Download A developmental switch induced by thyroid hormone: Xenopus

Review
TRENDS in Endocrinology and Metabolism
Vol.17 No.2 March 2006
A developmental switch induced by
thyroid hormone: Xenopus laevis
metamorphosis
J. David Furlow and Eric S. Neff
Section of Neurobiology, Physiology, and Behavior, University of California, Davis, One Shields Avenue, Davis, CA 95616-8519, USA
Thyroid hormone induces the complete metamorphosis
of anuran tadpoles into juvenile frogs. Arguably, anuran
metamorphosis is the most dramatic effect of a hormone
in any vertebrate. Recent advances in pharmacology and
molecular biology have made the study of this remarkable
process in the frog Xenopus laevis attractive to developmental biologists and endocrinologists alike. In particular,
the availability of a straightforward transgenesis assay
and the near completion of the Xenopus tropicalis
genome are enabling significant advances to be made in
our understanding of the major remaining problems of
metamorphosis: the extraordinary tissue specificity of
responses, the precise timing of morphological changes,
the degree of cell autonomy of hormone responses and
developmental competence. We argue that X. laevis
metamorphosis presents an exciting opportunity for
understanding the role of thyroid hormone in vertebrate
development.
Introduction
Almost 100 years ago, J.F. Gudernatsch made the
remarkable discovery that equine thyroid extracts could
accelerate the metamorphosis of tadpoles into juvenile
frogs [1]. Subsequent work demonstrated that removal of
the tadpole thyroid gland or treatment with inhibitors of
thyroid hormone (TH) synthesis prevents metamorphosis.
Indeed, THs that are identical to those produced by the
human thyroid gland, thyroxine (T4) and the more potent
3,5, 3 0 -triiodothyronine (T3), are detected in increasing
amounts as tadpole metamorphosis proceeds [2].
A hypothalamic–pituitary–thyroid axis controls the production of TH, in the same way as in mammals, although
amphibians appear to use corticotropin-releasing hormone instead of thyrotropin-releasing hormone to induce
pituitary thyroid-stimulating hormone (TSH) secretion at
the onset of metamorphosis [3]. Since the findings of
Gudernatsch and others established TH as the developmental signal that triggers the onset of metamorphosis,
the system has attracted investigators from various
disciplines, including those studying life cycle evolution,
environmental toxicology and basic mechanisms of TH
action. The central question of how such a simple
Corresponding author: Furlow, J.D. ([email protected]).
Available online 7 February 2006
molecule, two coupled tyrosines decorated by iodines,
could coordinate such a wide range of functional and
morphological changes remains a fascinating issue.
Most anurans (frogs and toads) undergo indirect
development, hatching as free-living larvae capable of
substantial growth but incapable of reproduction or
survival in a nonaquatic environment. Upon metamorphosis into juvenile frogs, virtually every organ system
undergoes extensive morphological and/or functional
changes as circulating TH levels rise (Box 1). These
changes can be grouped into three categories: (i) death and
resorption of larval tissues used only by the tadpole; (ii) de
novo growth and differentiation of new adult tissues; and
(iii) remodeling of larval tissues to create a new adult
function [4]. The two most visually striking changes are
complete resorption of the tail and the formation of the
hind- and fore-limbs from nests of undifferentiated
mesenchyme. Most tissues and organs, such as the skin,
skeletal muscle, digestive tract and the central nervous
system, undergo an extensive remodeling process involving spatially and temporally coordinated apoptosis of the
larval cell component combined with proliferation and
differentiation of adult precursor cells. The degree of
‘transdifferentiation’ of larval cells into new adult cells has
not been unequivocally documented in amphibians to
date, although this clearly occurs in specific tissues during
insect metamorphosis [5].
In this review, we discuss the transcriptional control of
metamorphosis by TH via its well-conserved TH receptors
(TRs), with a focus on a comparison between the tail
resorption versus limb growth programs. Although
digestive tract remodeling has been the subject of
extensive investigation, this process has been reviewed
extensively elsewhere, so will not be discussed in detail
here [6,7]. We also revisit some major long-standing
questions that are now approachable using recently
developed molecular and pharmacological tools. Most of
this work has been performed using the South African
clawed frog Xenopus laevis. Owing to its complete external
development, conserved mechanism of TH action and
large collection of TH-response genes, X. laevis represents
an ideal system for ascertaining the developmental roles
of TH and its receptors. Importantly, a transgenesis
method in this organism has been developed that
promotes integration of DNA into sperm chromatin before
the first cleavage of the fertilized egg, so that all of the
www.sciencedirect.com 1043-2760/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tem.2006.01.007
Review
TRENDS in Endocrinology and Metabolism
Box 1. Functional and morphological changes in selected
TH-responsive tissues and organs
Target tissue or organ
Death and resorption
Example: tail
Growth and differentiation
Example: hindlimbs
Remodeling
Example: gastrointestinal tract
Morphological or functional
changes
Fibroblast invasion and
breakdown of collagen
bundles in notochord and
under skin; dissolution of myotendinous junctions; apoptosis
and fragmentation of skeletal
muscle fibers and engulfment
by phagocytes. Other example:
gills
Extensive cell proliferation in
undifferentiated limb buds;
angiogenesis; differentiation of
myofibers, cartilage, and bone;
migration of motor neurons
and formation of neuromuscular junctions; apoptosis of cells
between digits. Other
examples: forelimbs, lower jaw
Shortening of intestine by 75%;
proliferation of entire epithelial
layer, eventually restricted to
troughs of epithelial folds;
transient thickening of epithelial layer followed by apoptosis of luminal epithelia; loss
of single larval fold (typhlosole)
with development of extensive
adult epithelial folds;
thickening of connective tissue
and basal lamina; differentiation of stomach in anterior
region. Other examples:
central nervous system, skin,
kidney, liver
cells of the resulting tadpole carry the transgene in a
nonmosaic fashion [8]. Integrated reporter constructs
faithfully reproduce the pattern of expression of endogenous genes [9]; a variety of mammalian promoters are
active in transgenic X. laevis [10], and transgenes are
faithfully transmitted through the germ line [11]. In
addition, genetic resources for the Xenopus community
continue to develop at a rapid pace, including complementary DNA (cDNA) sequencing and microarray production.
A particularly exciting development is the potential of a
close relative of X. laevis, Xenopus tropicalis, to be used in
traditional genetic experiments to identify key regulatory
genes during embryogenesis and metamorphosis [12].
X. tropicalis is a diploid organism with a genome half the
size of that of the tetraploid X. laevis and achieves sexual
maturity in about four months. A first draft of the entire
X. tropicalis genome will have been completed before the
end of 2005. Thus, the Xenopus metamorphosis model
combines well-understood endocrine physiology, ease of
hormonal manipulation of the externally developing
tadpole and potent molecular tools for the discovery
of important information about tissue- and developmental
stage-specific responses to TH.
www.sciencedirect.com
Vol.17 No.2 March 2006
39
Transcriptional control of metamorphosis by thyroid
hormone
Several lines of evidence demonstrate that most, if not all,
of the morphological and functional changes of metamorphosis are the result of changes in the transcription of
specific sets of genes induced by TH. In this section, we
summarize our current understanding of TH control of
transcription during metamorphosis via the thyroid
hormone receptors and associated coactivators.
TRs are structurally and functionally conserved
Early studies by Tata [13] showed that inhibitors of RNA
or protein synthesis could block TH-induced morphological changes in organ culture. Further evidence for the role
of transcriptional control of metamorphosis comes from
the fact that TRs belong to the nuclear receptor superfamily of transcription factor genes; lipophilic ligand
binding regulates the activity of many members of the
family [14]. X. laevis, in common with all vertebrates, has
two TR gene loci, designated TR-a and TR-b. Both
X. laevis TRs (xTRs) show over 85% amino acid identity
with their human homologs, with highest conservation in
their ligand-binding and DNA-binding domains [15]
(Figure 1a). The two xTRs are also highly similar (85%
identity), with the most significant difference being in the
amino-terminal activation function (AF) 1 domains. The
(a)
1
418
xTR-αA
hTR-α1 55%
DBD
LBD
95%
91%
1
369
xTR-β (xTR-βA1)
DBD
LBD
hTR-β1
96%
94%
1
(b)
488
DBD
xRXR-α
hRXR-α
(c)
63%
LBD
97%
99%
CoRNR boxes 2498
1
xNCoR
RDI
RDII
RDIII
hNCoR
84%
40%
70%
(d)
1
xSRC-3 bHLH–PAS
hSRC-3
86%
93%
NR boxes
60%
1519
CBP
67%
70%
TRENDS in Endocrinology & Metabolism
Figure 1. TRs and coregulatory proteins are conserved from frog to man.
Comparison of the X. laevis (x) and human (h): TRs (a), RXR-a (b), NCoR (c) and
SRC-3 (d), including amino acid similarity in certain highlighted domains. The
major expressed xTR-b isoform, TR-bA1, is shown. Black bars indicate the positions
of three CoRNR boxes mediating TR–NCoR interaction (c) and the three nuclear
receptor (NR) boxes (d), which are LXXLL motifs mediating TR–SRC-3 interaction.
Amino acid positions are indicated on top of the X. laevis proteins. Abbreviations:
DBD, DNA-binding domain; LBD, ligand-binding domain; RD, repression domain;
bHLH–PAS, basic helix–loop–helix–per-arnt-sim homology.
Review
40
TRENDS in Endocrinology and Metabolism
Relative expression
mammalian TR-b gene has two major splicing isoforms,
which create new amino-terminal transactivation
domains, designated TR-b1 and TR-b2. In X. laevis, the
major expressed TR-b isoform contains a very short
amino-terminal sequence and thus is more similar to the
avian TR-b0 isoform than to mammalian TR-b1 [15]. The
mammalian TR-a2 isoform does not bind ligand, and has
been described only in mammals. In X. laevis, xTR-a is
expressed after the embryo hatches, and mRNA levels in
whole tadpoles peak several days before the rise in TH at
metamorphosis [16] (Figure 2). xTR-a is detectable in all
cells examined so far, although the highest levels are
found in those cells that proliferate upon TH treatment,
such as the limb buds, basal layers of the skin, specific
head cartilages and subventricular zones of the brain [17].
xTR-b is expressed at low levels during embryogenesis
and premetamorphosis but then is strongly induced
during metamorphosis [16]. xTR-b is highly induced in
resorbing tissues such as the tail but at relatively low
levels in most proliferating tissues [18].
TRs regulate specific target genes by binding to cisacting DNA sites known as TH response elements (TREs).
TRs can bind to TREs as monomers or homodimers but
bind most efficiently as heterodimers with another family
of nuclear receptors, the retinoid-X receptors (RXRs) [14].
X. laevis has three RXR genes, a, b and g [19,20], which
are also highly conserved with their mammalian counterparts (Figure 1b). Each RXR gene shows distinct
expression patterns [12,20] (Figure 2). RXR-a expression
essentially mirrors that of TR-a expression – that is, rising
after hatching through metamorphosis – and is particularly highly expressed in the brain and limb buds; RXR-g
is only significantly expressed in the fertilized egg and
TH
xRXR-α
xTR-α
xRXR-β
xRXR-γ
0
xTR-β
22
35
45
50
60
66
Developmenal stage
TRENDS in Endocrinology & Metabolism
Figure 2. Differential expression of TRs and RXRs during X. laevis development.
Relative mRNA levels of xTR-a (red line), xTR-b (dashed red line), xRXR-a (blue line),
xRXR-b (dashed blue line) and xRXR-g (dotted blue line) in whole embryos and
tadpoles are shown during X. laevis development through to the end of
metamorphosis. TH levels are represented by the black line. Developmental stages
are shown on the y-axis [65]. Note that TR-a and RXR-a levels rise in concert after
hatching (stage 35) and well before TH levels begin to rise at the onset of
metamorphosis by stage 50. In addition, TR-b levels correlate closely with
circulating TH levels.
www.sciencedirect.com
Vol.17 No.2 March 2006
early embryos and RXR-b expression peaks during
neurulation and tail bud stages, when retinoids are
important for nervous system patterning. The prototypical TRE, a synthetic element selected for maximal
strength in transfection assays and TR-specific regulation,
consists of two AGGTCA half-sites in the same orientation, separated by four nucleotides. Many endogenous
TH-responsive genes contain consensus TREs, including
the X. laevis TR-b gene [21,22] and two other transcription
factor genes, encoding the X. laevis basic transcription
element binding protein BTEB [23] and the Xenopus
laevis homolog of the basic region leucine zipper protein
E4BP4 termed TH/bZIP [24]. Unlike the promoterproximal TR-b and TH/bZIP TREs, the BTEB gene TRE
was found to be located w6.5 kilobases from the
transcription start site in a highly conserved 200 base
pair region in both duplicated BTEB gene sequences [23].
Because genome duplication in the Xenopus genus
occurred 30 million years ago, this remote island of
homology implies the existence of an enhancer region
that is important for TH control of this gene. Distal and
evolutionarily conserved enhancer regions have been
discovered in mammalian estrogen-regulated genes, and
have led to the discovery of FoxA1 as an important
auxiliary factor for transcriptional control by the estrogen
receptor [25]. Similar analysis for a larger group of THregulated genes might reveal a combinatorial enhancer
code important for stage- and tissue-specific TH-mediated
transcriptional control during development.
TR corepressors and coactivators are structurally and
functionally conserved
Similarly to their mammalian counterparts, the xTRs
interact with coregulatory proteins that mediate their
potent transcriptional activation or repression functions
[14]. In vivo chromatin immunoprecipitation (ChIP)
assays using isolated tail or intestine nuclei demonstrated
that the unliganded xTR-a interacts with the TR-b and
TH/bZIP TREs in tadpole tissues before the onset of
metamorphosis [26] and is found in a complex with the
X. laevis homolog of the nuclear receptor corepressor
NCoR [27]. The TR complex also contains chromatinsilencing machinery such as histone deacetylases (e.g.
rpd3 or histone deacetylase 1) and accessory factors Sin3
and transducin-b-like protein 1 [28,29]. Addition of TH
disrupts recruitment of corepressors to target gene
promoters, and concomitant increases in acetylation of
histones H3 and H4 are observed. X. laevis NCoR shows
68% identity and 80% similarity overall compared with
human NCoR, including 17 of 18 amino acids forming each
TR-interacting corepressor nuclear receptor (CoRNR) box
(Figure 1c). The related silencing mediator of retinoid and
thyroid hormone receptors (SMRT) corepressor is also
expressed in X. laevis tissues before and during metamorphosis but only a partial complementary DNA sequence
has been reported.
Multiple proteins interact with the TRs upon ligand
binding and contribute to transcriptional activation by a
variety of mechanisms. Members of the p160/steroid
receptor coactivator (SRC) family of coactivators contain
intrinsic histone acetyltransferase activity and recruit
Review
TRENDS in Endocrinology and Metabolism
additional histone acetyltransferases such as cAMP
response element-binding protein-binding protein
(CBP)/p300 and p300/CBP-interacting protein (p/CIP).
SRC-3 and transcription intermediary factor 2 (also
known as SRC-2) have been identified in X. laevis
[30,31]; however, only SRC-3 has been analyzed in detail
at metamorphosis. X. laevis SRC-3 shows 77% overall
identity compared with its human counterpart, including
three conserved TR-interacting LXXLL motifs (Figure 1d).
The SRC-3 gene is induced by TH in the tail and intestine,
and this coactivator is recruited to the TH/bZIP promoter
TREs in a TH-dependent manner [32]. Recruitment of
SRC-3 to the TR-b TRE is TH dependent in the intestine
but not in the tail, despite significant increases in histone
acetylation on both promoters in both tissues [28,32]. The
significance of this promoter- and tissue-selective constitutive SRC-3 binding is unclear. For functional analysis
in vivo, a dominant negative form of SRC-3 (dnSRC-3)
containing only the central receptor-interacting LXXLL
motifs was expressed in transgenic tadpoles and competitively displaced endogenous SRC-3 from target genes,
inhibiting their induction [33]. The transgene strongly
inhibited TH-induced gill and tail resorption, intestinal
remodeling and Meckel’s cartilage (lower jaw) outgrowth,
and partially inhibited hindlimb development. Although
most animals died during spontaneous metamorphosis,
some survivors retained their tails for several months
after metamorphosis was complete [33]. Thus, coactivator
recruitment to the ligand-bound TR appears to be crucial
for larval tissue resorption programs and might also
contribute to the growth and differentiation of
adult tissues.
Additional coactivators and associated complexes could
have important roles in metamorphosis. Indeed, multiple
coactivators use the same interaction surface that would be
competitively occupied by the dnSRC-3 protein. For
example, the X. laevis homolog of the nucleosome-binding
Trip7 coactivator was identified as a TH-inducible gene in
the remodeling intestine, and it is induced by TH in a
variety of growing and remodeling tissues [34]. The
mediator complex is an important coactivator in mammalian cells, and promotes RNA polymerase II holoenzyme
assembly at target gene promoters [35]; however, it has not
yet been described in the Xenopus system. Thus, despite
recent progress on X. laevis corepressor and coactivator
expression and function, the relative contribution of
relieving ligand-independent gene repression, versus
ligand-induced gene activation, to various TH-regulated
morphogenetic programs remains to be established. In
addition, the requirement for various coactivator
complexes on distinct classes of TH-inducible promoters
in a tissue-specific or developmental stage-specific manner
requires further examination in all vertebrate systems.
TR activation is essential for metamorphosis
A role for TRs in mediating TH-induced metamorphosis
has been demonstrated by two different approaches. The
first approach used transgenic overexpression of a potent
dominant negative form of X. laevis TR-a (DNTR) that
lacks the last 17 amino acids of the ligand-binding domain
[36]. This truncation strongly inhibits ligand-binding
www.sciencedirect.com
Vol.17 No.2 March 2006
41
activity and prevents recruitment of coactivators that
rely on an intact AF-2 domain (i.e. the p160/SRC class). As
expected, the corepressor NCoR is retained on promoters
occupied by the DNTR even in the presence of TH, thus
maintaining target genes in a repressed state [37].
Ubiquitous expression of the DNTR prevents a broad
spectrum of TH-induced morphological changes, ranging
from proliferation of the jaw and brain, resorption of the
gills and tail, and remodeling of the intestinal tract [36]
(Figure 3). A twist on the dominant negative approach was
to attach both a strong viral activation domain and the
carboxy-terminus of NCoR containing only the receptor-
(a)
mCollagen α2(1)
Coll:GFP
GFP
2
mCollagen
α2(1)
GFP
(b)
401
Coll:DNTR
xTR-α DAF-2
(c)
NH 2
I
HO
O
H
I
CO 2 H
I
T3
O2N
O
CO 2 H
H3C
HO
CH3
NH-3
(d) nM T3:
µM NH-3:
0
0
5
0
5
0.5
5
1
5
2
MC
BA
MC
BA
nM T3:
µM NH-3:
15
0
15
2
0
2
TRENDS in Endocrinology & Metabolism
Figure 3. A dominant negative xTR-a–green fluorescent protein (GFP) fusion protein
and NH-3 both inhibit TH-induced metamorphosis. (a) Schematic of mouse
collagen a2(1) promoter-based expression constructs driving an AF-2 deleted
version of X. laevis TR-a fused to GFP (Coll:DNTR) or GFP alone (Coll:GFP). (b) Oneweek-old transgenic tadpoles created either with the Coll:DNTR construct or the
coll:GFP construct were treated with vehicle or 10 nM T3 for one week. (A) From top,
untreated coll:DNTR tadpole, T3-treated coll:GFP tadpole and T3-treated coll:DNTR
animal. (B–D) Fluorescently labeled tail musculature. (E–G) Tail notochord cross
section labeled with antisense collagenase-3 probe. B and E show untreated
coll:DNTR tadpole. C and E show T3-treated coll:GFP tadpole. D and G show T3treated coll:DNTR tadpole. Modified, with permission, from Ref. [36]. Note the
inhibition of gill and tail resorption, loss of tail skeletal muscle and upregulation of
collagenase-3 in the notochord in the T3-treated tadpoles expressing the DNTR. (c)
The structures of T3 and NH-3. (d) Premetamorphic tadpoles (stage 52) treated with
the indicated concentrations of T3 and/or NH-3 for one week. Arrows indicate
position of gills (or branchial arches, BA) and lower jaw (or Meckel’s cartilage, MC).
Modified, with permission, from Ref. [40]. Note that NH-3 competitively inhibits T3induced gill loss and Meckel’s cartilage growth, similar to the phenotype observed
in (b) with the tadpoles overexpressing DNTR. NH-3 inhibits T3-induced tail
resorption and collagenase-3 induction as well [40].
42
Review
TRENDS in Endocrinology and Metabolism
interacting domains to the AF-2 truncated receptor, with
the dual goal of interfering with NCoR recruitment and
promoting hormone-independent activation of target
genes [38]. Heat shock-inducible, ubiquitous expression
of this unique ‘dominant positive’ receptor resulted in
TH-independent gill resorption, hindlimb outgrowth,
intestinal remodeling and upregulation of several
TH-response genes.
A second approach to probe the importance of the
liganded TR in metamorphosis was to use a synthetic TR
antagonist, NH-3, which competitively inhibits THinduced transcription in transient transfection assays
[39]. NH-3 was designed rationally with a nitrophenyl
group extension to prevent proper helix-12 folding and
coactivator recruitment; indeed, the compound induces a
distinct conformation of the receptor, as judged by
protease protection assays [40]. Importantly, the compound strongly inhibits TH-induced morphological
changes in treated tadpoles as effectively as does DNTR
expression in transgenic tadpoles (Figure 3); it also
inhibits TH-induced transcription of endogenous genes
and completely and reversibly inhibits spontaneous
metamorphosis [40]. Therefore, NH-3 is the first TR
antagonist to show potent in vivo inhibition of TH action
and further demonstrates the clear dependence on TR
transactivation properties for metamorphosis to proceed.
Major questions of metamorphosis
Although significant progress has been made on the role of
TRs and associated coregulators in metamorphosis,
several key unanswered questions remain. Addressing
each of these questions has important ramifications for
hormone action in all animals.
Tissue specificity
Perhaps the most remarkable aspect of metamorphosis is
that the same simple molecule (i.e. TH) can induce
completely opposite morphological responses in distinct
tissues, ranging from growth and differentiation of limbs to
the complete resorption of the tail, even though both organs
contain similar tissue types (e.g. skin, nerves, blood vessels
and muscle). Differences in TR isoform expression suggest
specific roles in metamorphosis, and correlate with distinct
gene expression programs (Box 2). Supporting this
Box 2. TH signaling components in tadpole tail versus
hindlimbs at the onset of metamorphosis
Adult cell growth and differentiation (hindlimbs)
High TR-a and RXR-a levels
High NCoR and SMRT levels
High type II deiodinase levels
Corepressor release for proliferation (via TR-a)
Coactivator recruitment required for differentiation (via TR-a and/or
TR-b)?
Larval cell death and resorption (tail)
Low TR-a and RXR-a levels
Low coactivator and corepressor levels
High baseline of inducible type III deiodinase expression which
declines at climax
Inducible TR-b, SRC-3, type II deiodinase
Coactivator recruitment required for death and resorption (via TR-b)?
www.sciencedirect.com
Vol.17 No.2 March 2006
observation, the synthetic TR agonist GC-1 activates TR-b
with tenfold greater potency compared with TR-a, and
induces gill and tail resorption with little effect on limb
growth [41]. In Drosophila metamorphosis, distinct ecdysone receptor isoforms that differ only in their aminoterminal AF-1 domains control specific tissue responses, as
demonstrated by isoform-specific mutations [42,43]. Such
unequivocal genetic evidence for distinct TR function in
amphibians is not currently possible; even so, the exact
means of receptor isoform-specific control would still be
unclear, as it is in Drosophila. Indeed, tissue specificity
might be programmed during embryogenesis, long before
TR isoforms and TH first appear. Further dissection of
highly specific TR response genes using transgenic reporter
gene assays should reveal important enhancer and silencer
elements and lead to the identification of key transcription
factors involved in the exquisite degree of tissue-specific
responses induced by TH.
Timing
Timing is a crucial issue in metamorphosis. For example,
the animal must complete the differentiation of the hindand fore-limbs before the tail is lost, and the lungs must
become fully functional before the gills are resorbed. A
simple explanation for timing is based on the differential
sensitivity of tissues to TH [44]. Low doses of TH will
induce limb outgrowth, whereas several-fold higher
concentrations are required for tail resorption. Accordingly, during spontaneous metamorphosis, limbs grow
when there are low levels of circulating TH and the tail
resorbs only when TH levels are highest [14]. How is
differential tissue sensitivity achieved? One important
mechanism is the differential expression of type III
deiodinase (D3), which removes an inner-ring iodine
from both T4 and T3, creating inactive metabolites [44]
(Box 2). D3 is expressed in the tail before metamorphosis
and is induced by TH, followed by a decline at the onset of
resorption [45]. D3 expression is initially undetectable in
the limb buds and rises modestly as the limbs complete
morphogenesis. Dorsal retina-specific D3 expression
enables asymmetric proliferation of the ventral ciliary
marginal zone and ipsilateral projection of retinal
ganglion cells from that region to the optic tectum, thus
establishing binocular vision in the predatory frog [46].
Overexpression of D3 in transgenic X. laevis delays
metamorphosis, including ventral retinal proliferation,
and it is particularly effective at blocking tail resorption,
resulting in ‘tailed’ froglets [47]. By contrast, the type II
deiodinase (D2), which converts T4 to T3, is expressed at
high levels before metamorphosis in the subventricular
zones of the brain and the hindlimb buds. Thus, these
tissues are poised to respond to very low circulating T4
levels [48]. At metamorphic climax, when the final
changes of metamorphosis, such as tail resorption,
commence, D2 levels rise abruptly in the tail and in the
TSH-secreting cells of the pituitary [49]. Although the
capacity for negative feedback exists before climax [50],
the onset of D2 expression leads to a dramatic increase in
the ability of T4 to suppress pituitary TSH expression and
thus sets up a classical negative feedback loop in the adult.
D2 overexpression in transgenic tadpoles, either
Review
TRENDS in Endocrinology and Metabolism
ubiquitously or under specific promoters, has not yet
been reported.
Differences in TR and coregulator expression levels in
specific tissues might also contribute to precise timing
(Box 2). Limb buds express high levels of TR-a and SRC-3
before metamorphosis, whereas tails express much lower
levels of each [18,32]. Coincident induction of TR-b and a
coactivator in the tail would substantially increase target
gene responsiveness, leading to its timely demise. Perhaps
counter intuitively, NCoR and SMRT levels are also high
in the limb buds and low in the tails before metamorphosis
[27]. Unoccupied TRs bound to highly expressed NCoR
and SMRT corepressors might compete with low levels of
TH-occupied and coactivator-bound TRs at the beginning
stages of metamorphosis. Elevated corepressor levels
might be important to prevent precocious proliferation
without adequate amounts of TH, or might be used by
other nuclear receptors or transcription factors to maintain adult precursor cells in an undifferentiated state. In
addition, corepressor localization and receptor interaction
might be altered by growth factor activity [51,52]. The loss
of TR interaction with NCoR or SMRT in rapidly growing
cells would shift the TH dose–response curve leftward and
thereby amplify the sensitivity of the tissue.
Cell autonomy
A useful feature of amphibian biology is the ability to
culture various organs and tissues in vitro, or transplant
tissues onto various locations of the developing embryo
and tadpole. Such experiments first demonstrated the
tissue autonomy of TH responses; isolated tails in culture
will resorb, and isolated limb buds will grow and
differentiate, in a TH-dependent manner [14]. To determine the degree of cell autonomy, or whether the fate of
one cell is dependent on TH influence on adjacent cells,
DNTR was expressed in specific cell types using transgenic assays or by direct injection of expression vectors.
Epidermal larval cell-specific DNTR expression prevents
TH-induced cell death via apoptosis and repression of
larval skin-specific genes but enables adult skin to develop
normally underneath [53]. Using multiple tissue specificpromoters driving DNTR, several cell types in the
developing limbs were shown to develop autonomously.
Neural-specific DNTR expression resulted in an absence
of motor neurons and paralysis, with apparently normal
muscle, cartilage and bone formation; skeletal musclespecific DNTR expression also resulted in paralysis owing
to the complete lack of muscle despite normal limb size,
cartilage and bone [54]. The fate of motor neurons in the
absence of muscle fiber development was not reported.
However, most cell autonomy experiments have investigated whether skeletal muscle in the resorbing tail dies
via an intrinsic TH-induced program (‘suicide’) or as the
result of degradation of the surrounding extracellular
matrix by TH-responsive fibroblasts (‘murder’). The first
evidence that larval muscle dies autonomously came from
isolation of a myoblast cell line from tadpole tails [55].
These cells differentiated into myotubes that would
undergo features of apoptosis and induce caspase-3
expression upon TH treatment. Consistent with these
findings, skeletal muscle-specific DNTR expression by
www.sciencedirect.com
Vol.17 No.2 March 2006
43
transgenesis or direct plasmid injection strongly inhibits
muscle loss during metamorphosis [56,57]. Therefore,
larval muscle responds directly to TH to activate its own
death-inducing cascade, although complete extracellular
matrix breakdown and loss of the vasculature might
accelerate the process during the latter stages of tail
resorption [57]. The intrinsic larval muscle death program
involves activation of the apoptotic machinery including
multiple caspases and the proapoptotic gene bax [58,59].
Although the limb, skin and larval muscle results using
DNTR expression are fairly clear, other tissues and organs
might not show the same degree of cell autonomy.
Competence
A crucial concept in developmental biology is competence,
or the development of the ability of a cell to respond
maximally to a given inducer. Adequate receptor levels are
taken for granted as a primary mechanism of hormonal
competence. In X. laevis, many tadpole tissues, such as the
gills and the lower jaw, can respond to TH one week after
fertilization [60]. In general, the ability of tadpoles to
respond at least partially to exogenous TH correlates with
rising expression of xTR-a and xRXR-a [16]. Injection of
synthetic mRNAs encoding xTR-a and xRXR-a into
fertilized eggs enabled the precocious upregulation of
two early TH-response genes along with some associated
morphological changes; however, these changes were not
identical to those observed during natural metamorphosis
[61]. Tissue competence does not develop uniformly – tail
resorption competence develops later than that of most
other tissues. In addition, different gene classes acquire
transcriptional competence at distinct developmental
stages [18]. For example, when hindlimbs are still small
buds, the tail is poorly responsive to exogenous T3, and
only early genes such as TR-b can be significantly induced.
Several days later, when the hindlimbs are beginning to
grow and differentiate, TH addition causes a dramatic
increase in the tail resorption rate and the ability to
upregulate late-response genes such as that encoding
collagenase-3. Early genes are induced to the same degree
at both stages of development [18]. Therefore, adequate
receptor expression is necessary but not sufficient to
obtain maximal tissue sensitivity to TH, and late genes
might require the expression of an additional competence
factor for full induction. Stage-specific competence in
Drosophila metamorphosis requires a network of orphan
nuclear receptor activity [62]. Although the Xenopus
genome encodes most, if not all, of the vertebrate nuclear
receptor family, few have been studied at metamorphosis.
Certainly, other mechanisms for TR competence can be
envisioned, such as quantitative or qualitative differences
in coactivator expression and activity and/or rearrangement of target gene chromatin.
Conclusion
X. laevis has been a classical model system for embryogenesis for decades. With the application of advanced
molecular tools, detailed molecular dissection of vertebrate
postembryonic development, such as metamorphosis, is
now possible. The essential role of ligand-occupied TRs in
the induction of metamorphosis has been clearly
44
Review
TRENDS in Endocrinology and Metabolism
established by both ubiquitous and cell type-specific
expression of a dominant negative TR, and the ability of a
rationally designed TH antagonist, NH-3, to block induced
and spontaneous metamorphosis. Many cell types appear
to have autonomous, independent gene expression pathways leading to death or growth and differentiation in
response to TH. A large number of TH-responsive genes
have been identified in both death and growth pathways;
however, significant work to determine their functional
roles in metamorphosis remains to be carried out. One
TH-response gene that has been analyzed in detail – that
encoding the D3 enzyme – is a major contributor to the
timing of tissue responses to TH [47]. D3 function was
established through over- and mis-expression studies via
transgenesis and the use of pharmacological inhibitors.
Ultimately, full understanding of metamorphosis will be
enhanced by the development of efficient gene interference
assays suitable for use in tadpoles. Morpholino, or
traditional antisense oligonucleotides and RNAs, dilute
out rapidly or are degraded after injection into fertilized
eggs; however, successful direct injection of certain tadpole
tissues with these reagents has recently been demonstrated [58]. Tissue-specific and inducible expression of
RNA interference or ribozyme constructs would be useful
but such experiments are not yet routine in any transgenic
vertebrate system. The continued development of the
X. tropicalis model is especially important because it
could provide a powerful genetic handle on gene function
at metamorphosis [63,64]. Continued efforts using a
variety of approaches to investigate the major questions
of anuran metamorphosis outlined in this review (i.e. tissue
specificity, timing, cell autonomy and competence) will
have wide-ranging implications for important biological
problems as diverse as worldwide declines in amphibian
populations to the role of TH in human development.
Acknowledgements
The authors thank Julie Roessig for helpful comments on the manuscript.
J.D.F. dedicates this article to Don Brown on the occasion of his ‘emeritus’
status at the Carnegie Institution of Washington Department of
Embryology, and thanks him for many valuable conversations, encouragement and introduction to this fascinating system. Part of this work
was funded by NIH RO1 DK55511.
References
1 Gudernatsch, J.F. (1912) Feeding experiments on tadpoles. I. The
influence of specific organs given as food on growth and differentiation:
a contribution to the knowledge of organs with internal section. Arch
Entwicklungsmech Org 35, 457–483
2 Leloup, J. and Buscaglia, M. (1977) La triiodothyronine: hormone de la
metamorphose des amphibiens. C R Acad. Sci. 284, 2261–2263
3 Boorse, G.C. and Denver, R.J. (2004) Expression and hypophysiotropic
actions of corticotropin-releasing factor in Xenopus laevis. Gen. Comp.
Endocrinol. 137, 272–282
4 Dodd, M.H.I. and Dodd, J.M. (1976) The biology of metamorphosis. In
The biology of metamorphosis (Lofts, B., ed.), pp. 467–599, Academic
Press
5 Williams, D.W. and Truman, J.W. (2005) Remodeling dendrites during
insect metamorphosis. J. Neurobiol. 64, 24–33
6 Shi, Y.B. and Ishizuya-Oka, A. (2001) Thyroid hormone regulation of
apoptotic tissue remodeling: implications from molecular analysis of
amphibian metamorphosis. Prog. Nucleic Acid Res. Mol. Biol. 65,
53–100
www.sciencedirect.com
Vol.17 No.2 March 2006
7 Schreiber, A.M. et al. (2005) Remodeling of the intestine during
metamorphosis of Xenopus laevis. Proc. Natl. Acad. Sci. U. S. A. 102,
3720–3725
8 Amaya, E. and Kroll, K.L. (1999) A method for generating transgenic
frog embryos. Methods Mol. Biol. 97, 393–414
9 Kroll, K.L. and Amaya, E. (1996) Transgenic Xenopus embryos from
sperm nuclear transplantations reveal FGF signaling requirements
during gastrulation. Development 122, 3173–3183
10 Lim, W. et al. (2004) The mouse muscle creatine kinase promoter
faithfully drives reporter gene expression in transgenic Xenopus
laevis. Physiol. Genomics 18, 79–86
11 Marsh-Armstrong, N. et al. (1999) Germ-line transmission of
transgenes in Xenopus laevis. Proc. Natl. Acad. Sci. U. S. A. 96,
14389–14393
12 Hirsch, N. et al. (2002) Xenopus, the next generation: X. tropicalis
genetics and genomics. Dev. Dyn. 225, 422–433
13 Tata, J.R. (1966) Requirement for RNA and protein synthesis for induced
regression of the tadpole tail in organ culture. Dev. Biol. 13, 77–94
14 Shi, Y.B. (2000) Amphibian Metamorphosis: From Morphology to
Molecular Biology, John Wiley and Sons, Inc.
15 Yaoita, Y. et al. (1990) Xenopus laevis alpha and beta thyroid hormone
receptors. Proc. Natl. Acad. Sci. U. S. A. 87, 7090–7094
16 Yaoita, Y. and Brown, D.D. (1990) A correlation of thyroid hormone
receptor gene expression with amphibian metamorphosis. Genes Dev.
4, 1917–1924
17 Berry, D.L. et al. (1998) The expression pattern of thyroid hormone
response genes in remodeling tadpole tissues defines distinct growth
and resorption gene expression programs. Dev. Biol. 203, 24–35
18 Wang, Z. and Brown, D.D. (1993) Thyroid hormone-induced gene
expression program for amphibian tail resorption. J. Biol. Chem. 268,
16270–16278
19 Blumberg, B. et al. (1992) Multiple retinoid-responsive receptors in
a single cell: families of retinoid ‘X’ receptors and retinoic acid
receptors in the Xenopus egg. Proc. Natl. Acad. Sci. U. S. A. 89,
2321–2325
20 Marklew, S. et al. (1994) Isolation of a novel RXR from Xenopus that
most closely resembles mammalian RXR beta and is expressed
throughout early development. Biochim. Biophys. Acta 1218,
267–272
21 Ranjan, M. et al. (1994) Transcriptional repression of Xenopus TR beta
gene is mediated by a thyroid hormone response element located near
the start site. J. Biol. Chem. 269, 24699–24705
22 Machuca, I. et al. (1995) Analysis of structure and expression of the
Xenopus thyroid hormone receptor-beta gene to explain its autoinduction. Mol. Endocrinol. 9, 96–107
23 Furlow, J.D. and Kanamori, A. (2002) The transcription factor basic
transcription element-binding protein 1 is a direct thyroid hormone
response gene in the frog Xenopus laevis. Endocrinology 143,
3295–3305
24 Furlow, J.D. and Brown, D.D. (1999) In vitro and in vivo analysis of
the regulation of a transcription factor gene by thyroid hormone
during Xenopus laevis metamorphosis. Mol. Endocrinol. 13,
2076–2089
25 Carroll, J.S. et al. (2005) Chromosome-wide mapping of estrogen
receptor binding reveals long-range regulation requiring the forkhead
protein FoxA1. Cell 122, 33–43
26 Sachs, L.M. and Shi, Y.B. (2000) Targeted chromatin binding and
histone acetylation in vivo by thyroid hormone receptor during
amphibian development. Proc. Natl. Acad. Sci. U. S. A. 97,
13138–13143
27 Sachs, L.M. et al. (2002) Nuclear receptor corepressor recruitment by
unliganded thyroid hormone receptor in gene repression during
Xenopus laevis development. Mol. Cell. Biol. 22, 8527–8538
28 Havis, E. et al. (2003) Metamorphic T3-response genes have specific
co-regulator requirements. EMBO Rep. 4, 883–888
29 Tomita, A. et al. (2004) Recruitment of N-CoR/SMRT-TBLR1
corepressor complex by unliganded thyroid hormone receptor for
gene repression during frog development. Mol. Cell. Biol. 24,
3337–3346
30 de la Calle-Mustienes, E. and Gomez-Skarmeta, J.L. (2000) XTIF2, a
Xenopus homologue of the human transcription intermediary factor, is
required for a nuclear receptor pathway that also interacts with CBP
to suppress Brachyury and XMyoD. Mech. Dev. 91, 119–129
Review
TRENDS in Endocrinology and Metabolism
31 Kim, H.J. et al. (1998) Molecular cloning of xSRC-3, a novel
transcription coactivator from Xenopus, that is related to AIB1,
p/CIP, and TIF2. Mol. Endocrinol. 12, 1038–1047
32 Paul, B.D. et al. (2005) Tissue- and gene-specific recruitment of steroid
receptor coactivator-3 by thyroid hormone receptor during development. J Biol Chem. 280, 27165–27172
33 Paul, B.D. et al. (2005) Coactivator recruitment is essential for
liganded thyroid hormone receptor to initiate amphibian metamorphosis. Mol. Cell. Biol. 25, 5712–5724
34 Amano, T. et al. (2002) Thyroid hormone regulation of a transcriptional coactivator in Xenopus laevis: implication for a role in
postembryonic tissue remodeling. Dev. Dyn. 223, 526–535
35 Rachez, C. and Freedman, L.P. (2001) Mediator complexes and
transcription. Curr. Opin. Cell Biol. 13, 274–280
36 Schreiber, A.M. et al. (2001) Diverse developmental programs of
Xenopus laevis metamorphosis are inhibited by a dominant negative
thyroid hormone receptor. Proc. Natl. Acad. Sci. U. S. A. 98,
10739–10744
37 Buchholz, D.R. et al. (2003) A dominant-negative thyroid hormone
receptor blocks amphibian metamorphosis by retaining corepressors
at target genes. Mol. Cell. Biol. 23, 6750–6758
38 Buchholz, D.R. et al. (2004) Transgenic analysis reveals that thyroid
hormone receptor is sufficient to mediate the thyroid hormone signal
in frog metamorphosis. Mol. Cell. Biol. 24, 9026–9037
39 Nguyen, N.H. et al. (2002) Rational design and synthesis of a novel
thyroid hormone antagonist that blocks coactivator recruitment.
J. Med. Chem. 45, 3310–3320
40 Lim, W. et al. (2002) A thyroid hormone antagonist that inhibits
thyroid hormone action in vivo. J. Biol. Chem. 277, 35664–35670
41 Furlow, J.D. et al. (2004) Induction of larval tissue resorption in
Xenopus laevis tadpoles by the thyroid hormone receptor agonist GC1. J. Biol. Chem. 279, 26555–26562
42 Schubiger, M. et al. (1998) Drosophila EcR-B ecdysone receptor
isoforms are required for larval molting and for neuron remodeling
during metamorphosis. Development 125, 2053–2062
43 Schubiger, M. et al. (2003) Isoform specific control of gene activity
in vivo by the Drosophila ecdysone receptor. Mech. Dev. 120, 909–918
44 Brown, D.D. (2005) The role of deiodinases in amphibian metamorphosis. Thyroid 15, 815–821
45 St Germain, D.L. et al. (1994) A thyroid hormone-regulated gene in
Xenopus laevis encodes a type III iodothyronine 5-deiodinase.. Proc.
Natl. Acad. Sci. U. S. A. 91, 11282
46 Marsh-Armstrong, N. et al. (1999) Asymmetric growth and development of the Xenopus laevis retina during metamorphosis is controlled
by type III deiodinase. Neuron 24, 871–878
47 Huang, H. et al. (1999) Metamorphosis is inhibited in transgenic
Xenopus laevis tadpoles that overexpress type III deiodinase. Proc.
Natl. Acad. Sci. U. S. A. 96, 962–967
48 Cai, L. and Brown, D.D. (2004) Expression of type II iodothyronine
deiodinase marks the time that a tissue responds to thyroid hormoneinduced metamorphosis in Xenopus laevis. Dev. Biol. 266, 87–95
Vol.17 No.2 March 2006
49 Huang, H. et al. (2001) Timing of metamorphosis and the onset of the
negative feedback loop between the thyroid gland and the pituitary is
controlled by type II iodothyronine deiodinase in Xenopus laevis. Proc.
Natl. Acad. Sci. U. S. A. 98, 7348–7353
50 Manzon, R.G. and Denver, R.J. (2004) Regulation of pituitary
thyrotropin gene expression during Xenopus metamorphosis: negative
feedback is functional throughout metamorphosis. J. Endocrinol. 182,
273–285
51 Jonas, B.A. and Privalsky, M.L. (2004) SMRT and N-CoR corepressors
are regulated by distinct kinase signaling pathways. J. Biol. Chem.
279, 54676–54686
52 Hong, S.H. and Privalsky, M.L. (2000) The SMRT corepressor is
regulated by a MEK-1 kinase pathway: inhibition of corepressor
function is associated with SMRT phosphorylation and nuclear
export. Mol. Cell. Biol. 20, 6612–6625
53 Schreiber, A.M. and Brown, D.D. (2003) Tadpole skin dies autonomously in response to thyroid hormone at metamorphosis. Proc. Natl.
Acad. Sci. U. S. A. 100, 1769–1774
54 Brown, D.D. et al. (2005) Thyroid hormone controls multiple
independent programs required for limb development in Xenopus
laevis metamorphosis. Proc. Natl. Acad. Sci. U. S. A. 102,
12455–12458
55 Yaoita, Y. and Nakajima, K. (1997) Induction of apoptosis and CPP32
expression by thyroid hormone in a myoblastic cell line derived from
tadpole tail. J. Biol. Chem. 272, 5122–5127
56 Das, B. et al. (2002) Multiple thyroid hormone-induced muscle growth
and death programs during metamorphosis in Xenopus laevis. Proc.
Natl. Acad. Sci. U. S. A. 99, 12230–12235
57 Nakajima, K. and Yaoita, Y. (2003) Dual mechanisms governing
muscle cell death in tadpole tail during amphibian metamorphosis.
Dev. Dyn. 227, 246–255
58 Sachs, L.M. et al. (2004) Implication of bax in Xenopus laevis tail
regression at metamorphosis. Dev. Dyn. 231, 671–682
59 Rowe, I. et al. (2005) Apoptosis of tail muscle during amphibian
metamorphosis involves a caspase-9 dependent mechanism. Dev. Dyn.
233, 76–87
60 Tata, J.R. (1968) Early metamorphic competence of Xenopus larvae.
Dev. Biol. 18, 415–440
61 Puzianowska-Kuznicka, M. et al. (1997) Both thyroid hormone and 9cis retinoic acid receptors are required to efficiently mediate the effects
of thyroid hormone on embryonic development and specific gene
regulation in Xenopus laevis. Mol. Cell. Biol. 17, 4738–4749
62 King-Jones, K. and Thummel, C.S. (2005) Nuclear receptors – a
perspective from Drosophila. Nat. Rev. Genet. 6, 311–323
63 Noramly, S. et al. (2005) A gynogenetic screen to isolate naturally
occurring recessive mutations in Xenopus tropicalis. Mech. Dev. 122,
273–287
64 Grammer, T.C. et al. (2005) Identification of mutants in inbred
Xenopus tropicalis. Mech. Dev. 122, 263–272
65 Nieuwkoop, P.D. and Faber, J. (1956) Normal Table of Xenopus laevis,
North Holland Publishing
Free from Contents Direct
Elsevier offers free electronic tables of contents (e-tocs) for all its journals, alerting you when each new issue is published online.
These updates deliver full contents lists and direct links to each article to your inbox. The free e-toc is a great way of staying in touch
with the latest developments in your chosen field and allows you to see articles ahead of the print edition as soon as they are
published online.
To sign up for e-tocs visit: http://contentsdirect.elsevier.com/
www.sciencedirect.com
45