Pathways and consequences: Hedgehog signaling in human disease

Abstract

Signaling pathways that play a fundamental role during development are turning out to underlie many disease states when misregulated. Here, we review some of the recent findings in the Hedgehog (Hh) pathway and the role it plays in different human diseases. We present a summary of the diseases that result from the inactivation or inappropriate activation of the Hh pathway. The human phenotypes generally fit the findings in model organisms and help to identify some potential targets for therapy.

Full text

TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com 0962-8924/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0962-8924(02)02405-4
562 Review
The molecular genetic study of human developmental
pathways relies, in part, on the study of disease
phenotypes and on comparison with results obtained
in model organisms. This approach is possible
because most pathways are conserved during animal
development, although individual species may have
developed variations on the main theme. In the
absence of direct experimentation in humans, one has
to rely on a comparison of the phenotypes of different
diseases and the mutations associated with them, as
well as on the comparison of signal transduction and
other biochemical pathways among species. In this
review, we use the diseases caused by mutations in
the Hedgehog (Hh) signaling pathway to explore how
it might operate in humans, and to suggest which
proteins would, in principle, be good targets for therapy.
The Hedgehog pathway: an overview
Several of the components of the Hh pathway were
first identified in flies and later described in vertebrates
(reviewed in [1,2]). The interactions between these
different components constitute, in part, a chain of
consecutive repressive events that finally results in a
modification of gene transcription (see Fig. 1). Hh, an
extracellular ligand, is secreted by discrete subsets of
cells in many organs and is processed in the expressing
cell through an autocatalytic reaction that removes
the C-terminal domain and attaches a cholesterol
molecule to the C-terminus of the processed protein.
A palmitate molecule is then attached to the
N-terminus of the mature protein in a process
mediated by Sightless [Sit; also known as Skinny
hedgehog, Rasp and Central missing (Ski/Rasp/Cmn)].
These modifications appear to regulate Hh activity,
Signaling pathways that play a fundamental role during development are
turning out to underlie many disease states when misregulated. Here,we
review some of the recent findings in the Hedgehog (Hh) pathway and the role
it plays in different human diseases. We present a summary of the diseases that
result from the inactivation or inappropriate activation of the Hh pathway. The
human phenotypes generally fit the findings in model organisms and help to
identify some potential targets for therapy.
Published online: 5 November 2002
José L. Mullor
Pilar Sánchez
Ariel Ruiz i Altaba*
Developmental Genetics
Program and Dept of Cell
Biology, Skirball Institute,
NYU School of Medicine,
540 First Avenue, New
York, NY 10016, USA.
*e-mail: ria@
saturn.med.nyu.edu
Pathways and consequences:
Hedgehog signaling in human disease
José L. Mullor, Pilar Sánchez and Ariel Ruiz i Altaba
29 Thinakaran, G. et al. (1997) Evidence that levels
of presenilins (PS1 and PS2) are coordinately
regulated by competition for limiting cellular
factors. J. Biol. Chem. 272, 28415–28422
30 Mumm, J.S. and Kopan, R. (2000) Notch
signaling: from the outside in. Dev. Biol.
228, 151–165
31 Levitan, D. and Greenwald, I. (1995) Facilitation
of lin-12-mediated signalling by sel-12, a
Caenorhabditis elegans S182 Alzheimer’s disease
gene. Nature377, 351–354
32 Shen, J. et al. (1997) Skeletal and CNS defects in
Presenilin-1-deficient mice. Cell89, 629–639
33 Wong, P.C. et al. (1997) Presenilin 1 is required
for Notch1 and DII1 expression in the paraxial
mesoderm. Nature387, 288–292
34 Donoviel, D.B. et al. (1999) Mice lacking both
presenilin genes exhibit early embryonic
patterning defects. Genes Dev. 13, 2801–2810
35 Schroeter, E.H. et al. (1998) Notch-1 signalling
requires ligand-induced proteolytic release of
intracellular domain. Nature393, 382–386
36 De Strooper, B. et al. (1999) A presenilin-1-
dependent γ-secretase-like protease mediates
release of Notch intracellular domain. Nature
398, 518–522
37 Lopez-Schier, H. and Johnston, D.S. (2002)
Drosophilanicastrin is essential for the
intramembraneous cleavage of Notch. Dev.
Cell2, 79–89
38 Chung, H-M. and Struhl, G. (2001) Nicastrin is
required for presenilin-mediated transmembrane
cleavage in Drosophila. Nat. Cell Biol.
3, 1129–1132
39 Mizutani, T. et al.(2001) Conservation of the
biochemical mechanisms of signal transduction
among mammalian Notch family members.
Proc. Natl. Acad. Sci. U. S. A. 98, 9026–9031
40 Saxena, M.T. et al.(2001) Murine Notch homologs
(N1-4) undergo presenilin-dependent proteolysis.
J. Biol. Chem. 276, 40268–40273
41 Ni, C.Y. et al. (2001) γ-Secretase cleavage and
nuclear localization of ErbB-4 receptor tyrosine
kinase. Science294, 2179–2181
42 Lee, H.J. et al. (2002) Presenilin-dependent
γ-secretase-like intramembrane cleavage of
ErbB4. J. Biol. Chem. 277, 6318–6323
43 Marambaud, P. et al. (2002) A presenilin-1/
γ-secretase cleavage releases the E-cadherin
intracellular domain and regulates disassembly
of adherens junctions. EMBO J. 21, 1948–1956
44 Lammich, S. et al. (2002) Presenilin-dependent
intramembrane proteolysis of CD44 leads
to the liberation of its intracellular domain
and the secretion of an Aβ-like peptide.
J. Biol. Chem. 10.1074/jbc.M206872200
(www.jbc.org)
45 Ebinu, J.O. and Yankner, B.A. (2002) A RIP tide
in neuronal signal transduction. Neuron
34, 499–502
46 Cao, X. and Südhof, T.C. (2001) A
transcriptionally active complex of APP with
Fe65 and histone acetyltransferase Tip60.
Science293, 115–120
47 Baek, S.H. et al. (2002) Exchange of N-CoR
corepressor and Tip60 coactivator complexes
links gene expression by NF-κB and β-amyloid
precursor protein. Cell110, 55–67
48 Sastre, M. et al. (2001) Presenilin-dependent
γ-secretase processing of β-amyloid precursor
protein at a site corresponding to the S3 cleavage
of Notch. EMBO Rep. 2, 835–841
49 Gu, Y. et al. (2001) Distinct intramembrane
cleavage of the β-amyloid precursor protein family
resembling γ-secretase-like cleavage of Notch.
J. Biol. Chem. 276, 35235–35238
50 Yu, C. et al.(2001) Characterization of a
presenilin-mediated APP carboxyl terminal
fragment CTFγ: evidence for distinct mechanisms
involved in γ-secretase processing of the APP and
Notch1 transmembrane domains. J. Biol. Chem.
276, 43756–43760
51 Weidemann, A. et al. (2002) A novel ε-cleavage
within the transmembrane domain of the
Alzheimer amyloid precursor protein
demonstrates homology with Notch processing.
Biochemistry41, 2825–2835
52 Okochi, M. et al. (2002) Presenilins mediate a
dual intramembraneous γ-secretase cleavage of
Notch-1. EMBO J. 21, 5408–5416

diffusion and potency, and might, in some cases,
shape a signaling gradient. The requirement of such
modifications might depend on the context in which
Hh proteins act. After secretion, Hh molecules form
multimeric complexes in which the hydrophobic
moieties have been proposed to cluster together in an
inner core, leaving Hh proteins on the outside ready
to interact with molecules (e.g. proteoglycans) that
facilitate their transport. Because Hh proteins have
been detected and act far from their sources (reviewed
in [1,2]), such interactions could explain how Hh
diffuses through a hydrophilic environment despite
having two attached hydrophobic moieties.
Additional proteins are involved in Hh release
and diffusion (reviewed in [2]). These include
Dispatched (Disp) and Tout-velu (Ttv). Disp is a
12-transmembrane-domain protein with a
sterol-sensing domain required for Hh release.
By contrast, Ttv regulates synthesis of
proteoglycans and functions to allow movement of
Hh. Proper Hh signaling is also affected by the
glycoprotein Hedgehog-interacting protein (Hip),
described in mice, to which Hh binds with high
affinity. Hip is induced in responding cells and acts
as an attenuator of the Hh response as it inhibits
Hh binding to its receptor.
Two membrane proteins function to receive the
Hh signal: Patched (Ptc), a 12-transmembrane-domain
protein, and Smoothened (Smo), a 7-transmembrane-
domain protein homologous to G-protein-coupled
receptors (reviewed in [1]). In the absence of Hh,
Ptc represses Smo (reviewed in [2]). Hh binding to Ptc
releases the basal repression of Smo by Ptc, and
Smo then signals intracellularly to transduce the
Hh signal to the nucleus. Ptc has homology with Disp
and the Niemann-Pick C1 protein (NPC1), a human
protein associated with the NPC disease. NPC1 has a
role as a transmembrane pump and participates in
endocytic events. Similarly, Ptc is involved in the
intracellular vesicle trafficking that internalizes Hh,
although it is unclear whether this internalization is
part of a process to eliminate these proteins from the
membrane or to trigger signaling from vesicles
(reviewed in [2,3]). Other regulating proteins
potentially involved in Sonic hedgehog (Shh) signaling
are GAS1 and Megalin (reviewed in [3]). Megalin
[also known as gp330 or low-density lipoprotein
receptor-related protein (LRP)-2] mediates endocytosis
of various ligands, and Megalinmutant mice show a
phenotype similar to that of Shh
−
/
−
mice, which display
cyclopia [4]. Both Gas1 and Megalin bind Shh, and
Megalin internalizes it without lysosomal targeting
([5]; reviewed in [3]). The exact roles of GAS1 and
Megalin in the Shh pathway remain to be clarified.
Inside the cell and downstream of Smo, a large
multimolecular network transduces the Hh signal to
modify the Gli proteins, zinc-finger transcription factors
that mediate Hh signaling. This complex involves
Costal2 (Cos2; a microtubule-associated kinesin-like
protein), the kinase Fused (Fu), and Suppressor of
fused [Su(Fu)], a PEST-domain-containing protein
that antagonizes Fu function. Other proteins involved
in modulating the cytoplasmic transduction of the Hh
signal include protein kinase A (PKA), casein kinase 1
(CK1), glycogen synthase kinase 3 (GSK3), the kinase
Dyrk1 [6] and the F-box protein Slimb that functions
in the ubiquitin pathway (see [1,2,23] for reviews).
The three vertebrate Gli transcription factors,
and their fly homolog Cubitus interruptus (Ci), are
associated with the Hh pathway. The three Gli proteins
have partially overlapping functions, and their activities
depend on the state of Hh signaling. In the absence of
Hh, cytoplasmic Gli3 and Ci proteins are processed into
smaller repressor forms lacking C-terminal sequences.
Gli2 also yields C-terminally deleted forms, but this
appears not to be regulated by Hh signaling. Upon Hh
binding to Ptc, full-length Ci is modified into a nuclear
activating form. In vertebrates, the three Gli proteins
may mediate Hh signaling. The state of Gli proteins,
therefore, dictates the transcriptional state of
Hh target genes (reviewed in [1,7,8].)
Hh acts to regulate the three Gli proteins in
different ways (Fig. 2). Gli1 appears to act as an
activator to mediate and/or amplify the Hh response
TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com
563
Review
TRENDS in Cell Biology
Hh
Disp
Sit/Ski/Rasp/Cmn
Cholesterol
Hh Hh
Ttv
Hip
PtcGas1 Smo
PKA
Cos2
Su(Fu)
Slimb
Rab23
Fu
CK1
GSK3
Gli1, Gli2, Gli3
Zic2
Target genes
Signaling cell Responding cell
Fig. 1. The canonical Hedgehog (Hh) pathway. Thin arrows depict the flow of information. Proteins
that activate the Hh pathway are shown in blue and those that repress or attenuate it are shown in red.
The Gli and Zic transcription factors are indicated in black owing to their multiple roles (see text). The
production of mature Hh proteins in the signaling cell requires the presence of cholesterol and the
action of Sightless/Skinny hedgehog/Rasp/Central missing (Sit/Ski/Rasp/Cmn) and Dispatched (Disp).
Transport requires Tout-velu (Ttv). In the responding cell, the Hh signal is received and transduced by
Patched (Ptc) and Smoothened (Smo) and attenuated by Hedgehog-interacting protein (Hip). Gas1 is a
GPI-linked protein that binds Shh and appears to affect signaling. Its mode of action, however, is
unclear. Members of a series of cytoplasmic proteins help or antagonize the transmission of the Hh
signal to the Gli transcription factors. Protein kinase A (PKA), Costal-2 (Cos2), Suppressor of fused
Su(Fu), Rab23 and Slimb act antagonistically, whereas Fused (Fu), Casein kinase 1 (CK1) and glycogen
synthase kinase 3 (GSK3) act positively. These proteins act to modulate the activity of Gli proteins and
influence their functions as nuclear activators or repressors of transcription. Zic2 might interact with
Gli proteins and affects their function.

and is transcriptionally induced by Hh signaling
in all contexts examined. By contrast, the situation
with Gli2 and Gli3 is more complex. Hh signaling
represses both the transcription of Gli3and the
proteolytic formation of Gli3 repressors. However,
the function of Gli2, and possibly Gli3, can be positive
or negative in relation to Hh signaling in different
contexts (reviewed in [1,7,8]). For example, there is
evidence that Gli3 acts as an activator and that Hh
signaling turns full-length Gli2 into a potent activator
(reviewed in [7–9]). Moreover, both Gli2and Gli3
could respond and/or participate in other signaling
pathways as they are often expressed far from Hh
sources. For example, fibroblast growth factor (FGF)
induces Gli2and Gli3, and these induce Wnt signaling
in early amphibian development (reviewed in [9]).
Thus, Hh pathway function relies both on Gli
activating function and on inhibiting Gli repressor
formation. For example, in the early mouse neural
tube, Hh signaling induces ventral neuronal
differentiation (reviewed in [10]) by impeding the
formation of Gli3 repressors. In other contexts,
including floor plate induction and tumorigenesis, the
positive action of Gli1/2 is crucial (reviewed in [1,11]).
This generalized overview of the pathway might
be applicable to many species and tissues, but
specific aspects can vary and the complete pathway
has not been described in all species. For example,
Sit/Ski/Rasp/Cmnhas only recently been described
in flies, and Cos2 is not yet characterized in humans.
By contrast, other genes such as Hiphave been found
in vertebrates but not in flies.
The role of Hh in animal development
Three homologs of Hh have been described in
mammals: Sonic (Shh), Indian (Ihh) and Desert
(Dhh) hedgehogs. Although most studies have
been performed with Shh, all three might work
similarly [12] – but in different contexts. Thus, it
is likely that mutations in any of the components of
the pathway will affect signaling from all three
Hh proteins. Different roles have been described for
Shh during development, acting as a morphogen,
mitogen or differentiation factor. For example, there
is a gradient of Shh signaling in the ventral neural
tube that directly induces the pattern of neuron and
glial identities along the dorsoventral axis (reviewed
in [10]). In flies, Hh acts as a morphogen in the larval
cuticle, the wings and the abdomen (reviewed in [1]).
Shh functions as a mitogen for neuron precursors in
the developing cerebellum, as a differentiation factor
for Bergmann glia and as a survival factor for different
cell types, including motor and dopaminergic neurons,
tooth and neural crest cells (reviewed in [1,9]). As
mentioned above, Ihh and Dhh have additional
functions in tissues other than those in which Shh
acts. For instance, Ihh regulates proliferation and
differentiation of chondrocytes and regulates
pancreatic development (reviewed in [1]).
In addition to functioning in the embryo,
Hh proteins and Hh signal-transduction components
are expressed in postnatal and adult tissues,
suggesting that they function in the mature
organism (e.g. [13–16]). Defects in Hh signaling
could, therefore, affect both the human embryo and
the adult. Nevertheless, an inhibitor of the Hh
pathway, the alkaloid cyclopamine, does not appear
to have an obvious effect on adult ewes even though
their progeny is cyclopic [17,18]. However, the effect
of the inhibition of Hh function in adult tissues, such
as in gastric gland morphogenesis [16] and neuronal
survival [13], might be difficult to assess in
unmonitored livestock.
Human genes and human diseases
The SHH gene in forebrain development and cancer
Patients carrying heterozygous mutations in SHH
present holoprosencephaly (HPE) and Shh
−
/
−
mice
display cyclopia (reviewed in [19,20]). HPE affects the
forebrain and face to various degrees, from the most
extreme lethal alobar type to milder microforms that
include small midline facial defects. Interestingly, a
missense mutation in the SHHgene is associated
with a variable phenotype within the same family,
from alobar HPE to subtle defects including attention
deficiencies [21]. This raises the possibility that
Shh signaling is needed for normal cortical
development in humans (consistent with results in
mice [15], reviewed in [9]) and therefore, brain function.
In contrast to the problems that derive from loss
of function of Shh, its deregulation or gain of function
in the epidermis is sufficient to induce basal cell
carcinomas (BCCs) of the skin. These could arise
TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com
564 Review
TRENDS in Cell Biology
Properties
Activator/repressor
of Hh targets.
Hh signaling inhibits
repressor formation.
Activator/repressor
of Hh targets.
*Hh signal may induce
the production of a
strong activator.
Activator of Hh
targets.
Hh
Hh*
?
Activator
Activator
Activator
Repressor
Repressor
Gli3
Gli2
Gli1
PKA
?
Modulator of
Gli function?
Repressor of Hh tragets.
Repressor?
Zic2
?
?
?
N′C′
Fig. 2. Diagrammatic representation of the generalized regulation of the Gli proteins by
proteolysis, yielding activator and C-terminally deleted repressor forms. Zic2 is also shown;
it encodes a small protein with a GLI-type zinc-finger domain (gray box) that may act as a repressor
of transcription. Variations on these themes might occur in different organs, tissues and species.
For example, the N-terminal part of human GLI2 is much smaller than its mouse or frog homologs,
suggesting that it may lack repressor function. The hatched area represents the zinc-finger domain.
All proteins are schematized as boxes with N-termini to the left and C-termini to the right. Arrows
represent positive effects and T bars inhibitory effects. See text for details. Abbreviations:
Hh, Hedgehog; PKA, protein kinase A.

from hair follicles as BCCs have properties of
follicular differentiation and the SHH signaling
pathway participates in follicular development ([22],
reviewed in [11]).
Cholesterol, Shh signaling and disease
Several experiments in model systems show that
cholesterol modification of SHH is necessary for its
correct maturation, distribution and signaling.
Mutations in the cholesterol biosynthetic pathway affect
SHH signaling. In humans, the Smith-Laemli-Opitz
syndrome (SLOS) is an autosomal recessive disease
caused by mutations in 7-dehydrocholesterol
reductase, an enzyme that functions at the end of the
cholesterol biosynthetic pathway. SLOS patients
present growth retardation, microcephaly, mental
retardation and other malformations, many of which
occur in tissues in which Hh signaling is active.
Consistent with a possible effect on SHH signaling,
~5% of SLOS patients develop HPE. In addition,
alterations in cholesterol homeostasis might also
affect intracellular trafficking, possibly altering
HH behavior (reviewed in [1,2,24]).
EXT and bone defects
Mutations in the human genes EXT-1 and EXT-2,
which are two of the three EXThomologs of the
ttvfly gene, are associated with hereditary multiple
exostoses [25], frequent skeletal dysplasias or benign
bone tumors that can develop into malignant
chondrosarcoma. EXT-1 and EXT-2 are required for
the synthesis of heparan sulfate [26,27]. The
homology with Ttv suggests that loss of function of
EXT proteins involves impairment of Hh signaling,
which then results in a deregulation of bone growth.
The effects of mutations in EXT-1 and EXT-2, limited
to bone development, suggest that they affect the
function of IHH, which is prominently expressed in
this tissue, but not that of SHH or DHH, which are
expressed elsewhere.
Diseases resulting from alterations in PTCH1 and SMO
30–40% of Gorlin syndrome patients (GS; also known
as basal cell nevus syndrome or nevoid basal cell
carcinoma syndrome) have familial loss-of-function
mutations in the PTCH1gene (which encodes the
human Ptc1 homolog), suggesting that activation of
the Hh pathway is linked with this syndrome. Clinically,
GS patients present congenital abnormalities with
variable penetrance that include skeletal defects
(e.g. general overgrowth, polydactyly, fused or bifid ribs),
early onset of multiple BCCs and a higher-than-normal
rate of other tumors, including medulloblastomas
of the cerebellum (reviewed in [11,28,29]). While
Ptc
−
/
−
mice die as embryos showing ectopic activity of
Hh target genes, Ptc
+
/
−
mice display a phenotype that
partly recapitulates that of GS patients. Ptc
+
/
−
mice
develop rhabdomyosarcomas and medulloblastomas
(reviewed in [11,29]), and the frequency of the latter
increases in a p53
−
/
−
background [30]. Why BCCs do
not develop in Ptc
+
/
−
mice is unclear, although
Ptc
+
/
−
mice treated with ultraviolet or ionizing
radiation develop BCCs and trichoblastomas [31].
Sporadic somatic mutations in PTCH1have also
been described in several BCCs, medulloblastomas, and
possibly in trichoepitheliomas, meningiomas, breast
carcinomas and esophageal carcinomas, among other
tumors. A limited number of mutations in PTCH1have
similarly been linked to HPE [32]. However, as HPE
can arise from loss of SHH signaling, these mutations
might reflect a loss of response of PTCH1 to SHH,
thus permanently shutting down the pathway even
in the presence of ligand. Mutations in SMO have also
been associated with sporadic BCCs and primitive
neuroectodermal tumors. In this case, the mutations
appear to constitutively activate SMO, which in turn,
activates the HH pathway (reviewed in [11,28,29]).
Because of the central role of SMO in transducing the
HH signal in model organisms, it is expected that
mutations in SMOthat abolish or decrease its function
in humans will also lead to some form of HPE, if these
are viable to stages when HPE can be diagnosed.
Genes possibly linked to vesicles and transport
Although human diseases associated with Hh release
or trafficking have not yet been described, Rab23
−
/
−
(open brain) mutant mice ectopically activate the
Shh pathway and display exencephaly, indicating
that Rab23 function is needed to maintain pathway
repression [33]. Rab23 belongs to a family of small
Rab GTPases that can mediate organelle trafficking
and fusion events in the cell (reviewed in [2,34]).
While it is not clear how Rab23 affects Hh signaling,
the phenotype of Rab23
−
/
−
mice raises the possibility
that HPE and other diseases caused by alterations in
the HH pathway may also derive from mutations in
intracellular trafficking components.
Suppressor of fused in disease
Su(Fu) interacts with Gli proteins and has a repressive
action on Gli function in model organisms. Su(Fu)
antagonizes the function of the Fu kinase to regulate Gli
subcellular localization [35], possibly by sequestering
Gli proteins in the cytoplasm and/or inhibiting their
processing to an activated form, and targeting them
instead through Slimb for ubiquitin–proteasome
degradation (reviewed in [1]). As a repressor of the
pathway, SU(FU) loss-of-function mutations might
lead to human diseases associated with ectopic or
increased activity of the pathway. SU(FU) is located
in a chromosomal region linked to several types of
tumors, including glioblastoma multiforme, prostate
cancer, malignant melanoma and endometrial
cancer [35], and a link has been established
between mutations in SU(FU) and predisposition to
desmoplastic medulloblastoma [36], a tumor that can
arise from inappropriate SHH pathway activation.
Indeed, GS patients, carrying PTCH1mutations, can
develop medulloblastomas at a higher frequency than
the general population.
TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com
565
Review

Nuclear activation of the pathway: the GLI proteins
and disease
Molecular, cellular and genetic analyses in model
systems suggest that the Hh signal acts in at least
two ways to regulate target genes. One is to activate
Gli proteins to induce gene transcription and the
other is to inhibit the formation of Gli repressors
(mostly those of Gli3) to derepress targets. For
example, deletion of Gli3 in mice partially rescues the
Shh
−
/
−
mutant phenotype, arguing that an important
role of Shh is to antagonize that of Gli3 [37]. If human
GLI3 were to act only as a repressor of Hh targets,
mutations in GLI3could be predicted to give rise to
phenotypes similar to those that activate the pathway,
such as mutations in PTCH1. GLI3mutations cause
several diseases, including Pallister–Hall syndrome
(PHS) and Greig cephalopolysyndactyly syndrome
(GCPS) that are not necessarily similar to those
observed following inappropriate activation or
inhibition of the HH pathway. Similarly, GLI3
TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com
566 Review
Table 1. Highlights of Hedgehog pathway components and associated diseasesa
Hedgehog
pathway
component
Human
gene Nature and role in the pathway Diseases and
malformations
associated with
increased function
Diseases and
malformations associated
with decreased function
Gli1 GLI1 Zinc-finger transcription factor Basal cell carcinomasb in
Regulator of HH targets frogs and mice, brain
tumorsc in frogs
Gli2 GLI2 Zinc-finger transcription factor Basal cell carcinomasbLung, skeleton, limb, facial and other
Regulator of HH targets in mice malformations in mice
Gli3 GLI3 Zinc-finger transcription factor Pallister–Hall syndrome
Regulator of HH targets Greig cephalopolysyndactyly syndrome
Exencephaly, limb, skeletal and other
defects in mice
Zic2 ZIC2 Zinc-finger transcription factor Holoprosencephaly
Possible regulator of HH targets Holoprosencephaly-like and cerebellar
defects in mice
Supressor of SU(FU) PEST-domain-containing protein Medulloblastoma
Fused Contributes to regulate the state and activity
of Gli proteins
Rab23 Rab GTPase Exencephaly in mice
Inhibitor of the pathway, possibly
involved in vesicle transport
Megalin Low-density lipoprotein receptor-related Defective forebrain development in mice
protein
Endocytic SHH receptor
Smoothened SMO 7-transmembrane-domain protein Basal cell carcinomas Cyclopia and multiple defects including
Transducer of the HH signal Possibly brain tumors those in heart and gut in mice.
Patched-1 PTCH-1 12-transmembrane-domain protein Possibly Gorlin syndrome/basal cell nevus
HH membrane receptor, inhibits Smoothened holoprosencephaly syndrome (skeletal defects, tumors…)
Basal cell carcinomas, rhabdomyosarcomas,
medulloblastomas
Medulloblastomas and rhabdomyosarcomas
in mice
Tout-velu EXT-1/EXT-2 Putative glycosyltransferases (human Hereditary multiple exostoses (benign bone
homologs of Tout-velu) tumors)
Transport of Hh
Cholesterol Lipid Smitz–Lemli–Opitz syndrome
Regulation of Shh activity, diffusion and (growth problems, retardation and
potency holoprosencephaly in some cases)
Sonic SHH Secreted glycoproteins Basal cell carcinomasbHoloprosencephaly
Hedgehog Extracellular ligands in mice Holoprosencephaly, cyclopia and
multiple defects in many organs in mice
Desert DHH Secreted glycoproteins Abnormal testis development and peripheral
Hedgehog Extracellular ligands nerve sheath in mice
Indian IHH Secreted glycoproteins Abnormal bone formation in mice
Hedgehog Extracellular ligands
aPartial compilation of the identity, nature, function and associated human diseases of different components in the known Hedgehog (HH) pathway in humans. Results
in model systems are also incorporated. The similarity in the diseases produced by altering different elements of the pathway suggests its linearity from proximal
(bottom) to distal (top) components. See text for details.
bBasal-cell-carcinoma-like tumors in mice defined molecularly and histologically. In frog embryos, these are epidermal hyperplasias defined molecularly as basal cell
carcinoma-like.
cBrain hyperplasias with hyperproliferation of disorganized tissue containing precursor cells.

mutations have not yet been found in sporadic
medulloblastomas [38], which can carry PTCH1
mutations and consistently express GLI1(reviewed
in [11,15,28,29]). Diseases resulting from the ectopic
activation of the pathway are therefore likely to result
from both the inhibition of GLI3(and of its repressor
function) but also from the activation of GLI1/2.
Misregulation of GLI1/2 has been implicated in
HH pathway diseases. The involvement of GLI1 in
brain tumors has recently begun to emerge, despite
the original isolation of GLI1from a glioma line
15 years ago (reviewed in [11]). Gli1 can induce
tumors in experimental models, and many human
brain tumors express GLI1. Moreover, the
proliferation of several brain tumors/cells, including
medulloblastomas, can be inhibited by cyclopamine,
indicating the presence of an active pathway inducing
tumor growth [15]. Indeed, as the final positive
elements of the Hh pathway, activation of Gli1 was
proposed to be sufficient to mimic its activation by
ligand, and Gli1 misregulation was subsequently
shown to lead to the development of BCC-like tumors
in frog embryos and mice [39–41]. Misregulated Gli1
thus mimics the effects of ectopic Shh and loss of
PTCH1 ([22], reviewed in [1,2,11,29]). Because the
Hh–Gli pathway is active in precursor populations in
a variety of tissues and organs, such as granule cell
precursors in the cerebellum, deregulation of Gli1/2
function is likely to be involved in tumor development
in tissues that utilize the Hh pathway for their
development or maintenance. In contrast to the
gain-of-function phenotype of GLI1 and GLI2, there
are no human diseases associated with loss of GLI1 or
GLI2 function to date, and, in mice, Gli1 but not Gli2
appears to be redundant [42,43].
ZIC2: effects on HH signaling
ZIC2encodes a protein with a GLI-like zinc-finger,
and its mutation is associated with HPE ([44],
reviewed in [19,20]). In mice, decreasing Zic2function
yields a phenotype with signs of HPE, spina bifida
and alterations in cerebellar development [45]. Zic2
antagonizes activating Gli function in the frog neural
plate, and it can bind to Gli proteins, suggesting a
direct link with Hh signaling [46,47]. Although it
remains to be determined whether ZIC2 interacts
with GLI proteins in humans, its association with
HPE suggests that, in humans, ZIC2 might contribute
to the mediation of the SHH signal.
Pathways and therapies
Looking at the human HH pathway from a different
perspective, we can see whether mutations in the
various elements of the pathway cause the same
diseases (Table 1; Fig. 3). Clearly, alterations in
different components of the HH pathway can lead to
different phenotypes, although there is a good degree
of consistency, implying the linearity of the pathway.
For example, on the one hand, alterations in several
loci have been associated with HPE in humans and at
least five genes have been identified so far. These
genes are SHH, ZIC2, PTCH1, SIX3 and TGIF, of
which the first three are involved in or are likely to
affect Shh signaling ([32], reviewed in [19,20]). On
the other hand, diseases associated with growth
deregulation, such as tumors, can arise from a
gain of function of SHH, GLI or SMO proteins,
or loss of function of PTCH1, SU(FU) or EXT
proteins. As the Hh pathway is involved in many
developmental events, it will also likely be associated
with further human syndromes, defects and
malformations. For example, there is some evidence
that it may be implicated in the VACTERL syndrome,
characterized by vertebral defects, anal atresia,
cardiac malformations, tracheoesophageal fistula
with esophageal atresia, renal and limb anomalies [48].
Indeed, based on research in model systems, it is
expected that different PTCH1alleles will eventually
be associated with most diseases related to alterations
in HH signaling. Similarly, it is likely that mutations
in DISP, SMO, SUFU, GLIand other components
could be found in HH-related diseases, like HPE, if
these are not lethal.
Several therapeutic approaches to restore the
normal status of Hh signaling might be feasible.
For example, the oncogenic potential of human
squamous cell carcinoma cell lines can be reversed
TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com
567
Review
HH
HH
HH
Target genes
Signaling cell Responding cell
TRENDS in Cell Biology
Cholesterol
EXT-1
EXT-2 PTCH-1 SMO
SU(FU)
GLI1
GLI2
ZIC2
HPE
BCCs
SLOS, with
5% HPE
? HME GS, MBs, BCCs,
several tumors
MB
BCCs,
brain tumors
HPE
BCCs, brain
tumors
? HPE
Fig. 3. Elements of the Hedgehog (HH) pathway in humans reported to cause disease when altered.
Proteins that activate the pathway are shown in blue and those that repress it are shown in red. Also
shown are gain (upwards arrows) or loss (downwards arrows) of function mutations in the pathway
and their associated diseases (boxes). The effect of the mutations in
EXT
genes is not clear and are
thus shown in black. Mutations in PTCH1 associated with HPE are assumed to be gain-of-function
mutations that inactivate SMO even in the presence of HH. Overexpression of HH, GLI1 or GLI2 lead
to BCCs and other tumors in experimental systems, and GLI1 is consistently expressed in a variety
of human tumors, including BCCs and brain tumors such as MBs. Abbreviations: HME, hereditary
multiple exostoses; HPE, holoprosencephaly; MB, medulloblastoma; BCC, basal cell carcinoma;
GS, Gorlin’s syndrome; PTCH-1, Patched-1; SLOS, Smith–Laemli–Opitz syndrome; Smo, Smoothened.

by introducing a wild-type form of PTCH1[49]; and
introduction of Shhin skin cells rescues alopecia
induced by chemotherapy in mice [50,51]. These
results are encouraging and suggest that transient
expression of inhibitors of the HH pathway introducing
exogenous DNA might be feasible in some cases.
However, the development of drugs that agonize or
antagonize different negative or positive components
of the HH pathway is perhaps more attractive. One
such drug could be the purified SHH molecule itself.
Shh was used to alleviate symptoms in a rat model of
Parkinson’s disease [52], a stable form of Shh was
injected to treat nerve injury [53] and it has also been
shown to ameliorate neural crest defects induced by
ethanol [54]. However, diseases caused by changes in
elements downstream of Shh will not respond to its
administration. The small molecule cyclopamine, its
derivatives or functional analogs could be good
therapeutic agents to fight diseases caused by
activation of the Hh pathway at the receptor level.
For example, oncogenic mutations in SMOand
PTCH1can be reversed by the administration of
cyclopamine [55], and it inhibits the growth of brain
tumors, including medulloblastomas [15,23]. More
generally the development of other small molecules
antagonizing the function of the last mediators of the
pathway, the GLI proteins, holds great promise for
the treatment of human diseases caused by the
activation of the HH pathway at any level. Moreover,
if GLI1were dispensable in humans, as it appears to
be in mice [42], there might be few side effects to its
therapeutic inhibition. Even though making
small-molecule inhibitors to conventional transcription
factors has proven very difficult, the fact that GLI
proteins interact with a series of regulatory proteins in
the cytoplasm provides cause for encouragement.
TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com
568 Review
Acknowledgements
We thank Barbara Stecca,
Yorick Gitton,
Jessica Treisman,
Verónica Palma and
Van Nguyen for comments
on the review and
discussion. Owing to space
limitations, we could not
reference the important
work of many colleagues.
References
1 Ingham, P.W. and McMahon, A.P. (2001)
Hedgehog signaling in animal development:
paradigms and principles. Genes Dev.
15, 3059–3087
2 Ho, K.S. and Scott, M.P. (2002) Sonic hedgehog
in the nervous system: functions, modifications
and mechanisms. Curr. Opin. Neurobiol.
12, 57–63
3 Nybakken, K. and Perrimon, N. (2002) Hedgehog
signal transduction: recent findings. Curr. Opin.
Genet. Dev. 12, 503–511
4 Willnow, T.E. et al.(1996) Defective forebrain
development in mice lacking gp330/megalin.
Proc. Natl. Acad. Sci. U. S. A. 93, 8460–8464
5 McCarthy, R.A. et al. (2002) Megalin functions as
an endocytic sonic hedgehog receptor. J. Biol.
Chem.277, 25660–25667
6 Mao, J. et al. (2002) Regulation of Gli1
transcriptional activity in the nucleus by Dyrk1.
J. Biol. Chem. 277, 35156–35161
7 Aza-Blanc, P. and Kornberg, T.B. (1999) Ci:
a complex transducer of the hedgehog signal.
Trends Genet.15, 458–462
8 Ruiz i Altaba, A. (1999) The works of GLI and the
power of hedgehog. Nat. Cell Biol. 1, E147–E148
9 Ruiz i Altaba, A. et al. (2002) Hedgehog-Gli
signalling and the growth of the brain. Nat. Rev.
Neurosci.3, 24–33
10 Jessell, T.M. (2000) Neuronal specification in the
spinal cord: inductive signals and transcriptional
codes. Nat. Rev. Genet. 1, 20–29
11 Ruiz i Altaba, A. et al. (2002) Gli and hedgehog in
cancer: tumours, embryos and stem cells.
Nat. Rev. Cancer 2, 361–372
12 Pathi, S. et al. (2001) Comparative biological
responses to human Sonic, Indian, and Desert
hedgehog. Mech. Dev. 106, 107–117
13 Miao, N. et al. (1997) Sonic hedgehog promotes
the survival of specific CNS neuron populations
and protects these cells from toxic insult In vitro.
J. Neurosci. 17, 5891–5899
14 Traiffort, E. et al. (1999) Discrete localizations
of hedgehog signalling components in the
developing and adult rat nervous system.
Eur. J. Neurosci. 11, 3199–3214
15 Dahmane, N. et al. (2001) The Sonic
Hedgehog–Gli pathway regulates dorsal brain
growth and tumorigenesis. Development
128, 5201–5212
16 van den Brink, G.R. et al. (2001) Sonic
hedgehog regulates gastric gland morphogenesis
in man and mouse. Gastroenterology
121, 317–328
17 Cooper, M.K. et al. (1998) Teratogen-mediated
inhibition of target tissue response to Shh
signaling. Science280, 1603–1607
18 Incardona, J.P. et al. (1998) The teratogenic
Veratrum alkaloid cyclopamine inhibits sonic
hedgehog signal transduction. Development
125, 3553–3562
19 Roessler, E. and Muenke, M. (2001) Midline and
laterality defects: left and right meet in the
middle. BioEssays23, 888–900
20 Muenke, M. and Beachy, P.A. (2000) Genetics
of ventral forebrain development and
holoprosencephaly. Curr. Opin. Genet. Dev.
10, 262–269
21 Heussler, H.S. et al. (2002) Extreme variability of
expression of a Sonic Hedgehog mutation:
attention difficulties and holoprosencephaly.
Arch. Dis. Child. 86, 293–296
22 Oro, A.E. et al. (1997) Basal cell carcinomas in
mice overexpressing sonic hedgehog. Science
276, 817–821
23 Berman, D.M. et al. (2002) Medulloblastoma
growth inhibition by Hedgehog pathway
blockade. Science 297, 1559–1561
24 Incardona, J.P. and Roelink, H. (2000) The role of
cholesterol in Shh signaling and teratogen-induced
holoprosencephaly. Cell. Mol. Life Sci.
57, 1709–1719
25 Bellaiche, Y. et al. (1998) Tout-velu is a Drosophila
homologue of the putative tumour suppressor
EXT-1 and is needed for Hh diffusion. Nature
394, 85–88
26 Lind, T. et al.(1998) The putative tumor suppressors
EXT1 and EXT2 are glycosyltransferases required
for the biosynthesis of heparan sulfate. J. Biol.
Chem.273, 26265–26268
27 Simmons, A.D. et al. (1999) A direct interaction
between EXT proteins and glycosyltransferases
is defective in hereditary multiple exostoses.
Hum. Mol. Genet. 8, 2155–2164
28 Taipale, J. and Beachy, P.A. (2001) The Hedgehog
and Wnt signalling pathways in cancer. Nature
411, 349–354
29 Goodrich, L.V. and Scott, M.P. (1998) Hedgehog
and Patched in neural development and disease.
Neuron21, 1243–1257
30 Wetmore, C. et al. (2001) Loss of p53 but not ARF
accelerates medulloblastoma in mice heterozygous
for patched. Cancer Res. 61, 513–516
31 Aszterbaum, M. et al. (1999) Ultraviolet and
ionizing radiation enhance the growth of BCCs
and trichoblastomas in patched heterozygous
knockout mice. Nat. Med. 5, 1285–1291
32 Ming, J.E. et al. (2002) Mutations in PATCHED-1,
the receptor for SONIC HEDGEHOG, are
associated with holoprosencephaly. Hum. Genet.
110, 297–301
33 Eggenschwiler, J.T. et al. (2001) Rab23 is an
essential negative regulator of the mouse Sonic
hedgehog signalling pathway. Nature412, 194–198
34 Jeong, J. and McMahon, A.P. (2001) Developmental
biology. Vesicles and the spinal cord. Nature
412, 136–137
35 Stone, A. et al. (1999) Characterization of the
human suppressor of fused, a negative regulator
of the zinc-finger transcription factor Gli. J. Cell
Sci.112, 4437–4448
36 Taylor, M.D. et al. (2002) Mutations in SUFU
predispose to medulloblastoma. Nat. Genet.
31, 306–310
37 Litingtung, Y. and Chiang, C. (2000) Specification
of ventral neuron types is mediated by an
antagonistic interaction between Shh and Gli3.
Nat. Neurosci. 3, 979–985
38 Erez, A. et al. (2002) GLI3 is not mutated
commonly in sporadic medulloblastomas. Cancer
95, 28–31
39 Dahmane, N. et al. (1997) Activation of the
transcription factor Gli1 and the Sonic hedgehog
signalling pathway in skin tumours. Nature
389, 876–881
40 Nilsson, M. et al. (2000) Induction of basal cell
carcinomas and trichoepitheliomas in mice
overexpressing GLI-1. Proc. Natl. Acad. Sci.
U. S. A. 97, 3438–3443
41 Grachtchouk, M. et al. (2000) Basal cell carcinomas
in mice overexpressing Gli2 in skin. Nat. Genet.
24, 216–217
42 Park, H.L. et al. (2000) Mouse Gli1 mutants are
viable but have defects in SHH signaling in
combination with a Gli2 mutation. Development
127, 1593–1605
43 Mo, R. et al. (1997) Specific and redundant functions
of Gli2 and Gli3 zinc finger genes in skeletal
patterning and development. Development
124, 113–123

TRENDS in Cell Biology
Vol.12 No.12 December 2002
http://tcb.trends.com 0962-8924/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0962-8924(02)02402-9
569
Review
Owen Pornillos
Jennifer E. Garrus
Wesley I.Sundquist*
Dept of Biochemistry,
University of Utah, Salt
Lake City, UT 84132, USA.
*e-mail: wes@
biochem.utah.edu
Enveloped RNA viruses include pathogens with
exceptional histories of morbidity, such as HIV
(60 million people infected, and 20 million people
killed since 1981 [1]) and influenza (about 500 million
infections each year, and 40 million people killed in
the 1918 pandemic alone [2]), as well as important
emerging pathogens such as the Ebola virus. Their
medical importance and fascinating biology has made
them (and indeed all viruses) the subject of intense
study, which has revealed recurring themes that
underlie many viral replication strategies. First, the
limited coding capacity of RNA viruses forces them to
use host cell factors to extend their capabilities.
Second, viral proteins often achieve this by mimicking
the structures and functions of cellular proteins.
Third, different viruses, as well as cells themselves,
frequently use similar mechanisms to accomplish
difficult molecular transformations. Finally, whenever
possible, viruses modify their cellular environment to
maximize replication and to minimize cellular
responses (whereas cells do the converse).
These principles are nicely illustrated by the process
of enveloped RNA virus entry, which is initiated when
viral envelope proteins bind to cell-surface receptors.
These interactions frequently mimic ligand–receptor
interactions and can thereby initiate signaling
cascades that trigger endocytosis and/or increase the
permissivity of the host cell environment. Analyses of
the molecular basis of virus entry have identified the
unexpected conservation of a three-stranded, coiled-coil
architecture in viral envelope proteins that mediates
membrane fusion [3] and have also revealed surprising
similarities between this machinery and the SNARE
(soluble NSF attachment protein receptor) complexes
that mediate cellular vesicle fusion [4,5].
Here we review our emerging understanding of
the reciprocal process: that is, how HIV and other
enveloped RNA viruses exit cells. We also speculate
on how these processes, although not yet fully
elucidated, can be understood ultimately in terms of
the principles of viral replication outlined above.
HIV assembly
Late in the infectious cycle of HIV [6], the viral Gag
polyprotein captures the RNA genome, binds to the
plasma membrane and assembles into spherical,
enveloped particles that bud from the cell [7]. Gag is
organized into four distinct regions (see Fig. 1),
which carry out different primary functions in the
coordinated processes of viral assembly and egress: the
N-myristoylated MA domain targets Gag to the plasma
membrane, CA makes important protein–protein
interactions that are required for particle assembly,
NC captures the viral RNA genome and couples RNA
binding to particle assembly, and p6 recruits cellular
proteins that function in the final stages of virus release.
Although Gag is processed by the viral protease
to produce infectious virions, extracellular particles
are still produced in the absence of the viral protease,
To spread infection, enveloped viruses must bud from infected host cells.
Recent research indicates that HIV and other enveloped RNA viruses bud by
appropriating the cellular machinery that is normally used to create vesicles
that bud into late endosomal compartments called multivesicular bodies.
This new model of virus budding has many potential implications for cell
biology and viral pathogenesis.
Published online: 5 November 2002
Mechanisms of enveloped RNA
virus budding
Owen Pornillos, Jennifer E.Garrus and Wesley I. Sundquist
44 Brown, S.A. et al. (1998) Holoprosencephaly due
to mutations in ZIC2, a homologue of Drosophila
odd-paired. Nat. Genet. 20, 180–183
45 Aruga, J. et al. (2002) Zic2 controls cerebellar
development in cooperation with Zic1.
J. Neurosci. 22, 218–225
46 Brewster, R. et al. (1998) Gli/Zic factors pattern
the neural plate by defining domains of cell
differentiation. Nature393, 579–583
47 Koyabu, Y. et al. (2001) Physical and functional
interactions between Zic and Gli proteins. J. Biol.
Chem.276, 6889–6892
48 Kim, P.C. et al. (2001) Murine models of
VACTERL syndrome: Role of sonic hedgehog
signaling pathway. J. Pediatr. Surg. 36, 381–384
49 Koike, C. et al. (2002) Introduction of wild-type
patched gene suppresses the oncogenic potential
of human squamous cell carcinoma cell lines
including A431. Oncogene 21, 2670–2678
50 Sato, N. et al. (1999) Induction of the hair growth
phase in postnatal mice by localized transient
expression of Sonic hedgehog. J. Clin. Invest.
104, 855–864
51 Sato, N. et al. (2001) Effect of adenovirus-
mediated expression of Sonic hedgehog gene
on hair regrowth in mice with chemotherapy-
induced alopecia. J. Natl. Cancer Inst.
93, 1858–1864
52 Tsuboi, K. and Shults, C.W. (2002) Intrastriatal
injection of sonic hedgehog reduces behavioral
impairment in a rat model of Parkinson’s disease.
Exp. Neurol. 173, 95–104
53 Pepinsky, R.B. et al. (2002) Long-acting forms of
Sonic hedgehog with improved pharmacokinetic
and pharmacodynamic properties are efficacious
in a nerve injury model. J. Pharm. Sci.
91, 371–387
54 Ahlgren, S.C. et al. (2002) Sonic hedgehog rescues
cranial neural crest from cell death induced by
ethanol exposure. Proc. Natl. Acad. Sci. U. S. A.
99, 10476–10481
55 Taipale, J. et al. (2000) Effects of oncogenic
mutations in Smoothened and Patched
can be reversed by cyclopamine. Nature
406, 1005–1009