Acta Pharmacologica Sinica
(2017) 38: 614–622
© 2017 CPS and SIMM All rights reserved 1671-4083/17
www.nature.com/aps
Review
New knowledge of the mechanisms of sorafenib
resistance in liver cancer
Yan-jing ZHU
1, #
, Bo ZHENG
1, #
, Hong-yang WANG
1, 2,
*
,
Lei CHEN
1, 2,
*
1
International Co-operation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical
University, Shanghai 200438, China;
2
National Center for Liver Cancer, Shanghai 201805, China
Abstract
Sorafenib is an oral multikinase inhibitor that suppresses tumor cell proliferation and angiogenesis and promotes tumor cell apoptosis.
It was approved by the FDA for the treatment of advanced renal cell carcinoma in 2006, and as a unique target drug for advanced
hepatocellular carcinoma (HCC) in 2007. Sorafenib can signicantly extend the median survival time of patients but only by 3–5
months. Moreover, it is associated with serious adverse side effects, and drug resistance often develops. Therefore, it is of great
importance to explore the mechanisms underlying sorafenib resistance and to develop individualized therapeutic strategies for coping
with these problems. Recent studies have revealed that in addition to the primary resistance, several mechanisms are underlying
the acquired resistance to sorafenib, such as crosstalk involving PI3K/Akt and JAK-STAT pathways, the activation of hypoxia-inducible
pathways, and epithelial-mesenchymal transition. Here, we briey describe the function of sorafenib, its clinical application, and the
molecular mechanisms for drug resistance, especially for HCC patients.
Keywords:
sorafenib; hepatocellular carcinoma; targeted therapy; drug resistance; individualized treatment
Acta Pharmacologica Sinica (2017) 38: 614–622; doi: 10.1038/aps.2017. 5; published online 27 Mar 2017
Introduction
Since the molecular revolution of the 1980s, knowledge of
the etiology of cancer has increased dramatically, leading to
the discovery and development of targeted therapies tailored
to inhibit cancer-specific pathways. Among the numerous
molecular targeted drugs, sorafenib, an oral multi-target
kinase inhibitor that is also known as Nexavar, was developed
by the Bayer and Onyx companies. Due to its effect on renal
cell carcinoma and hepatocellular carcinoma, sorafenib has
been clinically approved for the treatment of advanced renal
cell carcinoma and hepatocellular carcinoma (HCC)
[1]
. More-
over, several clinical trials, including those for NSCLC
[2]
, meta-
static thyroid cancer
[3]
, steroid-refractory prostate cancer
[4]
,
and metastatic breast cancer
[5]
, with sorafenib as the potential
therapeutic strategy are now ongoing.
Being a multi-target kinase inhibitor, sorafenib can block
tumor cell proliferation by inhibiting the activity of Raf-1,
B-Raf and kinases in the Ras/Raf/MEK/ERK signaling path-
#
These authors contributed equally to this work.
*
To whom correspondence should be addressed.
E-mail chenlei@smmu.edu.cn (Lei CHEN);
hywangk@vip.sina.com (Hong-yang WANG)
Received 2016-11-21 Accepted 2017-01-19
way. Additionally, sorafenib can inhibit angiogenesis through
targeting of the hepatocyte factor receptor (c-Kit), Fms-like
tyrosine kinase (FLT-3), vascular endothelial growth factor
receptor (VEGFR)-2, VEGFR-3, platelet-derived growth factor
receptor (PDGFR-β) and other tyrosine kinases
[6, 7]
(Figure 1).
Preclinical studies have also found that sorafenib is effective
in various tumor cells, such as breast cancer MDA-MB-231
(containing G463V b-raf and k-ras gene mutations), melanoma
LOX, and pancreatic BxPC3 cells, as well as colon cancer
HCT116, DLD-1 and Colo-205 cells
[8, 9]
and other tumor cell
lines.
Although sorafenib has opened a window of hope after
decades of searching for effective agents to treat HCC, the
overall outcomes are far from satisfactory. Its side effects have
hampered its use; adverse events that have been reported
[10]
in patients receiving sorafenib are predominantly gastroin-
testinal, constitutional, or dermatologic in nature, including
diarrhea, weight loss, and hand–foot skin reactions. In severe
cases, it can lead to hypertension, abdominal pain and even
discontinuation of therapy. In addition, due to the genetic
heterogeneity of HCC, some patients are initially resistant
to sorafenib, which has led to the identication of predictive
biomarkers for primary resistance to sorafenib. It has recently
been reported that basal pERK levels, JNK and VEGFA may
615
www.chinaphar.com
Zhu YJ et al
Acta Pharmacologica Sinica
be candidate predictors of sorafenib response in HCC. The
acquired resistance to sorafenib has also drawn attention.
Several mechanisms are involved in the acquired resistance
to sorafenib, such as crosstalk involving the PI3K/Akt and
JAK-STAT pathways, the activation of hypoxia-inducible
pathways and epithelial-mesenchymal transition, as well as
others. Here, we attempt to describe the function of sorafenib,
its clinical application, and molecular mechanisms for drug
resistance, especially for HCC patients.
Clinical application in HCC
Monotherapy
Hepatocellular carcinoma (HCC) is the most common primary
malignant tumor of the liver; its annual diagnosis rate ranks
fth in cancers around the globe, and it is also the third major
leading cause of cancer-related death
[11]
. With nearly 700 000
new cases every year, China accounts for over 50% of newly
diagnosed HCC worldwide
[12]
. Because the number of HBV
carriers (>120 000 000) in China is large, chronic hepatitis-
induced cirrhosis and hepatocyte malignant transformation is
the major pathogenic factor for HCC. Due to the lack of effec-
tive targeted drugs, there are no standard adjunct therapies
following surgical resection at present. According to previ-
ous clinical trials, sorafenib displays signicant efcacy in the
treatment of advanced HCC patients.
In 2008, the SHARP (Sorafenib HCC Assessment Random-
ized Protocol) trial, a randomized controlled phase III trial
for the treatment of HCC in international multi-centers was
published in the New England Journal of Medicine
[1]
. In total,
602 advanced HCC patients who had not received systematic
treatment were randomly divided into two groups: one with
sorafenib treatment of doses of 400 mg, bid and the other as
placebo. The results showed that sorafenib was signicantly
associated with a higher median OS (10.7 versus 7.9 months;
P=0.0006) and a higher median TTP (5.5 versus 2.8 months;
P=0.000007) compared with the placebo group. This showed
that sorafenib as a rst-line drug for advanced HCC could pro-
long the median OS and radiological progression by 3 months
compared with the placebo
[13]
(Figure 2). Interestingly, Cheng
et al
[13]
reported that the Asian-Pacic portion of the phase III
Figure 1.
Cellular targets of sorafenib. Sorafenib blocks receptor tyrosine kinase signaling (VEGFR, PDGFR, c-Kit and RET) and inhibits downstream Raf
serine/threonine kinase activity to prevent tumor growth by anti-angiogenic, antiproliferative and/or pro-apoptotic effects
[10]
(from Bayer website).
Figure 2.
Kaplan-meier analysis of overall survival in the SHARP trial
[1]
.
616
www.nature.com/aps
Zhu YJ et al
Acta Pharmacologica Sinica
double-blind, randomized, controlled clinical trial of sorafenib
treatment also demonstrated that sorafenib could prolong
the OS of advanced HCC patients by 2.3 months, consistent
with the SHARP trial (Figure 3). Based on this result, the FDA
approved sorafenib for the treatment of unresectable HCC.
Combination with other drugs
Although single-agent sorafenib has an effect on HCC, its
adverse effects, such as rash, diarrhea, high blood pressure,
and hand-foot syndrome, limit high-dosage use of sorafenib
[14]
.
Therefore, it is necessary to combine sorafenib with other
drugs to lower its onset concentration. Additionally, although
sorafenib is a multi-targeted agent, it has a relatively single
mechanism compared with the complex growth mechanism
of tumors and needs to be used with other anti-tumor agents.
Both basic and clinical studies suggest that sorafenib in com-
bination with other medicines has a favorable effect and pro-
vides a new approach for clinical tumor treatment.
Capecitabine is an oral cytotoxic agent that has selective
activity against tumor cells and can be transformed to a cyto-
toxic drug called 5-uorouracil by adenylate in tumor tissues
to inhibit the phosphorylation of Akt and induce the expres-
sion of RKIP. Due to the over-expression of multi-resistant
and P-glycoprotein genes in HCC cells, they may have potent
resistance to capecitabine
[15]
. Sui et al
[16]
randomly divided 42
HCC patients into 2 groups, one subjected to a combination
of sorafenib and capecitabine at doses of S200 mg bid+C1500
mg/(m
2
·
d) and the other subjected to monotherapy with
capecitabine at the same dose. Those results showed that the
combination therapy could prolong the TTP (6.8 versus 4.3
months) and MST (10.9 versus 7.2 months) without additional
toxic reactions. In addition, Awada
et al
[17]
designed a series
of combined therapies with sorafenib and capecitabine and
found that the combination of sorafenib and capecitabine at
doses of 400 mg bid and 850 mg/d bid, respectively, achieved
a higher tumor inhibition rate with fewer adverse effects in the
treatment of advanced solid tumors, including HCC.
Doxorubicin, a type of anthracycline, can intercalate into
DNA and inhibit the synthesis of nucleic acids, and it has
broad-spectrum activity against tumors. Abou-Alfa et al phase
II clinical trial
[18]
suggested that the combination of sorafenib
with doxorubicin could synergistically inhibit the prolifera-
tion of tumor cells and the formation of new blood vessels in
tumor nodules through the “hyperchromic effect.” Further-
more, it is necessary to properly optimize the dose intensity of
doxorubicin because the AUC (the area under the curve) and
C
max
(maximum concentration) of doxorubicin in the human
body are limited
[19]
.
Fluorouracil has the ability to inhibit metabolites. It is
transformed into 5F-dUMP in tumor cells and can inhibit
thymidylate synthase (TS), disturb the synthesis of DNA, and
attenuate the proliferation of tumor cells. Petrini et al initi-
ated a study to evaluate the clinical potential of a combination
drug of sorafenib and low-dose uorouracil in the treatment
of 38 HCC patients, which showed that the mPF and OS were
9.6 and 12.2 months, respectively
[20]
. The drug had a positive
response and could effectively lower the dose intensity of uo-
rouracil and therefore reduce adverse effects.
Additionally, studies of the combination of sorafenib and
uracil-tegafur
[21]
or octreotide
[22]
reported a better curative
effect when they are used as a combination therapy than when
they are used as a monotherapy, which could reduce the dos-
age of sorafenib and the rate of side effects.
Combination with traditional treatment
According to recent studies, combining sorafenib with tradi-
tional treatment also has potential clinical utility. Sorafenib
can inhibit the synthesis of VEGF and the formation of new
blood vessels in tumors after surgical resection of HCC. It
also has been used to prevent the recurrence of tumors after
liver transplantation for HCC beyond the Milan criteria and
prolong the OS of patients with recurrent tumors. RFA (radio-
frequency ablation) is the most common clinical treatment for
localized physical ablation therapy of HCC, and it has several
advantages such as minimal invasion, quick recovery rate,
wide application and better repeatability
[23]
. Sun and his col-
leagues
[24]
found that the combination of RFA and sorafenib
could extend the median progression-free survival time fur-
ther than mono-RFA treatment for advanced-stage HCC (7.8
versus 4.6 months). Furthermore, patients receiving the drug
combination experienced an improvement in survival rate
within 6 months (80.0% versus 61.2%) and 12 months (53.3%
versus 30.4%) compared with the monotherapy. TACE (tran-
sarterial chemoembolization) is a well-used technique to treat
advanced HCC patients with no opportunity for surgical
resection. Chao et al
[25]
conducted a phase II clinical trial with
combination treatment of sorafenib and TACE for unresectable
HCC and found that the disease control rate was up to 91.2%
and that the combination treatment considerably increased the
survival time for moderate- and late-stage HCC patients.
However, one international study published recently in Lan-
cet Oncology concerning a randomized phase III clinical trial
(STORM)
[26]
surprised scholars both domestically and abroad.
This trial involved 1114 HCC patients who had already under-
Figure 3.
Kaplan-meier analysis of overall survival in the Asian-Pacific
trial
[13]
.
617
www.chinaphar.com
Zhu YJ et al
Acta Pharmacologica Sinica
gone resection and ablation. The STORM trial randomized
them into groups taking sorafenib or the placebo at 400 mg as
adjuvant therapy twice a day for 4 years. The result showed
that the RFS rates of the two groups were 33.4 and 33.8
months, which were considered statistically non-significant.
Their OS also demonstrated no difference. Thus, sorafenib
cannot improve the RFS and OS of HCC patients who have
already undergone resection and ablation. Additionally, as
patients have different hepatic injuries, adjuvant therapy with
sorafenib would cause frequent toxic reactions and intoler-
ance.
The results of the STORM trial challenged a wide range of
oncologists and raised many questions regarding the spe-
cific efficiency and target patients of sorafenib. The SHARP
and Asian-Pacific trials were all targeted at advanced HCC
patients who had not received systematic treatment to evalu-
ate the efficiency of sorafenib. However, the STORM trial
targeted patients who intended to receive curative resection or
localized ablation and those who were at an intermediate or
high risk of recurrence to evaluate the efciency of sorafenib
as an adjuvant therapy to prevent the recurrence of HCC. The
results showed that sorafenib as an adjuvant treatment did not
improve the RFS and OS of HCC patients who had previously
received curative treatment. Those three trials suggested that
more clinical trials with different subpopulations of HCC
patients and potential molecular classification markers are
needed to further explore the practical effect of sorafenib on
HCC. Additionally, there are ongoing studies investigating
the combination of sorafenib and TACE; the efficacy of this
treatment has not yet been demonstrated
[27, 28]
(Table 1). Alter-
natively, clinical trials of adjuvant treatment with sorafenib
in combination with ECOG1208 (NCT01004978), TACE-2
(NCT01324076) and TACTICS (NCT01217034) are still ongoing
(Table 1). Further research and joint efforts from clinicians are
also needed to implement combination therapy with sorafenib
and traditional treatment for advanced HCCs.
Mechanisms of drug resistance
Although sorafenib seems to be effective in prolonging median
survival time with limited side effects in HCC patients, it may
cause resistance in many patients, which has become an obsta-
cle to extending the overall survival time for HCC patients. In
addition, the results of the STORM trial suggest that sorafenib
cannot prolong the survival of liver cancer patients who have
accepted radical treatment, which posed a challenge to the
application of sorafenib. At present, there are studies on
the mechanisms of drug resistance to sorafenib, which may
include the following aspects.
Primary resistance
Because of the genetic heterogeneity of HCC, some HCC
cells and patients are initially resistant to sorafenib, which is
termed primary resistance
[29]
. However, the exact mechanism
remains unclear.
EGFR is the expression product of the proto-oncogene
c-erbB1, which resides on the surface of epithelial cells. After
binding with ligand, EGFRs can activate a series of down-
stream signaling pathways, thus regulating cell growth and
proliferation. More than half of HCC patients have EGFR
over-expression and abnormal activation
[30]
. Studies have
shown that the abnormal activation of EGFR/HER3 and the
overexpression of both EGFR and its ligand (especially double
adjustable ligand) can inhibit the antitumor effect of sorafenib.
Nonetheless, when the phosphorylation of EGFR/HER-3 was
inhibited by RNA interference and combined with sorafenib,
the anti-tumor proliferation and pro-apoptotic abilities of
sorafenib were improved
[31]
. Another report demonstrated
that EGFR activation may be the potential determinant of pri-
mary resistance of HCC cells to sorafenib, suggesting that the
overexpression of EGFR or ligand in HCC cells may lead to
Table 1.
Clinical trial of sorafenib combined with TACE for intermediate stage HCC.
*
The level of signicance for the primary endpoint TTP was set at
15%, which was a signicant result for sorafenib. However, the period of time until TACE ineligibility was 95 d for sorafenib and 224 d for placebo [HR
1.586 (95% CI 1.200–2.096), P=0.999], and as there was a good result with placebo, the efcacy of the TACE and sorafenib combination therapy was
not promising.
Trials Design Phase Primary endpoint
SPACE
[27]
Post-TACE
[28]
NCT010004978 (ECOG1208)
NCT01324076 (TACE-2)
NCT01217034 (TACTICS)
TACE+sorafenib vs TACE+placebo (TACE; DEB/scheduled)
TACE+sorafenib vs TACE+placebo (TACE; conventional/1 or
2 sessions)
TACE+sorafenib vs TACE+placebo (TACE; conventional or
DEB/scheduled)
TACE+sorafenib vs TACE+placebo (TACE; DEB/on demand)
TACE+sorafenib vs TACE alone (TACE; conventional/on
demand)
r-II
III
III
III
II
Median TTP 169 vs 166 d
HR 0.79, 95% CI 0.588–1.080, P=0.072
Median TTP 5.4 vs 3.7 months
HR 0.87, 95% CI 0.70–1.09, P=0.252
PFS
PFS
Time to untreatable progression
*
Negative
Ongoing
Ongoing
Ongoing
618
www.nature.com/aps
Zhu YJ et al
Acta Pharmacologica Sinica
sustained activation of EGFR downstream signaling and drug
resistance to sorafenib
[32]
.
In addition to EGFR and its ligand, downstream signaling
molecules, especially Ras/Raf/MEK/ERK, might contribute
to sorafenib resistance. In HCC, the Ras/Raf/MEK/ERK
signaling pathway is often activated, and the MAPK level
affects the sensitivity of HCC to sorafenib. Studies have found
that down-regulation of pERK may be associated with HCC
sorafenib resistance
[33]
. It has recently been reported that the
c-Jun N-terminal kinase (JNK), another member of the MAPK
family, can serve as a biomarker to predict the sensitivity to
sorafenib
[34]
.
VEGFR is also one of sorafenib’s cellular targets. Stud-
ies have shown that VEGFA stimulates paracrine secretion
of hepatocyte growth factor by stromal cells, which induces
tumor progression
[35]
. HCC patients with VEGFA amplifi-
cation are distinctly sensitive to sorafenib, indicating that
VEGFA might be a candidate predictor of sorafenib response
in HCC.
Thus, it is urgent to identify predictive biomarkers for pri-
mary resistance to sorafenib, and then apply the concept of
individualized treatment or seek therapeutic strategies such
as combining sorafenib with other anticancer agents to treat
HCC.
Acquired resistance
PI3K/Akt pathway and sorafenib resistance
PI3K/Akt is an important pathway involving cell apoptosis
and chemotherapeutic drug resistance. By inhibiting the
expression of Akt, we can sensitive cells to sorafenib-induced
apoptosis
[36]
. It has been reported that sorafenib can activate
SHP-1 and then negatively regulate the expression of pSTAT3
and inhibit the JAK/STAT signaling pathway. In acquired
drug-resistant HCC cell lines, we identied abnormal changes
in the JAK/STAT pathway, such as high expression of pSTAT3
and its downstream pro-apoptotic proteins, Mcl-1 and cyclin
D1, and lowered the expression of SHP-1 and pSHP-1
[37]
, sug-
gesting that sorafenib-related resistance may partly result
from the abnormal activation of STAT3.
Autophagy and sorafenib resistance
Autophagy is the body’s self-protective mechanism under var-
ious stress-induced signals, and currently, its role in HCC cells
is quite controversial. This mechanism may promote cancer
growth because it enables cells to survive nutrient deprivation.
Shimizu et al
[38]
found that sorafenib treatment led to accumu-
lation of autophagosomes and activation of autophagic flux,
as evidenced by increased LC3 lipidation and a clear decline
of the autophagy substrate p62 in Huh7, HLF and PLC/
PRF/5 cells, thus promoting HCC cell survival and limiting
sorafenib efficiency. However, by using chloroquine, 3-MA
or RNA interference that targets autophagy-related genes, the
anti-tumor effect of sorafenib was signicantly improved. In
addition, studies
have demonstrated that autophagy induced
by PSMD10, also known as gankyrin or p28, promotes tumor
progression
[39]
. PSMD10 augmented autophagic ux to resist
sorafenib or conventional chemotherapy, and inhibition of
autophagy suppressed PSMD10-mediated resistance.
Never-
theless, excessive stimulation may lead to programmed cell
death instead of survival
[40]
. Several studies have also shown
that excessive autophagy can promote apoptosis in tumor cells
and decrease tumor size. When sorafenib was combined with
pemetrexed, a folate anti-metabolite that stimulates autoph-
agy, the treatment increased the rate of autophagy and cell
death in vitro and suppressed tumor growth in vivo
[41]
. Thus,
autophagy can either enable cell survival or promote cell
death
[42]
, and further in-depth research is necessary for clari-
cation.
Epithelial-mesenchymal transition and sorafenib resistance
Epithelial-mesenchymal transition (EMT) can occur in tumor
cells under conditions of hypoxia and other stimulating
factors. In cancer, EMT is associated with poor patient sur-
vival
[43, 44]
because it is a key step in the development of metas-
tasis. In EMT, cell adhesion molecules (such as E-cadherin) are
lost, and mesenchymal cell markers such as vimentin (VIM)
are gained, resulting in the loss of polarity and cell-to-cell con-
tacts and enhancement of tumor cell migration and invasion.
Therefore, the tumor cells become more motile and insensitive
to antitumor drugs, including sorafenib
[45–47]
. In HCC, differ-
ent studies have demonstrated that sorafenib resistance mech-
anisms may involve EMT
[45, 48]
. In a study conducted by van
Malenstein and colleagues
[45]
, five resistant human liver cell
lines were developed through long-term exposure to sorafenib.
The cells changed in appearance, lost E-cadherin and KRT19
and showed high expression of vimentin, indicating epithelial-
to-mesenchymal transition. Resistant cells showed reduced
adherent growth, became more invasive and lost liver-specic
gene expression. However, the sensitivity to sorafenib after
development of resistance can partially be restored using
PI3K/Akt- or BCRP/Hedgehog-inhibitors in vitro. In addition
to hepatocytes, noncellular tumor components may also play
a role
[49]
. They manipulate hepatocellular carcinoma invasion
and metastasis by facilitating epithelial-mesenchymal transi-
tion, increasing the proteolytic activity of matrix metallopro-
teinases, and regulating antitumor immunity. Although the
exact mechanism between EMT and sorafenib resistance is still
unknown, and it is uncertain whether EMT is the trigger or
the result, further study investigating strategies for restoration
of sensitivity to sorafenib are needed.
Tumor microenvironment and sorafenib resistance
The tumor microenvironment plays an important role in the
occurrence and development of tumors. Anti-angiogenic
drugs can cause tumor blood vessel contraction and reduce
blood flow, resulting in a lack of oxygen to the tumor. It is
widely accepted that hypoxia in solid tumors is associated
with chemotherapy failure, selection of more invasive and
resistant clones, and poor prognosis
[50, 51]
. Hypoxic cells inside
solid tumors are extremely resistant to therapies, as their sur-
vival ability is increased due to the cellular adaptive response
to hypoxia, which is primarily controlled by hypoxia induc-