ArticlePDF AvailableLiterature Review

RAS‐targeted cancer therapy: Advances in drugging specific mutations

Wiley
MedComm
Authors:

Abstract and Figures

Rat sarcoma (RAS), as a frequently mutated oncogene, has been studied as an attractive target for treating RAS-driven cancers for over four decades. However, it is until the recent success of kirsten-RAS (KRAS)G12C inhibitor that RAS gets rid of the title "undruggable". It is worth noting that the therapeutic effect of KRASG12C inhibitors on different RAS allelic mutations or even different cancers with KRASG12C varies significantly. Thus, deep understanding of the characteristics of each allelic RAS mutation will be a prerequisite for developing new RAS inhibitors. In this review, the structural and biochemical features of different RAS mutations are summarized and compared. Besides, the pathological characteristics and treatment responses of different cancers carrying RAS mutations are listed based on clinical reports. In addition, the development of RAS inhibitors, either direct or indirect, that target the downstream components in RAS pathway is summarized as well. Hopefully, this review will broaden our knowledge on RAS-targeting strategies and trigger more intensive studies on exploiting new RAS allele-specific inhibitors.
This content is subject to copyright. Terms and conditions apply.
Received:  September  Revised: April  Accepted:  April 
DOI: ./mco.
REVIEW
RAS-targeted cancer therapy: Advances in drugging specific
mutations
Cen Liu1Danyang Ye1Hongliu Yang1Xu Chen1Zhijun Su1Xia Li2,
Mei Ding2,Yonggang Liu1,
Beijing University of Chinese Medicine,
Beijing, China
Institute of Genetics and Developmental
Biology, Chinese Academy of Sciences,
Beijing, China
Correspondence
Xia Li and Mei Ding, Institute of Genetics
and Developmental Biology, Chinese
Academy of Sciences, Beijing, China.
Email: xia.li_@genetics.ac.cn,
mding@genetics.ac.cn
Yonggang Liu, Beijing University of
Chinese Medicine, Beijing , China.
Email: liuyg@.com
Funding information
Beijing Municipal Natural Science
Foundation, Grant/Award Number:
; National Natural Science
Foundation of China, Grant/Award
Number: NSFC ; Central
Universities, Grant/Award Number:
-JYB-ZDGG-
Abstract
Rat sarcoma (RAS), as a frequently mutated oncogene, has been studied as an
attractive target for treating RAS-driven cancers for over four decades. However,
it is until the recent success of kirsten-RAS (KRAS)GC inhibitor that RAS gets
rid of the title “undruggable”. It is worth noting that the therapeutic effect of
KRASGC inhibitors on different RAS allelic mutations or even different cancers
with KRASGC varies significantly. Thus, deep understanding of the character-
istics of each allelic RAS mutation will be a prerequisite for developing new RAS
inhibitors. In this review, the structural and biochemical features of different
RAS mutations are summarized and compared. Besides, the pathological charac-
teristics and treatment responses of different cancers carrying RAS mutations are
listed based on clinical reports. In addition, the development of RAS inhibitors,
either direct or indirect, that target the downstream components in RAS path-
way is summarized as well. Hopefully, this review will broaden our knowledge
on RAS-targeting strategies and trigger more intensive studies on exploiting new
RAS allele-specific inhibitors.
KEYWORDS
allelic RAS mutation, cancer, personalized therapy, RAS inhibitor
1 INTRODUCTION
Rat sarcoma (RAS) proteins belong to the small-molecule
G protein family. As binary molecular switch, RAS
GTPases cycle between guanosine triphosphate (GTP)-
bound active form and guanosine diphosphate (GDP)-
bound inactive form, facilitating “on” or “off” state
in signal transduction. There are three canonical RAS
genes, kirsten-RAS (KRAS), neuroblastoma-RAS (NRAS),
and harvey-RAS (HRAS), encoding four RAS proteins
(KRASA, KRASB, NRAS, and HRAS). Among RAS
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the
original work is properly cited.
©  The Authors. MedComm published by Sichuan International Medical Exchange & Promotion Association (SCIMEA) and John Wiley & Sons Australia, Ltd.
genes, KRAS could encode two proteins (KRASA and
KRASB) due to alternative RNA splicing. RAS and its
downstream pathways control a variety of activities, such
as cell proliferation, survival, migration, etc.
Since the discovery of mutated oncogenic RAS in human
cancer in , it became a common sense that RAS mainly
drives the formation and development of tumor through
point mutations.– Oncogenic RAS family mutations have
been found in about % of all human cancers–; thus,
intensive research effort has been made to study RAS pro-
tein structure and biochemistry and develop anti-mutated
MedComm. ;:e. wileyonlinelibrary.com/journal/mco 1of27
https://doi.org/./mco.
2of27 LIU  .
RAS drugs for cancer therapy. However, the study is full
of failures and obstacles for the following reasons: () RAS
protein structure is smooth globular with shallow depres-
sions, which lack well-defined hydrophobic pockets for
targeting study. () As small GTPases, there are also stud-
ies that attempt to competitively inhibit RAS binding to
GTP, based on the fact that small G protein RAS needs to
bind to GTP to form an activated form. However, there is
picomolar binding strength between RAS and GTP, and
the concentration of GTP in the cytoplasm reaches .
mmol/L.– Therefore, the development of competitive
GTP-binding inhibitors is almost impossible. Thus, the
development of oncogenic RAS inhibitors had been stuck
in bottleneck until recent success in specific KRASGC
inhibitors.–
Subsequent clinical data manifest that KRASGC
inhibitor treatment efficacy differs in cancers with differ-
ent oncogenic RAS mutations or even in different cancer
types with KRASGC mutation. Therefore, knowledge of
the molecular structural property, biochemistry, and biol-
ogy will lay a foundation for designing an allelic-specific
mutated RAS inhibitor for personalized cancer treatment.
Based on the mutation sites and important domains
affecting RAS function, this review summarizes the struc-
tural biochemical differences of different mutations, so as
to provide reference for the development of inhibitors tar-
geting specific mutations. Furthermore, in order to achieve
better clinical treatment with inhibitors, the review also
lists the clinicopathological features among cancers com-
monly carrying RAS mutations and responses to treatment
strategies, providing reference for the appropriate selection
of personalized treatment for cancers with RAS mutations.
2PHYSIOLOGICAL ACTIVITIES OF
FUNCTIONAL RAS DOMAINS
RAS is a member of the small-molecule G protein family
that is activated by GTP binding. Different RAS paralogs
share similar structural compositions: a highly conserved
G domain (aa –) and a hypervariable region (aa –
/).–
The G domain is composed of an effector lobe (aa –)
and an allosteric lobe (–). The former contains two
switch regions: switch I and II, along with a P-loop.–
The latter contains regulatory sites that result in confor-
mational change in switch II once it binds calcium or
acetate.
Before playing a role, the C-terminal CAAX box of RAS
(comprising cysteine (C) and aliphatic (a) and variable
(X) amino acids) (CAAX) tail of RAS needs to undergo
a series of modifications.– As shown in Figure ,
after being modified by several enzymes, such as RAS
and a-factor converting enzyme (RCE), prenylcysteine
carboxyl methyltransferase (pcCMT), and palmitoyltrans-
ferases (PAT), CAAX is able to associate with the cyto-
plasmic leaflet of cellular membrane, and then RAS
is recruited to the plasma membrane and activated by
receptor tyrosine kinase (RTK) signals and subsequently
recruits downstream effector molecules to achieve sig-
nal transduction.– In the process, the cysteine in the
CAAX tail is first linked to a farnesyl group (-carbon)
via the thioether bond by FTase. Alternatively, the cysteine
of CAAX would also be modified with a geranyl group
(-carbon) by geranylgeranyltransferase I (GGTase I) if
FTase was blocked by inhibitors. RCE then excises the
AAX amino acids by recognizing farnesyl cysteine and
conjugates cysteine with a carboxyl group. The carboxyl
group is subsequently modified by pcCMT to produce car-
bomethoxy, which increases the lipophilicity of RAS and
helps RAS bind to the membrane. After this, the modifi-
cation processes of different paralogs begin to distinguish
each other. In addition to farnesyl cysteine, the remaining
cysteine residues of NRAS and HRAS would be palmitoy-
lated by PAT, which is located in the Golgi to facilitate their
transport to membrane. Unlike the former, KRASB will
not be modified by PAT because there is no remaining cys-
teine residue available. However, the polybase region of
KRASB will bind membrane through electrostatic inter-
action. In addition, KRASA has both a discontinuous
polybase region and excess cysteine available for palmitoy-
lation, which help KRASA choose whether or not to be
modified with PAT.
After binding to the plasma membrane, the RAS pro-
tein would be activated by the upstream RTK signal and
then converted into an active form for downstream signal-
ing. RAS activity switches between RAS–GTP (activated)
and RAS–GDP (inactive) binding forms, which is regu-
lated by guanine nucleotide exchange factors (GEFs) and
GTPase-activating proteins (GAPs). Recently, however,
RAS has been found to be activated by cytoplasmic RTK
granules rather than RTKs located in the plasma mem-
brane. Tulpule et al. found that oncogenic RTKs, losing
their lipid membrane-targeting sequences, could form
membrane-less cytoplasmic protein granules, in which
the RTK granules activate RAS in a lipid membrane-
independent manner. Besides, RAS activity can be regu-
lated by other modifications, such as phosphorylation,–
ubiquitination,– lysine acetylation,– and lysine
methylation.
In cells, once an upstream signal is received, GEF (such
as SOS) forms a complex with RAS binding to GDP while
promoting the dissociation of GDP from RAS. Then, GTP
would replace GEF in the complex to form the active
GTP-bound RAS to pass the signal downstream., Nor-
mally, the activation of RAS signaling pathway is transient
LIU  . 3of27
FIGURE 1 The process of Rat sarcoma (RAS) binding to plasma membrane after post-translational modification. CAAX motifs at
C-terminal of RAS, consisting of cysteine, aliphatic amino acids, and a variable amino acid, help RAS localize to specific plasma membrane
microdomains and subsequently pass signals to the downstream. The post-translational modification of CAAX is achieved by farnesylation,
hydrolysis by RAS and a-factor converting enzyme (RCE) and carboxymethylation by prenylcysteine carboxyl methyltransferase (pcCMT),
followed by palmitoylation by palmitoyltransferases (PAT) in Golgi (harvey-RAS (HRAS), neuroblastoma-RAS (NRAS), KRASA). The
difference between KRASB and other paralogs is that the CAAX of KRASB is able to bind to membrane depending on lysine residues.,
The figure was made using Biorender.
under physiological conditions. After the signal transmis-
sion is complete, GAP could boost RAS intrinsic GTPase
activity to dissociate from GTP and restore it to GDP-
bound form, thereby closing related signaling pathways.
However, when a specific point mutation of RAS occurs,
the intrinsic GTPase activity or GAP-binding ability of
RAS would be attenuated. Thus, the existence duration
of RAS–GTP, along with the activation duration of the
downstream signaling pathway, is prolonged, resulting in
abnormal cell proliferation as well as tumor occurrence
and development.–
3GENERAL DAMAGE INDUCED BY
RAS MUTATIONS AT HOT-SPOT SITES
There are mainly three RAS paralogs with different muta-
tion frequencies in cancer, including KRAS (%), NRAS
(%), and HRAS (%)., When oncogenic amino acid
substitution mutations occur, changes in RAS protein
structure and function lead to persistently activated down-
stream signaling pathways. Clinically, particular isoform,
mutation frequency, and amino acid substitution prefer-
ence at each site complicate RAS mutation occurrence in
various cancers. The types of RAS mutations and their
frequency in related cancers are shown in Figure .
RAS mutations occur with high frequency at G, G,
and Q,– causing unique changes in RAS domain
conformation, and facilitating persistence of RAS–
GTP-binding form. Extensive studies had been con-
ducted on the structural and functional influences of these
special sites on RAS, such as GDP/GTP binding, intrinsic
GTPase activity, and effector interaction. The biochemical
function of each hot-spot mutation site and the damage to
physiological activities of RAS caused by site-related muta-
tions are summarized below, including G, G, and Q.
4of27 LIU  .
FIGURE 2 The frequency of RAS mutations in human cancers. The data come from TCGA PanCancer Atlas Studies. The activating
mutations of RAS occur predominantly at codons , , and  and kirsten-RAS (KRAS) has the most tendency to be mutated in the three
paralogs. Clinically, RAS mutations are most prevalent in pancreatic adenocarcinoma (PC) (%), colorectal cancer (CRC) (%), and
non-small cell lung cancer (NSCLC) (%). (A) PC; (B) colorectal adenocarcinoma; (C) lung adenocarcinoma; (D) uterine corpus endometrial
adenocarcinoma; (E) uterine carcinosarcoma; (F) stomach adenocarcinoma; (G) testicular germ cell tumors; (H) cholangiocarcinoma; (I)
cervical squamous cell carcinoma; (J) skin cutaneous melanoma; (K) acute myeloid leukemia; (L) thyroid carcinoma; (M)
pheochromocytoma and paraganglioma; (N) thymoma.
LIU  . 5of27
3.1 GTP hydrolysis rate
RAS forms a transition state during either GAP-mediated
or RAS-intrinsic hydrolysis for GTP.– Although the
transition state form during RAS–GTP hydrolysis is still
controversial, current studies support that G, G
(located in the P-loop), and Q (located in the switch II
region) once mutated, the transition state form would be
disturbed to some extent, impairing the ability of RAS to
hydrolyze GTP. This is likely to be one of the reasons for the
oncogenic properties of mutated RAS. The related mecha-
nisms of GTP hydrolysis rate changes caused by G and
G mutations are complicated.
In the normal GAP-mediated hydrolysis, Arg (also
known as arginine finger) of GAP interacts with the car-
bon atom in the main chain at G of RAS.– However,
when G is mutated to other amino acids, the extra side
chains of alternative conflict with the interaction, which
stops the formation of transition state complex of G
mutant and GAP, resulting in a lower rate of the GAP-
dependent hydrolysis of GTP compared to wild type.,
Similarly, G mutants with large side chains are unable
to form transition state complexes with GAP, while the
mutants with small side chains are able to form tran-
sition states. In addition, there are data showing that
mutations in G and G (GD) affect the interaction
between hydrolysis catalytic residues Q of RAS, R
of GAP and GTP, thereby weakening GAP-mediated GTP
hydrolysis., Another report mentions that the GV
mutant could hinder fast hydrolytic steps by increasing the
flexibility of the region of RAS binding to GTP γ-phosphate
group, catalytic water and Arg. Surprisingly, Rabara
et al. found that KRAS mutation at G is sensitive to GTP
hydrolysis stimulated by NF (a GAP), which is different
from G and Q.
The intrinsic RAS GTPase activity, relative to GAP-
dependent GTP hydrolysis activity, may determine the
interaction duration between RAS and downstream objec-
tive response rate (RAF)-related pathways. The affinity
of RAF for KRAS is higher than that of PGAP and NF,
and more importantly, their binding protein domains of
RAS overlap. At the same time, these differences in affini-
ties may also be one of the reasons for biological behavior
differences among mutant forms. When RAS binds to
RAF–ras binding domain (RBD), the combination of Y
residue located in switch I of RAS and Q residue inter-
acts with the γ-phosphate group in GTP through a bridged
water molecule to form a transition state complex.,
Although the model of specific transition state formation
has not yet been determined, it is found that once other
amino acids, such as alanine and valine, replace glycine
at -position, the steric hindrance caused by oversized
side chain would hinder the formation of transition states
or reduce the stability. Moreover, according to previ-
ous reports, G is also involved in the hydrogen bond
interaction in GTP hydrolysis, and the large side chain
due to mutation would cause instability of the transition
state.
Compared with G and G, Q plays a direct and
important role in both intrinsic and GAP-mediated GTP
hydrolysis. Q regulates nucleophilic water molecules
close to the γ-phosphate of GTP, and the intrinsic
hydrolysis rate of RAS significantly reduces once Q is
mutated., Thus, the hydrolysis rates in Q mutants
are the lowest. However, unlike the G mutations, the
hydrolysis rate of the Q mutations could be signifi-
cantly increased by GAP, although the mechanism is still
unclear.
3.2 Nucleotide exchange rate
Compared with wild type, G and Q mutated RAS, the
nucleotide exchange rate in the RASGD is significantly
higher, and it is more important that this rapid nucleotide
exchange is spontaneous through an SOS-independent
way. Chen et al. GD mutation causes the destabiliza-
tion of the hydrogen bond formed between D in SWI
switch and GDP, facilitating RAS–GDP to GTP transforma-
tion. Additionally, an increase in the structural flexibility of
the switch domains SW and SW in mutant may result in
a decrease in the stability of the nucleotide-binding pocket.
For the reasons described above, nucleotide exchange with
a rapid rate starts up. It is found that the GD mutation is
rare in HRAS and NRAS compared with other mutations,
and there is a speculation that KRAS may have a specific
residue background different from the other paralogs to
support the instability of the GD active site. Unfortu-
nately, there are few data for other types of mutations at
G.
As for G and Q mutations, Smith et al. found that
without SOS mediation, the nucleotide exchange rate of
GV is . times slower than that of wild type, while
the exchange rate of QL is . times faster, and the for-
mer requires more SOS assistance to reach the wild-type
exchange rate.
3.3 Affinity differences in downstream
effectors
When mutated, the binding affinity of RAS to differ-
ent downstream effectors changes, which means that the
biases for downstream signaling may change.
6of27 LIU  .
There are data showing that the affinities of ARAF
and BRAF to wild-type RAS are stronger than the affin-
ity of GV mutant, while the binding preference of
RGL and RALGDS (two GEFs of RAL GTPases) for
RASGV increases. In addition, compared to KRASWT-
overexpressing cells, the Akt signaling of HBECsiP
cells with KRASGC overexpression is lower and Ral sig-
naling and anchorage-independent growth are stronger,
whereas the phosphorylated Akt levels are higher and Ral
activation is lower when KRASGD is overexpressed.
In conclusion, signal transduction bias and variation in
downstream signal stringency arising from different RAS
mutations suggest that it is necessary to selectively use
inhibitors targeting mutation-preferred pathways in facing
different mutations in clinical treatment.
4 CHARACTERISTICS OF SPECIFIC
RAS MUTATIONS
Although the previous section shows the general impact
of RAS mutations associated with respective hot-spot sites,
each mutation has specific biochemical characteristics.
For example, the development of GC inhibitor is based
on the feature that GC is in the GDP-binding form
for a longer time than KRASWT or other mutations.
Therefore, we introduce the biochemical characteristics
of common oncogenic RAS mutations in this section. In
addition, in order to emphasize the significance of specific
mutation, pathological differences brought by different
mutations, such as the tendency of co-mutations, transfor-
mation capacity, and clinicopathological features, are also
listed.
4.1 Differences in structure and
biochemical signature
The function of protein depends on its structure. Table
summarizes the structural differences of the following
RAS mutations, expecting to be helpful for the design of
allele-specific inhibitors.
By comparing various biochemical indicators of KRAS
mutants, Hunter et al. found that small structural
differences due to mutation led to specific biochem-
ical signatures. For example, alteration in nucleotide
exchange rate is most notable in KRASGD. According
to Table ,KRAS
GD enables faster nucleotide exchange
independent of GEFs. As for the intrinsic hydrolysis rate,
KRASGA/GR/QH/QL mutation brings about a drop to
the lowest level. On the contrary, unlike other mutants
with severely impaired hydrolysis capacity, KRASGC has
no significant effect on intrinsic hydrolysis rate compared
to wild type. Alternatively, it is more likely to develop the
GDP-binding form KRASGC than other mutations, which
gives opportunities for the generation of direct inhibitors.
Overall, biochemical differences among RAS mutations
offer an opportunity to design specific inhibitors of RAS
mutants.
4.2 Differences in co-mutation
Rabara et al. reported that the KRASGX is more sen-
sitive to GTP hydrolysis stimulated by NF (a GAP) than
G mutants and Q mutants and partially dependent on
upstream signaling from epidermal growth factor receptor
(EGFR) or other RTK, which may enable EGFR inhibitors
to prevent the development of colorectal cancer (CRC)
with GD mutation. In addition, Rabara et al. also
pointed out that cells with low-frequency and weakly
oncogenic KRAS mutations are more prone to additional
mutations. In other words, their genomes have higher
genetic instability. For example, KRASGX and NRASAT
are occasionally associated with NF mutations, while
NF mutations are rarely observed in cells with KRASGX
and KRASQX mutations. According to the above, EGFR
inhibitors may be an option that is worth considering in
the treatment of cancers with KRASGx and normal NF
activity.
4.3 Differences in transformation
capacity
The transformation capacities of different mutants vary.
By comparing a large number of c-Ha-ras mutants, See-
burg et al. showed that the transformation abilities of
HRASGR and HRASGV mutations are stronger than
those of other mutations at codon . Additionally, by
transfecting NIHT cells with plasmids expressing wild-
type or mutated KRAS, Smith et al. observed that the
transformation ability of mutations at codon  was slightly
stronger than that of mutations at codons , , , and
.
4.4 Clinicopathological varieties in
different RAS mutations
In addition to the above, the clinicopathologic features of
cancers carrying different RAS mutations are also differ-
ent. Table summarizes the pathological differences in
various RAS mutation-induced cancers.
LIU  . 7of27
TABLE 1 Structural differences of RAS mutations.
Allele Difference Potential biochemical impact References
KRASGC Cysteine insertion may improve the
electrostatic environment around
γ-phosphate of GTP
Intrinsic hydrolysis rate close to wild
type

More exposed nucleotide binding site An increased propensity for the
conformational transition of KRAS
from inactive to active

R of GAP away from GTP in
KRASGC–GTP–GAP complex
A decrease in the rate of GAP-mediated
hydrolysis
,
Broken hydrogen bond between Q and the
γ-phosphate of GTP
KRASGD The atom OE of Q side chain away from
GTP in KRASGD–GTP–GAP complex
A decrease in the rate of GAP-mediated
hydrolysis

More exposed nucleotide-binding site An increased propensity for the
conformational transition of KRAS
from inactive to active

Deviation of SII residues toward the αhelix
and disruption of the hydrogen bonding
network in SII
Slow GTP hydrolysis rate 
The electrostatic repulsion between the
carboxylate group on the side chain of
aspartic acid and the γ-phosphate of GTP
The binding force of KRASGD to
GppNp is weaker than that of
KRASGV or KRASWT

KRASGV The atom OE of Q side chain away from
GTP in KRASGV–GTP–GAP complex
A decrease in the rate of GAP-mediated
hydrolysis

KRASGR The flexibility of Q affected and the
cooperation of Q with nucleophilic water
broken
A decrease in GAP-dependent
hydrolysis rate

The side chain of arginine creates a steric
conflict with the Y residue, and displaces
adjacent coordinated water
A decrease in intrinsic hydrolysis rate 
KRASGA Alanine insertion may change the electrostatic
environment around the γ-phosphate of GTP
and reorder the solvent
A decrease in intrinsic hydrolysis rate 
KRASGD The arginine finger of GAP is blocked from
accessing phosphate of GTP
A decrease in GAP-dependent
hydrolysis rate

KRASGR The arginine finger of GAP is blocked from
accessing phosphate of GTP
A decrease in GAP-dependent
hydrolysis rate

KRASQH The atom ND of H side chain away from
GTP in KRASGV–GTP–GAP complex
A decrease in the rate of GAP-mediated
hydrolysis

HRASQL An increase in the flexibility of the SII domain,
while in the opposite with RAF–RBD binding
Reduced intrinsic hydrolysis rate and
stronger RAF/MEK/ERK signal
,
Abbreviations: GTP, guanosine triphosphate; HRAS, harvey-RAS; KRAS, kirsten-RAS; NRAS, neuroblastoma-RAS; RAF, Serine/threonine-protein kinase RAF;
RAS, rat sarcoma; RBD, RAS binding domain.
5DEVELOPMENT OF DIRECT
MUT-RAS INHIBITORS
5.1 Non-mutation-specific inhibition
ways targeting oncogenic RAS
The structure of RAS has been studied in depth. How-
ever, it was not until recent years that inhibitors that bind
directly and block the abnormal activities of oncogenic
RAS were developed. The first hurdle in targeting RAS
mutations is the lack of deep binding pockets. RAS is
similar to globular protein with smooth surface and shal-
low depressions, lacking a clear hydrophobic pocket and
obscuring allosteric site. The other strategy is reducing
the active state of RAS–GTP by competing for the GTP-
binding region. However, the binding strength of RAS with
8of27 LIU  .
TABLE 2 Clinicopathological differences among cancers.
Cancer Mutation Difference References
PC KRASGD Lower survival rate and poorer prognosis after first-line gemcitabine therapy
and shorter OS in patients with locally advanced and/or metastatic PDAC

Little response to the combination of cobimetinib and gemcitabine in PDAC 
A strong association with early distant metastasis after radical tumor
resection and shorter postoperative OS and PFS in PDAC

KRASGV Little response to the combination of cobimetinib and gemcitabine in PDAC 
CRC KRASGD Poor response to FOLFOX treatment and high risk of disease recurrence 
Shorter OS compared with patients with KRASWT 
KRASGX For patients with CRLM undergoing radical liver resection, KRAS G
mutations is associated with shorter OS than G mutations, especially
GV or GS. Besides, GD and GV are related to shorter OS and PFS
,
KRASGV A higher risk of relapse and death ,
Poorer survival in BRAFWT CRC 
KRASGD Multiple metastatic sites as the disease progressed 
A tendency to lead lymph node metastasis, more common in advanced
cancer and associated with higher PFS

KRASGX Associated with mucus histotype and favoring signaling pathways involved
in regulating mucin production in colonic mucosal cells

KRASGC Shorter OS, higher basal EGFR activation, and reduced immune profile 
NSCLC KRASGC Shorter PFS treated with PD-L inhibitors among patients with high PD-L
expression

KRASGV Worse OS, PFS, and recurrence rates 
Shorter survival, shorter duration of response to initial chemotherapy, and
shorter OS after immunotherapy in patients with advanced NSCLC with
KEAP/NFEL co-mutation

More pleural–pericardial metastases after thoracic surgery compared with
other mutations

Abbreviations: CRC, colorectal cancer; CRLM, colorectal liver metastases; EGFR, epidermal growthfactor receptor; NSCLC, non-small cell lung cancer; OS, overall
survival; PC, pancreatic adenocarcinoma; PDAC, pancreatic ductal carcinoma; PD-L, programmed death--ligand ; PFS, progression-free survival.
GTP reaches picomolar level and the concentration of GTP
in the cytoplasm reaches as high as . mmol/L; thus, it is
too difficult to develop drugs that compete with RAS–GTP
binding.,
After decades of efforts, some “hidden weaknesses” of
RAS have gradually been uncovered. Lu et al. found
that the formation of activated RAS experiences multiple
sub-states, and there is a new allosteric druggable site P
hidden in the protein, which could regulate the GTPase
activity of RAS. More importantly, this site is located at the
interface of the RAS domain, which interacts with down-
stream effectors, so targeting the interaction of RAS and
downstream effectors becomes a new strategy to design
drugs. In addition, some research interests have focused on
two RAS regions, switch II pocket and the junction area of
the switch I/II pocket, both of which are potential drug-
gable pockets. Theformerwasfoundtobeanimportant
region for covalent binding of drugs during the develop-
ment of KRASGC inhibitors, which could lock KRASGC
in an inactive form after binding to drugs, also named
KRAS-off inhibitors.,, The latter, as a conserved struc-
ture shared by all three RAS paralogs, has more potential as
a pan-RAS inhibitor targeting region, but how to improve
the affinity of the drug for mutant RAS and make it much
higher than for the wild-type RAS is a thorny challenge.
More recently, “KRAS-on” inhibitors that target active
KRAS–GTP have also been developed to block interactions
of “on-state” RAS with downstream effectors. In this strat-
egy, cyclophilin-A, as an intracellular chaperone protein
of drug, binds to S-IIP of RAS–GTP to form a tri-complex
of KRAS, cyclophilin-A, and drugs, such as RMC-.
The tri-complex leads to the covalent RAS cross-linkage,
subsequently blocking downstream effector binding to
KRAS. Similarly, in the presence of the compound RM-,
chaperone proteins block downstream signal transduc-
tion by forming tri-complexes to block binding of RAS
with RAF. This strategy provides more flexibility for
investigating ways to specifically target RAS mutations.
In addition, recent studies have revealed that there are
missense mutations at the same site or certain differences
LIU  . 9of27
among mutations at different RAS sites, which may have
implications for the development of drugs targeting spe-
cific RAS mutations. Unfortunately, there are few studies
comparing the differences between paralogs at the same
point. For example, a report shows significant phenotypic
differences as well as divergence in signaling pathways of
KRASGD and NRASGD in CRC.
5.2 Specific inhibitors targeting
mutated RAS
In addition to the non-specific mutation inhibitors, the
recent success in developing KRASGC inhibitors has
revived interest in developing KRAS inhibitors that either
directly target KRAS mutations or target key steps required
for KRAS activation. The research progress for specific tar-
geted inhibitors of KRASGx mutations (such as GC) is
still ahead of other mutant KRAS. Regrettably, efficacious
targeted inhibitors of GD and QR have not been found,
nor have specific inhibitors of oncogenic NRAS and HRAS
mutants.
.. Drugs targeting KRASGC
Introduction of KRASG12C inhibitors
KRASGC with a high incidence in cancers such as
non-small cell carcinoma is a common type of KRAS
mutation. Recently, researchers have found that there
is a new pocket adjacent to the mutated cysteine residue
(cys) on the switch II region of the GDP-binding
KRASGC protein. The developed small-molecule targeted
drugs could bind to the allosteric pocket by forming irre-
versible covalent compounds with cys. Moreover, other
inhibitors preferentially bind to the KRASGC–GDP and
block SOS-mediated nucleotide exchange and thereby
inhibit the hyperactivation signals of RAS. The successes
in clinical trials of these inhibitors make a breakthrough
in the development of KRASGC-specific compounds with
anticancer activity in vivo for targeting the SII pocket. The
information on the structure, covalent bond, and medica-
tion guidance of KRASGC inhibitors is summarized in
detail in several reviews. Dunnett-Kane et al. provide a
detailed description of the therapeutic activity of several
registered KRASGC-specific inhibitors and summarizes
their clinical trial stages and trial designs. In addition,
interestingly, in cancer cells that are insensitive or resistant
to a covalent KRASGC inhibitor ARS, Zhang et al.
reported that ARS-modified peptides in major histo-
compatibility complex I (MHC-I) complex could serve as
neoantigens. Subsequently, these neoantigens are recog-
nized by recombinant antibody PA. The PA induces
cytolytic T-cell response, so ARS-resistant KRAS GC
mutant cells are killed in vitro. The strategy provides
inspiration for reducing clinical resistance of single GC
inhibitors.
Therapeutic effect of RAS G12C inhibitors
Since the successful development of GC mutation
inhibitors was reported, various GC inhibitors have been
rapidly put into clinical trials. Among GC inhibitors,
sotorasib (AMG) and adagrasib (MRTX) are rep-
resentative agents that have been shown to have direct
therapeutic effects in various cancers with GC muta-
tions. Although targeted inhibitors of KRASGD have also
been developed, there are no clinical trial data. Therefore,
this section will focus on the clinical progress of GC
inhibitors.
On the one hand, sotorasib was approved by the Food
and Drug Administration as a second-line treatment for
locally advanced or metastatic non-small cell lung can-
cer (NSCLC) containing the KRASGC mutation. In the
CodeBreaK phase I trial,  mg was identified as the
recommended phase II dose by estimating the response of
approximately  patients in the dose-escalation cohort.
After . months as the median follow-up time, .% of
the patients in the subgroup with NSCLC had a confirmed
response with a disease control rate of .%. The median
progression-free survival (mPFS) for patients with NSCLC
was . months. In the phase II portion of CodeBreaK,
among the  enrolled patients, .% of the patients had
a confirmed response with a disease control rate of .%
with a median follow-up of . months. The mPFS was
. months and the median overall survival (mOS) was
. months. The rate of treatment-related adverse events
was .%, including .% probability of occurrence of
grade event and .% probability of occurrence of grade
event.

Also, sotorasib has shown monotherapy clinical activ-
ity in KRASGC-mutated CRC in the CodeBreaK phase
I trial, although the treatment is less effective. With a
median follow-up of . months, the response rate among
patients was .% with a disease control rate of .%.
On the other hand, the phase I/II KRYSTAL-
trial explored the therapeutic effect of adagrasib
monotherapy. A dose of  mg twice daily was iden-
tified as the recommended phase II dose by estimating
the response of  patients in the dose-escalation cohort.
Among  evaluable patients with KRASGC NSCLC
treated with adagrasib  mg bid, .% of the patients
had a confirmed response, with a median duration of
response of . months. The mPFS was . months and
the -month survival rate was .%.
Adagrasib is also being studied in CRC with
KRASGC. With a median follow-up of . months,
10 of 27 LIU  .
the response rate among patients was %, with a disease
control rate of %, and the mPFS was . months.
Resistance mechanism of G12C inhibitor monotherapy
In the clinical application of GC inhibitors, it has been
observed that there are differences in drug-resistance
involved pathways in different cancer types. In NSCLC,
upregulation of MEK and ERK is found to develop for
resistance of sotorasib. Differently, in CRC, the phos-
phorylation of EGFR leads to the resistance to GC
inhibitors.,, This suggests that for CRC carrying
GC, the combination of GC inhibitors and EGFR
inhibitors may be clinically more effective than GC
inhibitor monotherapy.
.. Drugs targeting KRASGD
There are several studies on specific inhibitors target-
ing KRASGD.KRAS
GD commonly occurs in digestive
system cancers, especially pancreatic cancer and CRC.
The lack of a highly reactive residue, such as the cys-
teine residue at position , renders targeted KRASGD
drug design in a distinct approach. Vatansever et al.
performed computational simulations on the kinetic data
of KRASGD to explore the changes in the domain of
mutated RAS, aiming to lay a foundation for subsequent
targeted drug research. Similarly, through computational
analysis of protein structures, Feng et al. found P,
the connection site adjacent to proline , and a small
molecule named KAL- that binds to this site
(KD= μM). In addition, the compound  devel-
oped by Stockwell et al. could target KRASGD with
micromolar affinity. Interestingly, when screening small-
molecule compounds targeting KRAS switch I/II pocket,
Kessler et al. reported that a compound named BI-
, in contrast to covalent inhibitors, could bind to both
the active and inactive forms of KRAS with nanomolar
affinity. Recently, a specific inhibitor targeting KRASGD,
namedMRTX,hasbeenshowntobindKRAS
GD in a
non-covalent binding form (KD=. pM). And the selec-
tivity of MRTX binding to KRASGD is -fold higher
than that of binding to KRASWT. In addition to small-
molecule compounds, there has been some progress in the
development of peptides as another form of targeted drugs.
For example, KRpep-d and its derived peptide KS- are
developed to selectively bind to the KRASGD–GDP with
sub-nanomolar KDvalues and inhibit nucleotide exchange
of KRASGD. Table summarizes the characteristics of
mutated KRAS inhibitors (except GC).
The structure of the chemical formula in Table is
drawn with ChemDraw.
.. Drugs targeting KRASGV
Currently, the clinical trials for specific inhibitors of
KRASGV are rare. Han et al. developed a derived
peptide based on H-REV in vivo that could interact
with KRASGVGDP to form a stable complex, thereby
blocking the activating function of KRAS and inhibiting
phosphorylation level of MEK/ERK.
In addition, there is a report finding that tyrosine kinase
receptor  (TKR), a thiourea derivative, is able to inhibit
cell proliferation and induce apoptosis in A cells by
targeting KRASGV.
.. Drugs targeting other KRAS mutations
In the development of KRASGR inhibitors, Zhang et al.
screened α,β-diketoamide from common electrophilic lig-
ands based on the nucleophilicity of arginine residue in
KRASGR. The research reveals that the ligand is able to
selectively bind to KRASGR in a covalent form, provid-
ing reference for the development of inhibitors targeting
KRASGR.
In the development of KRASGS inhibitors, Zhang
et al. also chose the nucleophilicity of serine residue
as the weakness of KRASGS. The potential of β-lactone
derivatives as ligands for KRASGS is revealed, and Ser
in SIIP is acylated by the carbonyl group of β-lactone
during ligand formation of a covalent complex with
KRASGS.
6 INDIRECT MUT-RAS INHIBITION:
INTERVENTION IN DOWNSTREAM
PATHWAYS
RAS, as an important component in regulating cell
growth and proliferation, is involved in a large network
of related cascaded signaling pathways. In the develop-
ment of anticancer drugs, it is essential to understand the
signaling pathways associated with RAS., Overactiva-
tion of the RAF/MEK/ERK and phosphatidylinositol--
kinase (PIK)/protein kinase B (AKT)/mechanistic target
of rapamycin (mTOR) pathways enhances growth, sur-
vival, and metabolism of cancer cells. Thus, the signaling
pathways have been identified as promising therapeutic
targets for cancer therapy. This section focuses on the
RAS downstream pathways shown in Figure that are
closely associated with the development of cancers with
RAS mutation. The impact of pathways for specific cancer
types is summarized as well.
LIU  . 11 of 27
TABLE 3 Features of non-GC KRAS inhibitors.
Targeted
mutation Name Structure
Reactive
functional group
Targeted position
in RAS KDReference
GD KAL- Quinolinol and
piperazinyl group
Proline   μM
 N/A S,D,E,and
I
 μM
BI- Isoindolinone SI/II pocket  nM 
MRTX N/A S-II pocket . pM 
KS- Amide bond
cyclization and
main chain
cyclization
S-II pocket  nM 
GV H-REV N/A L, Y, D, G,
D, K, and
Y
SI/II pocket and
P-loop
 μM
TKR N/A Thiourea and
benzotrifluoride
group
N/A N/A 
GS GSi- β-Lactone S-II pocket (Ser)  μM
GR N/A α,β-Diketoamide S-II pocket (ε-N and
η-N of R)
N/A 
12 of 27 LIU  .
FIGURE 3 Downstream signaling pathway of KRAS. After receiving the signal of epidermal growth factor, receptor tyrosine kinases
(RTKs) such as EGFR will recruit RAS, targeting the membrane and activating it. Therefore, phosphorylation activation signals are passed in
the downstream cascades, which contain the RAF–MEK–ERK, PIK–protein kinase B (AKT), RAL, and TIAM pathways. These cascades
regulate cell proliferation, migration, and invasion. The figure was made using Biorender.
6.1 Intervention in RAF–MEK–ERK
(MAPK) pathway
.. Components of MAPK pathway
The RAS/mitogen-activated protein kinase (MAPK) path-
way, as a classical downstream signaling pathway of RAS,
is closely related to variety of cancers. MAPK is a phos-
phorylated activation effector cascade composed of RAF,
MEK, and ERK protein kinases, regulating cell growth,
survival, proliferation, and differentiation.– RAF is
the direct effector of activated RAS as the beginning of
cascade signals. And there are three proteins in RAF fam-
ily, ARAF, BRAF, and CRAF. In cascade signaling, RAF
is recruited through its RBD domain selectively binding
to the active GTP–RAS on the plasma membrane.,
Subsequently, the cysteine-rich domain (CRD) domain of
RAF selectively binds to the farnesyl groups of modified
RAS, which further promotes the recruitment of RAF
to cell membrane., On the basis of RBD binding to
RAS, CRD also interacts with phosphatidyl serine in the
plasma membrane. During the recruitment, affected by
the increase in local concentration and conformational
change of RAF,, its homologous dimerization and
heterodimerization with any other RAF family member
will be induced, which also promotes the activating phos-
phorylation of RAF. After activating the downstream
MEK/ERK, RAF kinase reverts to an inactive conforma-
tion due to complex regulation, such as dephosphorylation
by kinases, autoinhibition by amino-terminal domain,
and negative feedback regulation of ERK.– However,
once RAF is mutated, the above complex regulation of
RAF activation will be disrupted. Oncogenic BRAF muta-
tions are common in cancer. One class of BRAF mutants,
BRAF-V mutants, will lose its autoinhibitory activ-
ity and also be activated continuously without relying on
RAS or dimerizing. Another class of mutants will simi-
larly be RAS independently activated, which still relies on
dimerization for complete activation. The third mutants,
stimulating BRAFWT to form heterodimers, are kinase-
impaired proteins and need RAS-binding activation and
dimerization.–
Similar to RAF, a region of the N-terminal lobe (neg-
ative regulatory region) of MEK kinase also exists for
autoinhibition. In the process of MEK activation by
RAF phosphorylation, kinases such as kinase suppres-
sor of RAS (KSR) act as scaffold proteins to assist the
assembly of RAF–MEK–ERK complex and promote RAF
to activate MEK., After the complex forms, the ser-
ine residues of MEK (S and S of MEK, S
and S of MEK) are diphosphorylated rapidly for
MEK activation., Subsequently, tyrosine and threo-
nine residues of downstream ERK (T/Y of ERK
and T/Y of ERK) are dephosphorylated. Unlike
RAF kinase, MEK/ERK mutations are rare in human
tumors.,
LIU  . 13 of 27
.. The role of MAPK in cancer
In driving the development of RAS-mutated cancers, the
activity of MAPK kinases varies among different tumors.
Furthermore, RAF subtypes have different influences on
RAS-driven tumors. For example, BRAF is dispensable for
NSCLCwithKRAS
GV or KRASGD, caused by the other
RAF kinases presenting a compensatory effect without
increased expression. On the contrary, CRAF is essential
to mediate oncogenic signaling in the same cancer.,
However, CRAF deficiency does not affect tumor devel-
opment of pancreatic ductal carcinoma (PDAC) with
KRASGD or KRASGV, while concomitant ablation of
EGFR and CRAF completely prevent PDAC development
drivenbyKRAS
GV/Trp mutation., However, the
selection of CRAF as a therapeutic target tends to rely
on ablation rather than inhibition of kinase activity. On
the one hand, clinical inhibition of CRAF kinase activity
requires prevention of inhibition of BRAF kinase activ-
ity, thereby avoiding increased toxicity. On the other
hand, CRAF ablation regulates apoptotic pathway-related
proteins independent of kinase activity or MAPK path-
ways, which may support CRAF ablation to inhibit tumor
development. In addition to apoptosis-related proteins,
CRAF regulates many proteins in a kinase-independent
manner. For example, CRAF inhibits the kinase activity
of ROKα, and the ablation of RAF induces the regres-
sion of squamous tumors via ROKα-mediated cellular
differentiation.
Unlike RAF, systemic ablation of MEK or ERK will
induce the unacceptable toxicities in adult mice, although
preventing tumor development. Interestingly, in the
CRC model with KRASGV, heterogeneity in intracellular
ERK phosphorylation has been observed. ERK levels are
generally higher in cancer cells adjacent to stromal cells at
the invasive front and lower in more central areas of cancer
specimens.
.. Clinical effect of MAPK inhibitors
Although the MAPK is a linear cascade, the regulation
of each element involves multiple kinases. Therefore, in
the treatment with inhibitors, the regulation of feedback
pathways in vivo after drug administration should be
considered to avoid drug resistance.
Significant efforts have been made to develop inhibitors
targeting the MAPK pathway, especially BRAF and MEK
inhibitors. However, only the BRAF inhibitors vemu-
rafenib and dabrafenib and the MEK inhibitors trametinib
and cobimetinib have been approved only for BRAF-
VE/K metastatic melanomas. In KRAS-mutated
tumors, BRAF inhibitors promote heterodimerization of
BRAF and CRAF, thereby activating the MAPK pathway
and helping secondary tumor development.– Recently,
some novel pan-RAF inhibitors have been investigated to
solve the problem of promoting dimerization. LY
inhibits MEK/ phosphorylation by inhibiting kinase
activity in BRAF–CRAF heterodimers and retards the
development of tumors carrying KRAS mutations while
inducing a more significant dimerization. However,
when used as a single agent, the required dose of RAF
inhibitor being effective in KRAS-mutated models is sig-
nificantly higher than that in the BRAF-VE model.
Therefore, it is necessary to continue upgrading effective
drugs or treat with a combination of drugs.
Similar to RAF inhibitors, cobimetinib also fails to
inhibit KRAS mutation-carrying tumors due to RAF feed-
back activation of MEK phosphorylation. In contrast,
trametinib appears to inhibit the proliferation of KRAS
mutant A cells effectively by impairing the RAF–
MEK interaction. However, its resistance still exists in
KRAS mutant NSCLC, which is caused by compensatory
activation of FGFR. As a result, MEK inhibitors are
increasingly used in combination with other inhibitors
to avoid drug resistance due to the feedback regulatory
mechanism induced by their use alone.
As for ERK inhibitors, the drug resistance tends to
develop after monotherapy with ERK/ inhibitor over
a period of time., One of the mechanisms is the
compensatory effect of ERK in place of the inactivated
ERK/. In response, inhibitors against ERK have been
developed. On the other hand, ERK inhibition blocks
negative feedback of ERK and induces feedforward acti-
vation of upstream RTK, thereby inducing activation of
alternative pathways such as the PIK/AKT pathway to
maintain tumor cell survival. Therefore, inhibitors target-
ing both ERK/ and other kinases are also developed
for treatment. For example, Gao et al. found that
an indole-substituted pyrimidine derivative inhibits the
activities of AKT and ERK/, thereby inhibiting tumor
growth and extending the survival time of tumor-bearing
mice.
As a significant downstream signaling pathway of RAS,
MAPK has become a popular therapeutic target. However,
the efficacy of MAPK inhibitors in monotherapy of tumors
with RAS mutation is not ideal. According to the above,
one of the reasons is that the MAPK pathway is abnormally
activated after the suppression of components. The mech-
anisms of reduced effectiveness or resistance of inhibitors
vary according to the target and how the inhibitor works.
On the other hand, the toxicity caused by inhibitors is also
worthy of attention. Therefore, rather than monother-
apy, combinations of drugs targeting the MAPK pathway
have been studied more frequently in the clinical treatment
of cancers carrying RAS mutations.
14 of 27 LIU  .
6.2 Intervention in
PI3K/PTEN/AKT/mTOR pathway
.. Components of PIK pathway
Similar to MAPK pathway, PIK pathway is an effector
cascade dependent on phosphorylated activation. PIK is
divided into three classes: I, II, and III. Class I PIKs are
expressed in various cell types and are related to develop-
ment of cancers.– Therefore, class I PIKs are selected
to be mainly discussed in this section. Class I PIKs are
heterodimeric proteins that are grouped into two subtypes:
IA and IB. PIK IA proteins are composed of a regula-
tory subunit (pα,pβ,pα,pα,pγ)anda-kDa
catalytic subunit (pα,pβ,pδ). PIK IAs act as
downstream kinases of TKRs and G protein-coupled recep-
tors (GPCRs). PIK IBs have a pγcatalytic subunit
binding p or p as regulatory subunit. Different from
PIK IAs, PIK IBs are activated by GPCRs.– Class
I PIKs are activated through different upstream mecha-
nisms, which mainly contain: () the regulatory subunit
p binding to phospho-YXXM motifs (X indicates any
amino acid) of the RTK, thereby triggering activation of the
catalytic subunit p; () growth factor receptor-bound
protein (GRB) binding to phospho-YXN motifs of the
RTK in advance and to the scaffolding protein GAB, which
in turn can bind to p; and () GRB binding to RTK
and subsequently activating SOS, RAS in turn, and finally
activating p independently of p.
After being activated, catalytic subunit of PIKs trans-
fers the phosphate group to PIP to produce PIP. In
the course, phosphatase and tensin homolog deleted on
chromosome  (PTEN), a negative regulator, acts as a
phosphatase to dephosphorylate PIP and convert it back
to PIP. Back to the downstream signaling pathway, PIP
recognizes downstream proteins with a pleckstrin homol-
ogy domain, such as PDK and AKT, and recruits them to
the cell membrane., In the cascade, AKT is partly acti-
vated by phosphorylation of PDK at T. Subsequently,
mTORC will fully activate AKT through phosphorylation
of AKT at S., Then, many downstream effectors,
which are involved in protein synthesis, cellular prolif-
eration, apoptosis, and cell survival, such as mTOR, are
phosphorylated by AKT.–
.. The role of PIK pathway in cancer
PIK pathway activated by RAS is essential for lung car-
cinogenesis driven by KRASGD. However, the estab-
lished tumors are less dependent on PIK signaling
and PIK/mTOR inhibition only leads to partial tumor
regression., Similarly, pαof PIK inactivation dose
dependently can prevent mouse lethality and the occur-
rence of cancers induced by KRASGD.pαactivity is
also required for in vivo superactivation of KRASGD and
other signaling pathways.,
.. Clinical effect of PIK pathway
inhibitors
PIK pathway has also been selected as a target for can-
cer therapy due to the discovery of overactivation of PIK
in a variety of cancers and its significance for proliferation
and survival of cancer cells. However, in the course of treat-
ment, problems such as abnormal activation of feedback,
compensation activation, drug resistance, and toxicity of
PIK pathway inhibitors are found.–
6.3 Intervention in other pathways
As a downstream effector of RAS, RAL GTPase mediates
various cellular activities to regulate tumor invasion, pro-
liferation, and resistance to cell death, which are executed
by the RAL effectors, including RALBP, Sec, Exo,
Filamin, PLD, and ZONAB.– RAL is grouped into
two subtypes: RALA and RALB. The former plays a major
role in cancer development and metastasis. In NSCLC,
growth of cancer cells carrying KRASGC is more sensitive
to RAL depletion. As for in pancreatic cancers, RALA is
essential for tumor growth, and RALA and RALB are both
required for tumor invasion.,
In addition of RAL, TIAM can bind to RAS through
its RAS-binding domain, causing synergistic formation of
Rac-GTP in a PIK-independent manner, thereby acti-
vating Rac to induce activation of the NF-kappa B tran-
scription factor and promotion of cancer cell survival.
However, in epithelial MDCK cells, Tiam–Rac and RAS
signaling seemingly oppose each other, since RASGV-
induced epithelial–mesenchymal transition is negatively
affected by Tiam–Rac signaling.,
7 COMBINATION STRATEGIES
ADAPTED TO SPECIFIC MUTATIONS IN
DIFFERENT CANCERS
After a long period of struggle in developing RAS-targeted
drugs, researchers have turned their attention to the RAS
downstream signaling pathway inhibitors, in order to indi-
rectly inhibit RAS activity. RAS would be considered as
an important component of the signaling pathway that
regulates cell growth and proliferation. The upstream
activation network of RAS is quite complex, and the down-
stream signaling pathway network is highly complicated
as well, both of which lead to intricate cross-talk between
LIU  . 15 of 27
TABLE 4 Efficacy of clinical trials adapted to cancers with mutated KRAS in last years.
Cancer Mutation Drug
Study
phase
Response
rate (N)
DR
(months)
PFS
(months)
OS
(months) Reference
PC GC Sotorasib I/II % () N/A . . 
GR Selumetinib
sulfate
II % () N/A . N/A NCT
CRC GC Sotorasib I.% () N/A . N/A 
Adagrasib I/II % () . N/A N/A 
Adagrasib I/II % () . . N/A 
Adagrasib with
cetuximab
% () . .
JNJ-
(ARS-)
I% () N/A (safety deficiency) 
NSCLC GC AMG
(sotorasib)
I % () .–. N/A N/A 
Sotorasib I .% () . . N/A 
II .% () . . . 
Docetaxel III .% () . . . 
Sotorasib .% () . . .
Adagrasib I/IIb .% () . . N/A 
Adagrasib II .% () . . . 
JNJ-
(ARS-)
I % () N/A (safety deficiency) 
GD Bortezomib
with acyclovir
II .% () N/A  NCT
Abbreviations: CRC, colorectal cancer; DR, duration of response; NSCLC, non-small cell lung cancer; OS, overall survival; PC, pancreatic adenocarcinoma; PFS,
progression-free survival.
RAS-related pathways. Therefore, protein targeting in a
pathway in the RAS-related signaling network often fails
to achieve the anticipated therapeutic purpose. Indeed, it
is found in clinical treatment that specific targeted ther-
apy or specific drug combination has an unexpected effect
on the cancers caused by different RAS mutations. In this
section, we mainly list the treatment of several cancers
commonly carrying KRAS mutations, such as pancreatic
cancer, NSCLC, and CRC. Table summarizes the progress
of clinical trials on direct KRAS inhibitors for cancers car-
rying known KRAS mutations with published results in
the last years. In addition, as a supplement, Table shows
ongoing clinical trials for cancers with KRAS mutations
(except for KRASGC).
7.1 Therapy of KRAS-mutated
pancreatic cancer
.. Therapy of pancreatic adenocarcinoma
with KRASGD
First of all, considering the successful development of
inhibitors targeting KRASGD mutation, the therapeutic
effect of KRASGD inhibitors on pancreatic adenocarci-
noma (PC) is preferentially discussed. As a non-covalent
KRASGD inhibitor, MRTX inhibits phosphorylation
levels of ERK/ and cell viability in KRASGD mutant
cell lines (IC = nM). Based on this, MRTX exhib-
ited obvious tumor regression (%) in KRASGD mutant
PDAC models. In addition to the efficacy, it is also found
that combination of MRTX and inhibition of EGFR or
PIKα, which are members of potential feedback or bypass
pathways, improves the antitumor activity in PDAC. As
another means of targeting GD mutations, combination
of siGD-LODER with gemcitabine and FOLFIRINOX
exhibits good tolerance and certain inhibition of disease
progression in patients with locally advanced pancreatic
cancer with KRASGD. Similarly, silence of KRASGD
by CRISPR-CasRx appears to suppress the tumor growth,
enhance the sensitivity of gemcitabine in PDAC, increase
the survival of mice, and show obvious toxicity.
From the downstream pathway inhibitor perspec-
tive, trametinib (MEKi) and ruxolitinib (STATi) help
nivolumab (programmed death- [PD-] inhibitor)
improve antitumor activity and survival in mice carrying
Ptf1aCre/+,LSL-KrasG12D/+,andTgfbr2flox/flox tumors. More
importantly, the combination strategy results in a clinical
16 of 27 LIU  .
TABLE 5 Ongoing clinical trials for cancers with KRAS mutations (except for GC).
Trial ID Tested interventions Treatment setting Phase Status
NCT HRS- Advanced KRASGD
mutant solid tumors
I Recruiting
NCT MRTX KRASGD mutant solid
tumors
I/II Recruiting
NCT ELI- P PDAC with KRASGD
and KRASGR
I Recruiting
NCT KRAS-EphA--CAR-DC abraxane
cyclophosphamide PD- antibody
Solid tumors with
KRASGV
I Recruiting
NCT Cyclophosphamide fludarabine
PD- antibodyBiological:
GV-specific TCR transduced
autologous T cells
Advanced PDAC with
KRASGV
I/II Recruiting
NCT Avutometinib (VS-) or/and
defactinib
NSCLC with KRASGV II Recruiting
Note: The data originated from https://clinicaltrials.gov.
Abbreviations: NSCLC, non-small cell lung cancer; PD-, programmed death-; PDAC, pancreatic ductal carcinoma; TCR, T cell receptor.
benefit for a patient undergoing chemotherapy-refractory
metastatic PDAC.
Despite the above-mentioned therapies, immunother-
apy is also receiving much attention. A study revealed that
inhibition of interleukin- enhances the antitumor effect
of anti-programmed death--ligand (PD-L) checkpoint
inhibitor therapy. The treatment combination delayed
tumor progression (p<.) and increased OS in engi-
neered PDAC model with KRASGD (p=.).
Recently, antidiabetic drug metformin was exca-
vated to have antitumor activity in PDAC induced by
KRASGD.InLSL-KrasG12D/+,Trp53fl/+,andPdx1-Cre
(KPC) mouse model, metformin has ability of blocking
tumor growth, inhibiting the incidence of abdominal
invasion, and increasing the OS. In addition, Si-HSF
was found to increase chemosensitivity to gemcitabine
in vivo.
.. Therapy of PC with KRASGV
In terms of targeting GV drugs, Ghufran et al. have
designed peptides inhibiting KRASGV, which are specu-
lated to have the ability to inhibit GV activity and reduce
the progression of cancer. In spite of targeting RAS itself, a
dual FT and geranylgeranyltransferase- inhibitor named
FGTI- can inhibit the growth of xenografts derived
from four patients with pancreatic cancer with mutant
KRAS (two GD and two GV) tumors.
In the aspect of immunotherapy, it has been reported
that the combination of anlotinib and PD- inhibitor and
chemotherapy exhibits a long-term partial response and
good tolerance in a young patient suffering from PDAC
with liver metastasis.
In other ways, sequential administration of cell-cycle
kinases CDK and CDK inhibitors after taxane treatment
reduced cell proliferation in PDAC mouse model with
both KRASGV and Cdkna-null mutations. Diego-
González et al. found that lipoplexes of si-FOSL-
and si-YAP reduce the tumor growth in mice carry-
ing tumors induced by pancreatic tumoral cell lines
(KRASGV and p knockout) through peri-tumoral
injection.
.. Therapy of PC with KRASGR
Pancreatic cancers with KRASGR are studied less than
the former. Recently, it was found that KRASGR is defec-
tive for interaction with pαof PIK, while PIKγwill
support macropinocytosis in KRASGR mutant PDAC in
a compensatory way., As for the therapy taken in
PC with KRASGR, the administration of selumetinib as
MEK/ inhibitor appears to have little effect. This sug-
gests that single-agent MEK inhibition is unable to meet
the therapeutic needs of patients suffering from pancreatic
cancer with GR. And the combination of cobimetinib
(MEK/ inhibitor) and gemcitabine improves PFS and OS
after treatment in patients with KRASGR compared with
pancreatic cancer patients with KRASGD/GV.
7.2 Therapy of KRAS-mutated CRC
.. Therapy of CRC with KRASGD
Different from the therapeutic effect in pancreatic can-
cer, MRTX has a lower inhibitory effect in CRC
LIU  . 17 of 27
carrying KRASGD. In contrast to PDAC, KRAS
mutations are usually not considered an initial driver in
CRC, which may be one of the reasons for the limited effect
of KRASGD inhibitors in CRC patients. Similarly, the
GD-targeting pathway is peptide KRpep-d. Similarly,
peptide KRpep-d, a GD-targeting inhibitor, had no
significant antitumor effect on the PDX model, while oxali-
platin showed a significant inhibitory effect. The failure
of KRpep-d is suspected to be related to bioavailability
and stability. Additionally, the combination of binime-
tinib, hydroxychloroquine, and bevacizumab makes a %
reduction in lung metastasis size in FOLFOX-resistant
patients after weeks treatment with this combination
with FOLFOX.
.. Therapy of CRC with KRASGV
The combination of lowdose trametinib ( nM) and
ABT (BclxL inhibitor) was found to inhibit tumor
growth against a KRASGV xenograft. In spite of tar-
geting MEK and BclxL, miR- is also found to exhibit
a potent growth inhibitory and proapoptotic effect by
directly targeting KRASGV and AKT.
.. Therapy of CRC with KRASGD
Phase III clinical trial evidence suggests that CRCs with
the KRASGD may benefit from EGFR inhibitors, such
as cetuximab, in contrast to the other most common
KRAS mutations. Therefore, the therapy of CRC with
KRASGD mainly revolves around cetuximab.– Chu
et al. revealed that -AAQB could resensitize KRAS
mutant cells to cetuximab before cells were treated with
cetuximab. In another study, the combination of first-
line chemotherapy drugs such as FOLFOX and cetuximab
improved OS and PFS after treatment in chemotherapy-
refractory CRC patients with KRASGD, while no response
was observed in CRC cell lines with KRASGX/QX muta-
tions or KRASWT CRC cell lines with BRAF mutations or
no expression of PTEN or EGFR proteins.,
Resistance of EGFR inhibitors in CRC with KRASGD
tends to depend on tumor suppressor NF. NF can convert
GTP–KRASGD to GDP–KRASGD, and the resistance
may be caused by impaired interaction between KRASGD
and NF.,
.. Therapy of CRC with other KRAS
alleles
The combination of irinotecan and cetuximab is adminis-
tered in patients with KRASWT mCRC who responded to
first-line chemotherapy with cetuximab and developed a
certain therapeutic effect after cetuximab in second- and
third-line treatment.
The combination of KRASGC inhibitor and EGFR
inhibitor targets the rare CRC that carries the GC muta-
tion. There is a lower response rate to KRASGC inhibitors
alone in CRC patients than in NSCLC patients because of
RTK–SHP-mediated adaptive RAS reactivation. The com-
bination of cetuximab can enhance the efficacy of AMG
in KRASGC-mutated CRC PDX model.,
7.3 Therapy of KRAS-mutated NSCLC
.. Therapy of NSCLC with KRASGC
Benefiting from the successful development of KRASGC
inhibitors, the current therapeutic strategy for NSCLC with
GC mutation is to combine KRASGC inhibitors with
other related pathway inhibitors, targeting immune check-
points, EGFR, SHP, SOS, MEK, PIK, mTOR, CDK/ or
others, to improve the therapeutic effect.– Neverthe-
less, the problem of resistance to GC inhibitors needs to
be focused.
.. Therapy of NSCLC with KRASGV
In the inhibition of RAS-related pathway, it is showed
that cotreatment with trametinib and osimertinib resen-
sitizes the EGFR mutant NSCLC cell line with KRASGV
to osimertinib. A clinical study revealed that patients in
the treatment of taxane cooperating with bevacizumab had
higher objective response rate (ORR) than those treated
with taxane alone. In another study, the combination
of selumetinib and docetaxel improved PFS and ORR in
patients with locally advanced or metastatic KRASGV
NSCLC (stage IIIB/IV), while no significant trend differ-
ences were observed due to the small statistical sample
base.
.. Therapy of NSCLC with KRASGD
One patient with high tumor mutational burden and pos-
itive PD-L expression with EGFRLR and KRASGD
mutations received therapy with a combination of beva-
cizumab, camrelizumab, and pemetrexed and achieved
remission over  months., In another clinical study, a
series of therapy of panitumumab concomitant with radia-
tion therapy and concurrent chemotherapy with paclitaxel
and carboplatin was settled. During therapy, KRASGD
lung cancer clone of the patient with stage III NSCLC
appeared to disappear.
18 of 27 LIU  .
TABLE 6 Role of RAS pathways in related cancers.
Cancer Mutation Role of mutation in cancer
Essential
pathway
component Role of component in cancer
PC KRAS mutation Inducing early PC ––
Promoting tumor metastasis and
aggressiveness,
RAL Essential for tumor growth and
invasion,
Maintaining tumor ––
KRASGD/GV CRAF Dispensable for tumor
development,
KRASGC/GD/QX Increased autophagic flux after
suppression of KRAS,
ERK Increased autophagic flux after
suppression of ERK,
CRC KRASGX/QX KRAS mutations are usually not
considered an initial
driver,
MAPK and
PIK/Akt
Low activation states in the
primary tumors
KRASGD CRC with GD will response to
EGFR inhibitors
NSCLC KRASGC/GV RAL Increased activation
AKT Decreased growth
factor-dependent AKT
activation 
KRASGD/GV CRAF Essential for tumor
development,
KRASGC RAL Tumor inhibition after RAL
depletion
KRASGD PIK Increased activation
MEK No obvious activation
KRASQH Increased dependence of tumor
development on MAPK
RAF/MEK Increased activation and
essential for tumor
development
Endometrial cancer KRAS mutation Inducing early EC and
development
PIK Increased activation
Promoting type I EC
aggressiveness
MAPK
PIK
Increased FGFR-dependent
activation
Neuroblastoma NRASQV Promoting tumor development MAPK Increased activation
Breast cancer HRASGV Inducing proliferation signal PIK Increased activation
KRASGV Maintaining mesenchymal
characteristics and metastatic
behavior
mTORC Increased activation
Abbreviations: AKT, protein kinase B; CRC, EC, endometrial cancer; colorectal cancer; MAPK, mitogen-activated protein kinase; NSCLC, non-small cell lung
cancer; PC, pancreatic adenocarcinoma; PIK, phosphatidylinositol--kinase.
8CONCLUSIONS AND PERSPECTIVE
Cancer therapies performed by targeting allelic RAS muta-
tions in specific clinical contexts could be regarded as
the milestone in personalized cancer treatment. After
decades of exploration, there are advanced understand-
ings of the structural differences, biochemical charac-
teristics, and downstream signaling preferences of RAS
mutations. These studies provide a solid foundation for
designing effective targeted therapy for cancers harboring
specific RAS mutations. Especially, successful develop-
ment of KRASGC inhibitors demonstrates the feasibility
of developing specific therapeutic strategies for each RAS
mutation. However, there are few studies on the structure
and biochemical differences of other RAS mutants, such as
KRASQL,