-
The endoplasmic reticulum (ER) is an essential organelle in
eukaryotic cells for calcium storage and regulated release and as
the entrance to the secretory pathway, for which approximately
one-third of all cellular proteins traffic enroute to their proper
intracellular or extra-cellular location. Numerous environmental,
physio-logical and pathological insults, as well as nutrient
fluctuations, disrupt the ER protein-folding environ-ment to cause
protein misfolding and accumulation of misfolded proteins, referred
to as ER stress. The unfolded protein response (UPR) is a
collection of sig-nalling pathways that evolved to maintain a
productive ER protein-folding environment. Both ER stress and UPR
activation are involved with the pathology of many, if not all,
degenerative diseases. Moreover, ER stress and UPR activation are
documented in the development of many cancer types (TABLE1; see
Supplementary infor-mation S1 (table)), and evidence suggests that
they have important roles in every aspect of cancer
development.
As cancer usually arises and progresses in a stressful
microenvironment, transformed cells may use UPR acti-vation as a
survival strategy. In fact, numerous studies demonstrate crucial
roles for UPR signalling in tumour growth and chemoresistance.
However, only recently has it been demonstrated that invivo UPR
activation is a vital step during oncogenic transformation and
cancer development. Recent studies also suggest that UPR signalling
molecules interact with well-established oncogene and tumour
suppressor gene networks to modulate their function during cancer
development. It will be important to understand exactly how
these
signalling pathways regulate each other, their inter-dependence
and how interference with one affects the others. Aside from its
pro-survival role, prolonged UPR activation owing to severe or
unresolved ER stress leads to cell death. This greatly complicates
the development of cancer therapies that target UPR signalling. In
this Review, we highlight recent advances in our understanding of
how UPR activation has both tumour-supporting and tumour-
suppressive roles, and we discuss strategies that target UPR
components for cancer treatment.
ER stress and UPR activation in cancerThe ER is the organelle in
eukaryotic cells that is responsible for intracellular Ca2+
homeostasis, lipid biosynthesis and protein folding and transport.
Protein folding in the ER is exquisitely sensitive to changes in
the environment, such as altered Ca2+ levels, redox state, nutrient
status, increases in the rate of protein synthesis, pathogens or
inflammatory stimuli, which lead to disrupted protein folding to
cause accumula-tion of unfolded or misfolded proteins a condition
termed ER stress. Early studies demonstrated two key events that
occur when proteins misfold in the ER. First, misfolded proteins
bind and sequester the chaperone immunoglobulin heavy-chain binding
protein (BIP; also known as GRP78 and HSP5A)1 and, second,
reduc-tion in the level of free BIP activates signalling pathways
to induce transcription of BIP, as well as other genes encoding
protein chaperones2,3, now known as the UPR. The outcome of UPR
activation involves tran-sient attenuation of protein synthesis,
increased capacity
Degenerative Diseases Program, Center for Cancer Research,
Sanford-Burnham Medical Research Institute, 10901 N.Torrey Pines
Rd, La Jolla, California 92037, USA.Correspondence to R.J.K.
e-mail: [email protected]:10.1038/nrc3800
The impact of the endoplasmic reticulum
protein-foldingenvironment on cancer developmentMiao Wang and
Randal J.Kaufman
Abstract | The endoplasmic reticulum (ER) is an essential
organelle in eukaryotic cells for the storage and regulated release
of calcium and as the entrance to the secretory pathway. Protein
misfolding in the ER causes accumulation of misfolded proteins (ER
stress) and activation of the unfolded protein response (UPR),
which has evolved to maintain a productive ER protein-folding
environment. Both ER stress and UPR activation are documented in
many different human cancers. In this Review, we summarize the
impact of ER stress and UPR activation on every aspect of cancer
and discuss outstanding questions for which answers will pave the
way for therapeutics.
REVIEWS
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 581
2014 Macmillan Publishers Limited. All rights reserved
http://www.nature.com/nrc/journal/v14/n9/full/nrc3800.html#supplementary-informationhttp://www.nature.com/nrc/journal/v14/n9/full/nrc3800.html#supplementary-informationmailto:rkaufman%40sanfordburnham.org?subject=mailto:rkaufman%40sanfordburnham.org?subject=
-
ER-associated degradation(ERAD). A process by which misfolded
proteins in the endoplasmic reticulum (ER) are targeted by
retrotranslocation and ubiquitylation for subsequent degradation by
the proteasome.
for protein trafficking through the ER, protein fold-ing and
transport, and increased protein degradative pathways, including
ERassociated degradation (ERAD) and autophagy. If these adaptive
mechanisms cannot resolve the protein-folding defect, cells enter
apoptosis. This applies not only to normal cells but also to cancer
cells. Therefore, it is not surprising that UPR activation
contributes to both enhanced survival and induced
apoptosis in cancer cells depending on the context. It needs to
be pointed out that much of what we under-stand about UPR
activation in cancer is derived from mouse models, which have
limitations (BOX1).
UPR activation in transformed cells is attributed to both
intrinsic and extrinsic factors (BOX 2; BOX 3). Hyperactivation of
oncogenes or loss-of-function mutations in tumour suppressor genes,
such as
Table 1 | Evidence of ER stress and UPR activation in various
human cancer types*
Cancer Activation of UPR components
Prognosis
Classification Site
Carcinoma Lung BIP and CHOP BIP was correlated with low tumour
stage and longer survival CHOP was correlated with high tumour
stage and shorter survival
BIP and GRP94 Correlated with low grade of differentiation and
high tumour stage
Breast BIP Expressed more in oestrogen receptor-negative tumours
than in oestrogen receptor-positive tumours
BIP and XBP1 Correlated with oestrogen receptor expression
BIP, GRP94, GRP75, HSP60 and calreticulin
NA
Colon BIP Correlated with high cell malignancy
CHOP Correlated with high tumour stage
Gastric BIP and GRP94 Correlated with tumour size, depth of
invasion, lymphatic and venous invasion, lymph node metastasis, and
tumour stages, but not independent prognostic factors
BIP and GRP94 BIP was correlated with low tumour stage, high
grade of differentiation and longer survival
GRP94 was correlated with low tumour stage
Pancreas BIP and HSP90 NA
Liver BIP, ATF6 and XBP1 Correlated with high histological
grade
BIP, HSP27 and HSP70 Correlated with high tumour venous
infiltration
BIP Correlated with CD147, which inhibits apoptosis and induces
chemosensitivity
Prostate BIP Correlated with shorter survival
BIP Correlated with castration resistance, greater risk of
recurrence and shorter overall survival
Kidney BIP Correlated with higher tumour grade, advanced tumour
stage, lymphovascular invasion, regional nodal involvement, distant
metastases and shorter survival
BIP Correlated with larger tumour size and higher tumour
stage
Skin BIP Correlated with increased tumour thickness, metastases
and shorter survival in patients with melanoma
Uterus BIP, ATF6 and CHOP NA
Ovary BIP Correlated with higher cell malignancy, but not
associated with survival
Leukaemia Lymphoblast BIP, XBP1s and calreticulin
NA
BIP, XBP1s, CHOP and calreticulin
Correlated with lower relapse rate and longer overall and
disease-free survival
Lymphoma Bcells XBP1s Correlated with higher tumour grade,
therapy resistance and shorter survival
XBP1s Correlated with advanced plasma differentiation
GRP94, IRE1 and GADD34
GRP94 was correlated with therapy resistance and shorter
survival
BIP Correlated with shorter overall survival and lower
sensitivity to therapy
Glioma Brain BIP Correlated with shorter survival
ATF6, activating transcription factor 6; BIP, immunoglobulin
heavy-chain binding protein; CHOP, C/EBP homologous protein; ER,
endoplasmic reticulum; GADD34, growth arrest and DNA
damage-inducible protein 34; GRP94, 94 kDa glucose-regulated
protein; HSP, heat shock protein; IRE1, inositol-requiring protein
1; NA, not applicable; UPR, unfolded protein response; XBP1, X-box
binding protein 1; XBP1s, transcriptionally active XBP1. *Each row
of the table represents one study; see Supplmentary information S1
(table) for a version of this table with references.
R E V I E W S
582 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
http://www.nature.com/nrc/journal/v14/n9/full/nrc3800.html#supplementary-information
-
Mitochondria-associated ER membranes(MAMs). A specialized
endoplasmic reticulum (ER) membrane is directly juxtaposed to the
mitochondrion to coordinate efficient communication between these
two organelles.
PolysomeA cluster of ribosomes translating a single mRNA
molecule.
induction of oncogenic HRAS, MYC or the oncogenic latent
membrane protein 1 of EpsteinBarr virus47, or loss of tumour
suppressors tuberous sclerosis com-plex 1 (TSC1; also known as
hamartin), TSC2 (also known as tuberin), BRCA1 or PTEN811, increase
pro-tein synthesis and translocation into the ER owing to high
metabolic demand during oncogenic transforma-tion. Consequently,
the UPR is activated to increase the protein folding capacity. In
addition, some gene mutations, such as smoothened (SMO) mutants,
cause UPR activation owing to their intrinsic misfolding12.
Furthermore, UPR activation is required to pro-mote ER expansion
for division and transmission to daughter cells during mitosis13.
Certain types of can-cer cells are highly secretory and therefore
prone to constitutive UPR activation. For example, haemato-logical
malignancies such as multiple myeloma and other plasma cell
malignancies express high levels of immunoglobulins. Increased
mucin production is also documented in many solid cancers,
including pancre-atic, lung, breast, ovarian and colon cancers14.
During malignant progression, cancer cells activate pathways that
co-opt cells in the tumour microenvironment, such as immune cells
and endothelial cells, to support tumour growth15,16, which may
require UPR signal-ling to increase folding and the secretion of
cytokines, metallo proteinases, angiogenesis factors and
extra-cellular matrix components. Besides the intrinsic factors,
rapidly proliferating cancer cells frequently encounter a hostile
environment, which disrupts ER protein folding to activate the UPR
(BOX2).
The role of the UPR in tumorigenesisThe UPR comprises three
parallel signalling branches: PRKR-like ER kinase (PERK; also known
as eIF2AK3)eukaryotic translation initiation factor 2 (eIF2),
inositol-requiring protein 1 (IRE1; also known as ERN1)X-box
binding protein 1 (XBP1) and activating transcription factor 6
(ATF6) (FIG.1). Emerging evi-dence suggests that UPR activation is
required for onco-genic transformation. As the UPR exerts both
protective and deleterious effects on cell survival upon ER stress,
UPR activation may facilitate, as well as suppress, malig-nant
transformation (FIG.2). Therefore, there would be a selective
advantage for premalignant cells harbouring gene mutations that
suppress UPR-induced apoptosis or senescence.
The PERKeIF2 pathway in ER stress. PERK is a typeI transmembrane
protein enriched at mitochondria associated ER membranes (MAMs)17
with a cytosolic serine/threonine kinase domain. Under non-stress
con-ditions, heat shock protein 90 (HSP90) and BIP bind to the
cytoplasmic and ER luminal domains of PERK, respectively, to
stabilize and prevent activation18. Under conditions of ER stress,
BIP binds to unfolded proteins and misfolded proteins, permitting
the release of PERK for homodimerization and autophosphorylation,
leading to its activation19,20. Activated PERK then phosphorylates
eIF2 (a subunit of the heterotrimeric eIF2 complex) at S51 (REF.21)
to attenuate translation initiation due to limiting amounts of the
eIF2GTPtRNAmet ternary complex. The ternary complex binds the 40S
ribosome to generate a 43S species that binds the 5 end of the mRNA
to initiate scanning downstream. When the 43S species encounters an
AUG codon in an optimal con-text for initiation, eIF5 activates
eIF2-mediated GTP hydrolysis22. To perform another round of
initiation, eIF2B is required to promote GTP exchange for GDP on
eIF2 a reaction that is inhibited by eIF5 (REF.23). Phosphorylation
of eIF2 greatly increases the affinity of eIF2 for GDP, thereby
preventing the eIF2B-catalyzed exchange reaction and sequestering
eIF2B with eIF2 in an inactive complex24, as well as inhibiting the
anti-exchange activity of eIF5 (REF.23). As the eIF2B/eIF2 ratio is
generally less than 1 (approximately 1/7 in rab-bit reticulocyte
lysate and 1/2 in Ehrlich ascites cells25), it was proposed that
small increases (as little as ~20%) in the amount of eIF2
phosphorylation could shut down general protein synthesis26. The
transient inhibition of protein synthesis probably promotes
polysome disassem-bly to increase the number of ribosomes available
to bind newly transcribed mRNAs, which encode UPR adaptive
functions. Besides eIF2, other PERK substrates have been suggested
to affect cell survival, function and dif-ferentiation, including
nuclear factor erythroid 2-related factor 2 (NRF2; also known as
NFE2L2)27, forkhead box O (FOXO)28 and diacylglycerol29. However,
as the PERK-dependent changes in gene expression in mouse embryonic
fibroblasts (MEFs) can be prevented by eIF2 mutation at the PERK
phosphorylation site30, the physiological importance of these
alternative substrates remains in question.
Box 1 | Considerations in studying the UPR in cancer
The consequence of activation of the unfolded protein response
(UPR) is closely associated with different types, severity and
duration of endoplasmic reticulum (ER) stress, which needs
consideration for data interpretation when characterizing UPR
activation in cancer. Many studies are performed with cells in
which genes have been deleted or knocked
down. There is a considerable difference between gene deletion
versus knockdown. Both approaches require adequate controls:
multiple gene-deleted lines or gene rescue experiments; for
knockdowns or clustered regularly interspaced short palindromic
repeat (CRISPR)-mediated deletions, elimination of off-target
effects is essential.
A reduction in a protein may have a completely different effect
than complete removal of a protein. It is now evident that many
requirements for a protein or protein modification exhibit a
bell-shaped distribution in which either higher or lower levels of
protein or protein modification show similar deficiencies.
It is also necessary to carry out detailed kinetic studies
because deletion of a gene may be advantageous after 24hours but
extremely detrimental for long-term survival. Most studies analyse
a single time point after gene knockdown. This is especially
important in the analysis of growth and metastasis in
xenotransplant experiments.
Most convincing experiments require genetic correction of a
phenotype by altering a downstream component to demonstrate the
correction is dependent on the gene that is deleted.
Most studies of ER stress use pharmacological induction of
protein misfolding that is, tunicamycin, thapsigargin,
dithiothreitol, and so on. As these agents have many additional
effects on the cell, it is important to study how the synthesis of
a misfolded protein, an increase in general protein secretion in
cancer cells or other physiological challenges (hypoxia, nutrient
deprivation, redox changes, reactive oxygen species (ROS), and so
on) affect tumour cell survival to obtain more physiologically
meaningful results.
R E V I E W S
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 583
2014 Macmillan Publishers Limited. All rights reserved
-
While attenuating global mRNA translation, PERKeIF2 activation
paradoxically increases the translation of a growing number of
mRNAs, including those encoding ATF4 (also known as CREB2) and ATF5
(REF.26), as well as amino acid transporters31. ATF4 enters the
nucleus to activate ER stress response genes that are responsible
for the antioxidant response and amino acid biosynthesis and
transport to promote cell survival26. ATF4 activates tran-scription
of the growth arrest and DNA damage-inducible protein 34 (GADD34;
also known as PPP1R15A) to direct eIF2 dephosphorylation and
restore global mRNA translation. ATF4 also activates transcription
of C/EBP homologous protein (CHOP; also known as DDIT3 and
GADD153)26,32, which is required for ER-stress-mediated apoptosis
both invitro and invivo33,34. At early times after ER stress, PERK
activation induces miR-211 expression, which represses CHOP
transcription through histone methylation35. CHOP expression can
also be suppressed by Toll-like receptor (TLR)TIR-domain-containing
adapter-inducing interferon- (TRIF; also known as TICAM1)
signalling through protein phosphatase 2A (PP2A)-mediated serine
dephosphorylation of the eIF2B -subunit36. Under conditions of
chronic stress, consti-tutive PERK-mediated phosphorylation of eIF2
leads to apoptosis, as the IRE1XBP1 and ATF6 pathways are
attenuated3739. Therefore, PERK activation promotes both adaptive,
as well as apoptotic, responses depending on the severity of the
stress. It is likely that the particular response differs between
cell types and environments on the basis of different thresholds
for ER stress tolerance of thecell.
Although CHOP accumulation in the cell correlates with cell
death, both ATF4 and CHOP mRNAs and proteins have short half-lives;
therefore, a strong and chronic activation of PERK is necessary to
increase steady state levels of CHOP to promote cell death37.
Previous reports suggest that CHOP represses BCL-2
expression40, upregulates BCL-2-interacting mediator of cell
death (BIM; also known as BCL2L11) transcrip-tion41 and promotes
the translocation of BAX to mito-chondria42. CHOP was also shown to
directly bind and induce the promoters of p53 upregulated modulator
of apoptosis (PUMA; also known as BBC3)43, lipocalin 2 (LCN2)44,
tribbles homologue 3 (TRIB3)45 and death receptor 5 (DR5; also
known as TNFRSF10B)46,47. It was recently confirmed that
CHOP-mediated DR5 induction is responsible for ER stress-induced
apoptosis via cas-pase 8 in cancer cells48. However, chromatin
immuno-precipitation followed by sequencing (ChIPseq) studies did
not detect either ATF4 or CHOP occupying genes of the pro-apoptotic
family upon induction of ER stress in MEFs. Instead, ATF4 and CHOP
formed heterodimers that upregulated genes encoding functions in
the UPR, autophagy and, surprisingly, mRNA translation, leading to
increased protein synthesis, ATP depletion, oxidative stress and
cell death49. Indeed, compared with wild-type cells, ER stress
caused less ER protein aggregation and apoptosis in Chop-null
cells, which is consistent with the idea that CHOP increases
protein synthesis to cause protein misfolding and oxidative
stress33,34,50,51. Therefore, although ER stress-induced apoptosis
is indirectly medi-ated by CHOP, it is possible that, in tumour
cells, dif-ferent pathways are activated downstream of CHOP to
regulate survival.
The PERKeIF2 pathway in tumorigenesis. As the PERKeIF2 pathway
induces either survival or apop-tosis upon ER stress, it may
facilitate, as well as suppress, malignant transformation depending
on the context. Indeed, Perk deletion delays Neu-dependent mammary
tumour development and reduces lung metastases, whereas long-term
PERK inactivation increases suscep-tibility to spontaneous mammary
tumorigenesis owing to increased genomic instability52. PERK
activation also
Box 2 | The tumour microenvironment and the induction of ER
stress
Tumours (especially solid tumours) are often challenged by
hypoxia and a lack of glucose, as well as other nutrients, owing to
poor vascularization upon quick expansion of the tumour mass, which
results in severe endoplasmic reticulum (ER) stress in cancer
cells56,172174. To survive the hostile environment, the unfolded
protein response (UPR) is activated in cancer cells, which
ameliorates ER stress to promote cell survival and
growth103105.
How do environmental factors induce ER stress in cancer cells?
Some environmental factors, such as hypoxia, directly impact on
protein modification in the ER, leading to accumulation of
misfolded or unfolded protein. A major post-translational
modification of proteins synthesized in the ER is disulphide bond
formation, which is catalysed by the family of disulphide
isomerases. A recent finding showed that disulphide bonds that are
formed during protein synthesis are oxygen-independent, but those
formed during post-translational folding or isomerization in the ER
are oxygen-dependent175. This provides insight into how hypoxia
causes ER stress and UPR activation. Hypoxia also increases the
stability of some UPR components, such as activating transcription
factor 4 (ATF4), possibly because ATF4 is degraded by proline
hydroxylation, similar to hypoxia-inducible factors (HIFs)176,177.
Both ATF4 and X-box binding protein 1 (XBP1) further augment
HIF1-mediated upregulation of its downstream targets to promote
cell survival93,178. However, eukaryotic translation initiation
factor 2 (eIF2) phosphorylation is more important than HIF
signalling in promoting survival of therapy-resistant cancer
cells104. Blocking UPR activation significantly increases cancer
cell death under hypoxic conditions179. Glucose deprivation, often
coinciding with hypoxia, also disrupts protein folding in the ER.
Glucose metabolism supplies tumour cells with energy in the form of
ATP, building blocks for biosynthesis, and functions as a donor for
asparagine-linked glycosylation. Glucose shortage leads to
disturbed ERCa2+ homeostasis that is mediated by deficient
sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA) activity
owing to a reduced energy supply180 and protein misfolding caused
by improper protein glycosylation. Some other environmental factors
indirectly induce ER stress and UPR activation. Amino acid
deprivation activates general control nonderepressible 2 (GCN2;
also known as eIF2K4) to phosphorylate eIF2 (discussed in BOX3).
Growth factors that are present in the tumour microenvironment can
also contribute to UPR activation in cancer, independent of ER
stress (discussed in BOX3).
R E V I E W S
584 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
-
promotes MYC-induced cell transformation through autophagy7,53.
This may be related to PERK-mediated eIF2 phosphorylation and the
resulting increase in ATF4, CHOP and factors that activate
transcription of many autophagy genes54,55. However, it was
reported that CHOP induction in response to prolonged ER stress
causes death of premalignant cells to prevent neoplastic
progression. Deletion of Chop increased tumour inci-dence in a
KrasG12V-induced mouse model of lung can-cer, suggesting a
tumour-suppressive role of CHOP56. Additionally,
hepatocyte-specific Chop deletion increased tumorigenesis in a
mouse model of hepatocellular car-cinoma57. CHOP mutations were
reported in human tumours58, although it is unknown whether these
muta-tions alter protein expression or function and whether they
contribute to tumorigenesis.
The IRE1XBP1 pathway in ER stress. Mammals have two IRE1 genes,
IRE1A (also known as ERN1; which encodes IRE1) and IRE1B (also
known as ERN2; which encodes IRE1). I RE1A is ubiquitously and
constitutively expressed, whereas IRE1B expres-sion is restricted
to intestinal and lung epithelial cells.
Like PERK, IRE1 is a typeI transmembrane protein with a
cytosolic serine/threonine kinase domain. Under non-stress
conditions, both HSP90 and HSP72 bind the IRE1 cytosolic domain to
maintain its stabil-ity18,59, while BIP binds the luminal domain of
IRE1 to prevent dimerization. Upon ER stress, unfolded and
misfolded proteins bind and sequester BIP, thereby releasing IRE1
for oligomerization, autophosphoryla-tion and activation of its
kinase and endoribonuclease activities19,20. Membrane fluidity also
influences PERK and IRE1 oligomerization and activation60.
Structural studies suggest that short peptides could interact with
a major histocompatability complex classI (MHC classI)-like groove
to promote dimerization in yeast IRE1 (REF.61). Although the MHC
classI-type groove in the human IRE1 homodimer was not
solvent-exposed62, recent studies show that it can bind some
peptides63. Activated IRE1 cleaves Xbp1 mRNA to initiate removal of
a 26-base intron in the cyto-plasm to produce a translational
frame-shift creating a transcriptionally active form (Xbp1s) that
enters the nucleus to regulate target genes64,65. To expedite the
response, Xbp1u (unspliced) mRNA localizes to the ER
Box 3 | UPR activation independent of ER stress
A recent study revealed that vascular endothelial growth factor
A (VEGFA) activates the unfolded protein response (UPR) via
phospholipase C (PLC)-mediated crosstalk with mTOR complex 1
(mTORC1) in endothelial cells, in the absence of endoplasmic
reticulum (ER) stress158. Thus, it is possible that growth factors
in the tumour microenvironment activate the UPR in tumour cells.
Although this hypothesis needs further testing with cancer cells
and other growth factors, it is possible that UPR activation in
cancer cells and cells in the tumour microenvironment, such as
endothelial cells and macrophages, is ER stress-independent under
certain conditions.
Greater evidence suggests that UPR signalling can be activated
in the absence of ER stress. This is particularly evident in the
regulation of protein synthesis through phosphorylation of
eukaryotic translation initiation factor 2 (eIF2). In mammals,
there are three additional kinases that also phosphorylate eIF2
S51: general control nonderepressible 2 (GCN2; also known as
eIF2K4) induced by amino acid deprivation, the heme-regulated
inhibitor kinase (HRI; also known as eIF2K1) induced by oxidative
stress or heme deprivation and the double-stranded RNA
(dsRNA)-activated protein kinase (PKR; also known as eIF2AK2)
activated by dsRNA as part of the interferon antiviral response.
Importantly, the stress conditions that activate any single eIF2
kinase may have secondary effects on the cell that cause activation
of other eIF2 kinases. For example, ER stress activates PKR, as
well as PRKR-like ER kinase (PERK), and ultraviolet (UV) light
activates both GCN2 and PERK. Although activation of any eIF2
kinase causes translation attenuation, the cellular response varies
tremendously depending on which kinase phosphorylates eIF2, the
cell type and the environment. Generally, eIF2 phosphorylation that
is mediated by GCN2, HRI and PKR leads to apoptosis39, whereas
transient eIF2 phosphorylation by PERK promotes cell survival, and
chronic eIF2 phosphorylation by PERK promotes apoptosis. The
mechanism for the different outcomes remains unknown but may
involve decreased synthesis of inhibitors of apoptosis that have
short half-lives or different levels of growth arrest and DNA
damage-inducible protein 34 (GADD34) to direct eIF2
dephosphorylation. In the stressful tumour microenvironment, the
level of eIF2 phosphorylation might be high, and so other factors
need to be considered that impact the effect of eIF2
phosphorylation rate on protein synthesis. For example, GADD34 is a
regulatory subunit of protein phosphatase 1 (PP1) that promotes
eIF2 dephosphorylation181, and protein phosphatase 2A
(PP2A)-mediated dephosphorylation of the -subunit of eIF2B
activates exchange activity of eIF2B to bypass eIF2
phosphorylation36. As eIF2 phosphorylation attenuates mRNA
translation, less misfolded proteins accumulate in the ER21,26,
which promotes cell survival.
Similar to PERK, the activation of GCN2, HRI and PKR are
associated with cancer. PKR has both tumour-supportive and
tumour-suppressive roles in cancer development182. GCN2 is
upregulated in cancers and its inhibition delays tumour growth in
xenograft models183. GCN2 upregulation is also required for
angiogenesis by augmenting amino acid deprivation-induced
expression of VEGFA. More importantly, this effect does not depend
on PERK activation184. It was also reported that HRI expression is
reduced in ovarian epithelial cancer185, although its impact in
cancer remains unclear.
Inositol-requiring protein 1 (IRE1) can also be activated by
Toll-like receptor (TLR) signalling, independent of ER stress186,
through recruitment of the E3 ubiquitin ligase TNF
receptor-associated factor 6 (TRAF6). Interaction of TRAF6 with
IRE1 prevents interaction with PP2A via its adaptor receptor for
activated C kinase 1 (RACK1) to prevent IRE1 dephosphorylation and
inactivation73. TRAF6 and PP2A compete for binding to the same site
on IRE1. TRAF6 binding reduces PP2A-mediated dephosphorylation,
promoting IRE1 activation74. IRE1 signalling is attenuated through
proteasomal degradation that is mediated by TRAF6 ubiquitylation of
IRE1 via a K48 linkage74. As TLR signalling is important in cancer
development and drug resistance187, IRE1 could be one of the
mediators in this process.
R E V I E W S
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 585
2014 Macmillan Publishers Limited. All rights reserved
-
Nature Reviews | Cancer
Misfoldedprotein
ER
BIP
PERK
IRE1
ATF6
ATF6
P
BIP
BIP BIP
BIP
BIP BIP
BIP
eIF2
eIF2
GADD34
ATF4
ATF4
Translation
Translation
Autophagy
ROS
Apoptosis
CHOP
JNK
RIDD RNA decay
XBP1u XBP1s XBP1
S1PS2P
Increased ER proteinfolding capacity and ERAD
Survival
Golgi Nucleus
P
P
BIP
BIP P
P
Regulated IRE1-dependent decay(RIDD). A process in which
activated inositolrequiring protein 1 (IRE1) induces cleavage and
degradation of microRNAs and of mRNAs encoding membrane and
secreted proteins.
membrane to facilitate IRE1 interaction66. Genes that are
regulated by IRE1XBP1 enhance protein fold-ing, trafficking and
ERAD, thereby resolving protein misfolding38,67, and forced XBP1s
expression inhibits CHOP expression, thereby promoting cell
survival68. In addition, overexpression of XBP1 induces many genes
involved in secretory pathways and physically expands the ER, which
results in the characteristic phenotype of professional secretory
cells67. However, the idea that IRE1XBP1 promotes cell survival is
challenged by recent findings that small molecule inhibitors of
IRE1 did not sensitize cells to ER stress-induced apoptosis, but
rather prevented expansion of secretory capacity69.
ER stress immediately activates IRE170,71, whereas IRE1
activation is mostly attenuated upon chronic ER stress71,72. It is
unclear how IRE1XBP1 signalling is attenuated under conditions of
sustained ER stress, although dephosphorylation, ubiquitylation and
deg-radation are probably involved73,74. Although protein
disulphide isomerase family A member 6 (PDIA6), a resident ER
protein, forms a disulphide bond with C148 in the IRE1 luminal
domain to attenuate sig-nalling75, other results show that PDIA6 is
required for IRE1 activation76. Furthermore, Xbp1u can function as
a negative regulator of both the XBP1s and ATF6 pathways by direct
interaction to promote their degra-dation77, which possibly blocks
survival signals during chronic ERstress.
If IRE1 signalling is not attenuated, chronic IRE1 activation
signals apoptosis. Hyperactivated IRE1 cleaves many mRNAs, in
addition to Xbp1 (REFS70,78) and its own mRNA79, a process called
regulated IRE1dependent decay (RIDD)80. A recent study suggests
that RIDD is dependent on the oli-gomeric state of IRE181. RIDD
also reduces the expression of some microRNAs (mi RNAs), includ-ing
miR-17, miR-34a, miR-96 and miR-125b, which repress caspase 2
expression82. However, the impor-tance of caspase 2 activation in
ER stress-induced apoptosis remains in question83. Activated IRE1
kinase also binds TNF receptor-associated factor 2 (TRAF2), which
recruits apoptosis signal-regulating kinase 1 (ASK1; also known as
MAP3K5) and JUN N-terminal kinase (JNK)84, leading to activation of
BIM and inactivation of BCL-2. However, the importance of JNK
activation in ER stress-induced apoptosis has not been
demonstrated.
The IRE1XBP1 pathway in tumorigenesis. The role of IRE1XBP1 in
multiple myeloma has been intensively studied because mature Bcell
differen-tiation into plasma cells and Bcell-mediated defence
against infection require XBP1s8587. Increased levels of XBP1s are
frequently associated with human mul-tiple myeloma, and genetically
engineered mice that over express XBP1s under the control of
immuno-globulin VH promoter and E enhancer elements
Figure 1 | The unfolded protein response (UPR) signalling
pathways. Upon endoplasmic reticulum (ER) stress, unfolded and
misfolded proteins bind and sequester immunoglobulin heavy-chain
binding protein (BIP), thereby activating the UPR. The UPR
comprises three parallel signalling branches: PRKR-like ER kinase
(PERK)eukaryotic translation initiation factor 2 (eIF2),
inositol-requiring protein 1 (IRE1)X-box binding protein 1 (XBP1)
and activating transcription factor 6 (ATF6). The outcome of UPR
activation increases protein folding, transport and ER-associated
protein degradation (ERAD), while attenuating protein synthesis. If
protein misfolding is not resolved, cells enter apoptosis. CHOP,
C/EBP homologous protein; GADD34, growth arrest and DNA
damage-inducible protein 34; JNK, JUN N-terminal kinase; P,
phosphorylation; RIDD, regulated IRE1-dependent decay; ROS,
reactive oxygen species; XBP1s, transcriptionally active XBP1;
XBP1u, unspliced XBP1.
R E V I E W S
586 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
-
Nature Reviews | Cancer
Protein misfolding in the ER
Hostile environment Cancer therapy
Loss ofBRCA1
Lossof TSC MYC
Loss ofPTEN RAS
p53 and TSCdownregulated
BIPinduced
DecreasedATF4 and CHOPexpression
Inducedautophagy
Enhancedglycolysis
CHOPinduction
Survival OncogenictransformationSenescenceor apoptosis
Tumorigenesis
Regulated intramembrane proteolysisA process in which
endoplasmic reticulum (ER) transmembrane transcription factors are
cleaved within the plane of the membrane to release cytosolic
fragments that enter the nucleus to regulate gene
transcription.
exhibit features reminiscent of multiple myeloma
transformation88. IRE1A and XBP1 mutations have been identified in
tumour cells from patients with multiple myeloma89,90. More
importantly, analy-sis of human multiple myeloma tumour lines that
are resistant to proteasome inhibition identified loss-of-function
mutations in either IRE1A or XBP190. Apparently, proteasome
inhibitors select for cells that do not require ERAD: that is,
multiple myeloma cells that lose immunoglobulin expression and
display pre-plasmablast characteristics90. Mutations in IRE1A were
also reported in other human cancers58,91, some of which lose or
reduce kinase and/or endoribonu-clease function81. In addition,
loss of XBP1 function promotes tumorigenesis in mouse models of
intes-tinal cancer 92. Although these findings suggest a
tumour-suppressive role for IRE1XBP1, increased XBP1 mRNA
splicing was observed in human triple- negative breast cancers,
possibly indicating a requirement for XBP1 in cancer stem-like
cells93.
The ATF6 pathway in ER stress. ATF6 is a typeII trans-membrane
protein that contains a cytosolic cAMP-responsive element-binding
protein (CREB)/ATF basic leucine zipper (bZIP) domain. Under
non-stressed conditions, ATF6 is retained in the ER through
interaction with BIP. Upon accumulation of misfolded protein, ATF6
is released from BIP and traffics to the Golgi apparatus for
processing by the proteases S1P (also known as MBTPS1) and S2P
(also known as MBTPS2)94 to generate an active transcription
factor, in a process termed regulated intramembrane proteolysis.
There are two homologues of ATF6 in the mamma-lian genome. Cleaved
ATF6 mediates the adaptive response to ER protein misfolding by
increasing the transcription of genes that increase ER capacity and
the expression of Xbp1 (REFS95,96), whereas ATF6 may function as a
repressor of ATF6-mediated transcrip-tion and function97.
Presently, no genes have been iden-tified that require ATF6 for
expression. PERKeIF2 signalling facilitates ATF6 synthesis and
trafficking to accentuate ATF6 signalling98. To date, no
substantial evidence supports a role of ATF6 in ER stress-induced
apoptosis.
The ATF6 pathway in tumorigenesis. Some studies suggest that
ATF6 promotes hepatocarcinogenesis by regulation of target genes99.
A missense polymorphism in ATF6 that increases mRNA expression of
ATF6 and its downstream genes was associated with susceptibil-ity
to hepatocellular carcinoma100. More importantly, BIP, a downstream
transcriptional target of ATF6, was reported to serve as a
malignancy marker for cells. Upon induction of ER stress, ATF6
quickly induces BIP expression, which binds to unfolded protein and
misfolded protein to ameliorate ER stress. Under nor-mal
conditions, BIP is localized to the ER lumen, but upon
overexpression in many human cancers (TABLE1; see Supplementary
information S1 (table)), it becomes detectable on the cell
surface101. BIP expression not only correlates with cancer cell
proliferation and his-tological grade but also correlates with
response to therapies and prognosis102 (TABLE1; see Supplementary
information S1 (table)).
The UPR in tumour cellsUPR in autonomous cancer cell survival.
UPR acti-vation protects cancer cells from stress-induced cell
death103105. Acute UPR activation enhances the pro-tein folding
capacity to meet the need for increased protein synthesis, which
benefits cancer cell survival. Where chronic ER stress kills normal
cells, tumour cells use strategies that neutralize apoptosis when
challenged with ER stress. In response to chronic stress, normal
cells usually attenuate the IRE1XBP1 and ATF6 pathways, so that the
apoptotic signals dominate38. Some cancer cells, however, exhibit
constitutive activation of
Figure 2 | Crosstalk between the unfolded protein response (UPR)
components and oncogene or tumour suppressor gene networks in
cancer cells. Either hyperactivation of oncogenes or loss of tumour
suppressor genes can activate the UPR, promoting cell survival,
oncogenic transformation or cell senescence or apoptosis, depending
on gene mutations and the cellular context. The loss of BRCA1
function upregulates immunoglobulin heavy-chain binding protein
(BIP) expression to survive chronic endoplasmic reticulum (ER)
stress10, although the mechanism is unclear. The loss of tuberous
sclerosis complex (TSC) function increases protein synthesis and
the requirement for ER protein folding, thereby causing ER stress
and UPR activation, but expression of activating transcription
factor 4 (ATF4) and C/EBP homologous protein (CHOP) are mostly
compromised9. Phosphorylation of eukaryotic translation initiation
factor 2 (eIF2) is required for the anti-proliferative and
pro-apoptotic effects of PTEN190. Loss of PTEN induces UPR
activation and increases aerobic glycolysis (known as the Warburg
effect), which is associated with transformation11,191.
Downregulation of tumour suppressor candidate 3 (TUSC3), which
affects N-linked glycosylation, also causes ER stress to activate
the UPR and increase the malignancy of prostate cancer cells192.
Furthermore, UPR activation, on the one hand, increases expression
of genes that are involved in tumour initiation and progression,
such as the a disintegrin and metalloproteinase (ADAM) family, to
facilitate tumorigenesis193. On the other hand, the UPR decreases
the expression of some tumour suppressors, including p53
(REFS194,195), TSC1 and TSC2 (REF.126), to promote cell survival
and oncogenic transformation. In other cases, sustained UPR
activation in response to prolonged ER stress causes death of
premalignant cells to prevent neoplastic progression. For example,
HRAS induces UPR-mediated cell senescence in premalignant cells4.
Red boxes indicate oncogenes and green boxes indicate tumour
suppressors.
R E V I E W S
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 587
2014 Macmillan Publishers Limited. All rights reserved
-
the IRE1XBP1 pathway106,107 or overexpression of BIP10,108,
which are anti-apoptotic. Furthermore, CHOP, induced by chronic ER
stress, activates transcription of the AKT inhibitor TRIB3, which
blocks mTOR pathways45,109,110 to inhibit proliferation and
activate autophagy. UPR activation also represses cyclin D1
translation due to global transient translation inhibition induced
by eIF2 phosphorylation, leading to subse-quent cell cycle arrest
in the G1 phase111. This increases dormancy of the cancer cells,
permitting survival in the stressed environment until more
favourable conditions are encountered. In addition, oncogene and
tumour suppressor gene mutations inhibit ER stress-induced
apoptosis machinery (FIG.2). Mutations in UPR path-ways may also
directly contribute to enhanced cancer cell survival upon stress.
Some IRE1 mutants iden-tified in human cancers showed reduced
endoribo-nuclease function. Although still able to splice XBP1
mRNA, they cannot induce RIDD, thereby promot-ing cell survival81.
Hence, cancer cells escape from ER stress-induced apoptosis.
By contrast, persistent ER stress or UPR activation
(particularly by pharmacological intervention) induces cancer cell
death through similar apoptosis pathways that are used in normal
cells (FIG.1). Therefore, chronic ER stress or UPR
activation-induced cell death pathways are intact in at least some
tumour cells. It will be of great interest to determine whether
persistent ER stress or UPR activation can induce tumour cell death
through other mechanisms. It will also be important to predict
whether particular tumour types are dependent on UPR signalling for
survival.
UPR, autophagy and cell metabolism in cancer develop-ment and
progression. Early studies showed that protein folding and
processing in the ER and trafficking to other organelles and the
cell surface require a series of com-plex energy-requiring
reactions112,113. As a result, protein folding in the ER is
susceptible to energy fluctuations in the cell and protein
misfolding may serve as a sig-nal for nutrient, energy or oxygen
deprivation. On the one hand, conditions of low nutrient supply
(for exam-ple, glucose deprivation or hypoxia) induce ER stress and
UPR activation, to improve protein folding and transport, restore
energy homeostasis and render cells resistant to cell death114. On
the other hand, excess nutri-ents (fatty acids, cholesterol and
glucose) also induce ER stress and UPR activation115,116. The
integration of the UPR with cell metabolism is of special
importance to the cancer biology field, as tumour cells display ER
stress, UPR activation and nutrient shortage, which are probably
due to poor vascular supply and rapid cell proliferation (BOX2),
and tumours arise at higher rates and are more malignant in a
nutrient-rich environment compared with a normal
environment117,118.
Faced with a lack of nutrients and an inadequate ER protein
folding environment, cells activate autophagy a stress-adaptive
self-eating process in which cell-ular components are encapsulated
within autophago-somes and degraded by lysosomal hydrolases to
remove misfolded proteins, restore ER homeostasis
and supply cells with essential nutrients. Similar to the UPR,
autophagy can lead to both cell death and survival119. The
mechanisms by which the UPR acti-vates autophagy are only partly
understood. Early studies showed a requirement for eIF2
phosphory-lation for autophagy induction54,120. For example,
activation of the PERKeIF2 pathway, in response to the expression
of polyglutamine 72 (polyQ72) aggregates, induced ER stress, LC3
(also known as MAP1LC3A) conversion, autophagosome forma-tion and
survival120. Whether eIF2 phosphorylation is required to attenuate
protein synthesis to initiate autophagy or whether events
downstream of eIF2 phosphorylation are required for autophagy
remains a major question. Recent studies showed that ATF4 and CHOP
function independently, as well as together, to induce a large set
of autophagy genes55. One study suggested the IRE1JNK pathway is
also required for autophagy and cell survival upon ER stress121.
Analysis of XBP1-deficient mice suggested that cells compen-sate
for decreased ERAD by activation of autophagy122. In addition, ATF6
is also reported to be required for interferon- (IFN)-induced
autophagy123. ER stress was also associated with activation of a
novel protein kinase C (PKC) family member, PKC124, and activa-tion
of Ca2+/calmodulin-dependent kinase kinase- (CaMKK)125. CaMKK
activates AMP-activated pro-tein kinase (AMPK) while attenuating
AKTmTOR signalling to enhance autophagy 126. However, it remains
poorly understood whether the crosstalk between the UPR and
autophagy contributes to cancer development and metabolism. A pilot
study recently demonstrated that oncogenic ER stress induces
acti-vation of the PERKeIF2ATF4 signalling pathway, increasing cell
survival via induction of cytoprotec-tive autophagy and enhanced
MYC-driven tumour transformation and growth7. Another study
indicated that PERK, ATF4 and CHOP protect human tumour cells
during hypoxia through autophagy 127. More studies are needed to
elucidate the complex crosstalk between these processes and reveal
their requirement in all stages of cancer development and
progression.
Excess nutrients also induce ER stress and acti-vate the UPR.
Exposure of cells to increased levels of free fatty acids (for
example, palmitate and stearate) induces ER stress115, probably
through aberrant protein palmitoylation128, increased accumulation
of reactive oxygen species (ROS) due to elevated fatty acid
oxida-tion and/or an increased protein folding load resulting from
hyperactivation of the mTOR anabolic signalling pathway.
Additionally, activation of AMPK or inhibi-tion of JNK prevented
palmitate-induced ER stress and UPR activation129. In response to
acute or physiological ER stress, the UPR pathways (PERKeIF2 and
IRE1XBP1) activate CCAAT/enhancer-binding proteins (C/EBPs), sterol
regulatory element-binding tran-scription factor 1 (SREBP1)130 and
SREBP2 (REF.131), two transcriptional activators of fatty acid and
cho-lesterol synthesis132, to accommodate the need for ER
expansion. SREBP-mediated lipogenic activity also maintains the
balance between the saturated and
R E V I E W S
588 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
-
Nature Reviews | Cancer
ATF6
ROS
CHOP
S1PS2P
Inflammation
Golgi
P
eIF2
eIF2
IB NF-B
NF-B
JNK AP1
ATF4
Translation
Misfoldedprotein
ER
BIP
PERK
IRE1
BIP
BIP
BIPBIP
BIP
BIP
BIP
BIP
BIP
P
P
P
P
InflammasomeA large intracellular multiprotein oligomeric
complex that is activated by pattern recognition receptors to
initiate an innate immune response by maturation of the
inflammatory cytokines interleukin1 (IL1) and IL18.
monounsaturated fatty acid pools to prevent lipo-toxicity in
tumour cells133. These findings, however, are complicated by the
observation that severe and persis-tent ER stress (for example, as
induced by tunicamycin treatment) may dampen lipogenesis and
increase fatty acid oxidation through mechanisms that are
depend-ent on UPR activation134. Besides fatty acids, cholesterol
loading of macrophages induces ER stress and elicits inflammatory
responses135, which promotes cancer development. Moreover, chronic
exposure to elevated glucose levels triggers ER stress and
glucotoxicity in cultured -cells136. Whether chronic exposure to
cho-lesterol and glucose induce ER stress in other cell types
remains unknown.
Obesity and type2 diabetes are frequently associ-ated with
higher levels of free fatty acids, cholesterol and glucose in the
circulation and overt ER stress in multiple tissues, as well as a
pro-inflammatory state. These parameters are associated with higher
risks of developing cancer, and the tumours generated are generally
more malignant. Although multiple mecha-nisms have been put forth
to explain this connection, it is possible that excess nutrients
associated with
metabolic disorders (for example, obesity and type2 diabetes)
trigger ER stress and UPR activation in pre-malignant and
transformed cells as well as stromal cells in the tumour
microenvironment, which affects cancer development and tumour cell
metabolism. For example, ER stress in cells produces
ROS49,50,137,138, which not only promotes genetic and epigenetic
altera-tions in cells but also induces inflammatory responses
(discussed below). Indeed, recent findings suggest that ER stress
alone is sufficient to generate pre-oncogenic cells, which leads to
hepatocellular carcinoma under conditions of a
high-fat-diet-induced inflammatory environment57.
The UPR in the tumour microenvironmentThe UPR: another
connection between inflammation and cancer. Chronic inflammation is
associated with and can contribute to all stages of cancer
development and progression. The inflammatory milieus of nor-mal
and neoplastic tissues can increase gene mutation rates and overall
genomic instability, promote cell pro-liferation, survival and
invasion, induce angiogenesis, facilitate evasion from immune
surveillance and ren-der tumour cells resistant to anticancer
therapies139. As shown in several pathological conditions, ER
stress and UPR activation are required for the signal transduc-tion
and transcriptional regulation of inflammatory mediators. It is
therefore anticipated that ER stress and UPR activation, aside from
the effects on tumour cell survival and proliferation, promote
cancer develop-ment and progression through activating inflammatory
responses.
ER stress is implicated in various chronic patho-logical
conditions (for example, obesity, diabetes, inflammatory bowel
diseases, atherosclerosis and neurodegenerative diseases) involving
inflamma-tion140. For example, loss of XBP1 function in Paneth
cells caused spontaneous enteritis, as a consequence of IRE1
hyperactivation141. Investigation of the pathogenic mechanisms
revealed a reciprocal regu-lation of ER stress and inflammation; in
which pro- inflammatory stimuli (for example, ROS, TLR ligands and
cytokines) trigger ER stress, which in turn initi-ates or amplifies
inflammatory responses142. Strikingly, all three UPR pathways lead
to activation of nuclear factor-B (NF-B), a master transcriptional
regula-tor of pro-inflammatory pathways (FIG.3). In -cells, ER
stress triggers activation of the NLR family pyrin domain
containing 3 (NLRP3) inflammasome and interleukin-1 (IL-1)
secretion through IRE1 and PERK-mediated induction of
thioredoxin-interacting protein (TXNIP), resulting in -cell
death143,144. By contrast, ER stress seems to activate the NLRP3
inflammasome in macrophages via a UPR-independent mechanism145.
The acute phase response (APR) is an innate sys-temic defence to
infection or injury. Pro-inflammatory cytokines (for example, IL-1,
IL-8 and tumour necrosis factor (TNF)) that are released by local
inflammatory cells travel through the blood and stimulate
hepatocytes to synthesize and secrete APR products. These APR
Figure 3 | The unfolded protein response (UPR) and inflammation.
The three UPR pathways augment the production of reactive oxygen
species (ROS) and activate nuclear factor-B (NF-B) and activator
protein 1 (AP1) pathways, thereby leading to inflammation. NF-B,
which is a master transcriptional regulator of pro-inflammatory
pathways, can be activated through binding to the
inositol-requiring protein 1 (IRE1)TNF receptor-associated factor 2
(TRAF2) complex in response to endoplasmic reticulum (ER) stress,
leading to recruitment of the IB kinase (IKK), IB phosphorylation
(P) and degradation, and nuclear translocation of NF-B196.
Moreover, the IRE1TRAF2 complex can recruit apoptosis
signal-regulating kinase 1 (ASK1) and activate JUN N-terminal
kinase (JNK), increasing the expression of pro-inflammatory genes
through enhanced AP1 activity197. The PRKR-like ER kinase
(PERK)eukaryotic translation initiation factor 2 (eIF2) and
activating transcription factor 6 (ATF6) branches of the UPR
activate NF-B through different mechanisms. Engaging PERKeIF2
signalling halts overall protein synthesis and increases the ratio
of NF-B to IB, owing to the short half-life of IB, thereby freeing
NF-B for nuclear translocation198,199. ATF6 activation following
exposure to the bacterial subtilase cytotoxin that cleaves
immunoglobulin heavy-chain binding protein (BIP) leads to AKT
phosphorylation and consequent NF-B activation109,200.
R E V I E W S
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 589
2014 Macmillan Publishers Limited. All rights reserved
-
Type M2 macrophageA subset of activated macrophages that are
involved in immunosuppression and tissue repair.
MHC classI pathway(Major histocompatability complex classI
pathway). A pathway by which cells present peptides from cytosolic
proteins to Tcells.
products can, in turn, amplify inflammation to eliminate
infection and restore tissue homeostasis. Upon ER stress in liver,
ATF6 and CREBH (also known as CREB3L3) are proteolytically released
from the membrane, traffic into the nucleus and form homodimers or
heterodimers to induce expression of APR genes, including
C-reactive protein (CRP) and serum amyloid P component (SAP; also
known as APCS)142. These findings were important because they were
the first to show a link between ER stress and inflammation. This
is a poorly studied subject that needs further investigation.
Although this scenario that ER stress and UPR activation
promotes cancer development and progres-sion through modulating
inflammatory responses remains mostly unexplored, a few studies
support this idea. It was recently reported that ER stress shortens
the lifespan of myeloid-derived suppressor cells in the periphery
and promotes their expansion in bone mar-row146. ER stress in
prostate cancer cells initiates tran-scription of pro-inflammatory
cytokines147. ER-stressed tumour cells also secrete soluble factors
that initiate ER stress responses and upregulate the expression of
pro-inflammatory cytokines in macrophages15. ER stress in
macrophages promotes the type M2 macrophage phenotype148 that in
turn supports tumour growth. In addition, ER stress-induced
expression of CHOP, in combination with TLR agonists, enhances
dendritic cell expression of IL-23 (REF.149) that favours
development of T helper17 (TH17) cell-mediated inflammation and
tumour growth150.
The UPR in immune defence. Immune effector cells with high
protein secretion capacity, as well as high protein synthesis and
turnover rates owing to rapid cell proliferation, are prone to ER
stress. As a consequence, the UPR is required for immune cell
differentiation and function. ER function is also essential for
anti-gen presentation by innate immune cells for adaptive immunity,
especially through the MHC classI pathway. MHC classI is
synthesized and loaded with peptides inside the ER prior to
trafficking to the plasma mem-brane. Therefore, altered ER
homeostasis disrupts MHC class I antigen presentation151.
Calreticulin (CRT), an ER chaperone that is induced by ER stress,
facilitates antigen processing and peptide loading of MHC classI
molecules152,153. Moreover, misfolded proteins that accumulate
within the ER are translo-cated to the cytosolic proteasomes for
degradation into peptides for antigen presentation. Phosphorylation
of eIF2 upon ER stress reduces synthesis of MHC mol-ecules, as well
as the overall peptide pool for MHC loading, and consequently
impairs antigen presenta-tion. Hence, ER stress can either aid or
impede anti-gen presentation pathways, depending on the cellular
context.
One hallmark of cancer is the evasion of can-cer cells from
immune surveillance154. ER stress-induced inflammation and
perturbed ER homeostasis in immune cells may interfere with the
function of immune cells to combat cancer. Although this hypothesis
requires further testing, recent studies
suggest a novel, tumour-suppressive mechanism of ER stress and
UPR activation in tumour cells. In pre-malignant and neoplastic
cells, ER stress and the UPR can initiate signalling cascades that
function as pro-phagocytosis and immunogenic signals for clearance
of cancer cells by the immune system155. In response to ER stress
caused by physiological conditions or pharmacological intervention,
several ER proteins, including CRT, ERp57 (also known as PDIA3) and
HSPs are translocated to the plasma membrane prior to cancer cell
death. Cell surface exposure of these ER proteins can lead to
activation of antitumour immune responses and the repression of
tumour growth155.
The UPR stimulates tumour angiogenesis. ER stress and UPR
activation in both tumour cells and endothe-lial cells stimulate
tumour angiogenesis (FIG.4). UPR activation not only protects
cancer cells from apop-tosis induced by hypoxia, as well as by lack
of glucose and other nutrients (BOX2), but also tips the balance
from anti-angiogenic factors (for example, throm-bospondin 1
(THBS1), CXC chemokine ligand 14 (CXCL14) and CXCL10) to
pro-angiogenic factors (for example, vascular endothelial growth
factor A (VEGFA), fibroblast growth factor 2 (FGF2), IL-1, IL-6 and
IL-8)16. ATF4 and XBP1s directly bind to the VEGFA promoter to
initiate VEGFA transcription. VEGFA mRNA stability is also
increased in response to UPR activation, via activation of AMPK156.
The endothelial cell compartment in the tumour micro-environment
also experiences ER stress and UPR activation owing to the
accumulation of misfolded proteins157 or the presence of VEGFA158.
A defec-tive UPR is associated with reduced endothelial cell
proliferation, survival and migration159. Knockdown of XBP1 or IRE1
decreases endothelial cell pro-liferation via suppression of AKT
and glycogen synthase kinase 3 (GSK3) phosphorylation, -catenin
nuclear translocation and E2F2 expres-sion160. Moreover,
heterozygous ablation of Bip in the tumour microenvironment
substantially inhibits tumour growth and angiogenesis161;
meanwhile, BIP also confers endothelial cell
chemoresistance162.
The UPR in cancer therapyTherapeutic interference can induce
severe ER stress, leading to cell death (TABLE2; see supplementary
infor-mation S2 (table)). Moreover, ER stress in the tumour
microenvironment modulates the function of cancer-supporting
stromal cells, such as endothelial cells161, and suppresses tumour
growth. However, as UPR acti-vation has both pro-survival and
anti-survival effects on cells, caution is necessary in the design
of therapies that target UPR components and in the interpretation
of the results. It is possible that tumour cells require optimal
UPR signalling for survival and that either increased or decreased
UPR signalling may compro-mise survival of the tumour cell.
Meanwhile, ER stress and UPR activation may alter the cancer cell
response to adjunctive therapies, offering a target for
combi-nation therapy. Specific gene targeting experiments
R E V I E W S
590 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
http://www.nature.com/nrc/journal/v14/n9/full/nrc3800.html#supplementary-informationhttp://www.nature.com/nrc/journal/v14/n9/full/nrc3800.html#supplementary-informationsekhar8421
sekhar8421
sekhar8421
sekhar8421
sekhar8421
sekhar8421
sekhar8421
sekhar8421
-
Hostile environment
Cancertherapy
Nature Reviews | Cancer
Tumour Tumour microenvironment
Cancer cell
UPR activation
Cytokines Growth factors
Increased protein folding capacity Decreased ROS Reduced
proliferation Decreased MHC I expression Induced angiogenic switch
Increased stemness
Survival in hostile environment Compromised immunosurveillance
Therapy resistance Metastasis
Macrophage UPR activation M2 phenotype
Cytokines, IL-6, TNF, etc.
T cell
Suppressed T cell function
DC UPR activation
UPRactivation
TolerogenicDC
Endothelialcell
Proliferation Survival Metastasis Maintains VEGFA levels
Angiogenesis
Tumour cell survival and growth
are required to dissect the requirement for different UPR
transduction pathways in the tumour and the microenvironment.
Targeting the UPR through monotherapy. Owing to pre-existing ER
stress induced by intrinsic and extrin-sic factors in cancer cells
as discussed above, agents that augment ER stress should tip the
balance towards apoptosis. Indeed, bortezomib, the first proteasome
inhibitor for cancer therapy to be approved by the US Food and Drug
Administration (FDA) owing to the success in treating multiple
myeloma and mantle cell lymphoma, functions as an ER stress
inducer. More importantly, the sensitivity to proteasome
inhibi-tors correlates with the amount of immunoglobulin sub units
that are retained within multiple myeloma cells163, and low XBP1
(or XBP1s) or ATF6 levels pre-dict poor response to bortezomib in
patients with mul-tiple myeloma164. It was recently shown that the
loss of IRE1 or XBP1 function causes resistance to protea-some
inhibitors owing to selection for cells that do not synthesize high
levels of immunoglobulin that is, pre-plasmablasts90. This suggests
that UPR activation can also function as a prognostic indicator of
thera-peutic outcomes. These findings indicate that highly
secretory cancer cells, such as multiple myeloma cells, will have a
lower threshold for ER stress-induced cell apoptosis, which
suggests that inducers of ER stress
may provide efficient cancer therapies in these can-cer types
(TABLE2; see Supplementary information S2 (table)).
Some new drugs under study are designed to tar-get specific UPR
pathways to inhibit UPR activation, thereby augmenting ER stress in
cancer cells (TABLE2; see Supplementary information S2 (table)).
For exam-ple, the PERK inhibitor GSK2656157 inhibits growth of
multiple human tumour xenografts in mice owing to its direct impact
not only on tumour cells but also on the tumour environment165.
However, the effects of GSK2656157 are not solely dependent on PERK
and eIF2 phosphorylation166. An IRE1 RNase inhibitor (B-109)
suppresses leukaemic progression in a mouse model167. Importantly,
the UPR signalling pathways have not evolved to be constitutively
activated. They function as an adaptive response to a transient
require-ment to expand ER protein folding capacity, whether in the
context of cell differentiation or as a response to an insult
(pathogen, toxin, inflammation, and so on). Therefore, compounds
that target UPR components will selectively kill cells that
experience ER stress and require a functional UPR for survival. If
this is correct, there may be selective toxicity of cancer cells to
UPR antagonists compared to normalcells.
Another important finding, which shows that tar-geting the UPR
is a promising approach for cancer therapy, is that BIP is
expressed on the surface of
Figure 4 | The cancer-supporting role of the unfolded protein
response (UPR). In most cases, the activation of the UPR supports
tumour survival and growth. On the one hand, UPR activation adapts
cancer cells to the hostile environment and/or to cancer therapies.
On the other hand, UPR activation in cells in the tumour
environment, such as endothelial cells and immune cells, can also
facilitate tumour growth. DC, dendritic cell; IL-6, interleukin-6;
MHC I, major histocompatibility complex classI; ROS, reactive
oxygen species; TNF, tumour necrosis factor; VEGFA, vascular
endothelial growth factor A.
R E V I E W S
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 591
2014 Macmillan Publishers Limited. All rights reserved
-
Table 2 | Strategies to target the UPR components for cancer
treatment*
Strategy Drugs Other involved mechanisms Clinical trials as
cancer therapy
Proteasome inhibitors: target the chymotrypsin-like subunits in
both the constitutive proteasome and the immuno- proteasome,
leading to accumulation of ubiquitylated proteins and ER
stress-mediated apoptosis
Bortezomib Inhibits IRE1XBP1 pathway Suppresses activation of
NF-B pathway Induces NOXA expression Triggers immunogenic cell
death
FDA-approved for multiple myeloma and mantle cell lymphoma;
Phase1/2 in solid tumours
Carfilzomib Promotes atypical activation of NF-B Promotes
upregulation of pro-apoptotic BIK and anti-apoptotic MCL1 Induces
complete autophagic flux
Phase1/2 in haematopoietic malignancies and lung cancer; Phase 3
in multiple myeloma
Nelfinavir Inhibited HSP90 function Induced upregulation of
SREBP1 and ATF6 results from inhibition of S2P Activates caspase 3,
caspase 7 and caspase 8 Inhibits AKT signalling, resulting in
downregulation of VEGFA and
HIF1 expression
Phase1/2 in solid tumours and multiple myeloma
Marizomib Induces caspase 8 and ROS-mediated apoptosis Phase1 in
solid tumours and haematopoietic malignancies
MLN9708 Induces activation of caspase 3, caspase 8 and caspase 9
Increases p53, p21, NOXA, PUMA, and E2F expression Inhibits NF-B
signalling pathway
Phase1 in solid tumours; Phase1/2 in haematopoietic
malignancies; Phase 3 in multiple myeloma
NPI-0052 Blocks NF-B signalling Phase1 in solid tumours and
haematopoietic malignancies
Falcarindiol Interferes with proteasome function; mechanisms
remain unclear Preclinical phase
BIP inhibitors: inhibit BIP expression
DHA Inhibited total and surface GRP78 expression Augments the
expression of the ER resident factors ERdj5 and inhibits
PERK
Phase2/3 in solid tumours
PAT-SM6 A monoclonal IgM antibody with high avidity of its
interaction with multiple BIP on cancer cell surface
Phase1 in multiple myeloma
Arctigenin Specifically blocks the transcriptional induction of
BIP and GRP94 under glucose deprivation
Blocked the activation of AKT induced by glucose deprivation
Suppressed both constitutively activated and IL-6-induced STAT3
phosphorylation and subsequent nuclear translocation
Preclinical phase
HSP90 inhibitors: disrupt HSP90 function
Tanespimycin Suppression of chymotryptic activity in the 20S
proteasome Downregulated BRAF, leading to decreased cell
proliferation Inhibited FGF2 and VEGFA-induced HUVEC proliferation
and resulted
in apoptosis
Phase1/2 in solid tumours and haematopoietic malignancies; Phase
3 in multiple myeloma
IPI-504 Interacts with the HSP90 conserved ATP-binding site
Inactivates the transcription factors XBP1 and ATF6 and blocks
the
tunicamycin-induced eIF2 activation by PERK Prevents BIP
accumulation
Phase1/2 in solid tumours and haematopoietic malignancies;
Phase3 in gastrointestinal stromal tumours
Ganetespib Inhibits AKT signalling Reduces expression levels of
HIF1 (but not HIF2) and STAT3
Phase1/2 in solid tumours and haematopoietic malignancies;
Phase3 in non-small-cell lung cancer
AUY922 Suppresses the activity of AKT and ERK in PTEN-null
oesophageal squamous cancer cells, but not in PTEN-proficient
ones
Inhibits NF- B signalling Reduces the expression of
anti-apoptotic protein RAF1
Phase1/2 in solid tumours and haematopoietic malignancies
AT13387 Induces cellular senescence Reduces expression of
oncoproteins EGFR, AKT, CDK4 Restores the expression of p27
Phase1/2 in solid tumours
SNX-5422 NA Phase1 in solid tumours and haematopoietic
malignancies; Phase2 in HER2-positive cancers
PU-H71 Reduces expression levels of AKT, ERK, RAF1, MYC, KIT,
IGF1R, TERT and EWSFLI1 in Ewing sarcoma cells
Promotes degradation of IKK and activated AKT and BCL-XL
Phase1 in solid tumours and haematopoietic malignancies
XL888 Promotes degradation of CDK4 and WEE1 Inhibits AKT
signalling Increases BIM expression and decreases MCL1
expression
Phase1 in melanoma
DS-2248 NA Phase1 in solid tumours
Debio 0932 NA Phase1 in solid tumours and haematopoietic
malignancies
R E V I E W S
592 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
-
cancer cells but not normal cells101. Overexpression of BIP in
cancer cells correlates with chemotherapy resistance (TABLE 1; see
Supplementary informa-tion S1 (table)), which can promote cancer
cell survival by inhibiting p53-mediated expression of pro-
apoptotic BCL-2-antagonist/killer (BAK) and NOXA (also known as
PMAIP1)168. In addition, expression of BIP seems to be increased in
the tumour vascu-lature, suggesting that targeting BIP (TABLE2; see
supplementary information S2 (table)) will have an impact on both
cancer cells and the tumour microenvironment162.
Targeting the UPR in combination cancer therapies. One
therapeutic rationale is to induce ER stress and UPR activation to
activate death pathways in cancer cells. Alternatively, preventing
UPR activation could sensitize cancer cells to other therapies, as
the UPR pro-motes adaptation and drug resistance103105,169. An IRE1
inhibitor sensitized resistant human glioblastoma cells to
oncolytic virus therapy both invitro and invivo170. Inhibition of
PERK kills hypoxic tumour cells that are radioresistant invivo104.
Besides the effects on the actively proliferating cancer cells, a
recent report showed ER stress to be a mechanism for cancer
therapy-induced senescence that acquired the senescence-associated
secretory phenotype171, which indicates that blocking UPR
activation can be an effective therapy for cancer cells undergoing
senescence. Furthermore, inhibition
of UPR signalling may also sensitize cancer-supporting stromal
cells, such as endothelial cells, to traditional cancer therapies.
Therefore, combination therapies that include drugs targeting ER
stress and UPR activation may be one of the most promising
anticancer approaches.
Future prospectsWe have come a long way in understanding the
genetic defects that contribute to cancer; however, we have a long
way to go to translate these findings into clinical advances, and
many questions remain (BOX4). There is an extremely strong pressure
for a cancer cell to survive hostile environments or chemotherapy.
The long-term approach will probably involve combinato-rial
therapies that attack the tumour at multiple levels.
Anti-angiogenesis therapies are not adequate alone, but they may
show synergy in combination with anti-UPR agents. It is also
essential to identify the driver mutations for individual cancer
types to design selec-tive targeting agents. We know that mutations
in BRAF generate melanoma and that BRAF inhibitors generate
resistance. It is necessary to inactivate pro-cesses of drug
resistance, which often involve DNA damage and repair pathways, as
well as eliminating adaptive survival pathways, which provide
potential for the outgrowth of drug-resistant cells. Ever since the
discovery of gene amplification as a mechanism for methotrexate
resistance, it is evident that therapies
PERK inhibitors: inhibit PERK activation and eIF2
phosphorylation
6-shogaol The effect on IRE1 and ATF6 was not obvious
Preclinical stage
GSK2656157 An ATP-competitive inhibitor of PERK Also has eIF2
phosphorylation-independent effects
Preclinical stage
GSK2606414 Binds to PERK active site Preclinical stage
IRE1 inhibitors: inhibit IRE1 endonuclease activity
STF-083010 Inhibits IRE1 endonuclease activity without affecting
its kinase activity
Preclinical stage
MKC-3946 Inhibits IRE1 endonuclease domain, and significantly
enhances apoptosis induced by bortezomib and 17-AAG, associated
with increased levels of CHOP
Preclinical stage
WNT signalling inhibitors
Pyrvinium Suppresses the transcriptional activation of BIP and
GRP94 induced by glucose deprivation or 2-deoxyglucose; other UPR
pathways (for example, XBP1 and ATF4) were also found to be
suppressed
FDA-approved classical anthelmintic; preclinical stage as cancer
therapy
Pan-deacetylase inhibitors
Panobinostat Increases the levels of BIP, IRE1 phosphorylation,
eIF2 phosphorylation, ATF4 and CHOP
Increases the pro-apoptotic BIK, BIM, BAX, and BAK levels, as
well as caspase 7 activity
Phase1/2 in solid tumours and haematopoietic malignancies; Phase
3 in haematopoietic malignancies
Anti-diabetic biguanides
Metformin Inhibition of XBP1 and ATF4 expression during glucose
deprivation FDA-approved anti-diabetes drug; Phase1/2 in solid
tumours and haematopoietic malignancies; Phase 3 in solid
tumours
ATF6, activating transcription factor 6; BIK, BCL-2-interacting
killer; BIM, BCL-2-interacting mediator of cell death; BIP,
immunoglobulin heavy-chain binding protein; CDK4, cyclin-dependent
kinase 4; CHOP, C/EBP homologous protein; EGFR, epidermal growth
factor receptor; eIF2, eukaryotic translation initiation factor 2;
ER, endoplasmic reticulum; FDA, US Food and Drug Administration;
FGF2, fibroblast growth factor 2; FLI1, Friend leukaemia
integration 1 transcription factor; GRP, glucose-regulated protein;
HER2, human epidermal receptor 2; HIF, hypoxia-inducible factor;
HSP90, heat shock protein 90; HUVEC, human umbilical vein
endothelial cell; Ig, immunoglobulin; IGF1R, insulin-like growth
factor 1 receptor; IKK, IB kinase; IL-6, interleukin-6; IRE1,
inositol-requiring protein 1; MCL1, induced myeloid leukaemia cell
differentiation protein; NA, not applicable; NF-B, nuclear
factor-B; PERK, PRKR-like ER kinase; ROS, reactive oxygen species;
SREBP1, sterol regulatory element binding transcription factor 1;
STAT3, signal transducer and activator of transcription 3; TERT,
telomerase reverse transcriptase; UPR, unfolded protein response;
VEGFA, vascular endothelial growth factor A; XBP1, X-box binding
protein 1. *See Supplementary information S2 (table) for a version
of this table with references. #See also ClinicalTrials.gov.
Table 2 (cont.) | Strategies to target the UPR components for
cancer treatment*
Strategy Drugs Other involved mechanisms Clinical trials as
cancer therapy
R E V I E W S
NATURE REVIEWS | CANCER VOLUME 14 | SEPTEMBER 2014 | 593
2014 Macmillan Publishers Limited. All rights reserved
http://www.nature.com/nrc/journal/v14/n9/full/nrc3800.html#supplementary-informationhttp://www.clinicaltrials.gov
-
need to be generated to hit the cancer cells hard, at high dose
and at multiple targets to eliminate the potential for drug
resistance. Targeting the UPR adap-tive pathways will provide one
asset of the armamen-tarium in these strategies, but they are
unlikely to be successful on their own. It is important that
clinical avenues are encouraged for testing of multiple agents in
single clinical studies, while at the same time recog-nizing the
potential for increased toxicities and drug
interactions. Certainly, with the advent of personal-ized
medicine and genome sequencing, therapeutic strategies will become
more patient-specific and could increase the success of remission,
especially in those cancers that are presently refractory to any
treatment. We have come a long way, and this is most evident by
remission rates in chronic myelogenous leukaemia (CML), but we have
a long distance to run to win therace.
1. Dorner,A.J., Bole,D.G. & Kaufman,R.J. The relationship of
N-linked glycosylation and heavy chain-binding protein association
with the secretion of glycoproteins. J.Cell Biol. 105, 26652674
(1987).
2. Kozutsumi,Y., Segal,M., Normington,K., Gething,M.J. &
Sambrook,J. The presence of malfolded proteins in the endoplasmic
reticulum signals the induction of glucose-regulated proteins.
Nature 332, 462464 (1988).
3. Dorner,A.J., Wasley,L.C. & Kaufman,R.J. Increased
synthesis of secreted proteins induces expression of
glucose-regulated proteins in butyrate-treated Chinese hamster
ovary cells. J.Biol. Chem. 264, 2060220607 (1989).
4. Denoyelle,C. etal. Anti-oncogenic role of the endoplasmic
reticulum differentially activated by mutations in the MAPK
pathway. Nature Cell Biol. 8, 10531063 (2006).This study unraveled
a potential tumour-suppressive function of the UPR through
mediating oncogenic HRAS-driven cellular senescence in melanocytes
independent of canonical senescence mediators such as p53 or p16
(also known as INK4A).
5. Lee,D.Y. & Sugden,B. The LMP1 oncogene of EBV activates
PERK and the unfolded protein response to drive its own synthesis.
Blood 111, 22802289 (2008).
6. Hsiao,J.R. etal. Endoplasmic reticulum stress triggers
XBP-1-mediated up-regulation of an EBV
oncoprotein in nasopharyngeal carcinoma. Cancer Res. 69,
44614467 (2009).
7. Hart,L.S. etal. ER stress-mediated autophagy promotes
Myc-dependent transformation and tumor growth. J.Clin. Invest. 122,
46214634 (2012).This study showed that the PERKeIF2 branch of the
UPR can promote cancer formation and progression in MYC-dependent
mouse lymphoma models by inducing cytoprotective autophagy.
8. Ozcan,U. etal. Loss of the tuberous sclerosis complex tumor
suppressors triggers the unfolded protein response to regulate
insulin signaling and apoptosis. Mol. Cell 29, 541551 (2008).
Box 4 | Ten unresolved questions regarding the impact of ER
stress on cancer development
The unfolded protein response (UPR) functions as a double-edged
sword by supporting or repressing cancer initiation and
progression. In premalignant cells, UPR activation prevents
oncogenic transformation by inducing cell death in response to
sustained endoplasmic reticulum (ER) stress and facilitates
clearance of damaged cells by the immune system. In cancer, UPR
activation provides a survival strategy for cells to thrive in a
stressful environment. Activation of the UPR may inhibit cancer
progression through induction of tumour cell death, in which UPR
inhibition may reduce tumour angiogenesis. Despite recent advances
in the field, many questions remain to be answered to encourage
therapeutic testing of UPR-directed agents for chemotherapy: Is
protein misfolding oncogenic? Protein misfolding in the ER induces
oxidative stress33,34,50,51. It is unknown whether this
oxidative stress is sufficient to cause DNA mutations to
activate oncogenes or inactivate tumour suppressor genes. Recent
studies suggest that protein misfolding in hepatocytes can be an
initiating event for hepatocellular carcinoma57.
UPR activation is both adaptive and pro-apoptotic. Why does UPR
activation promote cell survival under certain circumstances,
whereas persistent activation of the same UPR signalling pathway
causes cell death under other conditions?
Some UPR components are independent prognostic indicators of
therapeutic outcome. What is the status of the ER stress sensors
and downstream targets in a broad range of human patient cancer
samples compared to healthy adjacent tissue? Can the status of ER
stress and UPR activation serve as a prognostic indicator of
outcome?
What are the exact roles of individual UPR components in cancer
incidence and progression? Studies of transgenic and knockout mice
and characterization of specific small molecule inhibitors or
activators in mouse models are needed to determine whether
activation of a specific UPR signalling pathway is a rate-limiting
primary step or a secondary event during cancer initiation and/or
progression.
Cancer progresses in the presence of a stressful
microenvironment. How do cancer cells evade cell death upon chronic
UPR activation? Can UPR-targeted therapeutics be designed to
separate pro-survival and apoptotic responses?
Solid tumours are highly heterogeneous. Does the status of UPR
activation reflect such heterogeneity? Could UPR-mediated
pro-survival and pro-death signals coexist in different regions of
the same tumour?
UPR signalling pathways do not function in isolation. How do
different survival or death signalling pathways integrate with each
other to control the fate of tumour cells under unfavourable
conditions?
Cancer stem cells (CSCs) were recently identified to be
responsible for cancer metastasis. Indeed, X-box binding protein 1
(XBP1) is possibly required for the maintenance of this population
in triple-negative breast cancer93, whereas PRKR-like ER kinase
(PERK)eukaryotic translation initiation factor 2 (eIF2) is
activated during epithelial-to- mesenchymal transition, which is
required for invasion and metastasis188. The PERK pathway is also
required for rapid induction of detachment-induced autophagy, which
is crucial for the survival of detached cancer cells189. However,
more questions need be addressed. Does UPR activation promote
CSC-like properties and CSC-niche interactions to augment
metastasis? Which UPR component (or components) is essential for
CSC survival and differentiation?
Different tumour cells exhibit different degrees of protein
secretion. Does the level of protein secretion of a tumour cell
correlate with dependence on specific UPR pathways for survival?
Can protein secretion rate be used to stratify tumour cell
sensitivity to UPR-targeting agents?
Mutations in several UPR components have been identified in
human cancers. Are any of these driver mutations?
R E V I E W S
594 | SEPTEMBER 2014 | VOLUME 14
www.nature.com/reviews/cancer
2014 Macmillan Publishers Limited. All rights reserved
-
9. Kang,Y.J., Lu,M.K. & Guan,K.L. The TSC1 and TSC2 tumor
suppressors are required for proper ER stress response and protect
cells from ER stress-induced apoptosis. Cell Death Differ. 18,
133144 (2011).
10. Yeung,B.H. etal. Glucose-regulated protein 78 as a novel
effector of BRCA1 for inhibiting stress-induced apoptosis. Oncogene
27, 67826789 (2008).
11. Fang,M. etal. The ER UDPase ENTPD5 promotes protein
N-glycosylation, the Warburg effect, and proliferation in the PTEN
pathway. Cell 143, 711724 (2010).
12. Marada,S., Stewart,D.P., Bodeen,W.J., Han,Y.G. &
Ogden,S.K. The unfolded protein response selectively targets active
smoothened mutants. Mol. Cell. Biol. 33, 23752387 (2013).
13. Babour,A., Bicknell,A.A., Tourtellotte,J. & Niwa,M. A
surveillance pathway monitors the fitness of the endoplasmic
reticulum to control its inheritance. Cell 142, 256269 (2010).
14. Kufe,D.W. Mucins in cancer: function, prognosis and therapy.
Nature Rev. Cancer 9, 874885 (2009).
15. Mahadevan,N.R. etal. Transmission of endoplasmic reticulum
stress and pro-inflammation from tumor cells to myeloid cells.
Proc. Natl Acad. Sci. USA 108, 65616566 (2011).This study
demonstrated for the first time that ER stress can be transmitted
from tumour cells to macrophages to elicit a pro-inflammatory
response.
16. Wang,Y. etal. The unfolded protein response induces the
angiogenic switch in human tumor cells through the PERK/ATF4
pathway. Cancer Res. 72, 53965406 (2012).
17. Verfaillie,T. etal. PERK is required at the ER-mitochondrial
contact sites to convey apoptosis after ROS-based ER stress. Cell
Death Differ. 19, 18801891 (2012).
18. Marcu,M.G. etal. Heat shock protein 90 modulates the
unfolded protein response by stabilizing IRE1. Mol. Cell. Biol. 22,
85068513 (2002).
19. Bertolotti,A., Zhang,Y., Hendershot,L.M., Harding,H.P. &
Ron,D. Dynamic interaction of BiP and ER stress transducers in the
unfolded-protein response. Nature Cell Biol. 2, 326332 (2000).
20. Liu,C.Y., Schroder,M. & Kaufman,R.J. Ligand-independent
dimerization activates the stress response kinases IRE1 and PERK in
the lumen of the endoplasmic reticulum. J.Biol. Chem. 275,
2488124885 (2000).
21. Scheuner,D. etal. Translational control is required for the
unfolded protein response and invivo glucose homeostasis. Mol. Cell
7, 11651176 (2001).
22. Das,S., Ghosh,R. & Maitra,U. Eukaryotic translation
initiation factor 5 functions as a GTPase-activating protein.
J.Biol. Chem. 276, 67206726 (2001).
23. Jennings,M.D. & Pavitt,G. D. eIF5 is a dual function GAP
and GDI for eukaryotic translational control. Small GTPases 1,
118123 (2010).
24. Siekierka,J., Manne,V. & Ochoa,S. Mechanism of
translational control by partial phosphorylation of the subunit of
eukaryotic initiation factor 2. Proc. Natl Acad. Sci. USA 81,
352356 (1984).
25. Rowlands,A.G., Montine,K.S., Henshaw,E.C. & Panniers,R.
Physiological stresses inhibit guanine-nucleotide-exchange factor
in Ehrlich cells. Eur. J.Biochem. 175, 9399 (1988).
26. Harding,H.P. etal. Regulated translation initiation controls
stress-induced gene expression in mammalian cells. Mol. Cell 6,
10991108 (2000).
27. Cullinan,S.B. etal. Nrf2 is a direct PERK substrate and
effector of PERK-dependent cell survival. Mol. Cell. Biol. 23,
71987209 (2003).
28. Zhang,W. etal. ER stress potentiates insulin resistance
through PERK-mediated FOXO phosphorylation. Genes Dev. 27, 441449
(2013).
29. Bobrovnikova-Marjon,E. etal. PERK utilizes intrinsic lipid
kinase activity to generate phosphatidic acid, mediate Akt
activation, and promote adipocyte differentiation. Mol. Cell. Biol.
32, 22682278 (2012).
30. Lu,P.D. etal. Cytoprotection by pre-emptive conditional
phosphorylation of translation initiation factor 2. EMBO J. 23,
169179 (2004).
31. Yaman,I. etal. The zipper model of translational control: a
small upstream ORF is the switch that controls structural
remodeling of an mRNA leader. Cell 113, 519531 (2003).
32. Palam,L.R., Baird,T.D. & Wek,R.C. Phosphorylation of
eIF2 facilitates ribosomal bypass of an inhibitory upstream ORF to
enhance CHOP translation. J.Biol. Chem. 286, 1093910949 (2011).
33. Marciniak,S.J. etal. CHOP induces death by promoting protein
synthesis and oxidation in the stressed endoplasmic reticulum.
Genes Dev. 18, 30663077 (2004).
34. Song,B., Scheuner,D., Ron,D., Pennathur,S. &
Kaufman,R.J. Chop deletion reduces oxidative stress, improves cell
function, and promotes cell survival in multiple mouse models of
diabetes. J.Clin. Invest. 118, 33783389 (2008).
35. Chitnis,N.S. etal. miR-211 is a prosurvival microRNA that
regulates chop expression in a PERK-dependent manner. Mol. Cell 48,
353364 (2012).
36. Woo,C.W., Kutzler,L., Kimball,S.R. & Tabas,I. Toll-like
receptor activation suppresses ER stress factor CHOP and
translation inhibition through activation of eIF2B. Nature Cell
Biol. 14, 192200 (2012).
37. Rutkowski,D.T. etal. Adaptation to ER stress is mediated by
differential stabilities of pro-survival and pro-apoptotic mRNAs
and proteins. PLoS Biol. 4, e374 (2006).
38. Lin,J.H. etal. IRE1 signaling affects cell fate during the
unfolded protein response. Science 318, 944949 (2007).
39. Scheuner,D. etal. Double-stranded RNA-dependent protein
kinase phosphorylation of the -subunit of eukaryotic translation
initiation factor 2 mediates apoptosis. J.Biol. Chem. 281,
2145821468 (2006).
40. McCullough,K.D., Martindale,J.L., Klotz,L.O., Aw,T.Y. &
Holbrook,N.J. Gadd153 sensitizes cells to endoplasmic reticulum
stress by down-regulating Bcl2 and perturbing the cellular redox
state. Mol. Cell. Biol. 21, 12491259 (2001).
41. Puthalakath,H. etal. ER stress triggers apoptosis by
activating BH3-only protein Bim. Cell 129, 13371349 (2007).
42. Szegezdi,E., Logue,S.E., Gorman,A.M. & Samali,A.
Mediators of endoplasmic reticulum stress-induced apoptosis. EMBO
Rep. 7, 880885 (2006).
43. Galehdar,Z. etal. Neuronal apoptosis induced by endoplasmic
reticulum stress is regulated by ATF4-CHOP-mediated induction of
the Bcl-2 homology 3-only member PUMA. J.Neurosci. 30, 1693816948
(2010).
44. Hsin,I.L. etal. Lipocalin 2, a new GADD153 target gene, as
an apoptosis inducer o