*The Campbell Family
Institute for Breast Cancer
Research, University Health
Network, University of
Toronto, Toronto, Ontario
M5G 2C1, Canada.
‡Current address: Signal
Cancer Research UK London
Research Institute, 44
Lincoln’s Inn Fields, London
WC2A 3PX, UK.
§The Cancer Institute of New
Jersey and the Institute for
Advanced Study, New Jersey,
New Jersey, USA.
Correspondence to T.W.M.
2 February 2006
Beyond PTEN mutations: the PI3K
pathway as an integrator of multiple
inputs during tumorigenesis
Megan Cully*‡, Han You*, Arnold J. Levine§ and Tak W. Mak*
Abstract | The tumour-suppressor phosphatase with tensin homology (PTEN) is the most
important negative regulator of the cell-survival signalling pathway initiated by phosphati-
dylinositol 3-kinase (PI3K). Although PTEN is mutated or deleted in many tumours,
deregulation of the PI3K–PTEN network also occurs through other mechanisms. Crosstalk
between the PI3K pathways and other tumorigenic signalling pathways, such as those that
involve Ras, p53, TOR (target of rapamycin) or DJ1, can contribute to this deregulation. How
does the PI3K pathway integrate signals from numerous sources, and how can this
information be used in the rational design of cancer therapies?
The tumour suppressor PTEN (phosphatase with tensin
homology, which is deleted on chromosome 10) was
originally identified as a gene that is mutated in multiple
sporadic tumour types as well as in patients with can-
cer predisposition syndromes such as Cowden disease.
PTEN is a lipid phosphatase that negatively regulates
the phosphatidylinositol 3-kinase (PI3K) signalling
pathway1. The PI3K pathway is an important driver
of cell proliferation and cell survival, most notably in
cells that are responding to growth-factor–receptor
engagement. By opposing the effects of PI3K activa-
tion, PTEN functions as a tumour suppressor. So, the
PI3K–PTEN signalling network functions as a crucial
regulator of cell survival decisions. When PTEN is
deleted, mutated or otherwise inactivated, activation of
PI3K effectors — particularly the activation of the key
survival kinase protein kinase B (PKB, also known as
AKT) — can occur in the absence of any exogenous
stimulus, and tumorigenesis can be initiated. Numerous
types of tumours, both sporadic and those that arise as a
component of a cancer predisposition syndrome, show
alterations in PTEN2.
PTEN mutations — the tip of the iceberg?
When PTEN was first discovered, it seemed likely that
PTEN mutations would account for most of the cases
of PI3K pathway deregulation observed in tumours3–5.
However, 8 years of analysis have shown some intrigu-
ing discontinuities between PTEN mutations, the
tumour spectrum of Cowden disease4 and the activa-
tion of known downstream PI3K-pathway components
such as PKB. For example, spontaneous forms of breast
cancer rarely show loss of both PTEN alleles, but this
tumour type is commonly observed in patients with
germline PTEN mutations2,4. Immunohistochemical
staining of breast tumour samples has shown that
approximately half contain hyperactive PKB signal-
ling, but as few as 3% contain identifiable PTEN
mutations2,4,6. It has therefore become clear that there
must be mechanisms in addition to direct mutation
or deletion of PTEN by which the PI3K signalling
pathway can become constitutively activated. Within
the complex environment of an organism, cells are
continually integrating a plethora of signals to deter-
mine cell fate, survival and proliferation. It is in this
light that the PI3K–PTEN pathway can be considered
as a central integrator of a tangled web of signalling
networks with direct and indirect effects on each other.
This integratory role casts the PI3K–PTEN pathway as
an important arbiter of cell fate. Recent data supports
the idea that crosstalk among signalling pathways con-
tributes to a deregulation of PI3K–PTEN signalling
that can lead to tumorigenesis.
Biochemistry of the PI3K pathway
Identifying the component elements of the PI3K–PTEN
signalling network and determining how they are
regulated will improve our understanding of cancer
pathogenesis and lead to the rational development
of novel therapeutics. The core PI3K pathway has
been defined through both biochemical and genetic
experiments (FIG. 1).
184 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
A benign growth.
Loss of heterozygosity
The loss of the remaining
normal allele when one allele is
already lost or mutated.
In its active form, PI3K consists of a regulatory p85
subunit and a catalytic p110 subunit. When activated
by any one of a variety of mechanisms (BOX 1), PI3K
activation results in the generation of the second mes-
senger lipid phosphatidylinositol (3,4,5) triphosphate
(PIP3). PIP3 in turn recruits both phosphatidyli-
nositol-dependent kinase 1 (PDK1) and PKB to the
membrane, where PDK1 phosphorylates and activates
PKB. There are three highly homologous isoforms of
PKB that are transcribed from independent genes
and have overlapping but distinct functions7. In mice,
PKBα (also known as AKT1) mediates signals down-
stream of PI3K activation that promote cell survival
and proliferation. By contrast, PKBβ (also known as
AKT2) activation is associated with insulin-mediated
metabolic processes8,9. Pkbγ–/– (also known as Akt3)
mice have reduced brain size and weight, which might
be attributed to reduced cell size and cell number10.
The net result of the activation of all isoforms of PKB is
protection from apoptosis and increased proliferation
— events that favour tumorigenesis.
Several direct substrates of PKB phosphorylation
have crucial roles in cell-cycle regulation. These sub-
strates include the cell-cycle inhibitor p27 (also known
as KIP1), the forkhead box transcription factors (FOXO),
glycogen synthase kinase 3 (GSK3), serum- and gluco-
corticoid-induced kinase 1 (SGK1) and tuberous sclerosis
complex 2 (TSC2)11–13 (FIG. 1). Phosphorylation of p27 by
PKB results in p27 inactivation and thereby promotes
cell cycle entry. In addition, p27 expression is subject
to another level of regulation, which is mediated by
FOXO3A. When unphosphorylated, FOXO3A functions
as a selective transcription factor in the nucleus, inducing
the transcription of the genes that encode p27, the cell-
cycle-inhibitor cyclin G2 and the pro-apoptotic molecule
BIM14. Phosphorylation of FOXO3A by PKB results in
expulsion of FOXO3A from the nucleus and, therefore,
decreased transcription of the gene that encodes p27. In
addition, nuclear exclusion of FOXO3A increases cyclin
D1 expression, as unphosphorylated FOXO3A functions
as a transcriptional repressor for this gene (among several
others)15. Interestingly, Drosophila melanogaster Foxo and
mammalian FOXO1 have been found to transcription-
ally regulate expression of the insulin receptor when
the nutrient supply is limited. FOXO proteins can act
as insulin sensors that allow the rapid activation of the
insulin signalling pathway during times of low nutrient
levels16. FOXO could therefore mediate a key feedback-
control mechanism that regulates insulin signalling
Another key molecule inactivated by PKB phos-
phorylation is TSC2. When unphosphorylated, TSC2
hetero dimerizes with TSC1 to promote the GTPase
activity of RHEB, a Ras homologue that is highly
expressed in brain tissue17. Active, GTP-bound RHEB
promotes the activity of the kinase TOR (target of
rapamycin). TOR functions as a nutrient sensor that
integrates PI3K-mediated growth-factor signalling,
glucose availability and amino-acid availability18.
PKB activation inhibits the ability of the TSC1–TSC2
complex to act as a RHEB-GTPase activating protein
(RHEB-GAP), which therefore increases the amount of
RHEB-GTP present and consequently activates TOR.
Activated TOR exists in two complexes. The
rapa mycin-sensitive TOR complex contains rap-
tor (regulatory associated protein of TOR) and GβL
(G-protein β-subunit-like), and phosphorylates S6
kinase (S6K) and the EIF4E (eukaryotic transla-
tion-initiation factor 4E)-inhibitory binding protein
4EBP19–22. Phosphorylated S6K is active, and might
affect protein translation and cell size, although the
mechanisms remain controversial. In both mice and
D. melanogaster, deficiency for s6k results in decreased
cell size23,24. The second, rapamycin-insensitive TOR
complex contains rictor (rapamycin-insensitive com-
panion of TOR) and mediates signals to the cytoskel-
eton25–27. The rictor-containing TOR complex can also
phosphorylate and activate PKB28.
PI3K, PTEN and the TSC1–TSC2–TOR axis
The signalling axis that involves the TSC1–TSC2 complex
and TOR (TSC1–TSC2–TOR) has become a focal point
in studies of PI3K-mediated tumorigenesis for two rea-
sons. First of all, mutations in either TSC1 or TSC2 lead
to tuberous sclerosis, a hamartoma syndrome associated
with a predisposition to malignancy17. Notably, tuber-
ous sclerosis hamartomas generally display loss of
heterozygosity LOH at the mutant locus29. Secondly,
there is indirect evidence derived from studies of TOR
inhibition that a decrease in TOR activity prevents tumour
development in both humans and mice. For example,
chemical TOR inhibitors such as the rapamycin deriva-
tives CCI-779, RAD001 and AP23573 seem to have anti-
tumour activity for a wide range of malignancies, including
At a glance
• The phosphatidylinositol 3-kinase (PI3K)–phosphatase with tensin homology (PTEN)
signalling pathway is one of the most commonly altered pathways in human tumours.
However, mutations of the PTEN gene itself account for only a fraction of these
• The PI3K–PTEN pathway promotes cell survival and proliferation, increases in cell size
and chemoresistance. Each of these biological outcomes results from the interaction
of this pathway with other signalling networks.
• Ras and its downstream effectors can activate components of the PI3K–PTEN pathway
through numerous mechanisms. Each mechanism might be restricted to a particular
tumour type, allowing the design of a specific therapy that kills cancer cells but leaves
normal tissue unharmed.
• Crosstalk between the PI3K–PTEN and p53 pathways occurs at multiple nodes in
these pathways. When both PTEN and p53 are inactivated by mutations, malignancy is
promoted in a synergistic manner.
• The Ras, PI3K–PTEN and p53 pathways all converge either directly or indirectly on the
tumour suppressor TSC2, indicating a crucial role for this molecule in the integration
of multiple signals.
• DJ1 is a novel regulator of the PI3K–PTEN pathway and is associated with breast and
• The multiple pathways that influence the PI3K–PTEN signalling network do so through
a variety of mechanisms, providing numerous potential drug targets. Drugs that act
on these targets could be formulated to work either synergistically with agents that
act directly on PI3K or on elements that function downstream of mutated pathway
components. These drugs might offer an attractive additional or alternative approach
to combating PI3K-dependent tumours.
NATURE REVIEWS | CANCER
VOLUME 6 | MARCH 2006 | 185
© 2006 Nature Publishing Group
Apoptosis, cell-cycle arrest
Adrenal gland tumour.
non-small-cell lung cancer, breast cancer, renal-cell
carcinoma, anaplastic astrocytoma, mesothelioma,
soft-tissue sarcoma, cervical cancer and uterine can-
cer30. In particular, CCI-779 has proven to be of clinical
benefit to patients with breast or renal carcinomas31,32.
In Pten+/– mice, phaeochromocytomas and endometrial
tumours are rendered cytostatic by CCI-779 treatment33
(TABLE 1). Also, overexpression of PKB in a murine lym-
phoma model induces rapamycin-sensitive tumorigenesis
and drug resistance that can be phenocopied by over-
expression of EIF4E34. So, it seems that the maintenance of
tumours that harbour hyperactive PI3K–PKB signalling
requires intact TOR signalling.
The mechanism by which TOR activation drives
carcinogenesis remains elusive. Because rapamycin can
inhibit the formation of PKB-dependent cancers, targets
of the TOR–raptor complex must mediate at least some of
the observed tumorigenic effects. It is currently unclear
precisely how the TOR–raptor-mediated phosphoryla-
tion of S6K and 4EBP promotes tumour formation.
Although a correlation exists between increased transla-
tion and tumorigenesis, whether increased translation is
either necessary or sufficient for increased cancer suscep-
tibility has yet to be determined35. Correlative evidence
also indicates that increased cell size might be involved in
tumour formation, as brain lesions that arise in patients
Figure 1 | The PI3K–PTEN signalling network. The core phosphatidylinositol 3-kinase (PI3K) signalling pathway
(indicated by blue symbols) begins with PI3K activation (indicated by pink symbol) by receptor tyrosine kinases (RTKs)
(BOX 1). PI3K activity phosphorylates and converts the lipid second messenger phosphatidylinositol (4,5) bisphosphate
(PIP2) into phosphatidylinositol (3,4,5) triphosphate (PIP3), which recruits and activates phosphatidylinositol-dependent
kinase 1 (PDK1). PDK1 in turn phosphorylates and activates protein kinase B (PKB, also known as AKT), which inhibits the
activities of the forkhead (FOXO) transcription factors (which are mediators of apoptosis and cell-cycle arrest), resulting in
cell proliferation and survival. The tumour-suppressor phosphatase with tensin homology (PTEN) negatively regulates PI3K
signalling by dephosphorylating PIP3, converting it back to PIP2. The Ras signalling pathway (orange symbols) can be
triggered by a set of RTKs that are activated by growth factors. Ras can activate PI3K both directly and indirectly, as
described in BOX 1. The activation status of p53 can also affect the outcome of PI3K signalling by interacting with the PKB-
regulated FOXO transcription factors and with extracellular-regulated kinase 1 (ERK1) and ERK2. Other members of the
PI3K signalling pathway include SGK (serum- and glucocorticoid-induced kinase), TSC1/TSC2 (tuberous sclerosis 1 and 2),
RHEB (Ras homologue enriched in brain), TOR (target of rapamycin), 4EBP (eukaryotic initiation factor 4E (EIF4E)-binding
protein), p70S6K (ribosomal protein, S6 kinase 70kD), and PP2A (protein phosphatase 2A). Members of the RTK–Ras
signalling pathway include GRB2 (growth factor receptor-bound protein 2), SOS, Ras, Raf, MEK (mitogen-activated ERK
kinase) and ERK — activation of this pathway leads to cell proliferation.
186 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
with tuberous sclerosis often contain abnormally large
cells36. However, a definitive demonstration of causal-
ity here is also lacking. The generation and analysis of
mice that are deficient in S6k or Eif4ebp and hetero-
zygous for Tsc1, Tsc2 or Pten would establish whether
tumorigenic PI3K–TOR signalling acts through these
molecules. Interestingly, the rapamycin analogues used
in the tumour suppression studies should only affect
the raptor-containing complex. The efficacy of TOR
inhibitors might be substantially improved by the con-
comitant inhibition of PKB through pharmacological
inhibition of either PKB itself, its upstream activators,
or the TOR–rictor complex.
Between PI3K and TOR lies the TSC1–TSC2
complex on which numerous signalling pathways
converge. TSC2 is phosphorylated not only by PKB,
but also by extracellular-regulated kinase (ERK),
both of which function downstream of Ras37 (FIG. 1).
ERK-mediated phosphorylation occurs at a site
that is distinct from that phosphorylated by PKB.
Nevertheless, TSC2 phosphorylation by either PKB
or ERK disrupts the GAP activity of the TSC1–TSC2
complex. Interestingly, increased TOR activity is
found in tumours from patients with loss-of-function
mutations in the Ras-GAP NF1, and TOR is required
to sustain proliferation in these cells38. Mutations in
NF1 increase both the magnitude and duration of Ras
activation, and patients with germline NF1 mutations
have the tumour predisposition syndrome neurofi-
bromatosis type I (REF. 39). These observations indi-
cate that the PKB–TSC2–TOR pathway might have a
crucial role not only in PI3K-dependent tumours, but
also in Ras-dependent malignancies.
To what extent are tumours with mutations in Ras
or PI3K dependent on TSC1–TSC2 signalling for their
expansion and/or maintenance? Unexpectedly, expres-
sion of a mutated form of TSC2 that lacks all putative
PKB phosphorylation sites (and therefore cannot be
inhibited by PI3K signalling), or a mutant form of TSC2
that contains phospho-mimetic residues were both able
to rescue the cell size and lethality defects of D. mela-
nogaster gigas (the homologue of TSC2) mutants40. The
contribution of the ERK phosphorylation site was not
examined in this situation, and the role of this site in
development remains undetermined. During tumori-
genesis, however, the loss of one Tsc2 allele promotes
tumorigenesis in Pten+/– mice41,42. As is detailed below,
Ras can activate PI3K (and therefore PKB) as well as
ERK1/ERK2. So, abnormalities in the Ras signalling
pathway could affect TSC2 regulation through both
PKB and ERK1/ERK2. Experiments using knock-in
mice in which different TSC2 phosphorylation residues
are altered will help to dissect the in vivo contribution
of deregulated PKB and/or ERK to TSC2-associated
Between TSC1–TSC2 and TOR lies the Ras fam-
ily GTPase RHEB. Many other Ras family GTPases,
including Ras, Rac and Rho, act on more than one
downstream effector molecule. Accordingly, tumours
that are characterized by abnormal Ras signalling
often show the simultaneous activation of multiple
pathways. The same might be true for RHEB, and the
future discovery of RHEB effectors other than TOR
might shed light on the mechanism by which aberrant
RHEB activation causes tumours. Interestingly, nei-
ther mutations in TOR nor its downstream effectors
Box 1 | Mechanisms of PI3K activation
Phosphatidylinositol 3-kinase (PI3K) can become activated by at least three independent pathways, all of which start with
binding of ligand to receptor tyrosine kinases (RTKs). This causes RTKs to dimerize and undergo autophosphorylation (P)
at tyrosine residues, which allows them to interact with Src homology 2 (SH2)-domain-containing molecules47,82.
Specificity in binding between each specific phosphotyrosine residue and its cognate
SH2-domain-containing protein is achieved through residues that surround the phosphotyrosine83. In one PI3K activation
pathway (see left side of diagram), the 85 kDa regulatory subunit of PI3K (p85) binds directly to phospho-YXXM motifs (in
which X indicates any amino-acid) within the RTK84, triggering activation of PI3K’s 110 kDa catalytic subunit (p110). Other
PI3K signalling pathways depend on the adaptor protein GRB2 (growth factor receptor-bound protein 2, red horseshoe-
shaped symbol), which binds preferentially to phospho-YXN motifs of the RTK85. In the middle pathway of the diagram,
GRB2 binds to the scaffolding protein GAB (GRB2-associated binding protein), which in turn can bind to p85. GRB2 also
activates Ras (through the activation of SOS), and Ras
activates p110 independently of p85. GRB2 can also exist
in a large complex that contains both SOS, Ras, and GAB or
other scaffolding proteins, bringing these activators into
close proximity with p110 PI3K86. It is not clear precisely
which of these pathways predominates in different
physiological situations. There is evidence that Ras has to
function in concert with phosphotyrosine-bound p85 to
activate p110. For example, HRAS promotes the catalytic
activity of PI3K only when p85 is bound to phosphorylated
tyrosine residues87. So, Ras-mediated PI3K activation might
require two steps: the phosphorylation of YXXM motifs,
and the activation of small GTPases. Experiments that
directly address the in vivo relevance of Ras–PI3K
interactions will further elucidate the importance of Ras-
mediated PI3K activation in normal cell signalling and
tumorigenesis. PKB, protein kinase B.
NATURE REVIEWS | CANCER
VOLUME 6 | MARCH 2006 | 187
© 2006 Nature Publishing Group
S6K and 4EBP have been found in human cancers,
which begs the question of whether TOR is the sole
downstream effector of RHEB. If tumours that show
hyperactivated PKB- or Ras-signalling are neverthe-
less dependent on TOR, S6K or EIF4EBP for survival
or proliferation, the lack of mutations in these genes
might provide a unique opportunity for inhibiting a
crucial pathway without promoting drug resistance.
Imatinib (Glivec), the relatively specific BCR–ABL
kinase inhibitor that is currently used for the treat-
ment of chronic myeloid leukaemia, tends to lose
efficacy over the course of treatment as mutations
accumulate in the tumorigenic BCR–ABL fusion
gene43. If activating TOR mutations continue to be
rare, even in the presence of TOR-specific inhibi-
tors, these chemotherapeutic agents might have a
prolonged term of efficacy and might be useful even
in late-stage disease.
Similar to most other signalling pathways, the
PI3K pathway contains mechanisms by which it can
turn itself off. For example, a negative-feedback loop
probably exists downstream of the TOR effector S6K
because S6K activation results in both the transcrip-
tional repression and the inhibitory phosphorylation
of insulin receptor substrate (IRS) proteins44–46. IRS
proteins are adaptor molecules that are phosphory l-
ated in response to insulin or insulin-like growth
factors (IGFs). Phosphorylated IRS proteins relay
signals from receptors that bind these growth factors
to both PI3K and Ras47. IRS protein activation might
account for the different tumour spectra observed in
patients with germline PTEN mutations versus those
with germline TSC1 and TSC2 mutations. Mutations
in TSC1 and TSC2 presumably activate the nega-
tive-feedback loop, thereby activating only the TOR
pathway, whereas normal regulatory control of other
PI3K-mediated pathways would still be in place. On
the other hand, PTEN mutations presumably lead to
the hyperactivation of all PI3K-mediated pathways.
Accordingly, Cowden disease, which is associated
with germline PTEN mutations and therefore with
the activation of multiple PI3K-mediated signals,
is characterized by a much higher cancer risk than is
tuberous sclerosis (germline TSC2 mutations and TOR
activation)29,48. Furthermore, negative feedback from
a hyperactivated PI3K pathway would be predicted to
decrease the activity of other IRS-mediated responses
to insulin or IGFs. Consistent with this hypothesis,
IGF1 is unable to induce ERK phosphorylation in
Pten–/– mouse embryonic fibroblasts (MEFs) (M.C. and
T.W.M., unpublished observations). Pharmacological
inhibition of the TOR pathway could remove this
negative-feedback loop, thereby increasing the activ-
ity of other branches of the PI3K pathway as well as
additional molecules activated by IRS proteins. TOR
inhibitors might therefore be more effective as cancer
therapeutics when used in combination with inhibitors
that target either upstream molecules (such as growth
factor receptors) or other pathways activated by IRS
proteins, such as Ras or PI3K signalling pathways.
Table 1 | Tumour development in Pten+/– mice
Gene or drug
Pten+/– mice alone
Tumour tissue site
Lymph nodes, endometrium,
adrenal gland, prostate and breast
Adrenal gland and endometrium
Characterization of tumour phenotype
Tumours are associated with loss of heterozygosity of the Pten locus
CCI-779 (TOR inhibitor)
Cytostatic, decreased proliferation
Cdkn1b–/–Pten+/– mice develop carcinomas with complete penetrance,
whereas Pten+/– mice develop hyperplasias with ~50% penetrance
Decreased latency; altered tumour spectrum — Cdkn2a–/–Pten+/– mice
develop melanomas and squamous cell carcinomas
Increased penetrance of tumour phenotype
Adrenal gland, prostate and
Lymphoma (spontaneous or
Truncated SV40 large-
Increased invasiveness of tumours
Increased metastasis, high grade prostatic intraepithelial neoplasia
Transgenic expression of SV40 large-T antigen inactivates pRB,
p107 and p130 and induces astrocytomas; astrocytoma frequency is
increased in Pten+/– mice
Increased invasion (prostate) and penetrance (lymph node
Accelerated tumour formation
Prostate and lymph node 41,42
Mlh1–/– (involved in DNA
No change in tumour formation
Accelerated tumour formation in Pten+/– mice and in a model of breast
cancer based on a mutant form of polyoma middle-T antigen
Accelerated tumour formation; decreased frequency of RasV12
Dimethylbenzoic acid Skin carcinogenesis 55
Cdkn, cyclin-dependent kinase inhibitor; Grb2, growth factor receptor-bound protein 2; Nkx3.1, mouse homologue of NK homeobox, family 3, member A;
Pten, phosphatase with tensin homology; TACC1, transforming, acidic coiled-coil containing protein 1; TOR, target of rapamycin; Tsc2, tuberous sclerosis 2.
188 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
PI3K, PTEN and the Ras signalling pathwa y
Although PI3K is activated by the direct binding of its p85
regulatory subunit to a phosphorylated receptor tyrosine
kinase (RTK), additional mechanisms of PI3K activation
exist (BOX 1). The p110 (catalytic) subunit of PI3K can
also be activated through interaction with Ras49. In vivo
data support a functional link between the Ras and PI3K
pathways. In endometrial tumours, as well as in cell lines
derived from melanomas, mutation of the RAS and PTEN
genes is mutually exclusive50,51. Whereas RAS mutations
are common in pancreatic, lung and colon cancers, they
are rarely found in glioblastomas; the opposite is true for
PTEN mutations52–54. In mouse models of chemically-
induced skin carcinogenesis, Ras mutations typically
arise in Pten+/+ mice. In Pten+/– mice, there is a decreased
incidence of Ras mutations55. Furthermore, tumours that
lack Ras mutations tend to lose the second (wild-type)
Pten allele. These observations indicate that Ras activa-
tion and PTEN loss probably serve the same function
In addition to Ras itself, there are a number of cyto-
plasmic molecules that can activate PI3K, including the
IRS and GAB (growth factor receptor-bound protein 2
(GRB2)-associated binding protein ) families of adaptor
molecules. When RTKs engage ligand, IRS and GAB1/
GAB2 become phosphorylated at tyrosine residues, result-
ing in PI3K activation. Activation of GAB2 is required
for BCR–ABL-mediated leukaemogenesis in mice, and
Gab2–/– cells are resistant to BCR–ABL-induced trans-
formation56. Another adaptor molecule, GRB2, functions
upstream of both Ras and GAB1/GAB2, and therefore
has an important role in both Ras and GAB-mediated
activation of PI3K. Grb2+/– mice are partially resistant to
polyoma-middle-T-antigen-induced mammary tumori-
genesis, a system in which both Ras and PI3K are acti-
vated57. Breast cancer cells that overexpress the epidermal
growth factor (EGF) receptor ERBB2 depend on GRB2
activity for both proliferation and tumour formation58.
The identification of these multiple mechanisms
that can activate PI3K signalling might be of significant
therapeutic value, as drugs that target each of these
interactions could have differential effects, based on the
activity of each signalling element in different tissues.
For example, inhibition of GRB2 might be the most
effective way to block the PI3K signalling pathway in
mammary tumours, whereas inhibition of the Gab fam-
ily proteins could be a therapeutic target in the treatment
of leukaemia. Although protein–protein interactions are
notoriously difficult to disrupt through chemical inhibi-
tion, the in vivo delivery of inhibitory RNA molecules
could be a feasible approach to inhibiting this pathway59.
By choosing targets far upstream in the PI3K pathway,
it might be possible to change the activation status of
numerous downstream effectors and therefore reduce
their contributions to tumour formation.
PTEN and p53
The tumour suppressor p53 activates the transcription of
both PTEN and TSC2, and therefore functions as a nega-
tive regulator of the entire PI3K signalling pathway60–62.
The downregulation of PI3K signalling by p53 activation
is further enforced through p53-mediated transcrip-
tional repression of the gene encoding the p110 subunit
of PI3K63. p53 is activated in response to a wide variety of
cellular stress signals, including DNA damage, hypoxia,
mitotic spindle damage, heat and cold shock, inflamma-
tion, nitric oxide production and oncogene activation64–66.
These stresses all have the potential to decrease the fidel-
ity of cell-cycle progression and DNA replication, and
thereby increase mutation rates in cells. Transcriptional
regulation of the PI3K pathway by p53 eliminates PI3K-
mediated survival and proliferation signals, providing
an additional level of protection against continued DNA
replication in the presence of genotoxic stress.
Activation of p53 can also induce cellular senes-
cence65. The role of senescence in blocking cancer
formation has been vigorously debated for many
years, but a series of recent publications indicate
that senescence, induced by activation of either Ras
or p53, can prevent tumour progression67–70. Most
notably, prostate-specific disruption of Pten in mice
results in prostate neoplasias that are associated with
senescence markers. Additional deletion of Trp53 in
these mice prevents this senescence and results in the
formation of malignant tumours67.
p53 is one of the most commonly mutated genes in
human cancers. However, despite our extensive knowl-
edge of p53 function and regulation, significant difficulties
in gene delivery systems have limited our ability to restore
wild-type p53 expression to tumour cells. Inhibition of
PI3K signalling might offer a means of treating patients
who have tumours that carry p53 mutations. Because the
p53 apoptotic response requires the downregulation of
the PI3K pathway through the transcriptional activation
of PTEN60, simultaneous inhibition of the PI3K pathway
and activation of apoptosis downstream of p53 might
have synergistic effects. In addition, drugs that activate
senescence pathways downstream of p53 or Ras could
also synergistically increase the anti-tumour effects of
drugs that target PI3K or TOR.
PI3K, p53 and FOXO
Another example of intermolecular crosstalk exists
in the recently documented interaction between the
tumour suppressors p53 and FOXO3A. The activation,
inactivation and even the selection of genes to be tran-
scribed or repressed by p53 or FOXO is accomplished in
part by protein modifications. DNA damage triggers p53
activation, which in turn leads to the induction of SGK1
and the subsequent nuclear expulsion of FOXO3A71. In
this way, p53 regulates the expulsion of FOXO3A from
the nucleus in response to DNA damage. This seemingly
counter-intutitive phenomenon could be part of a posi-
tive-feedback loop required for complete p53 activation,
as nuclear FOXO3A can inactivate p53 transcriptional
activity (H.Y. et al., unpublished data). Translocation of
FOXO3A into the cytoplasm therefore allows p53 to act
in the nucleus. This reciprocal inhibition of p53 and
FOXO3A is avoided under normal circumstances — as
p53 concentrations increase in response to stress signals,
the induction of SGK1 results in the nuclear exclusion
of FOXO3A71. However, FOXO3A has either pro- or
NATURE REVIEWS | CANCER
VOLUME 6 | MARCH 2006 | 189
© 2006 Nature Publishing Group
anti-apoptotic roles that are context-dependent72, so
the biological significance of this crosstalk is currently
PI3K, PTEN and PARK7
Human breast and lung tumour cells rarely have PTEN
mutations, but hyperactivation of PKB is observed in
25–75% of these tumours73,74. Several mechanisms have
been proposed to drive this hyperactivation. For exam-
ple, lung cancers commonly carry activating mutations
in Ras, which would be expected to activate PKB through
both PI3K and TOR. The inactivation of p53 — another
frequent occurrence in these tumour types — should
decrease PTEN expression and result in increased levels
of phosphorylated PKB. A third mechanism arises from
the results of a recent study in which expression of the
Parkinson-disease-associated gene PARK7 (also known
as DJ1) was correlated with the presence of increased
phosphorylated PKB in both breast and lung cancer
samples. Moreover, increased DJ1 expression correlated
with an increased rate of relapse74. In the lung, this cor-
relation was strongest in early-stage tumours but weakest
in late-stage tumours that carried Ras mutations. How
DJ1 fits into the PI3K–PTEN signalling network and
negatively regulates PKB activation remains unclear.
DJ1 might inhibit PTEN or provide another type of
PKB-activating signal in early-stage tumours. At later
stages, genomic instability might result in mutations that
alleviate the selective pressure for DJ1 overexpression.
The precise function of DJ1 and its relationship to PI3K,
PKB and PTEN remain under investigation.
A new paradigm in cancer treatment is the rational
development of anticancer drugs, such as Herceptin, that
specifically target molecules that drive tumorigenesis.
Although this new generation of therapeutics holds great
promise, our currently incomplete understanding of the
molecular mechanisms that control key pathways in both
normal and cancer cells limits our ability to use these
drugs efficiently. Several studies have demonstrated the
difficulties in determining which subset of patients will
benefit from treatment with a targeted agent75–79. These
difficulties arise because we do not have biomarkers that
can identify patients who are most likely to respond to
therapy. Detailed dissection of tumorigenic pathways
will allow us to develop ‘molecular signatures’ that can
provide information on the activation status of the many
individual pathways within a cell. Analysis of these sig-
natures should aid us in predicting which drugs will be
effective for which patients. Conventional chemotherapy
often leads to the emergence of drug resistance, provid-
ing further evidence that the simultaneous targeting of
multiple key pathways will be the most effective strategy
for killing cancer cells.
Determining which factors to target requires a thor-
ough understanding of the crosstalk among numerous
signalling pathways, some or all of which could be
activated within a given tumour. Blanket inhibition of
multiple pathways is not an option because of the risk of
side-effects. For example, the inappropriate inhibition
of the PI3K pathway has been associated with diseases
as diverse as diabetes and schizophrenia80,81 (BOX 2). The
efficacy and safety of these inhibitors will depend on our
understanding of the role of PI3K in these disorders as
well as during tumour formation. Gene expression pro-
filing to select patients who are most likely to respond
to a certain inhibitor could help us to avoid testing the
efficacy of drugs in the wrong patient cohort. The selec-
tion of appropriate patients for a clinical trial is of para-
mount importance in judging a new drug’s true merits.
Pre-screening candidate patients by molecular signature
can identify a suitable subset of subjects for each new
drug and predict which drugs might have synergistic
effects. The experimental answers derived from this type
of work are only as good as the questions. By expanding
our understanding of the PI3K signalling network we
will not only be able to ask the right questions, but, more
importantly, find the right treatment for each patient.
Box 2 | Potential side-effects of PI3K inhibition
Because the phosphatidylinositol 3-kinase (PI3K) signalling pathway is an important
regulator of numerous normal cell processes, drugs that target this enzyme could
potentially have many side-effects. For example, PI3K activity is important for insulin
signalling and metabolism, so inhibition of PI3K signalling in an effort to control
tumour progression could lead to decreased insulin sensitivity and diabetes8,88.
Regulation of PI3K is also required for normal brain function, as decreased PI3K
signalling has been associated with schizophrenia80. Similarly, the successful treatment
of bipolar disorder with lithium, a glycogen synthase kinase 3 (GSK3) inhibitor, implies
that PI3K signalling cannot be reduced below a certain crucial level89. Parkinson
disease might be another potential consequence of drug-induced PI3K deregulation.
Homozygous mutations in the Parkinson-disease-associated gene PARK7 (also known
as DJ1) are associated with early-onset Parkinson disease90. Loss-of-function mutations
in Dj1 attenuated dopamine-dependent behaviours and increased sensitivity to
dopaminergic neuron loss induced by oxidative stress in mice91–93. Increased
expression levels of DJ1 have been reported in both lung and mammary tumours, and
DJ1 suppresses cell death in a PTEN (phosphatase with tensin homology)-dependent
manner in cultured mouse cells74. So, it is important to understand PI3K signalling in
both normal and diseased tissues before undertaking PI3K inhibition as an anticancer
strategy. Drugs that inhibit PI3K signalling for a short period of time in normal tissues
might have reversible or treatable side-effects, but such agents might not be useful for
the maintenance of stable or cytostatic disease. Nevertheless, because tumour cells
are often dependent on a hyperactive PI3K pathway, restoring normal levels of PI3K
signalling to these cells could be sufficient to slow tumour progression.
Pharmacological agents that decrease PI3K signalling only when this pathway is
hyperactive could therefore be of significant therapeutic value.
Stambolic, V. et al. Negative regulation of PKB/AKT-
dependent cell survival by the tumor suppressor
PTEN. Cell 95, 29–39 (1998).
Eng, C. PTEN: one gene, many syndromes. Hum.
Mutat. 22, 183–198 (2003).
Li, J. et al. PTEN, a putative protein tyrosine
phosphatase gene mutated in human brain,
breast, and prostate cancer. Science 275,
Liaw, D. et al. Germline mutations of the PTEN gene
in Cowden disease, an inherited breast and thyroid
cancer syndrome. Nature Genet. 16, 64–67 (1997).
Steck, P. A. et al. Identification of a candidate
tumour suppressor gene, MMAC1, at chromosome
10q23.3 that is mutated in multiple advanced
cancers. Nature Genet. 15, 356–362 (1997).
Shi, W. et al. Dysregulated PTEN–PKB and negative
receptor status in human breast cancer. Int. J. Cancer
104, 195–203 (2003).
Bellacosa, A., Testa, J. R., Moore, R. & Larue, L.
A portrait of AKT kinases: human cancer and animal
190 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
models depict a family with strong individualities.
Cancer Biol. Ther. 3, 268–275 (2004).
Garofalo, R. S. et al. Severe diabetes, age-dependent
loss of adipose tissue, and mild growth deficiency in
mice lacking Akt2/PKB β. J. Clin. Invest. 112,
Stiles, B. et al. Essential role of AKT-1/protein kinase
Bα in PTEN-controlled tumorigenesis. Mol. Cell. Biol.
22, 3842–3851 (2002).
10. Tschopp, O. et al. Essential role of protein kinase B γ
(PKB γ/Akt3) in postnatal brain development but not
in glucose homeostasis. Development 132,
11. Viglietto, G. et al. Cytoplasmic relocalization and
inhibition of the cyclin-dependent kinase inhibitor
p27Kip1 by PKB/Akt-mediated phosphorylation in
breast cancer. Nature Med. 8, 1136–1144 (2002).
12. Liang, J. et al. PKB/Akt phosphorylates p27, impairs
nuclear import of p27 and opposes p27-mediated G1
arrest. Nature Med. 8, 1153–1160 (2002).
13. Shin, I. et al. PKB/Akt mediates cell-cycle progression
by phosphorylation of p27Kip1 at threonine 157 and
modulation of its cellular localization. Nature Med. 8,
14. Accili, D. & Arden, K. C. FoxOs at the crossroads of
cellular metabolism, differentiation, and
transformation. Cell 117, 421–426 (2004).
15. Ramaswamy, S., Nakamura, N., Sansal, I., Bergeron, L.
& Sellers, W. R. A novel mechanism of gene regulation
and tumor suppression by the transcription factor
FKHR. Cancer Cell 2, 81–91 (2002).
16. Puig, O. & Tjian, R. Transcriptional feedback control of
insulin receptor by dFOXO/FOXO1. Genes Dev. 19,
17. Li, Y., Corradetti, M. N., Inoki, K. & Guan, K. L. TSC2:
filling the GAP in the mTOR signaling pathway. Trends
Biochem. Sci. 29, 32–38 (2004).
18. Hay, N. & Sonenberg, N. Upstream and downstream
of mTOR. Genes Dev. 18, 1926–1945 (2004).
19. Kim, D. H. et al. mTOR interacts with raptor to form a
nutrient-sensitive complex that signals to the cell
growth machinery. Cell 110, 163–175 (2002).
20. Kim, D. H. et al. GβL, a positive regulator of the
rapamycin-sensitive pathway required for the nutrient-
sensitive interaction between raptor and mTOR. Mol.
Cell 11, 895–904 (2003).
21. Fingar, D. C., Salama, S., Tsou, C., Harlow, E. & Blenis, J.
Mammalian cell size is controlled by mTOR and its
downstream targets S6K1 and 4EBP1/eIF4E. Genes
Dev. 16, 1472–1487 (2002).
22. Miron, M., Lasko, P. & Sonenberg, N. Signaling from
Akt to FRAP/TOR targets both 4E-BP and S6K in
Drosophila melanogaster. Mol. Cell. Biol. 23,
23. Shima, H. et al. Disruption of the p70(s6k)/p85(s6k)
gene reveals a small mouse phenotype and a new
functional S6 kinase. EMBO J. 17, 6649–6659
24. Montagne, J. et al. Drosophila S6 kinase: a regulator
of cell size. Science 285, 2126–2129 (1999).
25. Jacinto, E. et al. Mammalian TOR complex 2 controls
the actin cytoskeleton and is rapamycin insensitive.
Nature Cell Biol. 6, 1122–1128 (2004).
26. Loewith, R. et al. Two TOR complexes, only one of
which is rapamycin sensitive, have distinct roles in cell
growth control. Mol. Cell 10, 457–468 (2002).
27. Sarbassov, D. D. et al. Rictor, a novel binding partner
of mTOR, defines a rapamycin-insensitive and raptor-
independent pathway that regulates the cytoskeleton.
Curr. Biol. 14, 1296–1302 (2004).
28. Sarbassov, D. D., Guertin, D. A., Ali, S. M. & Sabatini,
D. M. Phosphorylation and regulation of Akt/PKB by
the rictor–mTOR complex. Science 307, 1098–1101
Identifies TOR as the kinase responsible for
phosphorylating and activating PKB.
29. Yeung, R. S. Multiple roles of the tuberous sclerosis
complex genes. Genes Chromosomes Cancer 38,
30. Chan, S. Targeting the mammalian target of rapamycin
(mTOR): a new approach to treating cancer. Br. J.
Cancer 91, 1420–1424 (2004).
31. Atkins, M. B. et al. Randomized phase II study of
multiple dose levels of CCI-779, a novel mammalian
target of rapamycin kinase inhibitor, in patients with
advanced refractory renal cell carcinoma. J. Clin.
Oncol. 22, 909–918 (2004).
32. Chan, S. et al. Phase II study of temsirolimus (CCI-
779), a novel inhibitor of mTOR, in heavily pretreated
patients with locally advanced or metastatic breast
cancer. J. Clin. Oncol. 23, 5314–5322 (2005).
33. Podsypanina, K. et al. An inhibitor of mTOR reduces
neoplasia and normalizes p70/S6 kinase activity in
Pten+/– mice. Proc. Natl Acad. Sci. USA 98,
One of the first demonstrations that rapamycin, a
pharmacological inhibitor of TOR, can inhibit
tumour growth in Pten-deficient mice.
34. Wendel, H. G. et al. Survival signalling by Akt and
eIF4E in oncogenesis and cancer therapy. Nature 428,
Shows that both tumour formation and the
chemoresistance found in PKB-overexpressing
tumour models depend on the 4EBP target eIF4E.
35. Ruggero, D. & Pandolfi, P. P. Does the ribosome
translate cancer? Nature Rev. Cancer 3, 179–192
36. Kwiatkowski, D. J. Rhebbing up mTOR: new insights
on TSC1 and TSC2, and the pathogenesis of tuberous
sclerosis. Cancer Biol. Ther. 2, 471–476 (2003).
37. Ma, L., Chen, Z., Erdjument-Bromage, H., Tempst, P. &
Pandolfi, P. P. Phosphorylation and functional
inactivation of TSC2 by Erk implications for tuberous
sclerosis and cancer pathogenesis. Cell 121,
38. Johannessen, C. M. et al. The NF1 tumor suppressor
critically regulates TSC2 and mTOR. Proc. Natl Acad.
Sci. USA 102, 8573–8578 (2005).
39. Cichowski, K., Santiago, S., Jardim, M., Johnson, B. W.
& Jacks, T. Dynamic regulation of the Ras pathway via
proteolysis of the NF1 tumor suppressor. Genes Dev.
17, 449–454 (2003).
40. Dong, J. & Pan, D. Tsc2 is not a critical target of Akt
during normal Drosophila development. Genes Dev.
18, 2479–2484 (2004).
41. Ma, L. et al. Genetic analysis of Pten and Tsc2
functional interactions in the mouse reveals
asymmetrical haploinsufficiency in tumor suppression.
Genes Dev. 19, 1779–1786 (2005).
Identifies the ERK phosphorylation sites on TSC2,
which indicates an important connection between
the Ras and TOR pathways.
42. Manning, B. D. et al. Feedback inhibition of Akt
signaling limits the growth of tumors lacking Tsc2.
Genes Dev. 19, 1773–1778 (2005).
43. Deininger, M., Buchdunger, E. & Druker, B. J. The
development of imatinib as a therapeutic agent for
chronic myeloid leukemia. Blood 105, 2640–2653
44. Simpson, L. et al. PTEN expression causes feedback
upregulation of insulin receptor substrate 2. Mol. Cell.
Biol. 21, 3947–3958 (2001).
45. Harrington, L. S. et al. The TSC1–2 tumor suppressor
controls insulin–PI3K signaling via regulation of IRS
proteins. J. Cell Biol. 166, 213–223 (2004).
46. Gual, P., Gremeaux, T., Gonzalez, T., Le Marchand-
Brustel, Y. & Tanti, J. F. MAP kinases and mTOR
mediate insulin-induced phosphorylation of insulin
receptor substrate-1 on serine residues 307, 612 and
632. Diabetologia 46, 1532–1542 (2003).
47. Pawson, T. & Nash, P. Assembly of cell regulatory
systems through protein interaction domains. Science
300, 445–452 (2003).
48. Marsh, D. J. et al. Mutation spectrum and genotype–
phenotype analyses in Cowden disease and
Bannayan–Zonana syndrome, two hamartoma
syndromes with germline PTEN mutation. Hum. Mol.
Genet. 7, 507–515 (1998).
49. Rodriguez-Viciana, P. et al. Phosphatidylinositol-3-OH
kinase as a direct target of Ras. Nature 370,
50. Tsao, H., Zhang, X., Fowlkes, K. & Haluska, F. G.
Relative reciprocity of NRAS and PTEN/MMAC1
alterations in cutaneous melanoma cell lines. Cancer
Res. 60, 1800–1804 (2000).
51. Ikeda, T. et al. Anticorresponding mutations of the
KRAS and PTEN genes in human endometrial cancer.
Oncol. Rep. 7, 567–570 (2000).
52. Liu, W., James, C. D., Frederick, L., Alderete, B. E. &
Jenkins, R. B. PTEN/MMAC1 mutations and EGFR
amplification in glioblastomas. Cancer Res. 57,
53. Simpson, L. & Parsons, R. PTEN: life as a tumor
suppressor. Exp. Cell Res. 264, 29–41 (2001).
54. Downward, J. Targeting RAS signalling pathways in
cancer therapy. Nature Rev. Cancer 3, 11–22
55. Mao, J. H. et al. Mutually exclusive mutations of the
Pten and ras pathways in skin tumor progression.
Genes Dev. 18, 1800–1805 (2004).
56. Sattler, M. et al. Critical role for Gab2 in transformation
by BCR–ABL. Cancer Cell 1, 479–492 (2002).
57. Cheng, A. M. et al. Mammalian Grb2 regulates
multiple steps in embryonic development and
malignant transformation. Cell 95, 793–803 (1998).
58. Tari, A. M. & Lopez-Berestein, G. GRB2: a pivotal
protein in signal transduction. Semin. Oncol. 28,
59. Leung, R. K. & Whittaker, P. A. RNA interference: from
gene silencing to gene-specific therapeutics.
Pharmacol. Ther. 107, 222–239 (2005).
60. Stambolic, V. et al. Regulation of PTEN transcription
by p53. Mol. Cell 8, 317–325 (2001).
61. Feng, Z., Zhang, H., Levine, A. J. & Jin, S. The
coordinate regulation of the p53 and mTOR pathways
in cells. Proc. Natl Acad. Sci. USA 102, 8204–8209
62. Inoki, K., Zhu, T. & Guan, K. L. TSC2 mediates cellular
energy response to control cell growth and survival.
Cell 115, 577–590 (2003).
63. Singh, B. et al. p53 regulates cell survival by inhibiting
PIK3CA in squamous cell carcinomas. Genes Dev. 16,
64. Levine, A. J. p53, the cellular gatekeeper for growth
and division. Cell 88, 323–331 (1997).
65. Jin, S. & Levine, A. J. The p53 functional circuit. J. Cell
Sci. 114, 4139–4140 (2001).
66. Vogelstein, B., Lane, D. & Levine, A. J. Surfing the p53
network. Nature 408, 307–310 (2000).
67. Chen, Z. et al. Crucial role of p53-dependent cellular
senescence in suppression of Pten-deficient
tumorigenesis. Nature 436, 725–730 (2005).
Indicates that senescence induced by Pten deletion
prevents incipient tumours from progressing to a
malignant state. Loss of p53 prevents this
senescence, thereby allowing prostate tumours to
form in Pten-null mice.
68. Michaloglou, C. et al. BRAFE600-associated
senescence-like cell cycle arrest of human naevi.
Nature 436, 720–724 (2005).
69. Collado, M. et al. Tumour biology: senescence in
premalignant tumours. Nature 436, 642 (2005).
70. Braig, M. et al. Oncogene-induced senescence as an
initial barrier in lymphoma development. Nature 436,
71. You, H. et al. p53-dependent inhibition of FKHRL1 in
response to DNA damage through protein kinase
SGK1. Proc. Natl Acad. Sci. USA 101, 14057–14062
72. You, H. & Mak, T. W. Crosstalk between p53 and FOXO
transcription factors. Cell Cycle 4, 37–38 (2005).
73. David, O. et al. Phospho-Akt overexpression in non-
small cell lung cancer confers significant stage-
independent survival disadvantage. Clin. Cancer Res.
10, 6865–6871 (2004).
74. Kim, R. H. et al. DJ-1, a novel regulator of the tumor
suppressor PTEN. Cancer Cell 7, 263–273 (2005).
75. Johnson, B. E. & Janne, P. A. Selecting patients for
epidermal growth factor receptor inhibitor treatment:
a FISH story or a tale of mutations? J. Clin. Oncol. 23,
76. Takano, T. et al. Epidermal growth factor receptor
gene mutations and increased copy numbers predict
gefitinib sensitivity in patients with recurrent non-
small-cell lung cancer. J. Clin. Oncol. 23, 6829–6837
77. Eberhard, D. A. et al. Mutations in the epidermal
growth factor receptor and in KRAS are predictive and
prognostic indicators in patients with non-small-cell
lung cancer treated with chemotherapy alone and in
combination with erlotinib. J. Clin. Oncol. 23,
78. Bell, D. W. et al. Epidermal growth factor receptor
mutations and gene amplification in non-small-cell
lung cancer: molecular analysis of the IDEAL/INTACT
gefitinib trials. J. Clin. Oncol. 23, 8081–8092
79. Hirsch, F. R. et al. Increased epidermal growth factor
receptor gene copy number detected by fluorescence
in situ hybridization associates with increased
sensitivity to gefitinib in patients with
bronchioloalveolar carcinoma subtypes: a Southwest
Oncology Group Study. J. Clin. Oncol. 23,
80. Emamian, E. S., Hall, D., Birnbaum, M. J.,
Karayiorgou, M. & Gogos, J. A. Convergent evidence
for impaired AKT1–GSK3β signaling in schizophrenia.
Nature Genet. 36, 131–137 (2004).
81. White, M. F. Insulin signaling in health and disease.
Science 302, 1710–1711 (2003).
82. Schlessinger, J. Ligand-induced, receptor-mediated
dimerization and activation of EGF receptor. Cell 110,
NATURE REVIEWS | CANCER
VOLUME 6 | MARCH 2006 | 191
© 2006 Nature Publishing Group
83. Songyang, Z. et al. SH2 domains recognize specific
phosphopeptide sequences. Cell 72, 767–778
84. Domchek, S. M., Auger, K. R., Chatterjee, S., Burke,
T. R. Jr & Shoelson, S. E. Inhibition of SH2 domain/
phosphoprotein association by a nonhydrolyzable
phosphonopeptide. Biochemistry 31, 9865–9870
85. Pawson, T. Specificity in signal transduction: from
phosphotyrosine-SH2 domain interactions to
complex cellular systems. Cell 116, 191–203
86. Ong, S. H. et al. Stimulation of phosphatidylinositol
3-kinase by fibroblast growth factor receptors is
mediated by coordinated recruitment of multiple
docking proteins. Proc. Natl Acad. Sci. USA 98,
87. Chan, T. O. et al. Small GTPases and tyrosine kinases
coregulate a molecular switch in the phosphoinositide
3-kinase regulatory subunit. Cancer Cell 1, 181–191
88. Cho, H. et al. Insulin resistance and a diabetes
mellitus-like syndrome in mice lacking the protein
kinase Akt2 (PKB β). Science 292, 1728–1731
89. Stambolic, V., Ruel, L. & Woodgett, J. R. Lithium
inhibits glycogen synthase kinase-3 activity and
mimics wingless signalling in intact cells. Curr. Biol. 6,
90. Bonifati, V. et al. Mutations in the DJ-1 gene
associated with autosomal recessive early-onset
Parkinsonism. Science 299, 256–259 (2003).
91. Goldberg, M. S. et al. Nigrostriatal dopaminergic
deficits and hypokinesia caused by inactivation of the
familial Parkinsonism-linked gene DJ-1. Neuron 45,
92. Kim, R. H. et al. Hypersensitivity of DJ-1-deficient
mice to 1-methyl-4-phenyl-1,2,3,6-
tetrahydropyrindine (MPTP) and oxidative stress.
Proc. Natl Acad. Sci. USA 102, 5215–5220
93. Chen, L. et al. Age-dependent motor deficits and
dopaminergic dysfunction in DJ-1 null mice. J. Biol.
Chem. 280, 21418–21426 (2005).
94. Stambolic, V. et al. High incidence of breast and
endometrial neoplasia resembling human Cowden
syndrome in pten+/– mice. Cancer Res. 60,
95. Di Cristofano, A., De Acetis, M., Koff, A., Cordon-
Cardo, C. & Pandolfi, P. P. Pten and p27KIP1 cooperate
in prostate cancer tumor suppression in the mouse.
Nature Genet. 27, 222–224 (2001).
96. You, M. J. et al. Genetic analysis of Pten and Ink4a/Arf
interactions in the suppression of tumorigenesis in
mice. Proc. Natl Acad. Sci. USA 99, 1455–1460
97. Mao, J. H. et al. Genetic interactions between Pten
and p53 in radiation-induced lymphoma development.
Oncogene 22, 8379–8385 (2003).
98. Freeman, D. J. et al. PTEN tumor suppressor
regulates p53 protein levels and activity through
mechanisms. Cancer Cell 3, 117–130 (2003).
99. Abate-Shen, C. et al. Nkx3.1; Pten mutant mice
develop invasive prostate adenocarcinoma and lymph
node metastases. Cancer Res. 63, 3886–3890
100. Xiao, A., Wu, H., Pandolfi, P. P., Louis, D. N. & Van
Dyke, T. Astrocyte inactivation of the pRb pathway
predisposes mice to malignant astrocytoma
development that is accelerated by PTEN mutation.
Cancer Cell 1, 157–168 (2002).
101. Wang, H. et al. DNA mismatch repair deficiency
accelerates endometrial tumorigenesis in Pten
heterozygous mice. Am. J. Pathol. 160, 1481–1486
102. Cully, M. et al. grb2 heterozygosity rescues
embryonic lethality but not tumorigenesis in pten+/–
mice. Proc. Natl Acad. Sci. USA 101, 15358–15363
103. Cully, M. et al. TACC1 promotes survival,
transformation and mammary tumorigenesis. Cancer
Res. 65 10363–10370 (2005).
Competing interests statement
The authors declare no competing financial interests.
The following terms in this article are linked online to:
ABL | BCR | eIF4E | ERK1 | ERK2 | FOXO3A | GAB1 | GAB2 |
GRB2 | NF1 | p27 | p53 | PDK1 | PKBα | PKBβ | Pkbγ | PTEN |
RHEB | SGK1 | TOR | TSC1 | TSC2
National Cancer Institute: http://www.cancer.gov
breast cancer | lung cancer
Nature signalling gateway:
Science signal transduction knowledge environment:
Biomolecular interaction network database: http://bind.ca/
Access to this interactive links box is free online.
192 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group