Halofuginone enhances the chemo-sensitivity of
cancer cells by suppressing NRF2 accumulation
Kouhei Tsuchida, Tadayuki Tsujita, Makiko
Hayashi, Asaka Ojima, Nadine Keleku-Lukwete,
Fumiki Katsuoka, Akihito Otsuki, Haruhisa
Kikuchi, Yoshiteru Oshima, Mikiko Suzuki, Masayuki Yamamoto
PII: S0891-5849(16)31142-X
DOI: http://dx.doi.org/10.1016/j.freeradbiomed.2016.12.041
Reference: FRB13147
To appear in: Free Radical Biology and Medicine
Received date: 12 September 2016
Revised date: 23 December 2016
Accepted date: 27 December 2016
Cite this article as: Kouhei Tsuchida, Tadayuki Tsujita, Makiko Hayashi, Asaka
Ojima, Nadine Keleku-Lukwete, Fumiki Katsuoka, Akihito Otsuki, Haruhisa
Kikuchi, Yoshiteru Oshima, Mikiko Suzuki and Masayuki Yamamoto, Halofuginone enhances the chemo-sensitivity of cancer cells by suppressing
NRF2 accumulation, Free Radical Biology and Medicine,

http://dx.doi.org/10.1016/j.freeradbiomed.2016.12.041

This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
www.elsevier.com
Halofuginone enhances the chemo-sensitivity of cancer cells by suppressing
NRF2 accumulation
Kouhei Tsuchida1,1, Tadayuki Tsujita*,1,2, Makiko Hayashi1
, Asaka Ojima1
, Nadine
Keleku-Lukwete1
, Fumiki Katsuoka3
, Akihito Otsuki4
, Haruhisa Kikuchi5
, Yoshiteru Oshima5
Mikiko Suzuki**,6 and Masayuki Yamamoto**,1, 3
1Department of Medical Biochemistry, Tohoku University Graduate School of Medicine,
Sendai 980-8575, Japan
2Department of Applied Biochemistry and Food Science, Saga University, Saga 840-8502,
Japan
Tohoku Medical Megabank Organization, Tohoku University, Sendai 980-8573, Japan
4Division of Medical Biochemistry, Faculty of Medicine, Tohoku Medical and Pharmaceutical
University, Sendai 981-8558, Japan
Laboratory of Natural Product Chemistry, Graduate School of Pharmaceutical Sciences,
Tohoku University, Sendai 980-8578, Japan
6Center for Radioisotope Sciences, Tohoku University Graduate School of Medicine, Sendai
980-8575, Japan
[email protected]
[email protected]
**To whom correspondence should be addressed: Masayuki Yamamoto, Department of
Medical Biochemistry, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi,
Aoba-ku, Sendai, Miyagi, 980-8575, Japan. Phone +81-22-717-8084; Fax +81-22-717-8090
**To whom correspondence should be addressed: Mikiko Suzuki, Center for Radioisotope
Sciences, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi, Aoba-ku,
Sendai, Miyagi, 980-8575, Japan. Phone +81-22-717-8088; Fax +81-22-717-8090
Abstract
The KEAP1-NRF2 system regulates the cellular defence against oxidative and xenobiotic
stresses. NRF2 is a transcription factor that activates the expression of cytoprotective genes
encoding antioxidative, detoxifying and metabolic enzymes as well as transporters. Under
normal conditions, KEAP1 represses NRF2 activity by degrading the NRF2 protein. When cells are exposed to stresses, KEAP1 stops promoting NRF2 degradation, and NRF2 rapidly

accumulates and activates the transcription of target genes. Constitutive accumulation of
NRF2 via a variety of mechanisms that disrupt KEAP1-mediated NRF2 degradation has been
observed in various cancer types. Constitutive NRF2 accumulation confers cancer cells with a
proliferative advantage as well as resistance to anti-cancer drugs and radiotherapies. To
suppress the chemo- and radio-resistance of cancer cells caused by NRF2 accumulation, we
conducted high-throughput chemical library screening for NRF2 inhibitors and identified
febrifugine derivatives. We found that application of the less-toxic derivative halofuginone in
a low dose range rapidly reduced NRF2 protein levels. Halofuginone induced a cellular amino
acid starvation response that repressed global protein synthesis and rapidly depleted NRF2.
Halofuginone treatment ameliorated the resistance of NRF2-addicted cancer cells to
anti-cancer drugs both in vitro and in vivo. These results provide preclinical proof-of-concept
evidence for halofuginone as an NRF2 inhibitor applicable to treatment of chemo- and
radio-resistant forms of cancer.
Keywords
NRF2, inhibitor, cancer, halofuginone, KEAP1
Introduction
Cancer cells often acquire defence systems against various endogenously and exogenously
generated stresses. While the dominant dogma concerning the essential cellular changes
occurring during preneoplasia and precancer stages regards these changes as abnormal or
foreign activities that evoke a basic host-parasite response, an alternative view of how cancer
develops considers these early and intermediate cellular changes as essentially physiologic
and adaptive [1]. This alternative concept introduces clonal adaptation as a basic response to
many genotoxic carcinogenic agents, including chemicals, radiation, and certain viruses. In
fact, this theory explains various phenomena, such as the observations that cancer cells acquire
constitutively active hypoxia-inducible factor (HIF) pathway activity to overcome hypoxic
stress [2] and upregulate programmed cell death-1 ligand-1 (PD-L1) expression to escape
from immune system attack [3]. Moreover, several cancer cells acquire defence mechanisms
against anti-cancer drug or radiation therapies. Resistance to chemotherapy and radiotherapy
is one of the major factors that worsens the prognosis for cancer patients. Importantly, one
cellular defence mechanism against these therapies adopted by several cancer cell types is
regulated by the Kelch-like ECH-associated protein 1 (KEAP1)- NF-E2-related factor like-2
(NRF2) system.
The KEAP1-NRF2 system serves as a cellular defence against oxidative and xenobiotic
stresses. NRF2 is a transcription factor that activates a set of cytoprotective genes in response
to such stresses [4]. In unstressed cells, the cysteine-rich protein KEAP1 serves as a sensor of
these stresses and strictly controls NRF2 activity [5]. KEAP1 is an adaptor protein connecting
NRF2 and Cullin-3 (CUL3), forming an ubiquitin E3 ligase complex [6]. Under normal
conditions, newly synthesized NRF2 protein is trapped by KEAP1 and degraded through the
ubiquitin-proteasome pathway, and NRF2 has a half-life of approximately 18 minutes or less
[7]. In contrast, cellular insults including exposure to reactive oxygen species (ROS) and
cytotoxic xenobiotics such as electrophiles modify cysteine residues and inactivate KEAP1.
Consequently, NRF2 escapes from rapid degradation, becomes stabilized and accumulates in
the nucleus.
In the nucleus, NRF2 forms heterodimers with small MAF (sMAF) proteins and activates
battery of cytoprotective genes, such as genes encoding antioxidative, detoxifying and
metabolic enzymes and transporters, by recognizing DNA sequences referred to as antioxidant
response elements (AREs) [8] or electrophile response elements (EpREs) [9], collectively
referred to as CNC-sMAF binding elements (CsMBEs) [10]. In addition to KEAP1-mediated
NRF2 degradation, a β-transducin repeat-containing protein (β-TrCP)-mediated pathway also
degrades NRF2 in the nucleus [11, 12]. This pathway is triggered by NRF2 phosphorylation
mediated by glycogen synthase kinase-3 (GSK-3) [13]. β-TrCP recruits phosphorylated NRF2
for degradation via the proteasome pathway within the nucleus. Notably, the β-TrCP-mediated
NRF2 degradation pathway plays an important role in NRF2 clearance, especially in cells in
which KEAP1-mediated NRF2 degradation is disrupted [14].
In several types of cancers, such as lung and oesophageal cancers, NRF2 protein has been
shown to constitutively accumulate in the nucleus. Accumulated NRF2 activates target genes
encoding antioxidative stress response proteins, detoxifying enzymes, glutathione
biosynthesis enzymes, ABC transporters and pentose phosphate pathway enzymes, which
enhance the resistance of tumours to chemo- and radio-therapies [15]. Furthermore, induction
of pentose phosphate pathway enzymes reprograms metabolism, thereby promoting cell
proliferation [16]. Indeed, patients with NRF2-accumulated cancer exhibit poor prognosis [17,
18]. Therefore, we hypothesize that NRF2 inhibitors are effective and necessary therapeutic
agents against cancer cells exhibiting NRF2 accumulation.
It has been shown that cancer cells induce constitutive NRF2 accumulation via multiple
mechanisms that lead to the disruption of KEAP1-mediated NRF2 degradation. First, NRF2
accumulates due to somatic mutations in the NFE2L2, KEAP1 and CUL3 genes encoding
NRF2, KEAP1 and CUL3, respectively [17, 19]. Loss-of-function somatic mutations in the
KEAP1 and CUL3 genes were frequently identified in lung adenocarcinoma specimens.
NFE2L2 gene mutations at its two sites of interaction with KEAP1 (i.e., in the DLG and ETGE
motifs of NRF2) have frequently been identified in lung squamous cell carcinoma and bladder
urothelial carcinoma samples [20]. Second, the phosphorylated form of p62, an
autophagy-related chaperone protein, inhibits the interaction between KEAP1 and NRF2,
leading to NRF2 accumulation in hepatocellular carcinoma [21, 22]. Third, fumarate, which is
known as an onco-metabolite, modifies cysteine residues of KEAP1 and activates NRF2 in
one type of renal cell carcinoma [23, 24]. In addition to these mechanisms, transcriptional
upregulation of NFE2L2 and downregulation of KEAP1 have been identified as causes of
NRF2 activation in cancer cells [25, 26]. Therefore, NRF2 inhibitors would be desirable for
the treatment of various types of malignant cancer cells harbouring accumulated NRF2 or
addicted to NRF2 activation.
In this study, to identify chemical compounds that inhibit NRF2 activity, we designed a
high-throughput screening system based on inhibition of NRF2 transcriptional activity using
an NRF2-addicted cancer cell line and conducted the chemical library screening. Utilizing this
system, we identified febrifugine and its derivatives as candidate NRF2 inhibitors.
Halofuginone is a less-toxic febrifugine derivative that rapidly suppresses the accumulation of
NRF2 protein in NRF2-addicted cancer cells. Halofuginone represses protein synthesis via an
amino acid starvation response elicited by inhibition of prolyl-tRNA synthetase. We found that
halofuginone treatment enhanced the sensitivity of NRF2-addicted cancer cells to anti-cancer
drugs both in vitro and in vivo. These results unequivocally demonstrate that halofuginone
serves as a chemo-sensitizer by inhibiting NRF2 accumulation in NRF2-addicted cancer cells.
Materials and methods
Reagents
Chemical compounds for screening were obtained from the Tohoku University Graduate
School of Pharmaceutical Sciences (kindly provided by Dr. Takayuki Doi). Halofuginone,
L-proline and doxorubicin were purchased from Sigma-Aldrich. Cisplatin was purchased
from Wako.
Cell culture
A549, KYSE70 and ABC1 cells were cultured in DMEM (Nacalai Tesque) supplemented with
10% activated-charcoal/resin-absorbed heat-inactivated fetal bovine serum (CF-FBS,
Equitech Bio) and antibiotics (100-U/mL penicillin and 100-μg/mL streptomycin). COR-L105
cells were cultured in RPMI-1640 medium (Nacalai Tesque) supplemented with 10% CF-FBS
and antibiotics. NCC16-P11 cells were cultured in MCDB153 medium (Sigma) supplemented
with 10% CF-FBS. BEAS-2B cells were cultured in BEGMTM bronchial epithelial cell growth
medium (Lonza) on dishes coated with bovine serum albumin, fibronectin and collagen. All
cells were cultured in a humidified atmosphere containing 5% CO2 at 37°C.
Generation of A549-ARE-Luc cells
The Nqo1-ARE-Luc reporter construct was described previously [27, 28]. To establish
A549-ARE-Luc cells, the Nqo1-ARE-Luc reporter plasmid and the pPGK-Puro plasmid [29]
were linearized and co-transfected into A549 cells using FuGENE6 transfection reagent
(Roche Molecular Biochemicals). Stable transformants were selected in DMEM containing
10% CF-FBS and 2-μg/mL puromycin.
siRNA transfection
NRF2 siRNA was purchased from Invitrogen (cat. no. HSS107128). The negative control
siRNA was purchased from IDT. Electroporation of siRNA into A549-ARE-Luc cells was
performed using the MP-100 MicroPorator (Digital Bio Technology).
High-throughput screening using A549-ARE-Luc cells
A549-ARE-Luc cells were seeded into white 384-well plates (Corning) and cultured for 16
hours. After incubation, the culture medium was replaced with fresh medium containing the
selected chemicals. After incubation for 24 hours, luciferase activity was measured using the
One-Glo luciferase assay system (Promega) with a PHERAstar FS microplate reader (BMG
Labtech). All plate and liquid handling was performed using a Biomek FXp
workstation
(Beckman Coulter) and a Multidrop Combi reagent dispenser (Thermo Fisher Scientific).
Luciferase activities of drug-treated samples were calculated by setting average value of
samples in each plate to 100%.
Cell viability assay
Cells were seeded into clear 384-well or 96-well plates and cultured for 12 hours. After
incubation, the culture medium was replaced with fresh medium containing the selected
chemicals. After culturing for 24-48 hours, the chemical- containing culture medium was
replaced with MTT assay medium (CellTiter 96 Aqueous One Solution; Promega) for 1-2
hours. After incubation, the OD490 of formazan dye from the MTT reagent was measured
using the PHERAstar FS microplate reader.
Immunoblotting
For analyses of cultured cells, whole-cell lysates were prepared by lysing the cells in SDS
buffer (0.25-M Tris-HCl (pH 6.8), 8% (w/v) SDS and 20% glycerol). For analyses of tumours,
nuclear lysate for immunoblot was prepared using NE-PER Nuclear and Cytoplasmic
Extraction Reagents (ThermoFisher Scientific). Protein concentrations were determined using
a BCA protein assay kit (Pierce) with bovine serum albumin as the standard. Anti-NRF2 [30],
anti-NQO1 [31], anti-α-tubulin (Sigma, T9026), anti-p-GCN2 (abcam, ab75836), anti-GCN2
(CST, 3302), anti-p-eIF2α (abcam, ab32157), anti-eIF2α (CST, 9722) and anti-lamin B
(Santa-Cruz, sc-6217) antibodies were used.
RNA extraction and reverse-transcription polymerase chain reaction (RT-PCR)
Total RNA was extracted with Sepazol RNA I Super G reagent (Nacalai Tesque) and
reverse-transcribed with a ReverTra Ace qPCR RT Kit (Toyobo) according to the
manufacturer’s instructions. The resulting cDNA was used as a template for quantitative
RT-PCR using Taqman or SYBR Green with a QuantStudio 6 Real-time PCR Analyzer
(Thermo Fisher Scientific). The primers used are described in Supplemental Table S1.
Analysis of protein synthesis
Cells were seeded in six-well plates and cultured to 80% confluence. Then, the cells were
treated with halofuginone or cycloheximide (CHX) for 15 minutes. The medium was removed,
and the cells were washed three times with PBS. Methionine-free medium was added for 30
minutes. The labelling compound L-azidohomoalanine (AHA, Thermo Fisher Scientific) was
added at a final concentration of 50 μM for 1 hour. Cells were washed three times with PBS,
and lysates were harvested in 200 μL per well of lysis buffer (1.0 % SDS, 50-mM Tris-HCl
(pH 8.0), and a protease inhibitor cocktail (Roche)). Click chemistry reactions were conducted
using 45 μg of input lysate, a biotinylated alkyne reagent (Thermo Fisher Scientific), and a
protein reaction buffer kit (Thermo Fisher Scientific) as per the manufacturer’s protocol.
Biotinylated proteins were analysed by Immunoblotting. Membranes were probed with an
anti-streptavidin-HRP antibody (Biolegend). Unprocessed lysates were also examined via
immunoblot and probed for α-tubulin expression.
Microarray analysis
Total RNA was prepared by using RNeasy Mini Kit (QIAGEN). The analysis was carried out
using whole human genome (4×44K) oligo microarray kit with SurePrint technology (Agilent
Technologies) following the manufacturer’s protocol. The expression data were normalized
and subjected to statistical analysis using GeneSpring software (Agilent Technologies). The
KEGG pathway analysis was performed using DAVID Bioinformatics Resource 6.8
http://david.abcc.ncifcrf.gov/). The KEGG pathway significantly enriched were defined as p
value < 0.05. The p values were corrected using Benjamini-Hochberg procedure.
Tumour xenograft experiment
KYSE70 cells (5×106
) with Matrigel (Corning) or A549 cells (1×107
) were subcutaneously
injected into the upper back region of 6-8-week-old male nude mice (BALB/C nu/nu mice).
The mice were treated with vehicle, halofuginone (0.25 mg/kg, every day), cisplatin (1.0
mg/kg, every other day) or both halofuginone and cisplatin. Tumour size was measured every
two days using calipers. The tumour volume was calculated using the following formula:
(major axis) × (minor axis) × (minor axis) × 0.5. Plasma levels of alanine aminotransferase
(ALT), alkaline phosphatase (ALP), blood urea nitrogen (BUN) and creatinine (CRE) were
measured utilizing a FUJI DRI-CHEM 7000 biochemical auto-analyser (Fujifilm). All mice
were handled according to the regulations of the Standards for Humane Care and Use of
Laboratory Animals of Tohoku University.
Statistical analysis
Data are presented as means ± standard deviation (S.D.). Student’s t-test (two-tailed) and
one-way ANOVA were used to calculate statistical significance (P), which is displayed as *
P<0.05 or ** P<0.01. Each cell viability curve was fitted to a four-parameter logistic equation.
Results
Identification of febrifugine derivatives as NRF2 inhibitors using a high-throughput
screening system
To discover chemical compounds that serve as NRF2 inhibitors, we developed a chemical
screening system based on the lung adenocarcinoma A549 cell line, which harbours a
homozygous KEAP1 loss-of-function mutation, resulting in constitutive accumulation of
NRF2 in the nucleus [32]. We utilized Nqo1-ARE-Luc, which is a luciferase reporter construct
driven by three copies of the regulatory module of the murine Nqo1 gene containing an ARE
or EpRE and the rabbit β-globin promoter, which was used to monitor NRF2 activity [27] (Fig.
1A). We introduced the reporter construct into A549 cells and isolated clone #8, which
exhibited high levels of luciferase activity. To assess whether luciferase activity in clone #8
was dependent of NRF2 activity, we transfected siRNA targeting NRF2 into clone #8 and
examined the luciferase activities of these cells. Luciferase activity was significantly
decreased in the NRF2 siRNA-transfected cells of clone #8, indicating that the luciferase
activity of clone #8 represents NRF2 activity (Fig. 1B). Therefore, we named clone #8 as
A549-ARE-Luc cells.
Through the screening system using A549-ARE-Luc cells, we screened 5,861 chemical
compounds accumulated by the Tohoku University Graduate School of Pharmaceutical
Sciences and identified eleven initial hit compounds exhibiting reduced luciferase activities to
less than 30% (Fig. 1C). Of the eight hit compounds, we found that febrifugine and its
derivative SH-168 were candidate NRF2 inhibitors (Fig. 1D). Febrifugine is a plant alkaloid
found in the roots of the blue evergreen hydrangea [33]. To validate the NRF2 inhibitory
activity and cellular toxicity of these two compounds, we examined their dose-dependent
effects on NRF2 activation and cell viability. We found that febrifugine and SH-168 repressed
luciferase activity in A549-ARE-Luc cells and that cell viability was not decreased by this

treatment, indicating that the inhibitory effects of these compounds were not caused by a
general toxic effect (Fig. 1E). These results indicate that febrifugine derivatives have NRF2
inhibitory activity.
Halofuginone decreases accumulation of the NRF2 protein level in cancer cells
One of the identified febrifugine derivatives, halofuginone, has been used as an antibiotic for
animals [34, 35]. Halofuginone is a racemic halogenated febrifugine derivative that was
artificially synthesized as a less-toxic form of febrifugine [36] (Fig. 2A). To examine whether
halofuginone inhibits NRF2, we treated human cancer cell lines exhibiting NRF2
accumulation with halofuginone and analysed the dose-dependent effects of this treatment on
the NRF2 protein levels in the nucleus. To this end, we utilized KYSE70 cells from human
oesophageal cancer harbouring a mutation in the NRF2 gene [15, 37] in addition to A549 cells
harbouring the KEAP1 gene mutation.
We found that the amount of NRF2 protein was significantly decreased in the presence of
halofuginone in both the KYSE70 and A549 cell lines (Fig. 2B). In contrast, total protein
levels were not substantially affected by halofuginone treatment (Fig. 2B). The half-maximal
inhibitory concentrations (IC50) of halofuginone for NRF2 were 22.3 and 37.2 nM in KYSE70
and A549 cells, respectively (Fig. 2C). NRF2 mRNA levels were not decreased, but rather
were increased after application of halofuginone to both KYSE70 and A549 cells (Fig. 2D),
indicating that halofuginone acts to decrease NRF2 protein expression at the
post-transcriptional level. Time-course studies showed that halofuginone decreased NRF2
protein expression in KYSE70 and A549 cells within 12 hours of treatment (Fig. 2E and 2F).
These results thus indicate that halofuginone depletes the NRF2 protein rapidly in a low
concentration range.
Halofuginone suppresses NRF2 protein synthesis via an amino acid starvation response
It has been shown that halofuginone directly binds to and inhibits prolyl-tRNA synthetase [38].
Halofuginone competitively inhibits prolyl-tRNA synthetase by occupying both the proline￾and tRNA-binding pockets of prolyl-tRNA synthetase [39, 40]. Supplementation of
prolyl-tRNA synthetase with excess L-proline restores the activity inhibited by halofuginone
[38]. To assess whether the observed NRF2-inhibitory activity of halofuginone is dependent
on inhibition of prolyl-tRNA synthetase activity, we added excess L-proline to KYSE70 and
A549 cell cultures and examined NRF2 protein levels. Of note, NRF2 inhibition was
abolished by L-proline treatment in both cell lines, supporting the notion that halofuginone
inhibits NRF2 via suppression of prolyl-tRNA synthetase activity (Fig. 3).
Halofuginone inhibits prolyl-tRNA synthetase and consequently induces the accumulation
of uncharged or non-aminoacylated tRNA, which activates the eIF2α-kinase GCN2 and
inactivates global translation [41-43]. This series of responses is referred to as the amino acid
starvation response [44]. To examine whether halofuginone induces the amino acid starvation
response in NRF2-addicted cancer cells, we analyzed phosphorylated forms of GCN2 and
eIF2α (p-GCN2 and p-eIF2α, respectively). We found that p-GCN2 and p-eIF2α were
increased within 1 hour of treatment (Fig. 4A). Of note, NRF2 protein levels also started to
decrease within 1 hour of treatment, suggesting that reduction of NRF2 protein by
halofuginone are associated with the amino acid starvation response (Fig. 4A).
To elucidate whether halofuginone blocks global protein translation, we performed
pulse-chase analysis of total proteins using a methionine analogue. The IC50 of halofuginone
for global protein synthesis was 22.6 and 45.7 nM in KYSE70 and A549 cells, respectively
(Fig. 4B-4E), and these values are comparable to the IC50 of halofuginone for NRF2 protein
levels in the nucleus (Fig. 2C). These results support the proposal that blocking protein
synthesis via the halofuginone-mediated amino acid starvation response rapidly decreases

NRF2 protein levels while maintaining the levels of most of the cellular proteins, given that
NRF2 is a very short-lived protein.
NRF2-addicted cancer cells are more susceptible to halofuginone treatment than normal
epithelial cells
To examine the toxic effects of halofuginone on normal epithelial cells, we treated
immortalized normal epithelial cells as well as NRF2-addictied cancer cells with halofuginone.
To this end, we used BEAS-2B and NCC16-P11 cells derived from human bronchial epithelial
cells and endocervical cells, respectively [45, 46]. We first confirmed that NRF2 protein was
not constitutively accumulated in BEAS-2B and NCC16-P11 cells (Fig. 5A). To analyse
toxicity of halofuginone, we treated these cells with halofuginone for 48 hours, which was
longer than incubation time for assay of NRF2 protein levels (24 hours), and examined cell
viabilities. The IC50 of halofuginone was 436.3 and 442.7 nM in BEAS-2B and NCC16-P11
cells, respectively (Fig. 5B, upper panels). On the other hand, the IC50 of halofuginone was
114.6 and 58.9 nM in KYSE70 and A549 cells (Fig. 5B, lower panels). These results indicate
that NRF2-addicted cancer cells are more susceptible to halofuginone treatment than
immortalized normal epithelial cells.
NRF2 target gene expression is decreased by halofuginone treatment
To analyse whether NRF2 target genes are downregulated by halofuginone treatment, we
performed microarray analysis using vehicle- and halofuginone-treated KYSE70. We
identified 2461 downregulated and 2329 upregulated genes by halofuginone treatment (Fig.
6A). We confirmed that genes associated with amino acid starvation response were
upregulated by halofuginone treatment (Fig. 6B). To annotate downregulated and upregulated
genes by halofuginone treatment, we preformed a KEGG pathway analysis and found that

glutathione metabolism was significantly enriched in genes downregulated by halofuginone
treatment (Fig. 6C). Furthermore, upon examination of the expression levels of known NRF2
target genes we found that genes encoding enzymes of glutathione synthesis and conjugation,
antioxidant enzymes, drug metabolizing enzymes and transporters, metabolic enzymes and
haem- and iron-metabolism enzymes were all suppressed by halofuginone (Fig. 6D). These
results indicate that multiple NRF2-mediated detoxification machineries are attenuated by the
halofuginone treatment. On the other hand, KEGG pathway analysis showed that both
“positive regulation of apoptosis process” and “negative regulation of cell proliferation” were
enriched in genes upregulated by the halofuginone treatment (Fig. 6C). These results may
explain the anti-cancer effects of halofuginone.
Halofuginone treatment abrogates the chemo-resistance of NRF2-addicted cancer cell
lines
As halofuginone treatment rapidly depletes NRF2 accumulation in NRF2-addicted cancer
cells, we hypothesized that halofuginone abrogates the resistance that NRF2-addicted cancer
cells have acquired against anti-cancer drugs. To test this hypothesis, we performed combined
administration of halofuginone and anti-cancer drugs. In addition, to determine whether the
anti-cancer effects of halofuginone depend on NRF2 inhibition, we used both NRF2-addicted
cancer cell lines (KYSE70 and A549) and non-NRF2-addicted cancer cell lines (ABC1 and
COR-L105). We first confirmed that the NRF2 protein was abundantly accumulated
specifically in KYSE70 and A549 cells but not in ABC1 or COR-L105 cells (Fig. 4A). The
expression levels of the NRF2 target gene NQO1 were also upregulated in KYSE70 and A549
cells but not in ABC1 or COR-L105 cells (Fig. 7A and 7B).
We next examined whether halofuginone enhances the cytotoxic effects of anti-cancer
drugs. To this end, we treated these cancer-derived cell lines with a combination of

halofuginone (50 nM) and anti-cancer drugs (cisplatin and doxorubicin). Notably, the
NRF2-addicted cancer cell lines exhibited higher IC50 of cisplatin and doxorubicin than the
non-NRF2-addicted cancer cell lines under conditions in which halofuginone was not applied
(vehicle treatment) (Fig. 7C and 7D). These results indicated that NRF2-addicted cancer cells
were resistant to anti-cancer agents (i.e., possessed chemo-resistance). Importantly, the IC50 of
cisplatin and doxorubicin was significantly reduced via halofuginone treatment specifically in
NRF2-addicted cancer cells. In contrast, non-NRF2-addicted cancer cell lines (ABC1 and
COR-L105 cells) exhibited comparable IC50 irrespective of halofuginone treatment. Moreover,
immortalized normal epithelial cells (BEAS-2B and NCC16-P11 cells) also exhibited
comparable IC50 irrespective of halofuginone treatment, indicating that halofuginone
treatment does not enhance toxicity of anti-cancer drugs in immortalized normal epithelial
cells (Supplemental Fig. S1A and S1B). These results thus demonstrate that halofuginone
treatment abrogates the chemo-resistance of NRF2-addicted cancer cells to anti-cancer drugs
via NRF2 inhibition.
Halofuginone enhances the anti-cancer effects of cisplatin in vivo
To assess whether halofuginone enhances the cytotoxic effects of anti-cancer drugs in vivo, we
established xenograft models by injecting KYSE70 cells into the flanks of nude mice. First,
we intraperitoneally injected 0.25-mg/kg halofuginone into the nude mice and confirmed that
NRF2 protein levels in tumours were indeed decreased by halofuginone treatments (Fig. 8A).
We then intraperitoneally injected vehicle, halofuginone (0.25 mg/kg, every day), cisplatin
(1.0 mg/kg, every other day) or both halofuginone and cisplatin into the nude mice. While the
tumour volumes did not change substantially between treatments with the vehicle,
halofuginone or cisplatin alone, combined treatment with halofuginone and cisplatin
significantly suppressed the tumour volume compared to treatment with halofuginone or

cisplatin alone (Fig. 8B). After 16 days of treatment with halofuginone and/or cisplatin, we
isolated tumours. The tumour size was reduced in the mice treated with the combination of
halofuginone and cisplatin compared to the mice treated with vehicle or with cisplatin alone
(Fig. 8C). These results indicate that halofuginone treatment increases the sensitivity of
tumours to cisplatin.
To examine whether halofuginone exhibits chemo-sensitizing effects for other
NRF2-addicted cancer cells, we performed a xenograft study using A549 cells. We injected
halofuginone into A549 xenograft tumours and confirmed that 0.25-mg/kg halofuginone
treatments decreased NRF2 protein levels (Supplemental Fig. S2A). We next intraperitoneally
injected vehicle, halofuginone (0.25 mg/kg, every day), cisplatin (2.0 mg/kg, every other day)
or both halofuginone and cisplatin into the nude mice. The combination of halofuginone and
cisplatin reduced tumour volumes efficiently also in A549 xenograft models compared to
vehicle, halofuginone or cisplatin alone (Supplemental Fig. S2B and S2C), indicating
chemo-sensitizing effects of halofuginone.
We also examined whether the halofuginone-mediated inhibition of NRF2 is toxic to
normal cells. To address this issue, we monitored body weight during halofuginone and
cisplatin treatment and observed an approximately 10% reduction in body weight among the
mice treated with the combination of halofuginone and cisplatin both in KYSE70 and A549
xenograft models (Fig. 8D and Supplemental Fig. S2D). To estimate tissue damage caused by
combined treatment with halofuginone and cisplatin, we examined indicators of liver and
renal function. The plasma levels of alanine aminotransferase (ALT), alkaline phosphatase
(ALP), and creatinine (CRE) in the mice treated with the combination of halofuginone and
cisplatin were comparable to those in vehicle-treated mice both in KYSE70 and A549
xenograft models (Fig. 8I and Supplemental Fig. S2E). While we observed an approximately
20% reduction in the plasma levels of blood urea nitrogen (BUN) among the mice treated with
halofuginone and cisplatin, these BUN values were within the normal range in mice [47].
These results indicate that halofuginone does not possess severe toxicity. Taken together,
halofuginone enhances the chemo-sensitivity of NRF2-addicted cancer cells to anti-cancer
drugs without increasing drug toxicity.

Discussion
Constitutive accumulation of NRF2 confers resistance to chemo- and radio-therapies,
resulting in poor prognosis among cancer patients [17, 18]. Here, we identified febrifugine and
its derivatives as NRF2 inhibitors using a high-throughput screen of the chemical library
established in the Tohoku University Graduate School of Pharmaceutical Sciences. We have
analysed one of these derivatives, halofuginone, which induces the amino acid starvation
response by inhibiting prolyl-tRNA synthetase activity and which results in the suppression of
global protein synthesis [38]. As NRF2 is a very short-lived protein, even in cancer cells in
which KEAP1-mediated degradation is disrupted, accumulated NRF2 protein is rapidly
depleted after halofuginone treatment, resulting in the reduction of cytoprotective enzyme
gene expression. Combined treatment with halofuginone and anti-cancer drugs enhanced the
cytotoxicity of the anti-cancer drugs specifically in NRF2-addicted cancer cell lines.
Moreover, halofuginone treatment promoted the anti-tumour effects of anti-cancer drugs in
vivo. As summarized in Figure 9, these results indicate that halofuginone is applicable for the
treatment of chemo- and radio-resistant forms of cancer associated with NRF2 activation.
Notably, halofuginone has previously been utilized as an antibiotic for animals [34, 35].
Recently, halofuginone has also been shown to exert anti-cancer effects via several
mechanisms. First, halofuginone inhibits angiogenesis in tumours by suppressing collagen
production in stromal cells [48, 49]. Second, halofuginone suppresses metastasis of cancer
cells by inhibiting matrix metalloproteinase-2 (MMP2) expression in tumours [50]. Third,
halofuginone suppresses tumour growth by inactivating TGFβ signalling [51]. Based on this
study, we add the important discovery that halofuginone blocks the cancer-promoting
functions of NRF2 activation. Halofuginone acts as a potent NRF2 inhibitor and abrogates the
chemo-resistance of NRF2-addicted cancer cells. As Phase II clinical trials of halofuginone
have been conducted for therapies of AIDS-related Kaposi sarcoma and fibrotic diseases [48,
52-55], our preclinical proof-of-concept evidence that halofuginone acts as anti-cancer
chemo-sensitizer by inhibiting NRF2 and suppressing chemo-resistance will promote future
clinical studies of treatment against NRF2-addicted chemo- and radio-resistant forms of
cancer.
We found that halofuginone inhibits NRF2 accumulation by suppressing global protein
synthesis via an amino acid starvation response. It should be noted that since the two cell lines
employed here harbour a somatic mutation in either KEAP1 (A549) or NRF2 (KYSE70), these
cells lack the KEAP1-NRF2 interaction [15, 32]. Consequently, NRF2 protein degradation via
the KEAP1-proteasome pathway should be minimal in these cells. Nonetheless, the NRF2
protein also possessed short half-life in KYSE70 and A549 cells. We surmise that NRF2
degradation in these cells depends on a KEAP1-independent pathway; one main NRF2
degradation pathway functioning in these cells is the b TrCP-proteasome pathway [11].
Consistent with the results of the present study, brusatol has been reported as an NRF2
inhibitor [56], and this compound is also an inhibitor of global protein synthesis [57]. As
NRF2 is a very short-lived protein even in cancer cells in which KEAP1-mediated degradation
is abolished, we surmise that blocking global protein synthesis may deplete NRF2
accumulation. Our present results show that halofuginone acts as an efficient NRF2 inhibitor
without inducing severe toxicity. We hypothesize that in terms of the inhibitory functions of
NRF2, there may be differences among the general protein synthesis inhibitors. However, this
hypothesis needs to be addressed further.
In contrast to the circumstances in NRF2-addicted cancer cells, NRF2 is well known to
protect normal cells against oxidative and xenobiotic stresses. Therefore, it appears reasonable
to raise concern that inhibiting NRF2 may worsen the cytotoxic effects of anti-cancer drugs on
normal cells. Here, we showed that the combination treatment of halofuginone and cisplatin
did not significantly increase the levels of tissue damage markers, indicating that halofuginone
treatment did not significantly enhance cytotoxicity. Nevertheless, an efficient drug delivery
system is important in delivering halofuginone to tumours. In addition, it has been shown that
NRF2 activation in the host, especially in myeloid-derived suppressor cells, suppresses
tumour progression and metastasis [58-60]. Therefore, one plausible hypothesis is that
specific administration of halofuginone to tumours in combination with anti-cancer agents and
concomitant systemic administration NRF2 inducers may achieve a maximal effect of cancer
chemotherapy.
In summary, it has been well recognized that acquisition of resistance to chemotherapy
worsens the prognosis of cancer patients and that some of the traditional anti-cancer agents
can encounter such resistance. Here, we have presented that inhibiting NRF2 using chemical
compounds weakens the chemo-resistance of cancer cells. Thus, NRF2 inhibitor
administration enables restoration of the use of well-established anti-cancer drugs. Therefore,
NRF2 inhibitors are expected to broaden treatment options and reduce medical expenditures.
Acknowledgments
We thank Drs. Sakae Saito, Akira Uruno and Yosuke Hirotsu, Ms. Satomi Gotoh, Sayoi
Inomata and Aya Goto, and the Biomedical Research Core of the Tohoku University Graduate
School of Medicine for providing technical support. This work was supported in part by
AMED-P-CREATE (to MY), AMED-P-DIRECT (to TT and MY) and AMED-CREST (to
MY) from the Japan Agency for Medical and Development (AMED), KAKENHI 26461395
from the Japan Society for the Promotion of Science (to MS), the NAITO Foundation,
Mitsubishi Life Science Foundation, the Takeda Science Foundation (to MY), the Platform
Project for Supporting in Drug Discovery and Life Science Research (Platform for Drug
Discovery, Informatics, and Structural Life Science) from the Ministry of Education, Culture,
Sports, Science and Technology (MEXT), and AMED (TT and MY).
References
[1] E. Farber, Is carcinogenesis fundamentally adversarial-confrontational or
physiologic-adaptive?, J Invest Dermatol 100(3) (1993) 251S-253S.
[2] G.L. Semenza, Targeting HIF-1 for cancer therapy, Nat Rev Cancer 3(10) (2003) 721-32.
[3] Y. Iwai, M. Ishida, Y. Tanaka, T. Okazaki, T. Honjo, N. Minato, Involvement of PD-L1 on
tumor cells in the escape from host immune system and tumor immunotherapy by PD-L1
blockade, Proc Natl Acad Sci U S A 99(19) (2002) 12293-7.
[4] K. Itoh, T. Chiba, S. Takahashi, T. Ishii, K. Igarashi, Y. Katoh, T. Oyake, N. Hayashi, K.
Satoh, I. Hatayama, M. Yamamoto, Y. Nabeshima, An Nrf2/small Maf heterodimer mediates
the induction of phase II detoxifying enzyme genes through antioxidant response elements,
Biochem Biophys Res Commun 236(2) (1997) 313-22.
[5] K. Itoh, N. Wakabayashi, Y. Katoh, T. Ishii, K. Igarashi, J.D. Engel, M. Yamamoto, Keap1
represses nuclear activation of antioxidant responsive elements by Nrf2 through binding to the
amino-terminal Neh2 domain, Genes Dev 13(1) (1999) 76-86.
[6] A. Kobayashi, M.I. Kang, H. Okawa, M. Ohtsuji, Y. Zenke, T. Chiba, K. Igarashi, M.
Yamamoto, Oxidative stress sensor Keap1 functions as an adaptor for Cul3-based E3 ligase to
regulate proteasomal degradation of Nrf2, Mol Cell Biol 24(16) (2004) 7130-9.
[7] K. Itoh, N. Wakabayashi, Y. Katoh, T. Ishii, T. O’Connor, M. Yamamoto, Keap1 regulates
both cytoplasmic-nuclear shuttling and degradation of Nrf2 in response to electrophiles,
Genes Cells 8(4) (2003) 379-91.
[8] T.H. Rushmore, M.R. Morton, C.B. Pickett, The antioxidant responsive element.
Activation by oxidative stress and identification of the DNA consensus sequence required for
functional activity, J Biol Chem 266(18) (1991) 11632-9.
[9] R.S. Friling, A. Bensimon, Y. Tichauer, V. Daniel, Xenobiotic-inducible expression of
murine glutathione S-transferase Ya subunit gene is controlled by an electrophile-responsive
element, Proc Natl Acad Sci U S A 87(16) (1990) 6258-62.
[10] A. Otsuki, M. Suzuki, F. Katsuoka, K. Tsuchida, H. Suda, M. Morita, R. Shimizu, M.
Yamamoto, Unique cistrome defined as CsMBE is strictly required for Nrf2-sMaf heterodimer
function in cytoprotection, Free Radic Biol Med 91 (2016) 45-57.
[11] P. Rada, A.I. Rojo, S. Chowdhry, M. McMahon, J.D. Hayes, A. Cuadrado,
SCF/{beta}-TrCP promotes glycogen synthase kinase 3-dependent degradation of the Nrf2
transcription factor in a Keap1-independent manner, Mol Cell Biol 31(6) (2011) 1121-33.
[12] M. Salazar, A.I. Rojo, D. Velasco, R.M. de Sagarra, A. Cuadrado, Glycogen synthase
kinase-3beta inhibits the xenobiotic and antioxidant cell response by direct phosphorylation
and nuclear exclusion of the transcription factor Nrf2, J Biol Chem 281(21) (2006) 14841-51.
[13] S. Chowdhry, Y. Zhang, M. McMahon, C. Sutherland, A. Cuadrado, J.D. Hayes, Nrf2 is
controlled by two distinct β-TrCP recognition motifs in its Neh6 domain, one of which can be
modulated by GSK-3 activity, Oncogene 32(32) (2013) 3765-81.
[14] K. Taguchi, I. Hirano, T. Itoh, M. Tanaka, A. Miyajima, A. Suzuki, H. Motohashi, M.
Yamamoto, Nrf2 enhances cholangiocyte expansion in Pten-deficient livers, Mol Cell Biol
34(5) (2014) 900-13.
[15] T. Shibata, A. Kokubu, S. Saito, M. Narisawa-Saito, H. Sasaki, K. Aoyagi, Y. Yoshimatsu,
Y. Tachimori, R. Kushima, T. Kiyono, M. Yamamoto, NRF2 mutation confers malignant
potential and resistance to chemoradiation therapy in advanced esophageal squamous cancer,
Neoplasia 13(9) (2011) 864-73.
[16] Y. Mitsuishi, K. Taguchi, Y. Kawatani, T. Shibata, T. Nukiwa, H. Aburatani, M.
Yamamoto, H. Motohashi, Nrf2 redirects glucose and glutamine into anabolic pathways in
metabolic reprogramming, Cancer Cell 22(1) (2012) 66-79.
[17] T. Shibata, T. Ohta, K.I. Tong, A. Kokubu, R. Odogawa, K. Tsuta, H. Asamura, M.
Yamamoto, S. Hirohashi, Cancer related mutations in NRF2 impair its recognition by
Keap1-Cul3 E3 ligase and promote malignancy, Proc Natl Acad Sci U S A 105(36) (2008)
13568-73.
[18] D. Inoue, T. Suzuki, Y. Mitsuishi, Y. Miki, S. Suzuki, S. Sugawara, M. Watanabe, A.
Sakurada, C. Endo, A. Uruno, H. Sasano, T. Nakagawa, K. Satoh, N. Tanaka, H. Kubo,
Motohashi, M. Yamamoto, Accumulation of p62/SQSTM1 is associated with poor prognosis
in patients with lung adenocarcinoma, Cancer Sci 103(4) (2012) 760-6.
[19] B. Padmanabhan, K.I. Tong, T. Ohta, Y. Nakamura, M. Scharlock, M. Ohtsuji, M.I. Kang,
A. Kobayashi, S. Yokoyama, M. Yamamoto, Structural basis for defects of Keap1 activity
provoked by its point mutations in lung cancer, Mol Cell 21(5) (2006) 689-700.
[20] C. Kandoth, M.D. McLellan, F. Vandin, K. Ye, B. Niu, C. Lu, M. Xie, Q. Zhang, J.F.
McMichael, M.A. Wyczalkowski, M.D. Leiserson, C.A. Miller, J.S. Welch, M.J. Walter, M.C.
Wendl, T.J. Ley, R.K. Wilson, B.J. Raphael, L. Ding, Mutational landscape and significance
across 12 major cancer types, Nature 502(7471) (2013) 333-9.
[21] M. Komatsu, H. Kurokawa, S. Waguri, K. Taguchi, A. Kobayashi, Y. Ichimura, Y.S. Sou,
I. Ueno, A. Sakamoto, K.I. Tong, M. Kim, Y. Nishito, S. Iemura, T. Natsume, T. Ueno, E.
Kominami, H. Motohashi, K. Tanaka, M. Yamamoto, The selective autophagy substrate p62
activates the stress responsive transcription factor Nrf2 through inactivation of Keap1, Nat
Cell Biol 12(3) (2010) 213-23.
[22] Y. Ichimura, S. Waguri, Y.S. Sou, S. Kageyama, J. Hasegawa, R. Ishimura, T. Saito,
Yang, T. Kouno, T. Fukutomi, T. Hoshii, A. Hirao, K. Takagi, T. Mizushima, H. Motohashi,
M.S. Lee, T. Yoshimori, K. Tanaka, M. Yamamoto, M. Komatsu, Phosphorylation of p62
activates the Keap1-Nrf2 pathway during selective autophagy, Mol Cell 51(5) (2013) 618-31.
[23] J. Adam, E. Hatipoglu, L. O’Flaherty, N. Ternette, N. Sahgal, H. Lockstone, D. Baban, E.
Nye, G.W. Stamp, K. Wolhuter, M. Stevens, R. Fischer, P. Carmeliet, P.H. Maxwell, C.W.
Pugh, N. Frizzell, T. Soga, B.M. Kessler, M. El-Bahrawy, P.J. Ratcliffe, P.J. Pollard, Renal

cyst formation in Fh1-deficient mice is independent of the Hif/Phd pathway: roles for
fumarate in KEAP1 succination and Nrf2 signaling, Cancer Cell 20(4) (2011) 524-37.
[24] A. Ooi, J.C. Wong, D. Petillo, D. Roossien, V. Perrier-Trudova, D. Whitten, B.W. Min,
M.H. Tan, Z. Zhang, X.J. Yang, M. Zhou, B. Gardie, V. Molinié, S. Richard, P.H. Tan, B.T. Teh,
K.A. Furge, An antioxidant response phenotype shared between hereditary and sporadic type
2 papillary renal cell carcinoma, Cancer Cell 20(4) (2011) 511-23.
[25] G.M. DeNicola, F.A. Karreth, T.J. Humpton, A. Gopinathan, C. Wei, K. Frese, D. Mangal,
K.H. Yu, C.J. Yeo, E.S. Calhoun, F. Scrimieri, J.M. Winter, R.H. Hruban, C.
Iacobuzio-Donahue, S.E. Kern, I.A. Blair, D.A. Tuveson, Oncogene-induced Nrf2
transcription promotes ROS detoxification and tumorigenesis, Nature 475(7354) (2011)
106-9.
[26] R. Wang, J. An, F. Ji, H. Jiao, H. Sun, D. Zhou, Hypermethylation of the Keap1 gene in
human lung cancer cell lines and lung cancer tissues, Biochem Biophys Res Commun 373(1)
(2008) 151-4.
[27] M.S. Yates, M. Tauchi, F. Katsuoka, K.C. Flanders, K.T. Liby, T. Honda, G.W. Gribble,
D.A. Johnson, J.A. Johnson, N.C. Burton, T.R. Guilarte, M. Yamamoto, M.B. Sporn, T.W.
Kensler, Pharmacodynamic characterization of chemopreventive triterpenoids as
exceptionally potent inducers of Nrf2-regulated genes, Mol Cancer Ther 6(1) (2007) 154-62.
[28] Y. Hirotsu, F. Katsuoka, K. Itoh, M. Yamamoto, Nrf2 degron-fused reporter system:
new tool for specific evaluation of Nrf2 inducers, Genes Cells 16(4) (2011) 406-15.
[29] K.L. Tucker, C. Beard, J. Dausmann, L. Jackson-Grusby, P.W. Laird, H. Lei, E. Li, R.
Jaenisch, Germ-line passage is required for establishment of methylation and expression
patterns of imprinted but not of nonimprinted genes, Genes Dev 10(8) (1996) 1008-20.
[30] A. Maruyama, S. Tsukamoto, K. Nishikawa, A. Yoshida, N. Harada, K. Motojima, T. Ishii,
A. Nakane, M. Yamamoto, K. Itoh, Nrf2 regulates the alternative first exons of CD36 in
macrophages through specific antioxidant response elements, Arch Biochem Biophys 477(1)
(2008) 139-45.
[31] M. McMahon, K. Itoh, M. Yamamoto, J.D. Hayes, Keap1-dependent proteasomal
degradation of transcription factor Nrf2 contributes to the negative regulation of antioxidant
response element-driven gene expression, J Biol Chem 278(24) (2003) 21592-600.
[32] A. Singh, V. Misra, R.K. Thimmulappa, H. Lee, S. Ames, M.O. Hoque, J.G. Herman, S.B.
Baylin, D. Sidransky, E. Gabrielson, M.V. Brock, S. Biswal, Dysfunctional KEAP1-NRF2
interaction in non-small-cell lung cancer, PLoS Med 3(10) (2006) e420.
[33] G.R. COATNEY, W.C. COOPER, W.B. CULWELL, W.C. WHITE, C.A. IMBODEN,
Studies in human malaria. XXV. Trial of febrifugine, an alkaloid obtained from Dichroa
febrifuga lour., against the Chesson strain of Plasmodium vivax, J Natl Malar Soc 9(2) (1950)
183-6.
[34] I. Villacorta, J.E. Peeters, E. Vanopdenbosch, E. Ares-Mazás, H. Theys, Efficacy of
halofuginone lactate against Cryptosporidium parvum in calves, Antimicrob Agents
Chemother 35(2) (1991) 283-7.
[35] M. Naciri, R. Mancassola, P. Yvoré, J.E. Peeters, The effect of halofuginone lactate on
experimental Cryptosporidium parvum infections in calves, Vet Parasitol 45(3-4) (1993)
199-207.
[36] J.F. Ryley, M.J. Betts, Chemotherapy of chicken coccidiosis, Adv Pharmacol Chemother
11 (1973) 221-93.
[37] Y. Shimada, M. Imamura, T. Wagata, N. Yamaguchi, T. Tobe, Characterization of 21
newly established esophageal cancer cell lines, Cancer 69(2) (1992) 277-84.
[38] T.L. Keller, D. Zocco, M.S. Sundrud, M. Hendrick, M. Edenius, J. Yum, Y.J. Kim, H.K.
Lee, J.F. Cortese, D.F. Wirth, J.D. Dignam, A. Rao, C.Y. Yeo, R. Mazitschek, M. Whitman,
Halofuginone and other febrifugine derivatives inhibit prolyl-tRNA synthetase, Nat Chem
Biol 8(3) (2012) 311-7.
[39] V. Jain, M. Yogavel, Y. Oshima, H. Kikuchi, B. Touquet, M.A. Hakimi, A. Sharma,
Structure of Prolyl-tRNA Synthetase-Halofuginone Complex Provides Basis for Development
of Drugs against Malaria and Toxoplasmosis, Structure 23(5) (2015) 819-29.
[40] H. Zhou, L. Sun, X.L. Yang, P. Schimmel, ATP-directed capture of bioactive herbal-based
medicine on human tRNA synthetase, Nature 494(7435) (2013) 121-4.
[41] H.P. Harding, I. Novoa, Y. Zhang, H. Zeng, R. Wek, M. Schapira, D. Ron, Regulated
translation initiation controls stress-induced gene expression in mammalian cells, Mol Cell
6(5) (2000) 1099-108.
[42] H.P. Harding, Y. Zhang, H. Zeng, I. Novoa, P.D. Lu, M. Calfon, N. Sadri, C. Yun, B.
Popko, R. Paules, D.F. Stojdl, J.C. Bell, T. Hettmann, J.M. Leiden, D. Ron, An integrated
stress response regulates amino acid metabolism and resistance to oxidative stress, Mol Cell
11(3) (2003) 619-33.
[43] R.C. Wek, H.Y. Jiang, T.G. Anthony, Coping with stress: eIF2 kinases and translational
control, Biochem Soc Trans 34(Pt 1) (2006) 7-11.
[44] C. Deval, C. Chaveroux, A.C. Maurin, Y. Cherasse, L. Parry, V. Carraro, D. Milenkovic,
M. Ferrara, A. Bruhat, C. Jousse, P. Fafournoux, Amino acid limitation regulates the
expression of genes involved in several specific biological processes through
GCN2-dependent and GCN2-independent pathways, FEBS J 276(3) (2009) 707-18.
[45] R.R. Reddel, Y. Ke, B.I. Gerwin, M.G. McMenamin, J.F. Lechner, R.T. Su, D.E. Brash,
J.B. Park, J.S. Rhim, C.C. Harris, Transformation of human bronchial epithelial cells by
infection with SV40 or adenovirus-12 SV40 hybrid virus, or transfection via strontium
phosphate coprecipitation with a plasmid containing SV40 early region genes, Cancer Res
48(7) (1988) 1904-9.
[46] Y. Ohta, K. Tsutsumi, K. Kikuchi, S. Yasumoto, Two distinct human uterine cervical
epithelial cell lines established after transfection with human papillomavirus 16 DNA, Jpn J
Cancer Res 88(7) (1997) 644-51.
[47] M.A. Schnell, C. Hardy, M. Hawley, K.J. Propert, J.M. Wilson, Effect of blood collection
technique in mice on clinical pathology parameters, Hum Gene Ther 13(1) (2002) 155-61.
[48] M. Elkin, I. Ariel, H.Q. Miao, A. Nagler, M. Pines, N. de-Groot, A. Hochberg, I.
Vlodavsky, Inhibition of bladder carcinoma angiogenesis, stromal support, and tumor growth
by halofuginone, Cancer Res 59(16) (1999) 4111-8.
[49] O. Zion, O. Genin, N. Kawada, K. Yoshizato, S. Roffe, A. Nagler, J.L. Iovanna, O. Halevy,
M. Pines, Inhibition of transforming growth factor beta signaling by halofuginone as a
modality for pancreas fibrosis prevention, Pancreas 38(4) (2009) 427-35.
[50] M. Elkin, R. Reich, A. Nagler, E. Aingorn, M. Pines, N. de-Groot, A. Hochberg, I.
Vlodavsky, Inhibition of matrix metalloproteinase-2 expression and bladder carcinoma
metastasis by halofuginone, Clin Cancer Res 5(8) (1999) 1982-8.
[51] P. Juárez, K.S. Mohammad, J.J. Yin, P.G. Fournier, R.C. McKenna, H.W. Davis, X.H.
Peng, M. Niewolna, D. Javelaud, J.M. Chirgwin, A. Mauviel, T.A. Guise, Halofuginone
inhibits the establishment and progression of melanoma bone metastases, Cancer Res 72(23)
(2012) 6247-56.
[52] M. Pines, A. Nagler, Halofuginone: a novel antifibrotic therapy, Gen Pharmacol 30(4)
(1998) 445-50.
[53] T. McGaha, T. Kodera, R. Phelps, H. Spiera, M. Pines, C. Bona, Effect of halofuginone on
the development of tight skin (TSK) syndrome, Autoimmunity 35(4) (2002) 277-82.
[54] M. Pines, D. Snyder, S. Yarkoni, A. Nagler, Halofuginone to treat fibrosis in chronic
graft-versus-host disease and scleroderma, Biol Blood Marrow Transplant 9(7) (2003) 417-25.
[55] H.B. Koon, B. Fingleton, J.Y. Lee, J.T. Geyer, E. Cesarman, R.A. Parise, M.J. Egorin, B.J.
Dezube, D. Aboulafia, S.E. Krown, Phase II AIDS Malignancy Consortium trial of topical
halofuginone in AIDS-related Kaposi sarcoma, J Acquir Immune Defic Syndr 56(1) (2011)
64-8.
[56] D. Ren, N.F. Villeneuve, T. Jiang, T. Wu, A. Lau, H.A. Toppin, D.D. Zhang, Brusatol
enhances the efficacy of chemotherapy by inhibiting the Nrf2-mediated defense mechanism,
Proc Natl Acad Sci U S A 108(4) (2011) 1433-8.
[57] S. Vartanian, T.P. Ma, J. Lee, P.M. Haverty, D.S. Kirkpatrick, K. Yu, D. Stokoe,
Application of Mass Spectrometry Profiling to Establish Brusatol as an Inhibitor of Global
Protein Synthesis, Mol Cell Proteomics 15(4) (2016) 1220-31.
[58] K. Hiramoto, H. Satoh, T. Suzuki, T. Moriguchi, J. Pi, T. Shimosegawa, M. Yamamoto,
Myeloid lineage-specific deletion of antioxidant system enhances tumor metastasis, Cancer
Prev Res (Phila) 7(8) (2014) 835-44.
[59] H. Satoh, T. Moriguchi, K. Taguchi, J. Takai, J.M. Maher, T. Suzuki, P.T. Winnard, V.
Raman, M. Ebina, T. Nukiwa, M. Yamamoto, Nrf2-deficiency creates a responsive
microenvironment for metastasis to the lung, Carcinogenesis 31(10) (2010) 1833-43.
[60] H. Satoh, T. Moriguchi, D. Saigusa, L. Baird, L. Yu, H. Rokutan, K. Igarashi, M. Ebina, T.
Shibata, M. Yamamoto, NRF2 Intensifies Host Defense Systems to Prevent Lung
Carcinogenesis, but After Tumor Initiation Accelerates Malignant Cell Growth, Cancer
Figure legends
Figure 1. Identification of febrifugine derivatives as NRF2 inhibitors using a
high-throughput screening system. (A) Schematic representation of the high-throughput
screening system used to analyse A549-ARE-Luc cells (clone #8). (B) Luciferase activities in
A549-ARE-Luc cells (clone #8) treated with control or NRF2 siRNA. (C) Luciferase activities
in A549-ARE-Luc cells treated with one of 5,861 chemical compounds accumulated at the
Tohoku University Graduate School of Pharmaceutical Sciences. (D) Chemical structures of
febrifugine (left panel) and SH-168 (right panel). (E) Luciferase activities and cell viabilities
of A549-ARE-Luc cells exposed to the indicated concentrations of febrifugine (left panel) or
SH-168 (right panel) for 24 hours.
Figure 2. Halofuginone decreases the accumulation of NRF2 protein in NRF2-addicted
cancer cells. (A) Chemical structure of halofuginone (HF). (B) The NRF2 and total protein
levels in whole-cell lysates of KYSE70 (upper panel) and A549 cells (lower panel) exposed to
the indicated concentrations of HF for 24 hours. (C) Quantification of NRF2 protein levels in
whole-cell lysates of KYSE70 and A549 cells exposed to the indicated concentrations of
halofuginone for 24 hours (n=3). The NRF2 protein levels were normalized to the α-tubulin
levels. IC50 is indicated. (D) Quantification of the NRF2 mRNA levels in KYSE70 and A549
cells exposed to the indicated concentrations of HF for 24 hours (n=4). NRF2 mRNA levels
were normalized to the HPRT mRNA levels. (E) NRF2 protein levels in KYSE70 and A549
cells exposed to 50 or 100 nM HF, respectively, for the indicated periods. (F) Quantification of
NRF2 protein levels in whole-cell lysates of KYSE70 and A549 cells exposed to 50 or 100 nM
HF, respectively, for the indicated periods (n=3). NRF2 protein levels were normalized to the
α-tubulin levels. Graphed data are presented as means ± SD.
Figure 3. Halofuginone inhibits NRF2 via suppression of prolyl-tRNA synthetase activity.
NRF2 protein levels in KYSE70 and A549 cells exposed to vehicle, proline, halofuginone
(HF) or a combination of halofuginone and L-proline (Pro) for 24 hours (n=3). NRF2 protein
levels were normalized to the α-tubulin levels. Graphed data are presented as means ± SD.
*P<0.05, **P<0.01.
Figure 4. Halofuginone suppresses NRF2 protein synthesis via an amino acid starvation
response. (A) Protein levels of phosphorylated GCN2, total GCN2, phosphorylated eIF2α,
total eIF2α and NRF2 in KYSE70 and A549 cells exposed to halofuginone (HF) for the
indicated periods (n=3). Protein levels were normalized to the α-tubulin levels. (B-E) Newly
synthesized protein levels in KYSE70 (B and C) and A549 cells (D and E) exposed to HF or
cycloheximide (CHX, as a positive control) for 1 hour (n=3). The cells were pretreated with
HF at the indicated concentrations for 15 minutes. Protein levels were normalized to the
α-tubulin levels. IC50 is indicated. Graphed data are presented as means ± SD.
Figure 5. NRF2-addicted cancer cells are more susceptible to halofuginone treatment
than immortalized normal epithelial cells. (A) NRF2 and NQO1 protein levels in
NRF2-addicted cancer cell lines (KYSE70 and A549) and immortalized normal epithelial
cells (BEAS-2B and NCC16-P11). Protein levels were normalized to the α-tubulin levels. (B)
Viabilities of normal epithelial cells (BEAS-2B and NCC16-P11) and NRF2-addicted cancer
cell lines (KYSE70 and A549) exposed to the indicated concentrations of halofuginone for 48
hours (n=3). IC50 is indicated. Graphed data are presented as means ± SD.
Figure 6. NRF2 target gene expression is decreased by halofuginone treatment. (A)
Scatter plots comparing transcript levels in vehicle- and halofuginone (HF)- treated KYSE70

cells. (B) Heat map showing the expression levels of genes associated with amino acid
starvation response (AAR). The colours of the heat map reflect log2-transformed transcript
abundance relative to the median expression level in each group, ranging from 3.0 (magenta)
to –3.0 (cyan). (C) KEGG biological pathways enriched in downregulated and upregulated
genes by HF treatment. (D) Heat map showing the expression levels of NRF2 target genes.
The colours of the heat map are shown in the same manner as in panel B.
Figure 7. Halofuginone ameliorates the resistance of NRF2-addicted cancer cells to
anti-cancer drugs. (A) NRF2 and NQO1 protein levels in NRF2-addicted cancer cell lines
(KYSE70 and A549) and in non-NRF2-addicted cancer cell lines (ABC1 and COR-L105)
(n=3). (B) The mRNA levels of the NRF2 target gene NQO1, which were normalized to the
HPRT mRNA levels (n=3). (C-D) Viabilities of KYSE70, A549, ABC1 and COR-L105 cells
exposed to the indicated concentrations of cisplatin (C) or doxorubicin (D) with or without 50
nM HF for 48 hours (n=3). IC50 is indicated. Graphed data are presented as means ± SD.
Figure 8. Halofuginone enhances the anti-cancer effects of cisplatin in vivo. (A) NRF2
protein levels in tumours of KYSE70 xenograft models treated with vehicle or halofuginone
(HF). (B) Tumour volume growth curve for xenograft model mice in the vehicle-,
halofuginone (HF)-, cisplatin- and combination-treated groups (n=5-6). (C) Representative
images of the excised tumours on day 16. (D) Body weight curve of the xenograft model mice
in the vehicle-, HF-, cisplatin- and combination-treated groups (n=3). (E) ALT, ALP, BUN and
CRE levels in plasma from xenograft model mice in the vehicle-, HF-, cisplatin- and
combination-treated groups (n=3). Graphed data are presented as means ± SD.
Figure 9. Schema of chemo-sensitization of NRF2-addicted cancer cells due to
halofuginone treatment. In NRF2-addicted cancer cells, KEAP1-mediated NRF2
degradation is disrupted. Therefore, NRF2 accumulates in the nucleus and activates target
genes, which participate in the chemo-resistance of cancer cells. Halofuginone inhibits
prolyl-tRNA synthetase (PRS), resulting in accumulation of uncharged tRNA. Accumulated
uncharged tRNA induces the amino acid starvation response and suppresses global protein
translation. As NRF2 is a short-lived protein, accumulated NRF2 is rapidly depleted via
halofuginone treatment, which confers cancer cells with enhanced sensitivity to anti-cancer
drugs such as cisplatin and doxorubicin.
Highlights
· High-throughput screening identified febrifugine derivatives as NRF2 inhibitors.
· The febrifugine derivative halofuginone (HF) suppresses NRF2 protein accumulation.
· HF represses protein synthesis via an amino acid starvation response.
· HF enhances the sensitivity of NRF2-addicted cancer cells to anti-cancer drugs.
· HF serves as a chemo-sensitizer of NRF2-addicted cancer cells.