106
M. Maden
humerus, ulna and 2 digits present; to (c), humerus, ulna, and 4 digits present; to (d) a
complete limb; and, at highest exposure levels, (e) a limb with preaxial polydactyly.
Rescued forelimb buds expressed tbx5, meis2 and fgf-1. Fgf-8 was expressed but in
a disrupted domain due to the disruption of AER formation (as seen in the chick).
These results, as well as others described below, suggested that forelimb development does require RA for its initiation, while hindlimb bud development does not.
Later studies with raldh2/− raldh3
−/− double mutants demonstrated that hindlimb
development was still normal when the mutants were rescued with RA (Zhao et al.
2009). The corresponding gene expression domains for fgf-8, shh, tbx4, meis2, and
hoxa-11 in the hindlimb buds of the RA-rescued double mutants were also relatively
normal.
A random mutagenesis screen identified a retinol dehydrogenase10 (rdh10)
mutant known as trex that had defective forelimbs and normal hindlimbs. In this
mutant, the forelimbs are phocomelic, often with a missing radius (Sandell et al.
2007). The genes affected in this mutant (tbx5, shh and fgf-8) are those altered in the
Raldh2
−/− rescued mutants. However, the trex embryos are not as severely affected as
the raldh2
−/− mutants and can live for a longer time (up to day 14.5 days vs 9.5 days
for the raldh2
−/− mutants). Further work suggested that the original rdh10
trex mutant
was a partial knockout with residual RDH10 activity, since targeted rdh10 knockout
embryos exhibited reduced or missing forelimb buds (Sandell et al. 2012; Rhinn
et al. 2011). Rdh10
trex mutants can be rescued by maternal supplementation with
RA; nevertheless, they show no RARE-lacZ activity, implying RA is not present.
The hindlimbs of rdh10
trex mutants express normal domains of tbx4, fgf-8, meis-2
and shh, but depict interdigital webbing due to loss of RA activity (Cunningham et al.
2011). Subsequent work by Cunningham et al. showed that both meis-1 and meis-2 are
expressed normally in both the forelimb and hindlimb buds of these mutants, leading
them to conclude these proximal genes are not established by RA (Cunningham
et al. 2013). Nevertheless, recent studies suggested that meis2 is indeed involved in
patterning of the PD axis, at least in the forelimb bud, and that the polarized levels of
meis2 (proximal high, distal low) are generated by an antagonism between polycomb
group factors and RA signaling (Yakushiji-Kaminatsui et al. 2018). It has been nicely
demonstrated that the meis2 promoter is bound by RARs in the proximal forelimb
bud.
To explain the developmental differences between the forelimb and hindlimb,
an early (1-10 somite stages) RA/FGF antagonistic relationship (exactly as seen in
the zebrafish) has been proposed (Zhao et al. 2009; Cunningham et al. 2013). In
raldh2
−/− and rdh10
−/− embryos, the anterior and posterior domains of sprouty2, a
marker of FGF signaling, expand towards each other into the forelimb bud region.
An ectopic domain of fgf-8 (present in the intermediate mesoderm adjacent to the
presumptive forelimb bud region) disappears if the mutants are rescued by extended
administration of RA which permits tbx5 expression and subsequent development
of a small forelimb bud. This inhibition of limb induction by FGF contrasts with its
ability to induce a limb bud when placed in the flank between the two limbs (Cohn
et al. 1995), but this may be an issue of timing.
M. Maden
humerus, ulna and 2 digits present; to (c), humerus, ulna, and 4 digits present; to (d) a
complete limb; and, at highest exposure levels, (e) a limb with preaxial polydactyly.
Rescued forelimb buds expressed tbx5, meis2 and fgf-1. Fgf-8 was expressed but in
a disrupted domain due to the disruption of AER formation (as seen in the chick).
These results, as well as others described below, suggested that forelimb development does require RA for its initiation, while hindlimb bud development does not.
Later studies with raldh2/− raldh3
−/− double mutants demonstrated that hindlimb
development was still normal when the mutants were rescued with RA (Zhao et al.
2009). The corresponding gene expression domains for fgf-8, shh, tbx4, meis2, and
hoxa-11 in the hindlimb buds of the RA-rescued double mutants were also relatively
normal.
A random mutagenesis screen identified a retinol dehydrogenase10 (rdh10)
mutant known as trex that had defective forelimbs and normal hindlimbs. In this
mutant, the forelimbs are phocomelic, often with a missing radius (Sandell et al.
2007). The genes affected in this mutant (tbx5, shh and fgf-8) are those altered in the
Raldh2
−/− rescued mutants. However, the trex embryos are not as severely affected as
the raldh2
−/− mutants and can live for a longer time (up to day 14.5 days vs 9.5 days
for the raldh2
−/− mutants). Further work suggested that the original rdh10
trex mutant
was a partial knockout with residual RDH10 activity, since targeted rdh10 knockout
embryos exhibited reduced or missing forelimb buds (Sandell et al. 2012; Rhinn
et al. 2011). Rdh10
trex mutants can be rescued by maternal supplementation with
RA; nevertheless, they show no RARE-lacZ activity, implying RA is not present.
The hindlimbs of rdh10
trex mutants express normal domains of tbx4, fgf-8, meis-2
and shh, but depict interdigital webbing due to loss of RA activity (Cunningham et al.
2011). Subsequent work by Cunningham et al. showed that both meis-1 and meis-2 are
expressed normally in both the forelimb and hindlimb buds of these mutants, leading
them to conclude these proximal genes are not established by RA (Cunningham
et al. 2013). Nevertheless, recent studies suggested that meis2 is indeed involved in
patterning of the PD axis, at least in the forelimb bud, and that the polarized levels of
meis2 (proximal high, distal low) are generated by an antagonism between polycomb
group factors and RA signaling (Yakushiji-Kaminatsui et al. 2018). It has been nicely
demonstrated that the meis2 promoter is bound by RARs in the proximal forelimb
bud.
To explain the developmental differences between the forelimb and hindlimb,
an early (1-10 somite stages) RA/FGF antagonistic relationship (exactly as seen in
the zebrafish) has been proposed (Zhao et al. 2009; Cunningham et al. 2013). In
raldh2
−/− and rdh10
−/− embryos, the anterior and posterior domains of sprouty2, a
marker of FGF signaling, expand towards each other into the forelimb bud region.
An ectopic domain of fgf-8 (present in the intermediate mesoderm adjacent to the
presumptive forelimb bud region) disappears if the mutants are rescued by extended
administration of RA which permits tbx5 expression and subsequent development
of a small forelimb bud. This inhibition of limb induction by FGF contrasts with its
ability to induce a limb bud when placed in the flank between the two limbs (Cohn
et al. 1995), but this may be an issue of timing.