pancreatic progenitors (Rodríguez -Seguel et al.
2013; Cebola et al. 2015).
RA is an essential, conserved signal necessary for patterning the posterior foregut and for induction of Pdx1expressing pancreas progenitors (Stafford et al. 2004; Chen
et al. 2004). Transcriptional profling in Xenopus has identifed direct RA targets, which include the TF hnf1b and the
Wnt-receptor frizzled 4 (fzd4) (Gere-Becker et al. 2018).
Hnf1b is a conserved regulator of liver and pancreatic development in vertebrates (Gere-Becker et al. 2018; Lokmane
et al. 2008; Poll et al. 2006; Kotalova et al. 2015) and heterozygous mutations in human HNF1B resulting in MODY
(OMIM #606391). RNA-seq studies in mouse also identif ed
Fzd4 expression to be enriched in pancreatic progenitors
(Rodríguez-Seguel et al. 2013), but it remains to be tested if
the non-canonical Wnt signal that promotes pancreatic fate
utilizes Fzd4. Finally, it is still unclear how the FGF, BMP,
RA, and non-canonical Wnt pathways interact in an epistatic
and combinatorial manner to regulate a pancreatic versus
hepatic GRN.
Repression of alternative fate by the master pancreatic
TFs Pdx1 and Ptf1a is also important during pancreas induction. Experimental overexpression of pdx1/ptf1a messenger
RNAs is suffcient to change the fate of Xenopus foregut
endoderm into pancreatic tissue (Afelik et al. 2006; Jarikji
et al. 2007); consistent with these observations, ChIP-seq
studies in human pancreatic progenitors revealed that PDX1
binds and represses hepatic genes to ensure pancreatic
lineage commitment (Teo et al. 2015; Wang et al. 2018).
Additional Xenopus studies have shown that the histone
methyltransferase Setd7 is necessary for pdx1 expression
and for establishment of active histone marks at pancreatic promoters (Kofent et al. 2016). How Setd7 is specif -
cally recruited to pdx1 and other pancreatic loci is currently
unclear, but an intriguing hypothesis is that FOX TF activity and/or RA signaling, both known to regulate enhancer
status, could be involved (Wang et al. 2015; Vinckier et al.
2020 ).
18.5.4. STOMACH
RA and FGF signaling during NF25–NF35 is required
for the formation of sox2-expressing stomach progenitors,
and these signals act in part via inducing expression of the
zinc-fnger transcriptional repressors Osr1 and Osr2, which
restrain BMP signaling (Rankin et al. 2012). Experimental
BMP/Smad1 gain-of-function during this developmental
window suppresses stomach progenitor induction (Rankin
et al. 2012). Conserved mechanisms that drive the curvature
of the stomach have also been identifed in Xenopus ( Davis
et al. 2017). During stomach morphogenesis from NF34–
NF39, the left stomach epithelium becomes polarized and
undergoes radial rearrangement; these asymmetries are
observed in both mouse and Xenopus and are dependent
on the left-right patterning genes foxj1, nodal1, and pitx2
(Davis et al. 2017).
18.5.5. INTESTINE
Intestinal fate is promoted by posteriorizing Wnt/BMP/FGF
signals, which cooperate to induce and maintain expression
of the master intestinal TF Cdx2 (Zorn and Wells 2009;
Stevens et al. 2017; Rankin et al. 2018). RA signaling during
NF14-NF25 acts as a molecular toggle between cdx2- positive/gata4-positive midgut and cdx2 -positive/satb2 -positive
hindgut fate (Figure 18.5B; Rankin et al. 2018; Múnera et al.
2017). Although Xenopus has a cloaca and not a colon, the
existence of gata4-positive midgut and satb2- positive distal hindgut progenitors demonstrates conserved molecular
pattern of the intestine amongst frog, mouse, and human;