You are on page 1of 16

REVIEWS

New perspectives for targeting RAF


kinase in human cancer
Zoi Karoulia1, Evripidis Gavathiotis2 and Poulikos I.Poulikakos1
Abstract | The discovery that a subset of human tumours is dependent on mutationally
deregulated BRAF kinase intensified the development of RAF inhibitors to be used as potential
therapeutics. The US Food and Drug Administration (FDA)-approved second-generation RAF
inhibitors vemurafenib and dabrafenib have elicited remarkable responses and improved survival
of patients with BRAFV600E/K melanoma, but their effectiveness is limited by resistance.
Beyond melanoma, current clinical RAF inhibitors show modest efficacy when used for colorectal
and thyroid BRAFV600E tumours or for tumours harbouring BRAF alterations other than the
V600 mutation. Accumulated experimental and clinical evidence indicates that the complex
biochemical mechanisms of RAF kinase signalling account both for the effectiveness of RAF
inhibitors and for the various mechanisms of tumour resistance to them. Recently, a number of
next-generation RAF inhibitors, with diverse structural and biochemical properties, have entered
preclinical and clinical development. In this Review, we discuss the current understanding of RAF
kinase regulation, mechanisms of inhibitor action and related clinical resistance to these drugs.
The recent elucidation of critical structural and biochemical aspects of RAF inhibitor action,
combined with the availability of a number of structurally diverse RAF inhibitors currently in
preclinical and clinical development, will enable the design of more effective RAF inhibitors
andRAF-inhibitor-based therapeutic strategies, tailored to different clinical contexts.

The family of RAF kinases (ARAF, BRAF and CRAF change from valine (V) to glutamic acid (E) in the acti-
(also known as RAF1)) constitute core components of vation segment of the kinase (V600E), promoting sever-
the RASRAFMEKERK signalling cascade (ERK al-fold kinase hyperactivation9,10. In addition toV600E,
signalling), a pathway that mediates signals from cell and other substitutions at V600 (such as V600K,
surface receptors to the nucleus to regulate cell growth, V600Dand V600R), a number of nonV600 missense
differentiation and survival14. Upon activation, RAF BRAF mutations have been found, mostly clustered in
kinases phosphorylate and activate the kinases MEK1 the activation segment or in the socalled glycine-rich
and MEK2, which in turn phosphorylate and activate loop of the kinase domain911. Point mutations are not
ERK1 and ERK2. Activated ERK promotes cell prolifera- the only alterations found in BRAF. Fusion proteins
1
Department of Oncological
Sciences and Department of tion and survival by phosphorylating multiple substrates resulting from translocations containing the catalytic
Dermatology, The Tisch both in the cytosol and in the nucleus4. domain of BRAF, as well as inframe deletions are also
Cancer Institute, Icahn School Deregulation of ERK signalling commonly occurs in found in certain tumour types, resulting in constitutive
of Medicine at Mount Sinai, cancer, frequently due to mutations of components of activation of BRAF and downstream ERK signalling 12,13.
New York, New York 10029,
USA.
the pathway 5. It has been known for more than four dec- BRAF is commonly mutated in melanomas (50%)14,
2
Department of Biochemistry, ades that mutations in the genes encoding for the RAS papillary thyroid cancers (45%)15, colon cancers (10%)16,
Department of Medicine, family of proteins are often found in human tumours6,7. non-small-cell lung cancers (NSCLCs) (10%)17, in vir-
Albert Einstein Cancer Center, RAS protein has been extremely difficult to target, and tually 100% of hairy cell leukemias18 and in 5060%
Albert Einstein College of
thus the downstream kinases RAF, MEK and ERK of patients with the idiopathic disorder Langerhans cell
Medicine, Bronx, New York
10461, USA. remain attractive therapeutic targets in such tumours, histiocytosis (LCH)19. These findings, along with preclini
Correspondence to P.I.P.
although current inhibitors of these kinases have only cal work demonstrating dependence of BRAFV600E
poulikos.poulikakos@ showed limited efficacy in RAS-mutant cancers8. BRAF tumours on BRAF and ERK signalling activity 20, sup-
mssm.edu mutations are present in approximately 8% of human ported the therapeutic benefit of targeting ERK signal-
doi:10.1038/nrc.2017.79 tumours9. The most frequent mutation in the BRAF gene ling in cancer and intensified efforts for the development
Published online 6 Oct 2017 is the 1799T>A substitution that results in an amino acid of selective ATP-competitive inhibitors of mutant BRAF

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 1



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

a trials. Both combinations (vemurafenib and cobimetinib


(Roche/Genentech) and dabrafenib and trametinib
(Novartis)) showed improved clinical efficacy compared
RAS-GTP RAS-GTP P P RAS-GTP with RAF inhibitor monotherapy 24,25. Two combina-
NTD NTD NTD 3 tions are now the standard of care for BRAFV600E/K
Kinase metastatic melanoma, although they are still not cura-
2 RAF active dimer
ATP ATP ATP tive treatments for the majority of patients. Importantly,
ATP
the most common mechanisms of resistance identified
P
P in tumours following relapse after combination RAF
NTD

RAF inactive 1 MEK


monomer and MEK inhibitor therapy are associated with recovery
P P
ofERK signalling and are similar to the mechanisms of
ERK resistance detected after RAF inhibitor monotherapy 2628;
this observation provides the rationale for the develop-
ment of future therapeutic strategies that could achieve
b BRAF active protomer c BRAF inactive monomer more potent inhibition of RAFERK signalling and may
therefore yield deeper and more durable responses.
The remarkable clinical efficacy of currently
N-lobe
R506 approved RAF inhibitors has been so far confined to
BRAFV600E/K melanoma. Besides the most prevalent
R506 N-lobe
mutations in melanoma, BRAFV600E (7080%) and
BRAFV600K (512%), additional BRAFV600 mutant
C-OUT
C-IN forms have been identified in patients with melanoma
DFG-IN
Hinge G-loop
Hinge
but are less common. Among them, BRAFV600R
DFG-IN
G-loop (37%) and the rare mutations BRAFV600M,
AS AS
BRAFV600D and BRAFV600G (<5% in total) have
been reported to be sensitive to approved RAF inhibitors
in preclinical2932 and clinical22,3335 studies. By contrast,
C-lobe preclinical and clinical data suggested that colorectal and
C-lobe thyroid BRAFV600E tumours, as well as tumours with
BRAF mutations other than V600 substitutions, are less
sensitive to current clinical RAF inhibitors3639.
In this Review, we highlight the most recent devel-
opments in our understanding of the biochemical
Figure 1 | RAF activation. a|Schematic showing canonical RAF activation. In conditions
of low RAS-GTP, RAF is monomeric and inactive in the cytosol dueNature Reviews | Cancer
to intramolecular basis of sensitivity and resistance to second-generation
interaction between the Nterminal domain (NTD; regulatory) and the Cterminal (kinase) RAF inhibitors in cancer therapy. We will also discuss
domain (1). Upregulation of RAS-GTP promotes the formation of the RAFRAS-GTP the ongoing development of third-generation RAF
complex in the membrane due to the high affinity of RAS-GTP for the RAS-binding inhibitors with new structural properties designed to
domain (RBD) present in the NTD of RAF (2). This step is followed by activating both address certain toxicities elicited by current clin-
phosphorylation and dimerization for full RAF activation (3). Chelix indicated by the red ical inhibitors and provide more durable therapeutic
wavy line in the kinase domain. b|Ribbon representation of one protomer of the active responses in patients with BRAF-mutant tumours.
BRAF dimer (Protein Data Bank (PDB) ID: 4MNE). The kinase domain consists of N and C
terminal lobes linked through a short flexible hinge. c | Ribbon representation of the Structural insight into RAF activation
inactive and monomeric BRAF kinase domain (PDB ID: 4RZV). In parts b and c, note
In normal cells and in the absence of upstream activity,
thedifferent positions of the Chelix (in green) and of the activation segment (AS,
magenta), which highlight the movement of the Chelix from the OUT to the IN position cytosolic RAF is thought to adopt a monomeric, closed
and the unfolding of the AS upon RAF activation. The R506 residue side chain in the IN and inactive conformation due to the intramolecular
(active) and the OUT (inactive) positions and the conformation of the glycine-rich loop interaction between the aminoterminal (regulatory)
(Gloop) and DFG motif in the active and inactive form are also depicted. and the carboxyterminal (kinase) domains40,41 (FIG.1a).
Canonical RAF activation is regulated by members of the
RAS family of small GTPases (KRAS, NRAS and HRAS),
kinase. Two second-generation RAF inhibitors, vemu- which, upon growth factor stimulation of upstream
rafenib and dabrafenib, showed remarkable clinical receptors, are converted from the inactive (RAS-GDP)
activity in patients with BRAFV600E/K melanoma to the active (RAS-GTP) form in the plasma membrane.
and received US Food and Drug Administration (FDA) The increase in the levels of RAS-GTP results in RAF
approval for the treatment of this disease21,22. However, translocation to the membrane and formation of the
despite prolonging patient survival, RAF inhibitor treat- RAFRAS-GTP complex, due to the high affinity of
Langerhans cell histiocytosis ment is rarely curative and is limited in most cases by RAS-GTP for the RAS-binding domain (RBD) in the
(LCH). A myeloid neoplasia the development of drug resistance and tumour relapse. N terminus of RAF42,43 (FIG.1a). RAF kinases bound to
characterized by inflammatory As preclinical and clinical evidence suggested that more RAS-GTP in the membrane become fully activated via
lesions containing pathological
dendritic cells, frequently
potent inhibition of ERK signalling in the tumour would priming (which includes phosphorylation at critical resi
harbouring the BRAFV600E yield improved responses23, combinations of RAF inhib- dues, such as S338 in CRAF) and homodimerization and
mutation. itors with MEK inhibitors were tested in subsequent heterodimerization through the kinase domain44,45.

2 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

RAF has a typical kinase domain architecture, with conformational changes promoted by the V600E substi-
the Nterminal lobe (Nlobe) and Cterminal lobe tution10,54. Recent studies suggested the importance of a
(Clobe) linked through a short flexible hinge. At the salt bridge between V600E in the activation segment and
interface of the two lobes lies the active site of the kinase K507 of the Chelix in stabilizing the active conforma-
that includes the nucleotide (ADP or ATP) binding tion of BRAFV600E as evidenced by crystal structures
site, the magnesium binding site (DFG motif ), and of the BRAFV600Edimer 54,55. However, cell-based and
the phospho-acceptor site (activation segment). The biochemical data established that BRAFV600 proteins in
motions between the lobes are controlled by the flexible the presence of low levels of RAS-GTP can be catalytically
hinge and enable recruitment of the substrate and release active monomers30,31,5153,56. The reason for this discrep-
of the product 46. The catalytically active conformationof ancy is presumably the absence of the Nterminal regu
the kinase domain requires a closed conformation latory domain of BRAF in available crystal structures; this
between the two lobes, an unfolded activation segment domain may be involved in the stabilization of an active
and the Chelix of the Nterminal lobe fixed into the IN monomeric conformation of V600mutant BRAF kinase.
position to bring the catalytic residues into a productive Further studies are needed to elucidate the structural
distance and orientation (FIG.1b). The active conforma- details of BRAF kinase activation. Structural understand-
tion in RAF is facilitated by dimerization. Dimerized ing of the interaction of the Nterminal regulatory domain
RAF has been visualized by several crystal structures with theRAF kinase domain, in eitherthe inactive con-
determined with BRAF or CRAF bound to inhibitors, formation or the active conformation of themonomer,
providing insights into the dimerization interface10,30,4750. would be a major step towards elucidatingthe regulation
The dimerization interface of each BRAF protomer of normal and oncogenicBRAF.
includes the Chelix of each kinase and specifically Beyond the physiological signalling mechanism
the interaction between the Cterminal R509 residue of of BRAF dimerization-induced activation through
each Chelix. Mutagenesis studies have established the activated RAS, cell-based studies showed that certain
critical role of R509 and adjacent residues in mediating oncogenic BRAF mutations promote spontaneous
BRAF dimerization5053. BRAF dimerization and activation, either by form-
Recently, two crystal structures of monomeric BRAF ing homodimers in the absence of RAS-GTP, such as
kinase bound to structurally related compounds30,54 and BRAFL597V or BRAFG466E11,30,31,51, or by further
the crystal structure of the BRAFMEK complex 55 have enhancing RAF dimerization promoted by RAS-GTP,
been determined, thereby providing important insight such as the catalytically inactive BRAFD594N mutant
on the conformational transitions and the structural or impaired activity mutants 57. As in the case of
basis of dimerization-dependent RAF activation. The BRAFV600 mutants, the exact mechanism by which
structure of monomeric BRAF shows the kinase in an these mutations induce dimerization of the kinase is not
inactive conformation characterized by an open config- well understood, due to the lack of structural informa-
uration between the Nlobe and Clobe of the kinase tion that includes the Nterminal domain of BRAF, or at
domain30,54. The Chelix is positioned in the OUT least the part that regulates dimerization of the catalytic
position, where it is stabilized by interactions with resi domain. Furthermore, several mutations are found in
dues of the folded activation segment, which has also residues in the activation segment or in the glycine-rich
been determined30,54 (FIG.1c). The structure of dimeric loop, which interacts with the activation segment 9. In
BRAF kinase bound to MEK enabled the visualization most cocrystal structures of BRAF kinase, several resi
of an active conformation of the BRAF structure and the dues of the activation segment are not resolved, pre-
closed configuration between the Nlobe and Clobe of cluding information of the position of these residues
the kinase55. In this conformation, the position of the and their interactions in the activated BRAF structure.
Chelix is shifted to the full IN position, and the acti- The recently reported structures of BRAF monomers
vation segment is extended as an unstructured loop that in the inactive conformation suggest that mutation
enables the shift of the Chelix to facilitate catalysis and in the activation segment such as V600E may disrupt
the interaction with the substrate55. The crystal struc- the hydrophobic packing between the inactive state of
tures show the DFG motif to adopt an IN conformation the activation segment and the Chelix through steric
in both the active and the inactive kinase conforma- clashes or unfavourable interactions; in turn, this can
tions30,55 (FIG.1b,c). These structural studies suggest that promote allosteric movement of the Chelix towards
Protomer
BRAF undergoes dynamic conformational transitions the IN position, which eventually favours dimeriza-
The structural unit of an within the Nlobe, Clobe, Chelix and activation seg- tion of the RAF kinase domain30,54. Additional crystal
oligomeric protein; in the case ment. These transitions are structurally interconnected structures of BRAF mutants other than the V600 sub-
of wild-type BRAF, a functional and reveal how dimerization promotes activation of the stitution are required for more definitive and detailed
BRAF dimer is composed of
kinasedomain. understanding of the mechanismsinvolved.
two protomers.
Despite a number of available crystal structures
Steric clashes of either wild-type (WT) BRAF or BRAFV600E, the Models of RAF inhibitor action
When atoms from different detailed structural mechanism of BRAF activation RAF inhibitors have unusual biochemical properties.
residues come into close by the most common BRAF mutation (V600 substi- Unlike most kinase inhibitors, which suppress their
proximity, the resultant
repulsion between the
tution) is still not well understood. Cocrystal struc- targets in all cells, current clinical RAF inhibitors sup-
atomsleads to a change tures of BRAFV600E with RAF inhibitors suggested press RAF activity and ERK signalling selectively in
inconformation. the potential for dimerization-induced activation by cells expressing mutant BRAF. In tumour and normal

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 3



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

a BRAF-VEM dimer BRAF-AZ dimer BRAF-SB dimer

C-OUT C-IN C-IN

AZ SB
VEM C-IN DFG-IN C-IN
DFG-OUT C-IN
DFG-IN DFG-IN DFG-IN
DFG-OUT SB
AZ

b C-OUT RAF inhibitor c C-IN RAF inhibitor

RAS-GTP RAS-GTP P P RAS-GTP RAS-GTP RAS-GTP P P RAS-GTP


NTD NTD NTD NTD NTD NTD
3 3
Kinase Kinase Inhibited
2 RAF active 2
INH INH ATP protomer INH INH INH RAF dimer
INH INH
P
P
NTD

NTD
RAF inactive 1 MEK RAF inactive 1 MEK
monomer monomer
P P
ERK ERK

Nature bound
Figure 2 | Mechanism of action of RAF inhibitors. a|Structural features of representative RAF inhibitors Reviewsto|aCancer
BRAF
dimer. The three structural types of RAF inhibitor bound to BRAF are shown as ribbon representations: Vemurafenib (VEM),
a typeI1/2 inhibitor (COUT/DFGIN), AZ628 (AZ), a typeII inhibitor (CIN/DFG-OUT), and SB590885 (SB), a typeI inhibitor
(CIN/DFGIN). Vemurafenib and other typeI1/2 inhibitors stabilize the Chelix (red) of the first BRAF protomer to the OUT
position and the Chelix (green) of the second BRAF protomer to the IN position (Protein Data Bank (PDB) ID: 3OG7).
AZ628 and other typeII inhibitors stabilize the Chelix of the first BRAF protomer (red) and the Chelix of the second
BRAF protomer (green) to the IN position (PDB ID: 4RZW). SB590885 and other typeI inhibitors stabilize both the Chelix
of the first BRAF protomer (red) and the Chelix of the second BRAF protomer (green) to the IN position (PDB ID: 2FB8).
AZ628 and other typeII inhibitors induce the DFG motif to an OUT conformation, while other inhibitors of the typeI and I1/2
groups maintain the DFG in an IN conformation. b|Schematic showing the COUT RAF inhibitor-induced paradoxical RAF
pathway activation and negative allostery. Most COUT RAF inhibitors stabilize the Chelix (indicated by the red wavy
line in the kinase domain) in the overall OUT position (1). However, the Cterminal Chelix residue R506 is displaced into
the IN position, which promotes the interaction of RAF with RAS-GTP (2). Subsequently, RAF is primed through
phosphorylation and dimerization. RAF inhibitor binding to one protomer and stabilization of the Chelix in the OUT
position sterically prevents inhibitor binding to the second monomer (negative allostery), leading to paradoxical pathway
activity (3). c|Schematic showing CIN RAF inhibitor-induced RAF activation and effective dimer inhibition. By stabilizing
the Chelix in the IN position (1), CIN RAF inhibitors promote the formation of the RAFRAS-GTP complex(2). This is
followed by RAF phosphorylation and dimerization. Due to low negative allostery, RAF inhibitors belonging to this group
effectively bind and inhibit both protomers in the RAF dimer (3). INH, inhibitor; NTD, N-terminal domain.

cells expressing WT BRAF, these RAF inhibitors do not differences in the position of the Chelix within each
inhibit but instead paradoxically activate RAF activity protomer (FIG.2a). A notable example is vemurafenib,
and downstream ERK signalling (the socalled RAF which binds the first protomer and stabilizes the C
inhibitor paradox)1,47,49,5663. The complex biochemical helix towards the OUT position. In the second protomer,
mechanism of action of RAF inhibitors has been the the IN position of the Chelix creates negative allostery
topic of intense investigation, and it was only recently, for binding a second vemurafenib molecule, because
with the integration of structural, biochemical and movement of the Chelix towards the OUT position
cell-based analysis of a large number of structurally in that protomer would cause disruption of the dimer
Negative allostery
Also known as allosteric diverse RAF inhibitors, that it has been understood in (Supplementary informationS1 (movie)). By contrast,
inhibition, occurs when binding moredetail. other RAF inhibitors, such as AZ628, that stabilize the
of one ligand to a substrate (in Chelix of the first protomer towards the IN position,
this case a BRAF protomer) Structural insight into RAF inhibitor activity. sterically allow occupancy of the second protomer by a
decreases the affinity of
another ligand at a different
Dependent on their chemical structure, RAF inhibitors second drug molecule (Supplementary informationS2
site (that is, the other stabilize the Chelix in a position between the IN and (movie)). Thus, RAF kinase inhibitors can be broadly
protomer). OUT conformations, forming imperfect dimers due to classified according to the conformation in which

4 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

they stabilize their target kinase CIN (either OUT position of the Chelix is mostly inferred based
ChelixIN/DFGIN (typeI) or ChelixIN/DFG- on structural and cell-based data and remains to be
OUT (typeII)) or COUT, the latter most commonly shown directly by biophysical methods, it helps explain
occurring as Chelix-OUT/DFGIN (typeI1/2)64, which why COUT inhibitors are selectively effective in cellu-
is the class to which the current clinical RAF inhibitors lar contexts in which RAF signals as a monomer (that is,
vemurafenib and dabrafenib belong (FIG.2a and TABLE1). BRAFV600 tumour cells with low RAS-GTP) and why
increased RAF dimerization is a common mechanism of
Linking structure to biochemical effects of RAF inhibi resistance to these drugs30.
tors in cells. The first studies focusing on the mechan Second, the basis for priming and RAS-dependent
isms underlying the unique biochemical properties of dimerization of RAF promoted by RAF inhibitors is the
RAF inhibitors were published during the preclinical stabilization of the Chelix towards the IN (active) posi-
and clinical development of inhibitors targeting mutated tion. CIN RAF inhibitor binding to RAF promotes the
BRAF47,56,57. Although not all studies concurred in their interaction of RAF with RAS-GTP, increasing activating
proposed models, they all agreed that paradoxical activa- phosphorylation and RAF dimerization, and essentially
tion of ERK signalling by RAF inhibitors in WT BRAF- hijacking the process of canonical growth factor-induced
expressing cells required active RAS. However, the details RAF activation30. The mechanism is reminiscent of
of this mechanism remained elusive. Furthermore, allosteric priming by inhibitor, which has been observed
in those and subsequent studies, RAF dimerization with other kinase inhibitors (such as those targeting
was shown to confer resistance to RAF inhibitors31,52, AKT69, protein kinase C (PKC)70, ribosomal 6kinase 1
but the structural basis of the phenomenon remained (S6K1)71 and Janus kinase 2 (JAK2)72), although in these
unknown. More recently, detailed analysis of crystal- cases virtually all the primed and phosphorylated kinase
lographic data in combination with cell-based studies molecules are drug-bound and inhibited. Analysis of RAF
enabled the structural properties of RAF inhibitors to be priming by various inhibitors, led to the hypothesis that
linked to their biochemical effects in cells30,49 (FIG.2b,c). it is the area around a specific amino acid of the Chelix
Specifically, thebiochemical effect of RAF inhibitors is (R506) and not the overall position of the Chelix that
the outcome of distinct allosteric mechanisms, deter- determines the extent of the RAFRAS-GTP interaction
mined by the combination of the specific conformation induced by the inhibitor 30. According to this model,
in which the RAF kinase is stabilized by the inhibitor CIN inhibitors, such as GDC0879 or SB590885, dis-
and by the cellular levels of RAS-GTP. place R506 from the OUT position and thus promote
First, resistance of dimeric RAF to inhibitors is the RAFRAS-GTP interaction30. Importantly, dabrafenib
consequence of stabilization of the Chelix in the OUT and to a lesser extent vemurafenib, although they stabi-
(inactive) position by the inhibitor, because this position lize the Chelix in an overall OUT position, also pro-
of the Chelix is not sterically allowed for both pro- mote RAFRAS-GTP interaction and RAF dimerization
tomers in a RAF dimer. Thus, binding of an COUT due to displacement of the R506 residue towards the IN
inhibitor to the first protomer stabilizes the position position30. Members of a certain class of next-generation
of the Chelix of the second protomer towards the inhibitors (such as PLX7904 and PLX8394; see below
IN position, reducing the affinity of a second COUT for more details) stabilize the entire Chelix, includ-
inhibitor for the second protomer (negative allostery). ing the R506 residue, in the OUT position and thus do
Such inhibitors are predicted to be ineffective inhibi- not promote paradoxical activation of ERK signalling
tors of dimeric RAF30 (FIG.2b). Certain COUT RAF in cells30,73. Although further experimental data are
inhibitors, such as dabrafenib, stabilize the Chelix in required to substantiate this hypothesis, it may explain
a less OUT position compared to vemurafenib in both the requirement of RAS-GTP for paradoxical activation
protomers, a position that allows the inhibitor to bind of ERK signalling by inhibitors and the concurrent find-
Therapeutic window both protomers, as seen in the crystal structure with ing of mutant-RAS expression in the squamous-cell car-
The range of concentrations
BRAF kinase. However, this might be an effect of crys- cinomas and keratoacanthomas, commonly induced as
ofa drug in the patient that
provides safe effective therapy. tallization forcing the two protomers into a distorted second-site tumours in patients with melanoma treated
The therapeutic window is wide dimer. In contrast, CIN RAF inhibitors bind both with second-generation RAF inhibitors74.
when the minimum toxic protomers of dimeric RAF with similar affinity and are Based on this model, paradoxical activation of ERK
concentration is much higher thus predicted to be equipotent inhibitors of monomeric signalling by RAF inhibitors is the consequence of inhib-
than the minimum effective
concentration.
and dimeric RAF30 (FIG.2c). Because CRAF signals as an itor binding to RAF, which in turn promotes RAF bind-
obligatory dimer 50,65, COUT inhibitors fail to suppress ing to RAS-GTP and results in an increase in active RAF
Allosteric priming its activity. Thus, CIN RAF inhibitors, by suppressing dimers. Overall, RAF dimers are ineffectively inhibited
A phenomenon recently both BRAF and CRAF dimers can also be considered by RAF inhibitors due to negative allostery for binding
observed with several
pan-RAF inhibitors6668 and may be effective when used to the second protomer (FIG.2b). CIN RAF inhibitors
small-molecule inhibitors, by
which binding of the inhibitor in combination in contexts such as RAS-mutant tumour promote the RASRAF interaction at lower concentra-
induces the active cells, in which CRAF is known to be a major activator of tions and effectively suppress ERK signalling at mod-
conformation of the target ERK signalling downstream of RAS. However, a poten- erate and high concentrations, due to lower negative
kinase, which results in its tially narrow therapeutic window and allosteric priming (see allostery (FIG.2c). In contrast, COUT inhibitors pro-
increased interaction with
upstream regulators and
below) are predicted to limit the effectiveness of CIN mote the RASRAF interaction at moderate and high
consequent kinase priming and RAF inhibitor monotherapy in RAS-mutant tumours30. concentrations and fail to suppress ERK signalling at
activating phosphorylation. Although the model of negative allostery due to the these concentrations due to higher negative allostery 30.

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 5



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Table 1 | Select RAF inhibitors and their stage of development


RAF inhibitor Class Stage of development Chemical structure Clinical trial Ref.
SB590885 CIN/DFGIN Preclinical N

N
N H3C
HO N
H N
CH3
O

GDC0879 CIN/DFGIN Preclinical OH


N
HO N
N

Vemurafenib COUT/DFGIN Approved for F


BRAFV600E melanoma Cl O
O
N S
F H O
N N
H

Dabrafenib COUT/DFGIN Approved for NH2


N
BRAFV600E/K
F
melanoma N
O
S
N
S H O F
N F

LGX818 COUT/DFGIN PhaseIII NCT01909453 104,125


N and
N NCT02928224
F
N
O HN NH
S O
N
H C O
3 HN
Cl O CH3

AZ628 CIN/DFG-OUT Preclinical


H
O N

H3C N N
HN
N
O

LY3009120 CIN/DFG-OUT PhaseI F NCT02014116 112


O

N N N
H H
H3C
N N N
H

TAK632 CIN/DFG-OUT Preclinical N


H
S O N CF3
HN
N O
F
O

MLN2480 NA (presumably PhaseI NH2 NCT01425008, 108110


(TAK580) CIN/DFG-OUT) Cl NCT02723006
N CF3 and
N
O
N HN
NCT02327169
Cl
HN N
S O

6 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Table 1 (cont.) | Select RAF inhibitors and their stage of development


RAF inhibitor Class Stage of development Chemical structure Clinical trial Ref.
BGB659 NA (presumably Preclinical
N O
CIN/DFG-OUT)
HN NH2
NH
O
O
CF3

BGB283 CIN/DFG-OUT PhaseI NCT02610361 114


N O

N
HN
N CF3
O O H

CCT196969 NA (presumably Preclinical


CIN/DFG-OUT) O F
O
N
N
N N N N
H H
HN

CCT241161 NA (presumably Preclinical CH3


CIN/DFG-OUT) O S
O
N
N
N N N N
H H
HN

CCT3833 NA (presumably PhaseI NA NCT02437227 107


(BAL3833) CIN/DFG-OUT)
PLX7904 COUT/DFGIN/ Preclinical H
N N
R506OUT
F H O
N S CH3
N N
O
N O
F

PLX8394 COUT/DFGIN/ PhaseI/II H NCT02428712 116


N N
R506OUT
F H O
N N S
N
O
N O
F
F

HM95573 NA PhaseI NA NCT03118817 121,122


and
NCT02405065
CEP32496 NA PhaseI CF3 NCT01877811 124
O N
(also known as NH N CH3
RXDX105) NH O
O
N O
O CH3

LXH254 NA PhaseI NA NCT02607813 115


RAF709 CIN/DFG-OUT Preclinical O

N O
H
N
F3C N
O O
N

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 7



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Table 1 (cont.) | Select RAF inhibitors and their stage of development


RAF inhibitor Class Stage of development Chemical structure Clinical trial Ref.
BI 882370 COUT/ Preclinical N
DFG-OUT N

F
O H N
N N N
S
O N
F
CH3

NA, not available.

Finally, there are data indicating that the position be unequivocally determined53,76. Nonetheless, the basis
ofthe Chelix and the DFG motif affect the inter for the wide clinical therapeutic window of the current
action of RAF with MEK. COUT RAF inhibitors (for clinical COUT RAF inhibitors vemurafenib and dab-
example, vemurafenib or dabrafenib) tend to disrupt the rafenib is believed to be their relative inability to inhibit
RAFMEK complex. Among CIN RAF inhibitors, dimeric RAF in normal cells, while potently inhibiting
those stabilizing a conformation closer to the native mutant BRAF in tumourcells.
active kinase (CIN/DFGIN, for example, GDC0879 Recent work has revealed important differences in
or SB590885) allosterically promote the formation of the mechanism by which BRAF mutations other than
the RAFMEK complex and thus display more pro- BRAFV600E promote catalytic activation of BRAF
nounced paradoxical ERK signalling activation than do (FIG.3). First, similar to BRAFV600E, other BRAFV600
CIN/DFG-OUT inhibitors30,55. point mutant forms, such as V600K, V600D and V600R,
It should be emphasized that despite the major con- are also catalytically active monomers in the absence of
tributions by a number of groups to our understanding RAS-GTP and are sensitive to inhibition by current
of the complex mechanism of action of RAF inhibitors, clinical COUT RAF inhibitors30,31. By contrast, the cat-
more work is required to further elucidate the highly alytic activity of nonV600 BRAF point mutants depends
dynamic structural rearrangement that RAF proteins either partially or fully on dimerization30,31,51. Thus,
undergo upon inhibitor binding. Adding to this com- nonV600 BRAF mutants, including BRAFL597V,
plexity is the variation in the details of the mechanism of BRAFK601E and BRAFG466E, signal as RAS-GTP-
action, even among inhibitors of the same class, and for independent RAF dimers by spontaneously dimerizing,
example, as a result of the mutational status of the RAF and they are thus resistant to vemurafenib and other
proteins, the cell-specific stoichiometry of the various COUT inhibitors11,30,31,51. However, this has been
RAF interacting proteins and the levels of RAS activity. shown only in cell-based experiments so far 30,31,51. The
Nonetheless, the mechanism of action of RAF inhib- structural basis for the difference between RAF inhibitor
itors has important clinical implications. First, COUT activity upon V600 and nonV600 BRAF point mutants
RAF inhibitors are predicted to be effective selectively will presumably require biochemical and crystallo
against monomeric forms of RAF, but any mechanism graphic studies including the Nterminal domain of
of RAF dimerization will cause drug resistance, a notion BRAF or at least the part of the Nterminal domain that
that has been confirmed both in the laboratory and in regulates RAF dimerization. Nonetheless, as with other
the clinic30,31,62,75. Second, CIN RAF inhibitors, by forms of dimeric RAF, COUT RAF inhibitors are not
inhibiting both monomeric and dimeric RAF to a simi predicted to be effective in tumours with nonV600
lar extent, will overcome resistance due to RAF dimer- mutant BRAF. Equipotent inhibition of monomeric
ization. However, they are predicted to have a fairly and dimeric RAF with CIN RAF inhibitors may be a
narrow therapeutic window, given that they will sup- promising alternative for this class of tumours30,31.
press ERK signalling in normal, WT BRAF-expressing In addition to point mutations, BRAF (and more
cells at concentrations similar to those used to sup- rarely CRAF) is found as part of fusion transcripts and
press ERK signalling in tumour cells, as suggested by proteins caused by translocations77. Such fusions are
preclinicaldata30. commonly found in paediatric gliomas78 but also in other
solid tumours, including lung cancers, advanced prostate
Activation of mutant BRAF cancers, gastric cancers, pancreatic cancers and mela-
Linking BRAF mutational alterations to predicted RAF noma77,79. Typically, translation of the fusion transcript
inhibitor response. A critical breakthrough in elucidat- results in the expression of RAF proteins either as the
ing the mechanism of action of RAF inhibitors was the intact kinase domain only or as the kinase domain fused
discovery that, unlike WT BRAF that requires dimeri- with a partner protein that has replaced the Nterminal
zation for its catalytic activation, mutant BRAFV600E domain of RAF77. Deletion of the Nterminal domain of
is able to signal as an active monomer in the absence of RAF has been shown to promote RAF activation by con-
RAS-GTP30,31,51,52,56. These studies were carried out using stitutive dimerization44,52; thus, RAF fusion proteins are
ectopically expressed proteins and mutational analysis; predicted to be resistant to COUT RAF inhibitors. As
thus, the cellular status and the composition of endo CIN RAF inhibitors potently suppress dimeric RAF,
genous BRAFV600E containing complexes have yet to they may be a therapeutic option in this context.

8 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Inframe deletions in BRAF, such as L485 in tumours12. This shortening of the 3/Chelix loop
P490deleted BRAF and N486P490deleted BRAF, results in a unique conformation that locks the Chelix
have also been found in a small percentage of certain in the IN position through the formation of an intact
tumours, with the highest prevalence in pancreatic car- K483toE501 salt bridge. Two recent s tudies focused on
cinomas (4.21%), as well as in lung, ovarian and thyroid the mechanism of activation of inframe BRAF deletions
tumours80. BRAF inframe deletions do not overlap with and their response to RAF inhibitors12,80. Chen etal.80
KRAS or other BRAF mutations80, suggesting that they proposed that the increased activity of these genetic alter-
are sufficient to drive ERK signalling hyperactivation in ations is caused by increasedBRAF homodimerization.
these tumours. The deletions are mapped within the con- Therefore, BRAF proteins carrying inframe deletions
served3/Chelix loop of the kinase domain of BRAF, are maintained in the active state that favours BRAF
and they are predicted to activate the kinase by shorten- homodimerization and were shown to be resistant to
ing the 3/Chelix loop in a fashion analogous to that vemurafenib but sensitive to the next-generation CIN
of inframe deletions in epidermal growth factor receptor RAF dimer inhibitor LY3009120 (REF.80) (see below for
(EGFR) and HER2 (also known as ERBB2), also found more detail). By contrast, Foster etal.12 suggested that
the activity of BRAF deletions is independent of RAF
RAF alteration Eective RAF inhibitor
dimerization and that these inframe-deletion mutants
a can signal as monomers, similar to BRAFV600E.
BRAF wild-type RAS RAS Structural analysis revealed that the ChelixIN active
High RAS-GTP C-IN conformation of the kinase domain of BRAF harbouring
N N
N-lobe
BRAF Figure 3 | Categories of RAF alterations found in tumours
C-lobe WT WT
and the types of RAF inhibitor predicted to be effective
against them. a|In the cellular context of RAS-activated
b N wild-type (WT) BRAF, BRAF, CRAF and ARAF form active
BRAF-V600 homodimers and heterodimers that are resistant to COUT
Low RAS-GTP C-OUT
V600 RAF inhibitors. CIN RAF inhibitors are predicted to
potently inhibit RAF dimers, although their effectiveness will
c be limited due to concurrent formation of active RAF dimers
BRAF-V600 by the inhibitor. b|Mutant BRAFV600 proteins in the
RAS RAS RAS RAS C-IN C-OUT
High RAS-GTP absence of RAS-GTP signal as a monomer and are potently
N inhibited by either COUT or CIN inhibitors. In this
N N
context, COUT inhibitors are predicted to be more
V600 effective treatments due to their wide therapeutic window.
WT WT c|In the context of tumours expressing BRAFV600 in the
presence of active RAS, ERK is activated by both
d BRAFV600E monomers and RAF dimers (homodimers and
BRAF-V600 heterodimers of BRAFV600E, WT BRAF, CRAF and ARAF).
splice variant C-IN Acombination of a COUT inhibitor with a CIN RAF
V600 V600 inhibitor is predicted to result in effective RAF inhibition in
the tumour while retaining a wide therapeutic window.
e d|ABRAFV600E splice variant that lacks part of the
N N Nterminal domain and constitutively dimerizes has been
Non-V600 BRAF C-IN
RAS-independent identified in tumours that have developed clinical resistance
to COUT RAF inhibitors. In this context, ERK is activated by
Non-V600 BRAFV600E homodimers, and CIN RAF inhibitors may be
an effective therapeutic option. e|For tumours expressing
mutant BRAF proteins other than BRAFV600 that signal as
f RAS-independent dimers, CIN RAF inhibitors may be an
Non-V600 BRAF RAS RAS C-IN effective therapeutic option. f|For tumours expressing
RAS-dependent
N N
mutant BRAF proteins other than V600 that require RAS
activity, CIN RAF inhibitors may be an effective
therapeutic option. g|Fusion transcripts of various genes
Non-V600 WT with the catalytic domain of either WT BRAF or, in some
cases, CRAF, lack the Nterminal domain that prevents RAF
g dimerization, and their protein products thus signal as
RAS-independent RAF dimers. For patients whose tumours
RAF fusion C-IN
WT WT depend on such fusion proteins, CIN RAF inhibitors may be
an effective therapeutic option. h|Inframe deletions in
BRAF that stabilize the Chelix in the active (IN) position
h have been found in tumours. For patients with such tumours,
BRAF in-frame N N CIN RAF inhibitors may be an effective therapeutic option.
C-IN
deletion C-lobe, C-terminal lobe; N, Nterminal domain; N-lobe,
N-terminal lobe; Del, inframe deletion; CIN, ChelixIN
Del WT Del Del
RAF inhibitor; COUT, Chelix-OUT RAF inhibitor.

Nature Reviews | Cancer


NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 9

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

a deletion, renders binding of vemurafenib incompatible as a result of relief of negative feedback3739,89 or due to
due to steric effects12. Although the two studies differ on cytokine and growth factor release from the tumour
the proposed mechanism of BRAF activation by inframe microenvironment9092 (FIG.4).
deletions, they both conclude that tumours expressing In tumour types other than melanomas harbouring
BRAF deletions will be resistant to current clinical RAF BRAFV600E mutations, such as in colorectal and thy-
inhibitors and likely to respond to equipotent inhibitors roid cancers, the response rates to second-generation
of monomeric and dimeric RAF12,80. RAF inhibitors are low 93,94. In these tumours, increased
Finally, there are a class of point mutations found in upstream RTK expression and activity at baseline com-
either the activation segment or the glycine-rich loop pared to melanoma (mostly through EGFR in colorectal
of BRAF that are not activating but that instead result cancers and HER2HER3 (also known as ERBB3) sig-
in a protein that either is catalytically inactive or shows nalling in thyroid cancers) has been reported to account
severely impaired activity compared to WT BRAF10,11,57. for the unexpected unresponsiveness3739. Consequently,
These forms of mutant BRAF with impaired activity feedback reactivation of upstream RTKs and RAS upon
have been found mainly in colorectal, melanoma and RAF inhibitor treatment is more pronounced in these
lung tumours81, require RAS activity and dimerization tumours, profoundly enhancing RAF dimerization and
to hyperactivate ERK signalling and transform cells11,31,57 thus limiting the effect of COUT inhibitors30.
and are commonly coexpressed with activated receptor
tyrosine kinases (RTKs) or upstream alterations of RAF Development of RAF inhibitors
signalling, such as RAS mutations or neurofibromin 1 First-generation (CIN) RAF inhibitors. The first
(NF1) loss82,83. As these mutant BRAF proteins form small-molecule ATP-competitive RAF inhibitors were
dimers, tumours harbouring them are predicted to be reported before the discovery of BRAF mutations
insensitive to current clinical COUT RAF inhibitors, and were developed with the goal of targeting CRAF
but CIN RAF inhibitors may be an effectiveoption. in cancer 95. The compound ZM336372 (likely to be
CIN/DFGIN) was identified from a chemical screen
Mechanisms of resistance to RAF inhibitors. The notion as both a CRAF and BRAF inhibitor invitro and was
that second-generation RAF inhibitors do not effectively associated with the first report of the phenomenon of
suppress dimeric RAF was emphatically confirmed paradoxical ERK activation induced by a RAF inhibitor
with the discovery of increased RAF dimerization as a in cells95. Sorafenib (Nexavar, BAY439006; Bayer/Onyx
common cause of clinical acquired resistance to RAF Pharmaceuticals) (CIN/DFG-OUT) is the only
inhibitors30,31,52,75,84. Various molecular mechanisms first-generation RAF inhibitor that advanced to the
have been identified that confer this clinical resistance clinic96 (and received FDA approval for the treatment
to second-generation RAF inhibitors. RAS activa- of advanced renal cell carcinoma and hepatocellular
tion75, either by RAS mutation or by overexpression of carcinoma97,98). Sorafenib shows weak activity against
upstream RTKs75,85, promotes RAF dimerization as well BRAFV600E in cells, and its clinical activity in the afore-
as activation of parallel survival pathways, such as PI3K mentioned tumours with WT BRAF is likely due to its
AKT signalling. BRAFV600E amplification has been multi-kinase profile. Other first-generation RAF inhibi-
found in a subset of resistant melanoma tumours28,84, tors that did not advance to the clinic include SB590885
with the resulting overexpression of the BRAFV600E (REF.99) and GDC0879 (REF.100) (both CIN/DFGIN)
protein predicted to promote resistance to RAF inhibi- and GW5074 (REFS101,102) and L779450 (REF.87) (there
tors by both increasing abundance of the target (that is, are currently no crystal structures for GW5074 and
BRAFV600E) and spontaneous BRAFV600E dimeri- L779450, but it is likely they are CIN/DFGIN).
zation31. Another common mechanism of clinical resist-
ance due to increased BRAFV600E dimerization is the Second-generation, BRAFV600Eselective (COUT)
expression of splice variants of BRAFV600E that lack RAF inhibitors. The discovery of BRAF mutations
exons encoding part of the N terminus of the protein, in 2002 (REF. 9) renewed interest in developing RAF
specifically those including the RBD, which prevents inhibitors, leading to second-generation compounds,
dimerization in the context of full-length BRAF52,86. which were identified after screening for inhibitors of
Finally, mutational activation of MEK has also been BRAFV600E48. The first such RAF inhibitor to enter
reported in a subset of BRAFV600E tumours that clinical trials was vemurafenib (Zelboraf, PLX4032;
developed resistance to RAF inhibitors87,88. Plexxikon/Genentech), developed by a structure-guided
In addition to acquired resistance, increased RAF discovery approach23,48. Vemurafenib was approved
dimerization may confer an adaptive response to by the FDA in 2011 for the treatment of patients with
RAF inhibitor therapy. In BRAFV600Eexpressing BRAFV600E metastatic melanoma22,33, followed by
tumour cells, any increase in RAS activity due to dabrafenib (Tafinlar, GSK2118436; GlaxoSmithKline),
upstream activation will cause formation of RAF dimers, the second RAF inhibitor to receive approval by the
including homodimers and heterodimers of WT and FDA in 2013 for the treatment of melanoma patients
Steric effects BRAFV600E,CRAF and BRAFV600E and ARAF and with BRAFV600E/K mutations21,103. The two inhibi-
A phenomenon arising as a BRAFV600E, thus hindering the effect of RAF inhib- tors showed similar clinical efficacy and induced similar
result of the fact that each
residue and its atoms occupy
itors30,89. A number of studies have proposed mecha- ontarget effects related to paradoxical activation ofERK
acertain amount of space in a nisms by which tumours may adapt to RAF inhibitor signalling in normal cells. Approximately 1426% of
defined conformation. therapy via upstream RTK and RAS activation, either patients with melanoma treated with second-generation

10 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

a Negative feedback b Microenvironment

HGF
Stromal broblast
MET receptor

RAS RAS RAS PI3K P P RAS


GAB1 GRB2
AKT
RAF RAF RAF
BRAFV600E BRAFV600E inhibitor RAF BRAFV600E RAF BRAFV600E
inhibitor inhibitor

P P
P P
MEK MEK MEK MEK

P P
P P
ERK ERK ERK ERK

RAF dimer

C-OUT

C-IN

c Structural constraints

Figure 4 | The interplay of mechanisms of adaptive response to RAF inhibitors. a|Relief of negative feedback.
Nature At steady
Reviews | Cancer
state, hyperactivated ERK downstream of mutant BRAF suppresses RAS due to negative feedback. Addition of a RAF inhibitor
results in relief of negative feedback and RAS activation, which in turn promotes increased flux through ERK signalling and
RAF dimerization41,73,74. Activation indicated by a green halo around the protein. b|Pathway reactivation due to paracrine
signalling by components of the tumour microenvironment (TME). Stromal cells in the TME release growth factors
(exemplified here by hepatocyte growth factor (HGF)) and cytokines, which promote resistance to RAF inhibitors by
activating parallel survival pathways, such as PI3K signalling, or by activating RAS7577. c|Structural constraints. RAF
dimerization promotes an adaptive response to COUT RAF inhibitors due to negative allostery for inhibitor binding to the
second protomer of a RAF dimer when the first is occupied by inhibitor. CIN RAF inhibitors are predicted to overcome this
resistance mechanism by binding with similar affinity to both protomers of the RAF dimer41. All of these mechanisms of
adaptive response are interconnected and converge upon structural constraints for inhibitor binding. Relief of negative
feedback promotes upstream receptor tyrosine kinase activation, which may be further enhanced by growth factors and
cytokines released in the TME. Both mechanisms promote increased flux in the pathway and levels of active RAF molecules.
They also promote RAF dimerization, which prevents effective RAF inhibition by current clinical COUT RAF inhibitors due
to negative allostery. GAB1, growth factor receptor-bound protein 2associated binding protein 1; GRB2, growth factor
receptor-bound protein 2.

RAF inhibitors were diagnosed with the secondary compound is similar to other COUT inhibitors. Ithas
cancers keratoacanthomas and squamous-cell carcino- been reported to show longer residence time (lower off-
mas, which can be treated by simple excision74. These rate) compared to vemurafenib or dabrafenib31, aprop-
second-site cancers are commonly associated with erty that may positively affect its clinical profile by
expression of a RAS mutation and have been almost elim- prolonging target inhibition.
inated with the combination of a RAF and a MEK inhibi
tor, which is the current standard of care (see below). Third-generation RAF inhibitors. Third-generation
Other side effects, such as photosensitivity caused by RAF inhibitors, which are currently under preclini-
Residence time
The period for which the target vemurafenib22 and fever caused by dabrafenib21, appear cal and clinical development, fall structurally and bio-
kinase is occupied by the to be more drug-specific, presumably related to different chemically into two main classes (TABLE1). One group
small-molecule inhibitor. off-target effects of the twodrugs. includes compounds that equipotently inhibit the
LGX818 (Encorafenib, Array BioPharma) is another monomeric and dimeric forms of RAF and are thus
Off-rate
The rate of dissociation of the
RAF inhibitor selective for mutant BRAF that has shown expected to overcome resistance due to RAF dimeriza-
small-molecule inhibitor from promising clinical activity and is currently in phaseIII tion. Among them are TAK632 (REF.105) and MLN2480
the kinase. clinical trials104. Preclinically, the cellular profile of the (TAK580)106 (Takeda/Millennium), LY3009120 (REF.68)

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 11



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

(Eli Lilly) and CCT3833 (BAL3833) (REF.107)(ICR/Royal behave biochemically in a similar manner to CIN RAF
Marsden NHS Foundation Trust). Crystallographic data inhibitors. In preclinical studies, both CCT196969 and
obtained with TAK632 or LY3009120 bound to BRAF CCT241161 inhibited the growth of BRAFV600E mel-
show a type II conformation (CIN/DFG-OUT), anomas as well as RAS-mutant melanomas and colo
which explains the similar affinity for both protomers in rectal tumours66. CCT196969 and CCT241161 were also
the RAF dimer, as well as the generally slow off-rate of effective in patient-derived xenograft (PDX) models that
thesecompounds67,68,105. included melanomas with intrinsic or acquired resist-
In preclinical studies, TAK632 demonstrated potent ance to second-generation RAF and MEK inhibitors66.
antiproliferative activities in multiple cancer cell lines The study of the safety and tolerability of the lead com-
harbouring BRAF, NRAS or KRAS mutations withthe pound of this class of inhibitors, CCT3833 (BAL3833),
potential to delay the emergence of resistance 30,67. is ongoing in a phaseI clinical trial for patients with
The cocrystal structure of TAK632 bound to WT advanced solid tumours107. Melanoma patients har-
BRAF revealed compound occupancy and similar con- bouring BRAFV600E (either treatment-naive or after
formations of both protomers of the BRAF dimer 105. progression on RAF inhibitor therapy) or RAS muta-
TAK632 exhibited equipotent inhibition of monomeric tions are also included in the trial. Finally, LXH254 and
and dimeric RAF in cells30,67 and is expected to be effec- RAF709, both currently under development by Novartis
tive in vemurafenib-resistant melanomas harbouring are also equipotent inhibitors of monomeric and dimeric
BRAF or NRAS mutations and BRAFV600E colorectal RAF. An ongoing phaseI clinical trial is recruiting
or thyroid tumours67. MLN2480, which has character- patients with tumours with ERK signalling alterations
istics similar to TAK632, is currently in a phaseI clin- to be treated with LXH254 (REF.115).
ical trial for patients with melanomas and other solid The other group of third-generation RAF inhibitors
tumours108. MLN2480 will also be evaluated in a phaseI includes the paradox breakers, which are COUT RAF-
clinical trial in combination with the programmed cell inhibitor derivatives with variable terminal sulfonamide
death 1 (PD1) monoclonal antibody (nivolumab) for and sulfamide substitutions. Both PLX7904 and its ana-
patients with melanoma109. Finally, a phaseI clinical trial logue PLX8394 (Plexxikon) are more potent inhibitors of
of MLN2480 in combination with inhibitors targeting BRAFV600E than vemurafenib, and unlike vemurafenib,
other families of kinases (MLN0128, an mTOR com- they do not promote paradoxical ERK activation in WT
plex1 (mTORC1) and mTORC2 inhibitor, and alisertib, BRAF cells or squamous-cell carcinomas in preclini-
an Aurora kinaseA inhibitor), chemotherapy regimens cal models73. These compounds stabilize the Chelix
(paclitaxel and irinotecan) or the EGFR monoclonal of BRAF in the OUT position, and structural analysis
antibody (cetuximab) will be conducted for patients with showed that the position of the R506 residue of the C
advanced solidmalignancies110. helix with PLX7904 has the furthest OUT position com-
LY3009120 is another CIN/DFG-OUT RAF pared to that with non-paradox breakers30,73. Thus, the
inhibitor that showed antitumour activity in preclin- paradox breakers are predicted to have fewer ontarget
ical studies against NRAS or KRAS mutant tumours toxicities than second-generation RAF inhibitors and may
and vemurafenib-resistant melanomas68,111. LY3009120 be an effective option for chronic administration in more
binds to both RAF protomers within the BRAF dimer, indolent neoplasias, such as in LCH patients with tumours
stabilizing them in similar conformations, and, like expressing BRAFV600E19. PLX8394 is in a phaseI/II
other CIN RAF inhibitors, it inhibits monomeric and clinical trial for the treatment of patients with advanced
dimeric BRAF with similar potency 30,41,93. LY3009120 unresectable solid tumours116. Preclinical s tudies using
was also effective against BRAF deletions, including paradox breakers have also suggested that they will be
BRAFN486P490 and other inframe BRAF deletions effective against vemurafenib-resistant tumours and
that have been identified in pancreatic and thyroid BRAF fusions that are common in paediatric astrocyto-
tumours80. Currently, LY3009120 is in a phaseI clinical mas, presumably due to their higher potency against RAF
trial for patients with advanced cancers112. when compared with vemurafenib117119. However, as they
A compound from BeiGene, BGB659, showed are COUT inhibitors, RAF dimerization is predicted to
properties of equipotent inhibitors of monomeric and limit the effectiveness of these compounds.
dimeric RAF31. In preclinical studies, BGB659 was effec- Finally, BI882370 is an COUT RAF inhibitor
tive in melanomas with V600E and nonV600 BRAF under development by Boehringer Ingelheim that
mutations, in vemurafenib-resistant BRAFV600E showed promising preclinical activity against BRAF-
melanomas, as well as against ectopically expressed mutant melanoma and colorectal tumours, both in cell
BRAF fusions31. A related compound, BGB283, acts line models and invivo120. This compound differs from
as a dual RAF and EGFR inhibitor 113. In preclinical other COUT RAF inhibitors by stabilizing a con-
studies, it suppressed the proliferation of cells harbour- formation in which it interacts with the Phe595 from
ing BRAFV600E or EGFR mutations and inhibited the DFG motif and keeps the DFG motif in an inactive
EGFR reactivation in BRAFV600E colorectal tumours. (OUT) position120. However, further studies are required
AphaseI clinical trial of BGB283 for patients with solid to determine potential functional consequences of
tumours is ongoing 114. thesedifferences.
CCT196969 and CCT241161 have been reported as Currently in clinical trials are additional RAF inhibitors
dual pan-RAF and SRC kinase inhibitors66. Although that cannot yet be classified due to a lack of structural data,
crystallographic data are not available, these compounds including HM95573 (Hanmi Pharmaceutical/Genentech)

12 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

and CEP32496 (RXDX105, Ambit Biosciences/Ignyta). in patients treated with vemurafenib alone) and was
HM95573 is currently in an expansion clinical trial for also approved by the FDA in 2015. Given their superior
solid tumours harbouring mutations in either BRAF, efficacy over RAF inhibitor monotherapies, these RAF
KRAS or NRAS genes121,122. CEP32496 (REF.123) is a quina- MEK inhibitor combinations are currently the standard
zoline BRAF inhibitor that has demonstrated multikinase of practice for the treatment of melanoma patients with
activity against BRAFV600E, BCR-ABL, RET and ephrin BRAFV600E/K mutations.
type A receptor 2 (EPHA2) and is reported to be effec- Clinical trials of drug combinations including the
tive in preclinical studies of melanomas and colorectal second-generation RAF inhibitor LGX818 are cur-
tumours harbouring BRAFV600E123. CEP32496 will be rently underway: one with the MEK inhibitor MEK162
evaluated in a phaseI clinical trial for the treatment of (binimetinib, Array BioPharma) in patients with
patients with solid tumours124. BRAFV600E melanomas (COLUMBUS)104, and another
Based on preclinical data, clinical efficacy of next- in combination with MEK162 and the EGFR antibody
generation RAF inhibitors may also be limited by either cetuximab in patients with colorectal BRAFV600E
ontarget toxicities or development of drug resistance. tumours (BEACON CRC) 125. Preliminary results
Third-generation COUT RAF inhibitors (such as the from the COLUMBUS trial presented on the Array
paradox breakers) are expected to retain the wide ther- BioPharma website indicate significant responses, includ-
apeutic window of second-generation COUT RAF ing increased median PFS and improvement in overall
inhibitors, without the ontarget toxicities of second-site response rate compared with vemurafenib monotherapy.
tumours due to paradoxical ERK activation in normal Despite the improvement of clinical outcome,
cells. However, RAF dimerization is predicted to limit responses to RAFMEK inhibitor combinations are vari
the effectiveness of these inhibitors. In contrast, CIN able, and resistance develops for the majority of patients
RAFinhibitors are predicted to overcome resistance due treated with these regimens. Most mechanisms of resist-
to RAF dimerization, but resistance to these compounds ance to RAFMEK inhibitor combination are similar to
is anticipated to be conferred by other mechanisms, such the ones identified with RAF inhibitor monotherapy and
as other mutations in BRAF or alterations that would are associated with recovery of ERK signalling in the
enable cell proliferation and survival in the absence of presence of inhibitors27,126. This suggests that strategies
ERK signalling. However, allosteric priming and a fairly that would more potently and durably suppress RAF
narrow therapeutic window due to inhibition of ERK ERK signalling in the tumour, while retaining a wide
signalling in normal cells at lower drug concentrations is therapeutic window, may yield even better responses
expected to limit the effectiveness of CIN RAF inhibi- and further forestall drug resistance. The wide thera-
tors. Therapeutic strategies combining structurally com- peutic window of COUT RAF inhibitors suggests that
plementary RAF inhibitors with inhibitors of downstream they will most likely remain components of any future
components of ERK signalling (MEK and ERK inhibitors) combinatorial strategy for BRAFV600E tumours.
may enable sufficient inhibition of ERK signalling in the A number of next-generation CIN RAF inhibi
tumour and thus provide durable responses and perhaps tors are currently in clinical trials (TABLE1). Such com-
even cures for patients with BRAF-mutant tumours. poundsare predicted to suppress dimeric RAF and thus
be more effective in a broad range of RAF-dependent
Combinatorial therapeutic strategies tumours, including tumours expressing nonV600 BRAF
As in both preclinical models and in patients, potency mutations, or perhaps RAS-mutant tumours. However,
of ERK inhibition in the tumour was found to correlate the narrow therapeutic window of these inhibitors may be
with response to RAF inhibitor treatment 23, RAF inhib- a critical factor limiting their efficacy as single agents. As
itors were combined with MEK inhibitors in melanoma aforementioned, adding a modest dose of an CIN RAF
patients with BRAFV600E tumours in order to achieve inhibitor to the current standard COUTMEK inhibi-
more potent and durable inhibition of ERK signalling tor combination may be an effective strategy for patients
in the tumour 24,25. RAFMEK inhibitor combinations with BRAFV600E tumours30. Preclinical work supports
demonstrated increased progression-free survival (PFS) the idea of including CIN RAF inhibitors in combi-
and objective responses compared to RAF inhibitor nation with COUT RAF inhibitors (and also MEK or
monotherapy 24,25. Moreover, the cutaneous toxicities ERK inhibitors) to overcome feedback recovery and thus
induced by RAF inhibitors were dramatically decreased achieve durable suppression of ERK signalling 30,127.
in the combination, further confirming that they were RAF inhibitors are also included in a number of trials
the result of paradoxical activation of ERK signalling in that are undertaking a combinatorial approach to target
normal tissue. The dabrafenib and trametinib combina- parallel pathways (for example, RTKs, cyclin-dependent
tion24 (Novartis) increased the response rate to 76% and kinase 4 (CDK4) and CDK6 and PI3K) in addition to
the median duration of response to 10.5months as com- ERK signalling 128131. Although in principle, these strat-
pared with a 54% response rate and 5.8 months duration egies are usually based on strong preclinical rationale,
of response with dabrafenib alone, and was approved additive ontarget and off-target toxicities tend to result
by the FDA in 2014. The combination of vemurafenib in lower dosing and thus limit the efficacy of the individ-
and cobimetinib25 (Roche/Genentech) demonstrated ual compounds in the clinic. Again, it is critical that the
similar responses (response rate 68% and duration of need for a wide therapeutic window be taken into account
response 9.9months as compared with response rate in the design of clinical trials assessing c ombinations of
of45% and median duration of response of 6.2 months targetedagents.

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 13



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Cross-resistance Finally, a great deal of interest is being directed towards has not yet been determined. Nevertheless, peptides
The development of tumour elucidating the combinatorial interaction of targeted generated from the dimerization interface of BRAF
resistance to a potential therapy with immunotherapy, especially in melanoma have been shown to prevent formation of homodimers
therapy after treatment of a treatment. Cytotoxic T lymphocyte-associated antigen4 and heterodimers of BRAFCRAF in a BRAF-mutant
patient with a different
therapeutic agent.
(CTLA4) and PD1 inhibitors have shown remarkable or RAS-mutant cellular context 51, suggesting the pos-
clinical efficacy and prolonged survival of patients with sibility of an alternative strategy for developing inhibi
melanoma132134 and are now FDA-approved for the treat- tors of RAF dimerization. Lastly, we mentioned that
ment of metastatic melanoma135. Trials with concomitant several BRAF mutations other than V600 can spontan
administration of immune checkpoint inhibitors and RAF eously induce BRAF dimerization through the kinase
and MEK inhibitors have so far been discontinued due domain30,31,51. Structures of such BRAF mutants favour-
to unacceptable adverse effects, although it is not incon- ing dimer formation are not available and should pro-
ceivable that certain combinations and scheduling may be vide further understanding about their mechanism of
better tolerated than others. If simultaneous treatment is activation, as well as serve as templates for tailored drug
not an option, whether immunotherapy or targeted ther- design to these mutations.
apy should be the first-line treatment for patients with Importantly, the structural basis of the interaction
BRAFV600E melanomas remains an open question, as of the Nterminal regulatory domain and the kinase
there is reported evidence for cross-resistance136. In current domain in the inactive conformation and the active
clinical practice, targeted therapy is the preferred first-line conformation remains elusive. Such information
therapy for rapidly growing tumours because it tends to would enable our understanding of the architectural
yield faster remissions than immunotherapy, but a num- organization of BRAF and the conformational transi-
ber of ongoing clinical trials are aimed at exploring this tions involved in the context of the full-length BRAF
question more broadly 137. from its inactive monomer conformation to the active
dimer and activemonomer, particularly in the case
Challenges and future perspectives of the BRAFV600 mutant. The effects of the current
Since the validation of BRAFV600E as a drug target, as inhibitors upon these inter-domain interactions in the
well as outcomes from sequencing projects highlight- active and inactive BRAF conformations are not well
ing the role of BRAFV600E and other BRAF mutations understood. Given the conformational changes in the
in several cancers, drug discovery and development of kinase domain and dimerization process observed with
BRAF inhibitors has progressed at a fast pace. However, the current BRAF inhibitors, it is likely that these same
there is a pressing need for more effective RAF signal- inhibitors will have an impact on inter-domain inter-
ling inhibition that would overcome resistance due actions. Such information will enable further improve-
to mechanisms of BRAF dimerization. As our under- ment of the current inhibitors but will also provide
standing of the structural and biochemical regulation unique opportunities for BRAF allosteric inhibitors
of BRAF progresses, novel and more specific targeting outside of the catalyticpocket.
strategies should be developed. Future efforts should The oncoprotein BRAF is a validated therapeutic
explore the possibility for improving current ATP- target in a large number of human tumours. Current
competitive COUT inhibitors to prevent negative clinical RAF inhibitors have improved survival of
allostery and bypass paradoxical activation. In addition, patients with melanoma whose tumours harbour BRAF
third-generation CIN inhibitors are more effective mutations, but resistance limits their effectiveness in
inhibitors of RAF dimers than are COUT inhibi- these patients, as well as in patients with BRAFV600
tors, although their selectivity for mutant or WT BRAF tumours other than melanomas, or harbouring BRAF
dimers versus BRAF monomers is not optimal. Finally, mutations other than V600. Our deeper understanding
one could envision the development of allosteric inhib- of the molecular and biochemical mechanisms of resist-
itors that effectively inhibit the mutant or WT BRAF ance to current clinical RAF inhibitors, combinedwith
dimer and/or BRAF monomer but that do not promote the development of next-generation RAF inhibitors
the RAF p riming that leads to paradoxical activation. withdifferent structural and biochemical properties, will
Key biochemical data suggest the importance of the enable the design of innovative, potentially more effec-
BRAFCRAF heterodimer in RAF oncogenic signal- tive, RAF-inhibitor-based therapeutic strategies tailored
ling 56,57,65,138, although the structure of this heterodimer to the clinical context.

1. Lavoie,H. & Therrien,M. Regulation of RAF protein 6. Young,A. etal. Ras signaling and therapies. biochemical insight into the mechanism of BRAF
kinases in ERK signalling. Nat. Rev. Mol. Cell Biol. 16, Adv.Cancer Res. 102, 117 (2009). activation by mutations.
281298 (2015). 7. Baines,A.T., Xu,D. & Der,C.J. Inhibition of Ras for 11. Garnett,M.J., Rana,S., Paterson,H., Barford,D. &
2. Dhillon,A.S., Hagan,S., Rath,O. & Kolch,W. MAP cancer treatment: the search continues. Future Med. Marais,R. Wild-type and mutant BRAF activate CRAF
kinase signalling pathways in cancer. Oncogene 26, Chem. 3, 17871808 (2011). through distinct mechanisms involving
32793290 (2007). 8. Caunt,C.J., Sale,M.J., Smith,P.D. & Cook,S.J. MEK1 heterodimerization. Mol. Cell 20, 963969 (2005).
3. Wellbrock,C., Karasarides,M. & Marais,R. The RAF and MEK2 inhibitors and cancer therapy: the long and 12. Foster,S.A. etal. Activation mechanism of oncogenic
proteins take centre stage. Nat. Rev. Mol. Cell Biol. 5, winding road. Nat. Rev. Cancer 15, 577592 (2015). deletion mutations in BRAF, EGFR, and HER2.
875885 (2004). 9. Davies,H. etal. Mutations of the BRAF gene in human CancerCell 29, 477493 (2016).
4. Matallanas,D. etal. Raf family kinases: old dogs have cancer. Nature 417, 949954 (2002). 13. Holderfield,M., Deuker,M.M., McCormick,F. &
learned new tricks. Genes Cancer 2, 232260 (2011). 10. Wan,P.T. etal. Mechanism of activation of the RAF- McMahon,M. Targeting RAF kinases for cancer
5. Roberts,P.J. & Der,C.J. Targeting the Raf-MEK-ERK ERK signaling pathway by oncogenic mutations of therapy: BRAF-mutated melanoma and beyond.
mitogen-activated protein kinase cascade for the BRAF. Cell 116, 855867 (2004). Nat.Rev. Cancer 14, 455467 (2014).
treatment of cancer. Oncogene 26, 32913310 This paper provides the first crystal structure of 14. Cancer Genome Atlas Network. Genomic classification
(2007). BRAFV600E bound to inhibitor, along with of cutaneous melanoma. Cell 161, 16811696 (2015).

14 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

15. Kimura,E.T. etal. High prevalence of BRAF mutations mutant colorectal cancers to RAF inhibition with 58. Holderfield,M. etal. RAF inhibitors activate the
in thyroid cancer: genetic evidence for constitutive vemurafenib. Cancer Discov. 2, 227235 (2012). MAPK pathway by relieving inhibitory
activation of the RET/PTC-RAS-BRAF signaling This paper shows that insensitivity of colorectal autophosphorylation. Cancer Cell 23, 594602
pathway in papillary thyroid carcinoma. Cancer Res. BRAFV600E tumours to RAF inhibitors is caused by (2013).
63, 14541457 (2003). relief of negative feedback and upstream activation 59. Halaban,R. etal. PLX4032, a selective BRAF(V600E)
16. Rajagopalan,H. etal. Tumorigenesis: RAF/RAS of RTK signalling, most commonly through EGFR. kinase inhibitor, activates the ERK pathway and
oncogenes and mismatch-repair status. Nature 418, 39. Montero-Conde,C. etal. Relief of feedback inhibition enhances cell migration and proliferation of BRAF
934 (2002). of HER3 transcription by RAF and MEK inhibitors melanoma cells. Pigment Cell Melanoma Res. 23,
17. Cancer Genome Atlas Research Network. attenuates their antitumor effects in BRAF-mutant 190200 (2010).
Comprehensive molecular profiling of lung thyroid carcinomas. Cancer Discov. 3, 520533 60. Joseph,E.W. etal. The RAF inhibitor PLX4032
adenocarcinoma. Nature 511, 543550 (2014). (2013). inhibits ERK signaling and tumor cell proliferation in a
18. Tiacci,E. etal. BRAF mutations in hairy-cell leukemia. This paper shows that insensitivity of thyroid V600E BRAF-selective manner. Proc. Natl Acad. Sci.
N.Engl. J.Med. 364, 23052315 (2011). BRAFV600E tumours to RAF inhibitors is caused USA 107, 1490314908 (2010).
19. Badalian-Very,G. etal. Recurrent BRAF mutations in by relief of negative feedback and upstream 61. Holderfield,M., Nagel,T.E. & Stuart,D.D.
Langerhans cell histiocytosis. Blood 116, 19191923 activation of RTK signalling, most commonly Mechanism and consequences of RAF kinase
(2010). through HER3. activation by small-molecule inhibitors. Br. J.Cancer
20. Solit,D.B. etal. BRAF mutation predicts sensitivity 40. Warne,P.H., Viciana,P.R. & Downward,J. Direct 111, 640645 (2014).
toMEK inhibition. Nature 439, 358362 (2006). interaction of Ras and the amino-terminal region 62. Poulikakos,P.I. & Rosen,N. Mutant BRAF melanomas
This paper provides preclinical evidence supporting ofRaf1 invitro. Nature 364, 352355 (1993). dependence and resistance. Cancer Cell 19, 1115
the targeting of BRAFERK signalling in tumours 41. Cutler,R.E.Jr, Stephens,R.M., Saracino,M.R. & (2011).
harbouring the mutation BRAFV600E. Morrison,D.K. Autoregulation of the Raf1 serine/ 63. Samatar,A.A. & Poulikakos,P.I. Targeting RAS-ERK
21. Hauschild,A. etal. Dabrafenib in BRAF-mutated threonine kinase. Proc. Natl Acad. Sci. USA 95, signalling in cancer: promises and challenges.
metastatic melanoma: a multicentre, open-label, 92149219 (1998). Nat.Rev. Drug Discov. 13, 928942 (2014).
phase 3 randomised controlled trial. Lancet 380, 42. Vojtek,A.B., Hollenberg,S.M. & Cooper,J.A. 64. Zhang,J., Yang,P.L. & Gray,N.S. Targeting cancer
358365 (2012). Mammalian Ras interacts directly with the serine/ with small molecule kinase inhibitors. Nat. Rev. Cancer
22. Chapman,P.B. etal. Improved survival with threonine kinase Raf. Cell 74, 205214 (1993). 9, 2839 (2009).
vemurafenib in melanoma with BRAF V600E 43. Moodie,S.A., Willumsen,B.M., Weber,M.J. & 65. Hu,J. etal. Allosteric activation of functionally
mutation. N.Engl. J.Med. 364, 25072516 (2011). Wolfman,A. Complexes of Ras. GTP with Raf1 and asymmetric RAF kinase dimers. Cell 154, 10361046
23. Bollag,G. etal. Clinical efficacy of a RAF inhibitor mitogen-activated protein kinase kinase. Science 260, (2013).
needs broad target blockade in BRAF-mutant 16581661 (1993). 66. Girotti,M.R. etal. Paradox-breaking RAF inhibitors
melanoma. Nature 467, 596599 (2010). 44. Weber,C.K., Slupsky,J.R., Kalmes,H.A. & that also target SRC are effective in drug-resistant
This paper reports on the discovery and early Rapp,U.R. Active Ras induces heterodimerization of BRAF mutant melanoma. Cancer Cell 27, 8596
clinical results of the first FDA-approved RAF cRaf and BRaf. Cancer Res. 61, 35953598 (2001). (2015).
inhibitor, vemurafenib. 45. Avruch,J. etal. Ras activation of the Raf kinase: 67. Nakamura,A. etal. Antitumor activity of the selective
24. Long,G.V. etal. Combined BRAF and MEK inhibition tyrosine kinase recruitment of the MAP kinase cascade. pan-RAF inhibitor TAK632 in BRAF inhibitor-resistant
versus BRAF inhibition alone in melanoma. N.Engl. Recent Prog. Horm. Res. 56, 127155 (2001). melanoma. Cancer Res. 73, 70437055 (2013).
J.Med. 371, 18771888 (2014). 46. Kornev,A.P. & Taylor,S.S. Dynamics-driven allostery 68. Peng,S.B. etal. Inhibition of RAF isoforms and active
25. Larkin,J. etal. Combined vemurafenib and in protein kinases. Trends Biochem. Sci. 40, 628647 dimers by LY3009120 leads to anti-tumor activities in
cobimetinib in BRAF-mutated melanoma. N.Engl. (2015). RAS or BRAF mutant cancers. Cancer Cell 28,
J.Med. 371, 18671876 (2014). 47. Hatzivassiliou,G. etal. RAF inhibitors prime wild-type 384398 (2015).
26. Carlino,M.S. etal. Antiproliferative effects of RAF to activate the MAPK pathway and enhance 69. Okuzumi,T. etal. Inhibitor hijacking of Akt activation.
continued mitogen-activated protein kinase pathway growth. Nature 464, 431435 (2010). Nat. Chem. Biol. 5, 484493 (2009).
inhibition following acquired resistance to BRAF This study provides insight into the unique 70. Cameron,A.J., Escribano,C., Saurin,A.T.,
and/or MEK inhibition in melanoma. Mol. Cancer Ther. biochemical mechanism of RAF inhibitor action. Kostelecky,B. & Parker,P.J. PKC maturation is
12, 13321342 (2013). 48. Tsai,J. etal. Discovery of a selective inhibitor of promoted by nucleotide pocket occupation
27. Moriceau,G. etal. Tunable-combinatorial mechanisms oncogenic BRaf kinase with potent antimelanoma independently of intrinsic kinase activity. Nat. Struct.
of acquired resistance limit the efficacy of BRAF/MEK activity. Proc. Natl Acad. Sci. USA 105, 30413046 Mol. Biol. 16, 624630 (2009).
cotargeting but result in melanoma drug addiction. (2008). 71. Pearce,L.R. etal. Characterization of PF4708671,
Cancer Cell 27, 240256 (2015). 49. Lavoie,H. etal. Inhibitors that stabilize a closed RAF anovel and highly specific inhibitor of p70 ribosomal
28. Villanueva,J. etal. Concurrent MEK2 mutation and kinase domain conformation induce dimerization. S6 kinase (S6K1). Biochem. J. 431, 245255 (2010).
BRAF amplification confer resistance to BRAF and Nat.Chem. Biol. 9, 428436 (2013). 72. Andraos,R. etal. Modulation of activation-loop
MEK inhibitors in melanoma. Cell Rep. 4, 10901099 50. Rajakulendran,T., Sahmi,M., Lefrancois,M., Sicheri,F. phosphorylation by JAK inhibitors is binding mode
(2013). & Therrien,M. A dimerization-dependent mechanism dependent. Cancer Discov. 2, 512523 (2012).
29. Yang,H. etal. RG7204 (PLX4032), a selective drives RAF catalytic activation. Nature 461, 542545 73. Zhang,C. etal. RAF inhibitors that evade paradoxical
BRAFV600E inhibitor, displays potent antitumor (2009). MAPK pathway activation. Nature 526, 583586
activity in preclinical melanoma models. Cancer Res. This paper provides mechanistic insight into the (2015).
70, 55185527 (2010). mechanism of RAF activation by dimerization. 74. Su,F. etal. RAS mutations in cutaneous squamous-cell
30. Karoulia,Z. etal. An integrated model of RAF 51. Freeman,A.K., Ritt,D.A. & Morrison,D.K. Effects carcinomas in patients treated with BRAF inhibitors.
inhibitor action predicts inhibitor activity against ofRaf dimerization and its inhibition on normal and N.Engl. J.Med. 366, 207215 (2012).
oncogenic BRAF signaling. Cancer Cell 30, 485498 disease-associated Raf signaling. Mol. Cell 49, 75. Nazarian,R. etal. Melanomas acquire resistance
(2016). 751758 (2013). toBRAF(V600E) inhibition by RTK or NRAS
This study provides a unified mechanistic model of 52. Poulikakos,P.I. etal. RAF inhibitor resistance is upregulation. Nature 468, 973977 (2010).
RAF inhibitor action and a roadmap for the clinical mediated by dimerization of aberrantly spliced This paper provides evidence identifying RAS
use of next-generation RAF inhibitors based on BRAF(V600E). Nature 480, 387390 (2011). mutation and RTK upregulation as major
their structural properties. This paper provides information identifying the mechanisms of acquired clinical resistance of
31. Yao,Z. etal. BRAF mutants evade ERK-dependent expression of splice variants of BRAFV600E, BRAFV600E melanomas to RAF inhibitors.
feedback by different mechanisms that determine whichcause clinical resistance to RAF inhibitors by 76. Diedrich,B. etal. Discrete cytosolic macromolecular
their sensitivity to pharmacologic inhibition. enhanced homodimerization, and confirming that BRAF complexes exhibit distinct activities and
CancerCell 28, 370383 (2015). RAF dimerization is a common mechanism of composition. EMBO J. 36, 646663 (2017).
32. Gentilcore,G. etal. Effect of dabrafenib on melanoma resistance to RAF inhibitors. 77. Ross,J.S. etal. The distribution of BRAF gene fusions
cell lines harbouring the BRAF(V600D/R) mutations. 53. Roring,M. etal. Distinct requirement for an intact in solid tumors and response to targeted therapy.
BMC Cancer 13, 17 (2013). dimer interface in wild-type, V600E and kinase-dead Int.J.Cancer 138, 881890 (2016).
33. Flaherty,K.T. etal. Inhibition of mutated, activated BRaf signalling. EMBO J. 31, 26292647 (2012). 78. Jones,D.T. etal. Tandem duplication producing a novel
BRAF in metastatic melanoma. N.Engl. J.Med. 363, 54. Thevakumaran,N. etal. Crystal structure of a BRAF oncogenic BRAF fusion gene defines the majority of
809819 (2010). kinase domain monomer explains basis for allosteric pilocytic astrocytomas. Cancer Res. 68, 86738677
34. McArthur,G.A. etal. Safety and efficacy of regulation. Nat. Struct. Mol. Biol. 22, 3743 (2015). (2008).
vemurafenib in BRAF(V600E) and BRAF(V600K) 55. Haling,J.R. etal. Structure of the BRAF-MEK complex 79. Palanisamy,N. etal. Rearrangements of the RAF
mutation-positive melanoma (BRIM3): extended reveals a kinase activity independent role for BRAF in kinase pathway in prostate cancer, gastric cancer and
followup of a phase 3, randomised, open-label study. MAPK signaling. Cancer Cell 26, 402413 (2014). melanoma. Nat. Med. 16, 793798 (2010).
Lancet Oncol. 15, 323332 (2014). 56. Poulikakos,P.I., Zhang,C., Bollag,G., Shokat,K.M. & 80. Chen,S.H. etal. Oncogenic BRAF deletions that
35. Klein,O. etal. BRAF inhibitor activity in V600R Rosen,N. RAF inhibitors transactivate RAF dimers and function as homodimers and are sensitive to inhibition
metastatic melanoma. Eur. J.Cancer 49, 10731079 ERK signalling in cells with wild-type BRAF. Nature by RAF dimer inhibitor LY3009120. Cancer Discov. 6,
(2013). 464, 427430 (2010). 300315 (2016).
36. Dahlman,K.B. etal. BRAF(L597) mutations in This paper provides insight into the unique 81. Zheng,G. etal. Clinical detection and categorization
melanoma are associated with sensitivity to MEK biochemical mechanism of RAF inhibitor action. ofuncommon and concomitant mutations involving
inhibitors. Cancer Discov. 2, 791797 (2012). 57. Heidorn,S.J. etal. Kinase-dead BRAF and oncogenic BRAF. BMC Cancer 15, 779 (2015).
37. Prahallad,A. etal. Unresponsiveness of colon cancer RAS cooperate to drive tumor progression through 82. Nieto,P. etal. A Braf kinase-inactive mutant induces
to BRAF(V600E) inhibition through feedback CRAF. Cell 140, 209221 (2010). lung adenocarcinoma. Nature 548, 239243 (2017).
activation of EGFR. Nature 483, 100103 (2012). This paper shows that mutant BRAF with impaired 83. Yao,Z. etal. Tumours with class 3 BRAF mutants are
38. Corcoran,R.B. etal. EGFR-mediated reactivation of kinase activity can cooperate with active RAS to sensitive to the inhibition of activated RAS. Nature
MAPK signaling contributes to insensitivity of BRAF promote tumour progression. 548, 234238 (2017).

NATURE REVIEWS | CANCER ADVANCE ONLINE PUBLICATION | 15



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

84. Shi,H. etal. Melanoma whole-exome sequencing 104. US National Library of Medicine. ClinicalTrials.gov 124. US National Library of Medicine. ClinicalTrials.gov
identifies (V600E)BRAF amplification-mediated https://clinicaltrials.gov/ct2/show/NCT01909453 https://clinicaltrials.gov/ct2/show/NCT01877811
acquired BRAF inhibitor resistance. Nat. Commun. 3, (2017). (2017).
724 (2012). 105. Okaniwa,M. etal. Discovery of a selective kinase 125. US National Library of Medicine. ClinicalTrials.gov
85. Villanueva,J. etal. Acquired resistance to BRAF inhibitor (TAK632) targeting pan-RAF inhibition: https://clinicaltrials.gov/ct2/show/NCT02928224
inhibitors mediated by a RAF kinase switch in design, synthesis, and biological evaluation of (2017).
melanoma can be overcome by cotargeting MEK C-7substituted 1,3benzothiazole derivatives. 126. Long,G.V. etal. Increased MAPK reactivation in early
andIGF1R/PI3K. Cancer Cell 18, 683695 (2010). J.Med.Chem. 56, 64786494 (2013). resistance to dabrafenib/trametinib combination
86. Shi,H. etal. Acquired resistance and clonal evolution 106. Sun,Y. etal. A brain-penetrant RAF dimer antagonist therapy of BRAF-mutant metastatic melanoma.
in melanoma during BRAF inhibitor therapy. for the noncanonical BRAF oncoprotein of pediatric Nat.Commun. 5, 5694 (2014).
CancerDiscov. 4, 8093 (2014). low-grade astrocytomas. Neuro Oncol. 19, 774785 127. Carlino,M.S. etal. Differential activity of MEK and
87. Emery,C.M. etal. BRAF-inhibitor associated (2017). ERK inhibitors in BRAF inhibitor resistant melanoma.
MEKmutations increase RAF-dependent and- 107. US National Library of Medicine. ClinicalTrials.gov Mol. Oncol. 8, 544554 (2014).
independent enzymatic activity. Mol. Cancer Res. https://clinicaltrials.gov/ct2/show/NCT02437227 128. US National Library of Medicine. ClinicalTrials.gov
http://dx.doi.org/10.1158/1541-7786.MCR-17-0211 (2017). https://clinicaltrials.gov/ct2/show/NCT02872259
(2017). 108. US National Library of Medicine. ClinicalTrials.gov (2017).
88. Wagle,N. etal. Dissecting therapeutic resistance https://clinicaltrials.gov/ct2/show/NCT01425008 129. US National Library of Medicine. ClinicalTrials.gov
toRAF inhibition in melanoma by tumor genomic (2016). https://clinicaltrials.gov/ct2/show/NCT01719380
profiling. J.Clin. Oncol. 29, 30853096 (2011). 109. US National Library of Medicine. ClinicalTrials.gov (2017).
89. Lito,P. etal. Relief of profound feedback inhibition of https://clinicaltrials.gov/ct2/show/NCT02723006 130. US National Library of Medicine. ClinicalTrials.gov
mitogenic signaling by RAF inhibitors attenuates their (2017). https://clinicaltrials.gov/ct2/show/NCT01543698
activity in BRAFV600E melanomas. Cancer Cell 22, 110. US National Library of Medicine. ClinicalTrials.gov (2017).
668682 (2012). https://clinicaltrials.gov/ct2/show/NCT02327169 131. US National Library of Medicine. ClinicalTrials.gov
90. Obenauf,A.C. etal. Therapy-induced tumour (2017). https://clinicaltrials.gov/ct2/show/NCT02159066
secretomes promote resistance and tumour 111. Henry,J.R. etal. Discovery of 1-(3,3dimethylbutyl)- (2017).
progression. Nature 520, 368372 (2015). 3-(2fluoro4methyl5-(7methyl2-(methylamino) 132. Topalian,S.L. etal. Safety, activity, and immune
91. Straussman,R. etal. Tumour micro-environment elicits pyrido[2,3- d]pyrimidin6yl)phenyl)urea correlates of antiPD1 antibody in cancer.
innate resistance to RAF inhibitors through HGF (LY3009120)as a pan-RAF inhibitor with minimal N.Engl.J.Med. 366, 24432454 (2012).
secretion. Nature 487, 500504 (2012). paradoxical activation and activity against BRAF 133. Hodi,F.S. etal. Improved survival with ipilimumab in
This paper shows that adaptive response to RAF orRAS mutant tumor cells. J.Med. Chem. 58, patients with metastatic melanoma. N.Engl. J.Med.
inhibitors is promoted by growth factors in the 41654179 (2015). 363, 711723 (2010).
tumour microenvironment and upstream RTK 112. US National Library of Medicine. ClinicalTrials.gov 134. Robert,C. etal. Pembrolizumab versus ipilimumab
activation. https://clinicaltrials.gov/ct2/show/NCT02014116 inadvanced melanoma. N.Engl. J.Med. 372,
92. Wilson,T.R. etal. Widespread potential for growth- (2017). 25212532 (2015).
factor-driven resistance to anticancer kinase inhibitors. 113. Tang,Z. etal. BGB283, a novel RAF kinase and EGFR 135. Hoos,A. Development of immuno-oncology drugs
Nature 487, 505509 (2012). inhibitor, displays potent antitumor activity in BRAF- from CTLA4 to PD1 to the next generations.
93. Kopetz,S. etal. Phase II pilot study of vemurafenib mutated colorectal cancers. Mol. Cancer Ther. 14, Nat.Rev. Drug Discov. 15, 235247 (2016).
inpatients with metastatic BRAF-mutated colorectal 21872197 (2015). 136. Hugo,W. etal. Genomic and transcriptomic features
cancer. J.Clin. Oncol. 33, 40324038 (2015). 114. US National Library of Medicine. ClinicalTrials.gov of response to antiPD1 therapy in metastatic
94. Hyman,D.M. etal. Vemurafenib in multiple https://clinicaltrials.gov/ct2/show/NCT02610361 melanoma. Cell 165, 3544 (2016).
nonmelanoma cancers with BRAF V600 mutations. (2017). 137. Mandal,M. & Massi,D. Immunotolerance as a
N.Engl. J.Med. 373, 726736 (2015). 115. US National Library of Medicine. ClinicalTrials.gov mechanism of resistance to targeted therapies in
95. Hall-Jackson,C.A. etal. Paradoxical activation of Raf https://clinicaltrials.gov/ct2/show/NCT02607813 melanoma. Handb. Exp. Pharmacol. http://dx.doi.org/
by a novel Raf inhibitor. Chem. Biol. 6, 559568 (2017). 10.1007/164_2017_5 (2017).
(1999). 116. US National Library of Medicine. ClinicalTrials.gov 138. Rushworth,L.K., Hindley,A.D., ONeill,E. & Kolch,W.
96. Lyons,J.F., Wilhelm,S., Hibner,B. & Bollag,G. https://clinicaltrials.gov/ct2/show/NCT02428712 Regulation and role of Raf1/BRaf heterodimerization.
Discovery of a novel Raf kinase inhibitor. Endocr.Relat. (2017). Mol. Cell. Biol. 26, 22622272 (2006).
Cancer 8, 219225 (2001). 117. Sievert,A.J. etal. Paradoxical activation and RAF
97. Escudier,B. etal. Sorafenib in advanced clear-cell inhibitor resistance of BRAF protein kinase fusions Acknowledgements
renal-cell carcinoma. N.Engl. J.Med. 356, 125134 characterizing pediatric astrocytomas. The authors are grateful to Bogos Agianian for generating the
(2007). Proc.NatlAcad. Sci. USA 110, 59575962 movies. E.G. is supported by NIH grant R01CA178394 and
98. Llovet,J.M. etal. Sorafenib in advanced (2013). awards from the Melanoma Research Alliance, the Pershing
hepatocellular carcinoma. N.Engl. J.Med. 359, 118. Le,K., Blomain,E.S., Rodeck,U. & Aplin,A.E. Square Sohn Cancer Research Alliance, the Irma T. Hirschl
378390 (2008). Selective RAF inhibitor impairs ERK1/2 Trust, the Gabrielles Angels Foundation for Cancer Research,
99. King,A.J. etal. Demonstration of a genetic phosphorylation and growth in mutant NRAS, and the Sidney Kimmel Foundation for Cancer Research. P.I.P.
therapeutic index for tumors expressing oncogenic vemurafenib-resistant melanoma cells. Pigment Cell is supported by NIH grant R01CA204314, the Sidney
BRAF by the kinase inhibitor SB590885. Cancer Res. Melanoma Res. 26, 509517 (2013). Kimmel Foundation for Cancer Research, the Melanoma
66, 1110011105 (2006). 119. Basile,K.J., Le,K., Hartsough,E.J. & Aplin,A.E. Research Foundation, the Dermatology Foundation and the
100. Hoeflich,K.P. etal. Antitumor efficacy of the novel RAF Inhibition of mutant BRAF splice variant signaling Melanoma Research Alliance.
inhibitor GDC0879 is predicted by BRAFV600E bynext-generation, selective RAF inhibitors.
mutational status and sustained extracellular signal- PigmentCell Melanoma Res. 27, 479484 Author contributions
regulated kinase/mitogen-activated protein kinase (2014). Z.K., E.G. and P.I.P. contributed equally to researching the
pathway suppression. Cancer Res. 69, 30423051 120. Waizenegger,I.C. etal. A novel RAF kinase inhibitor literature, to writing the article and to reviewing and editing
(2009). with DFG-out-binding mode: high efficacy in BRAF- the manuscript before submission.
101. Shelton,J.G. etal. Differential effects of kinase mutant tumor xenograft models in the absence of
cascade inhibitors on neoplastic and cytokine-mediated normal tissue hyperproliferation. Mol. Cancer Ther. Competing interests statement
cell proliferation. Leukemia 17, 15, 354365 (2016). The authors declare no competing interests.
17651782 (2003). 121. US National Library of Medicine. ClinicalTrials.gov
102. Chin,P.C. etal. The cRaf inhibitor GW5074 provides https://clinicaltrials.gov/ct2/show/NCT03118817 Publishers note
neuroprotection invitro and in an animal model of (2017). Springer Nature remains neutral with regard to jurisdictional
neurodegeneration through a MEK-ERK and Akt- 122. US National Library of Medicine. ClinicalTrials.gov claims in published maps and institutional affiliations.
independent mechanism. J.Neurochem. 90, https://clinicaltrials.gov/ct2/show/NCT02405065
595608 (2004). (2017).
103. Falchook,G.S. etal. Dabrafenib in patients with 123. James,J. etal. CEP32496: a novel orally active SUPPLEMENTARY INFORMATION
melanoma, untreated brain metastases, and other BRAF(V600E) inhibitor with selective cellular and See online article: S1 (movie) | S2 (movie)
solid tumours: a phase1 dose-escalation trial. Lancet invivo antitumor activity. Mol. Cancer Ther. 11, ALL LINKS ARE ACTIVE IN THE ONLINE PDF
379, 18931901 (2012). 930941 (2012).

16 | ADVANCE ONLINE PUBLICATION www.nature.com/nrc



2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.