Targeting lipid metabolism in multiple myeloma cells: Rational development of a synergistic strategy with proteasome inhibitors

Article abstract of DOI:10.1111/bph.15653

Background and Purpose: Aberrant lipid metabolism is recognized as a key feature of cancer cells. Our initial research on MS-based analysis of lipids in a multiple myeloma (MM) cell line showed a significant accumulation of lipids in multiple myeloma cells after proteasome inhibition. This finding prompted us to hypothesize that multiple myeloma cell survival depends on the maximal utilization of abnormally accumulated lipids. Therefore, we explored whether lipid metabolism-modulating agents would synergize with proteasome inhibitors. Experimental Approach: Lipid accumulation in multiple myeloma cells was measured by MS. Synergism between lipid regulators and proteasome inhibitors was assessed by cell viability and apoptosis. A novel stable derivative of fenofibrate (FCE) was synthesized and used to treat multiple myeloma cells in vitro and in vivo along with the proteasome inhibitor ixazomib. ChIP-seq, western blotting and RT-qPCR were performed to explore the potential mechanism(s) underlying the increase in lipid levels in multiple myeloma cells after proteasome inhibition. Key Results: Accumulation of lipids in multiple myeloma cells was induced by proteasome inhibition. Lipid-lowering drugs and MG-132 exerted a synergistic effect to kill multiple myeloma cells. FCE showed significant synergistic activity in vitro and in vivo with ixazomib. The abnormal lipid accumulation in multiple myeloma cells that was enhanced by proteasome inhibitors might be due to the elevated SREBP1/2 expression induced by ATF4. Conclusions and Implications: Our results provide a proof of principle and support for the further clinical evaluation of the combination of lipid-modulating drugs with proteasome inhibitors in the treatment of multiple myeloma.

Full text of DOI:10.1111/bph.15653

Yang Tai (Orcid ID: 0000-0002-2145-7435)
Targeting lipid metabolism in multiple myeloma cells: rational development of a synergistic
strategy with proteasome inhibitors
Running Title: Targeting lipid metabolism in multiple myeloma cells
Gaojie Xua1, Sheng Huanga1, Jian Penga1, Xiaofang Gaoa, Minhui Lia, Sisi Yuc, Zuofeng Liud,
Pengfan Qiea, Yu Wanga, Siqi Yua, Siyuan Liua, Hu Wena, Lijuan Sua, Ping Lia, Bin Guanga,b,
Renhan Donga,b, Jin Liua*, and Tai Yanga*
a School of Pharmacy, Chengdu Medical College, No. 783 Xindu Avenue, Xindu District,
Chengdu, Sichuan Province, China, 610500
b Chengdu Biobel Biotechnology Co., Ltd., No. 88 Keyuan South Road, New and High-tech
Zone, Chengdu, Sichuan Province, China, 610094
c Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, School of Medicine, University
of Electronic Science and Technology of China, Chengdu, Sichuan Province, China, 610041
d Department of Hematology, The Third People’s Hospital of Chengdu, Chengdu, Sichuan
Province, China, 610031
1 Contributed equally to this study.
Corresponding authors: Jin Liu and Tai Yang
Mailing address: Chengdu Medical College, No. 783 Xindu Avenue, Xindu District, Chengdu,
Sichuan Province, China, 610500
Tel.: +86-028-62739161; Fax: +86-028-62739161; Mobile phone: 13402839935
E-mail addresses: jinliu@cmc.edu.cn and taiyang@cmc.edu.cn
Data availability statement: The data that support the findings of this study are available from
the corresponding author upon reasonable request. Some data may not be made available
because of privacy or ethical restrictions.
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/bph.15653
This article is protected by copyright. All rights reserved.

Abstract
Background and purpose: Aberrant lipid metabolism is now recognized as a key feature of
cancer cells. Our initial research on mass spectrometry-based analysis of lipids in a multiple
myeloma (MM) cell line showed a significant accumulation of lipids in MM cells after
proteasome inhibition. This finding prompted us to hypothesize that MM cell survival depends
on the maximal utilization of abnormally accumulated lipids. Therefore, we explored whether
lipid metabolism-modulating agents would synergize with proteasome inhibitors (PIs).
Experimental approach: The abnormal massive lipid accumulation in MM cells was detected
using mass spectrometry. Cell viability and cell apoptosis were detected to assess the
synergistic effect of lipid regulators and PIs. Otherwise, a novel stable derivative (FCE) of
fenofibrate (FEN) was synthesized and used to treat MM cells in vitro and in vivo along with
ixazomib. ChIP-seq, western blotting and RT-qPCR were performed to explore the potential
mechanism underlying the increase in lipid levels in MM cells after proteasome inhibition.
Key results: The accumulation of lipids in MM cells was induced by proteasome inhibition.
Lipid-lowering drugs and MG-132 exerted a synergistic effect to kill MM cells. FCE showed
significant synergistic activity in vitro and in vivo with ixazomib. The abnormal lipid
accumulation in MM cells that was enhanced by proteasome inhibitors might be due to the
elevated SREBP1/2 expression induced by ATF4.
Conclusions and implications: In summary, the results provide a proof of principle and
rationale for the further clinical evaluation of the combination of lipid-modulating drugs with
proteasome inhibitors.
Keywords: Multiple myeloma, Proteasome inhibitors, Lipid, ATF4
Abbreviations: MM, multiple myeloma; PIs, proteasome inhibitors; FEN, fenofibrate; FCE,
A novel FEN cyclohexyl ester; FEC, fenofibric acid; ER, endoplasmic reticulum; EVs,
extracellular vesicles; Ixa, ixazomib citrate; CCK-8, cell counting kit-8 assay; CI, combination
index; PBS, phosphate-buffered saline; PI, propidium iodide; PVDF, polyvinylidene fluoride;
TBST, Tris-buffered saline containing 0.05% Tween-20; HRP, horseradish peroxidase; SCID,
severe combined immunodeficiency; i.g., intragastric gavage; HP-β-CD, hydroxypropyl beta
cyclodextrin; OPLS-DA, orthogonal projections to latent structures discriminant analysis; PCA,
principal component analysis; PLS-DA, partial least-squares discriminant analysis; TG,
triglycerides; PE, phosphatidylethanolamine; PS, phosphatidylserine; PA, phosphatidic acid;
UPR, unfolded protein response; ChIP-seq, chromatin immunoprecipitation followed by
sequencing; SREBP, sterol regulatory element binding protein; Fato, fatostatin; Orl, orlistat.
This article is protected by copyright. All rights reserved.

1. Introduction
According to recent studies, highly proliferative cancer cells take advantage of exogenous
lipids and lipoproteins or endogenous lipid synthesis to satisfy their strong requirements for
lipids and cholesterol (Beloribi-Djefaflia, Vasseur & Guillaumond, 2016; Tirinato et al., 2017).
A significant upregulation of several lipogenic enzymes has been observed in multiple types of
human cancers and is closely related to tumor progression. For example, the overexpression of
fatty acid synthase (FASN) induced by activated EGFR/ERK in pancreatic ductal
adenocarcinoma plays a role in tumor growth (Bian, Yu, Wang & Li, 2015); the upregulation
of ATP-citrate lyase by low-molecular-weight isoforms of cyclin E improves lipid droplet
formation, thus providing breast cancer cells with essential building blocks to support growth
(Lucenay et al., 2016). Furthermore, reprogramming of lipid metabolism occurs in many cancer
cells or in patients with cancer after treatment with drugs or resistance to anticancer agents. In
tamoxifen-resistant breast cancer cells, the levels of neutral lipids in lipid droplets were
significantly increased, and free cholesterol accumulated in the lysosomes to improve tumor
survival (Hultsch et al., 2018a). A striking upregulation of FASN was observed in pancreatic
cell lines with poor responsiveness to gemcitabine; therefore, the FASN inhibitor orlistat
showed considerable synergistic activity with gemcitabine to promote the apoptosis of
gemcitabine-resistant pancreatic cells (Tadros et al., 2017). Free fatty acid (FA) uptake was
improved in tumors, particularly in adipose-associated tumors under the pressure of
antiangiogenic drugs (Iwamoto et al., 2018b).
Recently, an increasing number of studies have suggested that lipid metabolism also plays
a potential role in the treatment of multiple myeloma (MM). Compared to statin nonusers with
MM, the risk of developing MM was reduced by 20-28% in patients using statins for 48-72
months prior to the diagnosis/index date (Epstein et al., 2017); MM-specific mortality and all-
cause mortality were decreased by 24% and 21%, respectively, in patients with MM receiving
statin therapy (Sanfilippo et al., 2016). In addition, the proliferation of MM cells was inhibited
after the decrease in β-oxidation and de novo FA synthesis induced by etomoxir and orlistat
(Tirado-Velez, Joumady, Saez-Benito, Cozar-Castellano & Perdomo, 2012). The inhibition of
acetyl-CoA-carboxylase expression, which is essential for FA synthesis, might be a useful
adjuvant therapy to treat MM (Masarwi, DeSchiffart, Ham & Reagan, 2019). Based on these
studies, lipid metabolism is important for the adaptive survival of MM cells, which was
eliminated by intervening in the pathways for lipid metabolism.
As a hematological tumor characterized by the malignant proliferation of clonal plasma
cells, MM cells secrete and accumulate numerous proteins, including misfolded and unfolded
This article is protected by copyright. All rights reserved.

proteins, triggering endoplasmic reticulum (ER) stress (Marino, Petrusca & Roodman, 2019;
Tabas & Ron, 2011). Hence, MM cells strongly rely on the ubiquitin–proteasome pathway,
which degrades damaged proteins that are toxic to cells. Proteasome inhibitors (PIs) disrupt the
balance of the ER-associated degradation system by inhibiting the activity of proteasomes,
resulting in excess ER stress and subsequent apoptosis induction in MM cells. In addition, MM
cells must generate a large number of extracellular vesicles (EVs) that secrete large amounts
of immunoglobulin into the extracellular space, which might relieve ER stress for MM cells to
maintain adaptive survival (Cenci & Sitia, 2007; Di Noto et al., 2014). Lipids are key
components in the biogenesis and functioning of EVs (Brzozowski et al., 2018; Record,
Silvente-Poirot, Poirot & Wakelam, 2018). Therefore, early findings support the hypothesis
that the large amount of undegraded protein induced by proteasome inhibition triggers an
increase in intracellular lipid levels to allow damaged proteins to be secreted and relieve ER
stress. Thus, strategies targeting the abnormal lipid metabolism induced by proteasome
inhibition may exert a significant synergistic effect on killing MM cells treated with PIs.
Here, we explored lipid metabolism in MM cells before and after proteasome inhibition
using nontargeted metabolomics, as detected with mass spectrometry. Based on the abnormal
lipid accumulation after proteasome inhibition, the present study elucidated the synergistic
effects of a combination of PIs and lipid regulators on inducing the death of MM cells. The
present study also identified potentially relevant genes and pathways involved in the abnormal
lipid metabolism induced by proteasome inhibition. This study proposes a therapeutic theory
and strategies designed to exploit the lipid-related metabolic dependence in MM to hopefully
develop a novel anti-MM strategy.
2. Materials and methods
2.1 Cell lines, culture, and reagents
The MM.1S and RPMI8226 cell lines were purchased from the American Type Culture
Collection (ATCC) and cultured in RPMI 1640 medium supplemented with 10% fetal bovine
serum, streptomycin (100 μg·mL-1), and penicillin (100 U·mL-1) in a humidified incubator at
37°C with 5% CO2. Ixazomib citrate (Ixa) was purchased from Selleck Chemicals (Houston,
TX, USA). The internal standard D5-TG (20:0-20:1-20:0, CAS: 958764-14-2) was obtained
from Sigma Aldrich (St. Louis, MO, USA). Fenofibrate (FEN) was obtained from Sigma
Aldrich (. A novel FEN cyclohexyl ester (FCE) was synthesized as described in the Supporting
Information (Method S1.1).
This article is protected by copyright. All rights reserved.

2.2 Nontargeted lipidomics detection
In the MG-treated group, 1×107 MM.1S cells (cell density, 1×106 cells mL-1) were treated
with the proteasome inhibitor MG-132 (100 nM final concentration). MM.1S cells in the ML-
treated group were simultaneously treated with 100 nM MG-132 and 15 μM lovastatin. MM.1S
cells in the MFL-treated group were simultaneously exposed to 100 nM MG-132, 15 μM
lovastatin, and 8 μM FEN. Equal amounts of MM.1S cells were mock-treated with vehicle
(DMSO, negative control group). Each set contained three replicate samples. After
approximately 24 h of drug treatment (cell viability was monitored at different time points
within 24 h), the cells were centrifuged at 300 × g for 5 min, and metabolites were extracted
from the cell pellet using the protocol described in a previous study (Sellick, Hansen, Stephens,
Goodacre & Dickson, 2011). A brief protocol is provided in the Supporting Information
(Method S2). The reverse-phase chromatographic separation conditions and the mass
spectrometry conditions are described in the Supporting Information (Methods 1.3 1.4 and
Tables S1 and S2). The methods for processing nontargeted lipid group data are described in
the Supporting Information (Method S4).
2.3 Cell proliferation assay
The cell counting kit-8 assay (CCK-8) bioassay was used to analyze MM cell proliferation
after treatment with MG-132 or lovastatin. Briefly, MM.1S and RPMI8226 cells were seeded
at a density of 40000 cells/well (2×105 cells mL-1) in 96-well plates, and then cells were
exposed to various concentrations of MG-132 (40 to 10000 nM) or lovastatin (200 to 50000
nM). An equal volume of DMSO was added to the other wells as a control. Following treatment
with the agent for 48 h, 10 μL of the CCK-8 solution were added to the culture medium and
incubated for an additional 2 h before the absorbance of the cells was measured at 450 nm. Cell
viability was calculated using the following formula: viability (% of control cells) = treated
cells × 100/control cells.
2.4 In vitro synergy test
The effect of the combination of MG-132 with a chemotherapeutic agent on proliferation
was assessed by calculating combination index (CI) values using CalcuSyn software (Biosoft,
Cambridge, UK). The drug treatments were administered individually or in combination at a
fixed ratio. Based on algorithms developed by Chou and Talalay, the CI provides a quantitative
measure of the degree of interaction between multiple agents. CI values < 1, = 1, and > 1 denote
synergistic, additive and antagonistic effects, respectively (Chou & Talalay, 1984).
This article is protected by copyright. All rights reserved.

2.5 Apoptosis detection
After the treatment of MM cells with either the control vehicle or various concentrations
of drugs for 8 h, the cells were washed twice with phosphate-buffered saline (PBS) and
resuspended in 100 μLof binding buffer containing annexin V-FITC and propidium iodide (PI).
After a 15-min incubation, the cells were acquired and analyzed using a NovoCyte cytometer
(ACEA Biosciences, CA, USA), and the data were analyzed using Novo Express (ACEA
Biosciences) for the presence of an annexin-positive/PI-negative apoptotic cell population.
2.6 Western blot analysis
The immuno‐related procedures used comply with the recommendations made by the
British Journal of Pharmacology (Alexander et al., 2018). Equal amounts of total cellular
proteins (100 μg/lane) were resolved by SDS-polyacrylamide gel electrophoresis and
transferred to polyvinylidene fluoride (PVDF) membranes (Roche Company, Basel,
Switzerland). Membranes were blocked with 5% milk dissolved in Tris-buffered saline
containing 0.05% Tween-20 (TBST) and probed with rabbit monoclonal anti-caspase-3
(1:1000; Cell Signaling Technology Cat# 14220, RRID: AB_2798429), rabbit monoclonal anti-
cleaved caspase-3 (1:1000; Cell Signaling Technology Cat# 9664, RRID: AB_2070042)),
rabbit monoclonal anti-PARP and anti-cleaved PARP (1:1000; Cell Signaling Technology Cat#
9532, RRID: AB_659884), anti-activating transcription factor-4 (ATF4) (1:1000; Cell
Signaling Technology Cat# 11815, RRID: AB_2616025), mouse monoclonal anti-SREBP1
(1:1000; Santa Cruz Biotechnology Cat# sc-56483, RRID: AB_831733), mouse monoclonal
anti-SREBP2 (1:1000; Santa Cruz Biotechnology Cat# sc-271615, RRID: AB_10709730),
rabbit anti-β-actin (1:1000; Cell Signaling Technology Cat# 8457, RRID: AB_10950489), and
rabbit anti-GAPDH (1:1000; Cell Signaling Technology Cat# 2118, RRID: AB_561053)
antibodies, followed by horseradish peroxidase (HRP)-conjugated secondary antibodies, and
then developed using a chemiluminescence-based detection system.
2.7 Animal model and in vivo study
The Ethics Commission of Chengdu Medical College approved the animal studies
described here. 24 five-week-old female CB17 mice with severe combined immunodeficiency
(SCID) were obtained from Weitonglihua Company in Beijing, China. Mice were housed for
1-2 weeks to allow them to adapt to the environment before the experiment started. A total of
2 × 107 MM.1S cells in 100 μL of PBS and 100 µL of Matrigel basement membrane matrix
(Corning, NY, USA) were subcutaneously injected into the right flank of CB-17 SCID mice,
This article is protected by copyright. All rights reserved.

and tumor growth was monitored using calipers. When the tumors grew to approximately 1000
mm3 in diameter, mice were randomized into an Ixa-treated group [Ixa group, n=8, 7 mg·kg-1,
intragastric gavage (i.g.), twice a week], an Ixa- and FEN-treated group (Ixa and FEN group,
n=8; Ixa: 7 mg·kg-1, i.g., twice a week; FEN: 300 mg·kg-1, i.g., once a day), and an Ixa- and
FCE-treated group (Ixa and FCE group, n=8; Ixa: 7 mg·kg-1, i.g., twice a week; FCE: 300
mg·kg-1, i.g., once a day). Ixa was formulated in 10% hydroxypropyl beta cyclodextrin (HP-β-
CD), and FEN and FCE were dissolved in 100% olive oil. Treatment was continued for up to
20 days from the date of randomization (day 0). Tumor sizes were measured at 2-day intervals.
Tumor volumes were calculated using a standard formula: (shortest diameter)2 × (longest
diameter) × 0.5. At the end of this experiment, mice were euthanized by cervical dislocation.
2.8 RNA analysis
The total RNA was extracted from the treated MM cells using RNAiso Plus TRIzol
(Takara, Japan) according to the manufacturer’s instructions. Then, the RNA was reverse
transcribed to cDNAs using the PrimeScript™ RT Reagent Kit (Takara) under the following
conditions: 42°C for 2 min to remove genomic DNA; 37°C for 15 min and 85°C for 5 s. Real-
time PCR quantification was performed using SYBR Green Master Mix (Takara) with the CFX
96 Real-time system (Bio-Rad, Hercules, CA). Real-time signals were normalized to GAPDH,
as indicated. The following primers were used to detect ATF4, SREBP1/2, and GAPDH: ATF4
(5’-ATGACCGAAATGAGCTTCCTG-3’
SREBP1 (5’-CGGAACCATCTTGGCAACAGT-3’ and 3’-CGCTTCTCAATGGCGTTGT-5’);
SREBP2 (5’-CCTGGGAGACATCGACGAGAT-3’ and 3’-
and
3’-GCTGGAGAACCCATGAGGT-5’);
TGAATGACCGTTGCACTGAAG-5’); and GAPDH (5’-TGCACCACCAACTGCTTAGC-3’
and 3’-GGCATGGACTGTGGTCATGAG-5’).
2.9 ChIP and ChIP-seq
MM.1S cells (1×106 cells per mL, 5×107 cells total) were plated, MG-132 (800 nM) was
added, and DMSO was used as a control. After 6 h of drug treatment, the cells were centrifuged
at 300 × g for 5 min and resuspended in 20 mL of fresh complete RPMI 1640 medium. Cells
were fixed with formaldehyde (final concentration of 1%) for 10 min at room temperature in a
shaker and quenched with 0.125ꢀM glycine for 5ꢀmin at room temperature. Cells were washed
twice with precooled PBS. The remaining experimental procedures were performed as
described in a previous study (Donato et al., 2017). Briefly, lysates were incubated overnight
at 4°C with Protein ADynabeads coupled with 5 μg of antibody (anti-ATF4, as described above)
This article is protected by copyright. All rights reserved.

after chromatin shearing by sonication. Beads were recovered using a magnet and washed after
immunoprecipitation. DNA was eluted, and cross-links were reversed at 65°C for 4 h and then
purified with a Qiagen kit. DNA was quantitated using the Qubit® dsDNA HS Assay Kit with
a Qubit® 2.0 Fluorimeter (Invitrogen, Carlsbad, CA). For ChIP-seq, 5 ng of purified ChIP
DNA were used to generate the sequencing library with a NEB kit and sequenced with the
Illumina HiSeq X Ten system. The processing and analysis of the ChIP-seq data are described
in the Supporting Information (Method S3 and Table S3).
2.10 Statistical analyses
The data and statistical analyses comply with the recommendations of the British Journal
of Pharmacology (Curtis et al., 2018). Statistical analyses were performed using Prism 7
software (GraphPad Software, La Jolla, CA, USA). Each result is presented as the mean ± SEM.
The differences between the means of the control group and treatment group were determined
using t-tests, and differences with a probability less than 0.05 were considered significant.
3. Results
3.1 PIs induce abnormal lipid accumulation in MM cells
MG-132 and control groups were completely separated in the orthogonal projections to
latent structures discriminant analysis (OPLS-DA) score plots (Fig. S1), indicating that lipid
metabolism was significantly altered in MM.1S cells. Then, an unsupervised principal
component analysis (PCA) model was applied to obtain an overview of variance between the
profiles of lipid species. As shown in Fig. 1A, the MG-treated groups and negative control
groups were significantly separated in the PCA model, indicating that the lipid metabolism of
the two groups differed substantially. Next, the partial least-squares discriminant analysis
(PLS-DA) score plot revealed that MG-treated and negative control groups were clearly
separated by PLS component 1 and PLS component 2, which explained 83% and 9.7% of the
variance, respectively (Fig. 1B). These results imply that various changes in lipidomic profiles
occur after proteasome inhibition. We constructed a heat map that provided the relative average
concentrations of the selected 106 lipid metabolites in the MG-treated and negative control
groups to further identify the potential roles of these lipid species (Fig. 1C). Among these
metabolites, the intracellular levels of 103 lipids increased significantly after MM.1S cells were
treated with a PI, and only three lipids were decreased. Similarly, the volcano plot analysis
(Fig. 1D) revealed that specific lipids shown in red were present at levels above the threshold
(FC > 2, and P < 0.05), and almost no substances were observed on the left side of the volcano
This article is protected by copyright. All rights reserved.

plot, indicating that lipids significantly accumulated in MM.1S cells after proteasome
inhibition. The overall increases in changes in the lipid profile between the MG-treated and
negative control groups are shown in the pie-doughnut chart in Fig. 1E. Triacylglycerides
(known also as triglycerides, TG) constituted 53.47% of the overall lipids, followed by the
bound fatty acid phosphatidyl ethanolamine. Taken together, the PIs induced a striking
accumulation of lipids. The aberrant lipid metabolism plays important roles in cancer cell
invasion and in the induction of tumor angiogenesis. These features prompted us to explore
whether lipid metabolism-modulating agents synergize with PIs.
3.2 Lovastatin synergizes with PIs
Lovastatin inhibits the proliferation of MM cells in a concentration-dependent manner.
The 50% growth inhibitory concentrations (GI50s) of lovastatin in MM.1S and RPMI8226
cells were 35.2 and 27.1 μM, respectively (Fig. 2A). This result was consistent with a previous
study showing that statins trigger apoptosis in certain cancer cells, including MM cells
(Mollinedo et al., 2010; Wong et al., 2007). As shown in Fig. 2B, the GI50s of MG-132 in
MM.1S and RPMI8226 cells were 54.0 and 165.0 nM, respectively. As expected, the inhibition
of cell proliferation increased significantly after treatment with the combination of MG-132
and lovastatin (Fig. 2C). This effect was evident even at a low concentration of lovastatin (15
μM for MM.1S cells and 10 μM for RPMI8226 cells), revealing a slight inhibitory effect on
MM cell proliferation. A CI analysis was performed to determine whether the combination of
lovastatin and MG-132 resulted in synergistic cytotoxic effects on MM.1S and RPMI8226 cells.
As shown in Fig. 2C, a synergistic interaction between lovastatin and MG-132 was indeed
observed in MM cells (CI＜1). As shown in Fig. S2, the addition of lovastatin also increased
the percentage of apoptotic cells compared with MG-132 alone. Similarly, the addition of
lovastatin resulted in the apparent generation of apoptosis markers, such as cleaved caspase-3
and PARP (Fig. S2B). These results implied that the accumulated lipids induced by the PI might
confer acquired resistance to MM cells; thus, lovastatin might exert synergistic antitumor-
proliferative effects with PIs by reducing intracellular lipid levels. Consequently, we compared
the changes in the overall lipid profiles of PI-treated MM cells and MM cells treated with the
combination of PIs and lovastatin.
3.3 Changes in the lipid profile induced by lovastatin
The OPLS-DA score plots (Fig. S3), PCA analysis (Fig. 3A), and PLS-DA score plots
(Fig. 3B) showed significant differences in lipid metabolite levels between MG-treated groups
This article is protected by copyright. All rights reserved.

and the groups supplemented with lovastatin. The heat map showed 249 differentially
expressed lipids identified in this experiment (Fig. 3C). Lovastatin decreased the levels of
approximately half of these lipids (Fig. 3E), which were mainly classified as phospholipids,
such as phosphatidylethanolamine (PE, 20.72%), phosphatidylserine (PG, 17.12%), and
phosphatidic acid (PA, 7.21%). Notably, the addition of lovastatin further increased the levels
of TG-based lipids that were originally elevated after proteasome inhibition (Fig. 3C). This
phenomenon prompted us to consider whether TG-modulating drugs would downregulate TG-
based lipids and synergize with lovastatin and PIs to maximize the killing efficacy of MM cells.
FEN is a fibric acid derivative indicated for use in the treatment of primary
hypercholesterolemia, mixed dyslipidemia, and hypertriglyceridemia in adults and may be
particularly beneficial for patients with particularly high TG levels (Keating & Croom, 2007;
Tonkin et al., 2012). As expected, the addition of FEN significantly reduced the TG-based lipid
levels in MM.1S cells by more than 90% (Fig. 3F). The combination of the three drugs induced
the highest apoptotic cell rates compared with either drug alone or the combination of MG-132
and lovastatin (Fig. S4). These results further indicate that the combination of lipid-modulating
drugs and PIs represents a promising strategy for targeting the lipid metabolism of MM cells.
3.4 Synergistic killing of MM cells by a novel FEN derivative
As revealed in Fig. 4A, simultaneous treatment with FEN and MG-132 led to significantly
lower viability of MM.1S cells than cells treated with either FEN or MG-132 alone. In more
detail, although FEN and MG-132 did not alter the survival of tumor cells at concentrations of
6 μM and 60 nM, respectively, the combination of MG-132 and FEN led to a striking decrease
in survival from approximately 100% to 25%, and this marked synergistic killing effect was
stronger than the combination of lovastatin and MG-132 at concentrations of 15 μM and 60
nM, respectively (Fig. 2B). Furthermore, the addition of FEN also increased the percentages
of apoptotic MM.1S cells compared with MG-132 alone (Fig. S5). Therefore, we decided to
treat a xenograft model with PIs in combination with FEN to evaluate the synergistic effects in
vivo. However, FEN is rapidly hydrolyzed by an esterase in vivo to fenofibric acid (FEC) after
oral administration (Tziomalos & Athyros, 2006). Because FEC exhibits low effectiveness in
triggering tumor cell death (Stalinska et al., 2019), we speculated that FEC is unable to
effectively synergize with PIs to kill tumor cells. As shown in the present study, FEC did not
induce significant synergistic activity at concentrations greater than 50 μM (Fig. 4B). This
result strongly suggested that the strong synergistic effect of the combination of FEN and PIs
on MM cells in vitro was not reproduced in vivo. The present results led us to find a solution
This article is protected by copyright. All rights reserved.

to prevent the rapid hydrolysis of FEN in vivo, which may enable the drug to synergistically
act on PIs in vivo. Since steric hindrance of the cyclohexyl group has been shown to increase
the stability of the ester group in vivo, the cyclohexyl ester was applied to substantially improve
stability (Dang et al., 2012; Hammer, Wu, Sastry & Bodor, 1991). Therefore, we used FEC and
hexyl alcohol to synthesize a more stable side chain ester ring (cyclohexyl), and the synthetic
route and characterization of FCE are shown in Scheme 1 and Fig. S6; the purity of FCE was
analyzed using HPLC (Fig. S7). As expected, the combination of FCE and MG-132
significantly synergistically killed MM cells. In particular, FCE (16 μM) or MG-132 (40 nM)
alone did not affect the survival of MM.1S cells; surprisingly, the combination of the two drugs
reduced cell viability by more than 90% (Fig. 4C), and FCE acted synergistically (CI < 1) with
MG-132 to inhibit growth at almost all combination doses tested (Fig. 4C, lower panel).
Furthermore, MG-132 (100 nM and 150 nM) displayed moderate cytotoxicity in MM.1S cells,
as the observed percentages of early apoptotic cells were approximately 7.39% and 16.91%,
respectively, while the percentages were synergistically increased (37.26% and 44.84%,
respectively) when this treatment was combined with 10 μM FCE (Fig. 4D).
FCE alone did not exert anti-tumor activity at 300 mg/kg in the MM.1S cell xenograft
model (Fig. S8). In the present study, the compounds were not administered until the xenografts
grew to approximately 1000 mm3 (Fig. 4E). Under this condition, the oral administration of
the PI Ixa alone did not obviously inhibit the growth of xenografts (Fig. 4E). No synergistic
effects were observed when the MM.1S cell xenograft model was treated with Ixa and FEN,
whereas the combination treatment with FCE and Ixa exerted a significant synergistic
therapeutic effect on the xenograft model (Fig. 4E and H). The average tumor weights of the
mice in the Ixa and Ixa combined with FEN groups were 0.87 ± 0.57 g and 1.09 ± 0.75 g,
respectively, whereas the tumor weight of the combined FCE- and Ixa-treated group was 0.36
± 0.35 g. (Fig. 4F). The combination treatment with FCE and Ixa did not significantly
decrease the body weight compared with treatment with Ixa alone or the combined FEN and
Ixa treatment (Fig. 4G), indicating that the drug combination had no obvious toxicity. In
summary, the results of the present study indicate that FCE has great potential as an anti-MM
compound, mainly due to its ability to exert synergistic effects with PIs in vitro and in vivo.
Moreover, we attempted to determine the pathway leading to the aforementioned abnormal
lipid metabolism in MM cells under the pressure of proteasome inhibition.
This article is protected by copyright. All rights reserved.

3.5 PIs induce abnormal lipid metabolism via the ATF4/SREBP pathway
The results of the study were consistent with previous studies showing that the translation
of ATF4, an important effector of the unfolded protein response (UPR), was enhanced by
proteasome inhibition (Fig. 5A) (Hu et al., 2012). Moreover, the expression of ATF4 was also
significantly increased after proteasome inhibition (Fig. 5B). ATF4, a downstream protein in
the UPR/PERK/eIF2α pathway, has also been reported to be involved in lipid metabolism(Han
& Kaufman, 2016; Xiao et al., 2013; Yeh, Lai, Lin, Hsu, Lo & Her, 2017). ChIP-seq was
performed for an exhaustive analysis of genes involved in the regulation of lipids that are
regulated byATF4 during ER stress induced by proteasome inhibition to further investigate the
correlations between abnormal lipid accumulation and ATF4 expression. The total peaks
detected in the MG-132 group and negative control group are shown in Table 1, and the feature
distributions of the ATF4 binding peaks of these two groups are also shown in Fig. S9. ATF4
was predominantly detected in the ±5-kb region surrounding transcription start sites of
SREBP1/2 (the binding data obtained using ChIP-seq are listed in Table 2), and the
significantly enriched peaks were identified in a genome-wide analysis (Fig. 5C and 5D). This
result prompted us to detect SRBEP1/2 expression in cells subjected to proteasome inhibition.
As expected, the expression of the SREBP1/2 mRNAs in MM.1S and RPMI8226 cells was
significantly elevated upon proteasome inhibition (Fig. 5E). The findings also showed
the increased expression of both the precursor (~120 kDa) and mature forms (60-78 kDa) of
SREBP proteins upon proteasome inhibition (Fig. 5F). ATF4 knockdown for 48 h by an siRNA
sequence suppressed SREBP1/2 gene expression, which also suggested that ATF4 contributes
to the expression of the SREBP1/2 gene (Fig. S10). Based on accumulating evidence, SREBPs
are key transcription factors that induce lipid synthesis and uptake, and elevated SREBP levels
are linked to cell proliferation and metastasis (Cheng, Geng, Cheng & Guo, 2018; Lewis et al.,
2015; Nickels, 2018). Thus, the abnormal lipid accumulation in MM cells subjected to
proteasome inhibition might be due to the elevated SREBP expression induced by ATF4.
Regarding the synergistic effect of classic lipid-regulating drugs and PIs, we speculated that
SREBP inhibitors interfere with lipid synthesis to synergize with PIs and MM cells. As
expected, the SREBP inhibitor fatostatin (Fato) and the lipid synthase inhibitor orlistat (Orl)
maximized the PI-induced killing of tumor cells (Fig. S11Aand B). Specifically, the synergistic
killing of tumor cells induced by Orl was highly significant; Orl and MG did not induce a
significant increase in apoptosis at 10 µM (2.77%) and 100 nM (8.69%), respectively, and
notably, the proportion of apoptotic cells induced by this combination reached 32.3% (Fig.
S9C).
This article is protected by copyright. All rights reserved.

4. Discussion
Numerous studies have confirmed that the accumulation of lipid droplets in cancer cells
serves as a pivotal adaptive response to harmful conditions, and approaches targeting lipid
metabolism in cancer cells are a promising therapeutic strategy (Liu, Luo, Halim & Song, 2017).
PI-resistant MM cell lines were characterized as displaying global mitochondrial metabolic
changes that led to increased protein folding and lipid restructuring as major drivers of
adaptation towards PIs (Besse et al., 2019). The most surprising findings of our present study
were that PIs induce the abnormal accumulation of large amounts of lipids, including
phospholipids and TG. The phospholipid synthesis rate increases with oncogenesis and tumor
progression, and emerging enzymes in phospholipid metabolism are potentially useful as
treatment targets for anticancer therapy (Cheng, Bhujwalla & Glunde, 2016). In addition, TGs
accumulate and induce significantly increased proliferation of prostate cancer cells (Schlaepfer
et al., 2015). Taken together, these studies suggest the functional importance of lipid in PI
resistance, which prompted us to explore whether lipid accumulation in MM cells induced by
PIs exerted a positive effect on their adaptive survival.
According to a recent study, antiangiogenic drugs trigger lipid-dependent metabolic
reprogramming in tumor cells and confer antiangiogenic drug resistance, leading to an effective
cancer therapy concept involving the combination of drugs that target angiogenesis and lipid
metabolism (Iwamoto et al., 2018a). Tamoxifen-resistant breast cancer T-47D cell variants
exhibit a striking increase in the levels of neutral lipids and free cholesterol in lysosomes,
indicating the applicability of cholesterol-lowering medication during adjuvant endocrine
therapy (Hultsch et al., 2018b). Treatment with the lipid-regulating drugs statins alone trigger
apoptosis in MM cells, and the combination of lovastatin and thalidomide has been observed
to exert strong antiproliferative and proapoptotic effects (Dmoszynska, Podhorecka, Klimek &
Grzasko, 2006; Holstein & Hohl, 2011; Liu et al., 2004). Furthermore, our research strongly
supports the rationale for the combined use of PIs and lovastatin. As expected, lovastatin and
PIs function synergistically to kill MM cells in vitro (Fig. 2C). For MM.1S cells, although
lovastatin did not affect the survival of tumor cells at a concentration of 20 μM, its addition led
to a striking decrease in survival from the 50% induced by MG-132 (80 nM) to 10%, and this
marked synergistic killing effect was also confirmed by the results of a proliferation test in
RPMI8226 cells. This result is also instructive for the use of statins in clinical trials. Due to the
inhibitory or proapoptotic effects of statins on MM cells, researchers have conducted clinical
trials applying statins to MM cells (NCT00281476 and NCT00399867). The purpose of one
study (NCT00281476) was to evaluate the effect of single high doses of simvastatin on bone
metabolism and biochemical markers of disease in MM cells. Our study might provide a
This article is protected by copyright. All rights reserved.

rational approach for future clinical trials examining the use of statins against MM, and
combination therapy with statins and PIs may result in a better response in patients with MM.
This result also provides clues that lipid accumulation may promote the adaptive survival of
MM cells under the pressure of proteasome inhibition.
In the present study, lovastatin significantly reduced the phospholipid level that was
originally elevated by PIs in MM cells, and interestingly, a large amount of TG-based lipids
accumulated in MM cells (Fig. 3C). This result suggests an abnormal metabolism of lipids
under the pressure induced by proteasome inhibition, and lovastatin only inhibited
phospholipid-based metabolic flux and then induced the feedback mechanism of increased
metabolic flux of TG-based lipids (Fig. 3D). Researchers have not conclusively determined
how TG-based lipids mediate the adaptive survival of tumor cells. A previous study suggested
the important function of TGs in mediating the adaptive survival of tumor cells. TG
accumulation protects breast cancer cells from Fato-induced apoptosis (Brovkovych et al.,
2018). Tumor cells release the excess lipids that they store as triglycerides within lipid droplets,
and this process provides cancer cells with a constant energy source for rapid tumor growth
(Cao, 2019; Nieman et al., 2011). Since lovastatin further increases the accumulation of TG-
based lipids in MM cells, the addition of triglyceride-lowering drugs may further enhance the
killing of MM cells. As expected, the addition of FEN led to the formulation of a combination
of three drugs, which further induce apoptosis in more tumor cells than the combination of
lovastatin and PIs (Fig. S4). Notably, more than half of the accumulated lipids induced by PIs
were TG-based lipids, while lovastatin reduced intracellular phospholipid levels and further
exacerbated the accumulation of TG-based lipids. Therefore, the TG-lowering drug FEN could
be combined with PIs to achieve the highly effective synergistic killing of MM cells.
The most striking result to emerge from Fig. 4 is that FEN and PIs did not exert inhibitory
effect on the proliferation of MM cells at concentrations of 6 μM and 60 nM, respectively, but
once the two drugs were applied in combination, the cell survival rate decreased from 100% to
approximately 25%. Although FEN has been tested for its antitumor effects on various cancer
cell lines, this compound is not a strong inhibitor of cancer cell proliferation, usually inhibiting
tumor cells with an IC50 of more than 50 μM or even greater than 200 μM (Li et al., 2014; You,
Hour, Chen, Luo, Hsu & Lee, 2019; Zak, Gelebart & Lai, 2010). This high concentration is
difficult to achieve in the blood to exert its antitumor effects. Thus, the combination strategy is
more reasonable because of the effective inhibitory activity of a relatively low concentration
of FEN in MM cells. However, upon reaching the general circulation, FEN is completely and
immediately hydrolyzed by esterases in plasma to FEC, which is practically ineffective at
This article is protected by copyright. All rights reserved.

triggering cancer cell death (Stalinska et al., 2019). Thus, we have incorporated chemical
modifications into the FEN structure to increase its stability and ultimately its anticancer
potential. As expected, the combination of fenofibrate derivatives (FCE) and Ixa achieved
significant results in animals. Conversely, the combination of FEN and Ixa failed to prevent
MM cell growth in mice. Although the antitumor activity of FEN, particularly its
antiproliferative effect on MM cells, has been recognized (Schmeel, Schmeel & Schmidt-Wolf,
2017), and although FEN was slated to be applied in clinical trials for MM (NCT01965834),
the use of FEN alone or in combination with PIs in the treatment of MM is not reasonable
because it is quickly converted to FA by blood and tissue esterases. Taken together, the
combination of drugs targeting lipid metabolism and PIs has important implications for the
exploration of a new therapeutic strategy for the management and treatment of MM. Next, we
attempted to reveal the cause of the abnormal accumulation of lipids induced by PIs in MM
cells.
The accumulation of unfolded or misfolded proteins induced by PIs in the ER lumen
activates the UPR to subsequently regulate the quantity of ER driven by the synthesis of lipids
(Wallington-Beddoe & Pitson, 2017). This study confirmed that the level of ATF4, an important
effector of the UPR, was substantially increased upon proteasome inhibition (Fig. 5A-B). ATF4
functionally promotes lipogenesis, and ATF4 transgenic zebrafish show increased weight gain
with larger fat pads due to adipocyte hyperplasia (Yeh, Lai, Lin, Hsu, Lo & Her, 2017). Lipid
accumulation caused by a high-fructose diet was also attenuated in Atf4−/− mice (Xiao et al.,
2013). Here, we chose to dissect the specific role of ATF4 in lipogenesis and to map the cellular
targets of ATF4 in MM.1S cells by performing a promoter ChIP screen. We identified numerous
target genes, and an analysis of their chromosomal distribution led to the interesting discovery
that SREBP1/2 was occupied by ATF4 in MM.1S cells (Fig. 5C-D). Accordingly, the
significantly increased expression of the SREBP1/2 mRNAs and proteins in MM.1S and
RPMI8226 cells treated with proteasome inhibitors was detected (Fig. 5E-F). SREBPs are
proteins that regulate the synthesis and uptake of lipids, including cholesterol (Lee et al., 2018).
SREBP1 is involved in global lipid synthesis and growth, and SREBP2 is involved in
cholesterol regulation (Shimano & Sato, 2017). Upon activation, SREBP is translocated to the
Golgi, and proteolytic cleavage produces a mature transcriptionally active 60-78-kDa fragment,
which induces the expression of several lipid biosynthetic genes in the nucleus (Horton,
Goldstein & Brown, 2002). Thus, the lipid accumulation induced by inhibition of proteasome
activity in MM cells is most likely due to the significant increase in the level of both precursor
and mature forms of SREBP1/2. The SREBP inhibitor Fato and the lipid synthase inhibitor Orl
This article is protected by copyright. All rights reserved.

were applied to determine whether they synergized with the PI to kill MM cells and to further
explore the functions of these abnormally accumulated lipids. The combination of Fato and Orl
with the PI MG separately maximized the killing of tumor cells compared with the single drugs
(Fig. S11). However, ATF4-SREBPs may not the only one pathway to induce the abnormal
lipids accumulation. After all, our study did not rule out other possible causes of abnormal lipid
accumulation, such as increased fat intake, decreased fat export, decreased fat oxidation and so
on. But the exploration of these mechanisms will be included in our following research.
In summary, the present study describes the abnormal accumulation of lipids in MM cells
upon proteasome inhibition, which may induce the expression of the transcription factor ATF4.
Then, ATF4 binds to the promoter region of SREBP to initiate its expression, which could lead
to abnormal lipid accumulation. Based on these results, abnormal lipid accumulation may
enable MM cells to resist the toxicity of PIs. A series of lipid-regulating drugs exerted
significant synergistic killing effects with PIs on MM cells. Our results may provide a new
treatment strategy based on the abnormal lipid accumulation induced by PIs.
Conflict of interest
The authors have no conflicts of interest to disclose.
Declaration of transparency and scientific Rigour
This Declaration acknowledges that this paper adheres to the principles for transparent
reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design
& Analysis, Immunoblotting & Immunochemistry and Animal Experimentation, and as
recommended by funding agencies, publishers, and other organizations engaged with
supporting research.
Acknowledgments
ChIP experiments and data analyses were technically supported by DIATRE
Biotechnology Co., Ltd. (Shanghai, China). This study was supported by the Scientific
Research Fund of the National Natural Science Foundation of China (No. 81201668), Sichuan
Science and Technology Program (No. 2021JDRC0037 and 20ZDYF2374), and Chengdu
Medical College (NO. CYZ19-02).
This article is protected by copyright. All rights reserved.

Authors’ contributions
G.X., S.H., and J.P. performed the most of experiments and analyzed the data. S.S.Y.,
X.F.G., M.H.L. and Z.F.L. performed synergy tests. P.F.Q., Y.W. and S.Q.Y. performed
apoptosis and western blot analyses. S.Y.L., H.W., L.J.S. and P.L. performed in vivo studies.
B.G. and R.H.D. synthesized FCE. J.L. and T.Y. conceived and designed the studies and wrote
the manuscript.
5. References
Alexander SPH, Fabbro D, Kelly E, Mathie A, Peters JA, Veale EL, et al. (2019). THE CONCISE
GUIDE TO PHARMACOLOGY 2019/20: Enzymes. Brit J Pharmacol 176: S297-S396.
Alexander SPH, Roberts RE, Broughton BRS, Sobey CG, George CH, Stanford SC, et al. (2018).
Goals and practicalities of immunoblotting and immunohistochemistry: A guide for
submission to the British Journal of Pharmacology. Br J Pharmacol 175: 407-411.
Beloribi-Djefaflia S, Vasseur S, & Guillaumond F (2016). Lipid metabolic reprogramming in
cancer cells. Oncogenesis 5: e189.
Besse L, Besse A, Mendez-Lopez M, Vasickova K, Sedlackova M, Vanhara P, et al. (2019). A
metabolic switch in proteasome inhibitor-resistant multiple myeloma ensures higher
mitochondrial metabolism, protein folding and sphingomyelin synthesis. Haematologica 104:
e415-e419.
Bian Y, Yu Y, Wang SS, & Li L (2015). Up-regulation of fatty acid synthase induced by EGFR/ERK
activation promotes tumor growth in pancreatic cancer. Biochem Bioph Res Co 463: 612-617.
Brovkovych V, Izhar Y, Danes JM, Dubrovskyi O, Sakallioglu IT, Morrow LM, et al. (2018).
Fatostatin induces pro- and anti-apoptotic lipid accumulation in breast cancer. Oncogenesis 7:
66.
Brzozowski JS, Jankowski H, Bond DR, McCague SB, Munro BR, Predebon MJ, et al. (2018).
Lipidomic profiling of extracellular vesicles derived from prostate and prostate cancer cell lines.
Lipids Health Dis 17: 211.
Cao Y (2019). Adipocyte and lipid metabolism in cancer drug resistance. J Clin Invest 129:
3006-3017.
Cenci S, & Sitia R (2007). Managing and exploiting stress in the antibody factory. FEBS Lett 581:
3652-3657.
Cheng C, Geng F, Cheng X, & Guo D (2018). Lipid metabolism reprogramming and its potential
This article is protected by copyright. All rights reserved.

targets in cancer. Cancer Commun (Lond) 38: 27.
Cheng M, Bhujwalla ZM, & Glunde K (2016). Targeting Phospholipid Metabolism in Cancer.
Front Oncol 6: 266.
Chou TC, & Talalay P (1984). Quantitative analysis of dose-effect relationships: the combined
effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul 22: 27-55.
Curtis MJ, Alexander S, Cirino G, Docherty JR, George CH, Giembycz MA, et al. (2018).
Experimental design and analysis and their reporting II: updated and simplified guidance for
authors and peer reviewers. Br J Pharmacol 175: 987-993.
Dang HT, Lee YM, Kang GJ, Yoo ES, Hong J, Lee SM, et al. (2012). In vitro stability and in vivo
anti-inflammatory efficacy of synthetic jasmonates. Bioorg Med Chem 20: 4109-4116.
Di Noto G, Chiarini M, Paolini L, Mazzoldi EL, Giustini V, Radeghieri A, et al. (2014).
Immunoglobulin Free Light Chains and GAGs Mediate Multiple Myeloma Extracellular Vesicles
Uptake and Secondary NfkappaB Nuclear Translocation. Front Immunol 5: 517.
Dmoszynska A, Podhorecka M, Klimek P, & Grzasko N (2006). Lovastatin and thalidomide have
a combined effect on the rate of multiple myeloma cell apoptosis in short term cell cultures.
Eur J Clin Pharmacol 62: 325-329.
Donato E, Croci O, Sabo A, Muller H, Morelli MJ, Pelizzola M, et al. (2017). Compensatory RNA
polymerase 2 loading determines the efficacy and transcriptional selectivity of JQ1 in Myc-
driven tumors. Leukemia 31: 479-490.
Epstein MM, Divine G, Chao CR, Wells KE, Feigelson HS, Scholes D, et al. (2017). Statin use and
risk of multiple myeloma: An analysis from the cancer research network. International Journal
of Cancer 141: 480-487.
Hammer RH, Wu W, Sastry JS, & Bodor N (1991). Short acting soft mydriatics. Curr Eye Res 10:
565-570.
Han J, & Kaufman RJ (2016). The role of ER stress in lipid metabolism and lipotoxicity. J Lipid
Res 57: 1329-1338.
Holstein SA, & Hohl RJ (2011). Isoprenoid biosynthetic pathway inhibition disrupts monoclonal
protein secretion and induces the unfolded protein response pathway in multiple myeloma
cells. Leuk Res 35: 551-559.
Horton JD, Goldstein JL, & Brown MS (2002). SREBPs: activators of the complete program of
cholesterol and fatty acid synthesis in the liver. J Clin Invest 109: 1125-1131.
This article is protected by copyright. All rights reserved.

Hu J, Dang N, Menu E, De Bruyne E, Xu D, Van Camp B, et al. (2012). Activation of ATF4
mediates unwanted Mcl-1 accumulation by proteasome inhibition. Blood 119: 826-837.
Hultsch S, Kankainen M, Paavolainen L, Kovanen RM, Ikonen E, Kangaspeska S, et al. (2018a).
Association of tamoxifen resistance and lipid reprogramming in breast cancer. Bmc Cancer 18.
Hultsch S, Kankainen M, Paavolainen L, Kovanen RM, Ikonen E, Kangaspeska S, et al. (2018b).
Association of tamoxifen resistance and lipid reprogramming in breast cancer. BMC Cancer 18:
850.
Iwamoto H, Abe M, Yang Y, Cui D, Seki T, Nakamura M, et al. (2018a). Cancer Lipid Metabolism
Confers Antiangiogenic Drug Resistance. Cell Metab 28: 104-117 e105.
Iwamoto H, Abe M, Yang YL, Cui DM, Seki T, Nakamura M, et al. (2018b). Cancer Lipid
Metabolism Confers Antiangiogenic Drug Resistance. Cell Metabolism 28: 104-+.
Keating GM, & Croom KF (2007). Fenofibrate: a review of its use in primary dyslipidaemia, the
metabolic syndrome and type 2 diabetes mellitus. Drugs 67: 121-153.
Lee JH, Phelan P, Shin M, Oh BC, Han X, Im SS, et al. (2018). SREBP-1a-stimulated lipid synthesis
is required for macrophage phagocytosis downstream of TLR4-directed mTORC1. Proc Natl
Acad Sci U S A 115: E12228-E12234.
Lewis CA, Brault C, Peck B, Bensaad K, Griffiths B, Mitter R, et al. (2015). SREBP maintains lipid
biosynthesis and viability of cancer cells under lipid- and oxygen-deprived conditions and
defines a gene signature associated with poor survival in glioblastoma multiforme. Oncogene
34: 5128-5140.
Li T, Zhang Q, Zhang J, Yang G, Shao Z, Luo J, et al. (2014). Fenofibrate induces apoptosis of
triple-negative breast cancer cells via activation of NF-kappaB pathway. BMC Cancer 14: 96.
Liu QP, Luo Q, Halim A, & Song GB (2017). Targeting lipid metabolism of cancer cells: A
promising therapeutic strategy for cancer. Cancer Letters 401: 39-45.
Liu ZL, Luo JM, Dong ZR, Wang FX, Zhang XJ, Yang JC, et al. (2004). [In vitro effects of mevastatin
on the proliferation and apoptosis in human multiple myeloma cell line U266]. Zhongguo Shi
Yan Xue Ye Xue Za Zhi 12: 340-345.
Lucenay KS, Doostan I, Karakas C, Bui TY, Ding ZY, Mills GB, et al. (2016). Cyclin E Associates
with the Lipogenic Enzyme ATP-Citrate Lyase to Enable Malignant Growth of Breast Cancer
Cells. Cancer Research 76: 2406-2418.
Marino S, Petrusca DN, & Roodman GD (2019). Therapeutic targets in myeloma bone disease.
Br J Pharmacol.
This article is protected by copyright. All rights reserved.

Masarwi M, DeSchiffart A, Ham J, & Reagan MR (2019). Multiple Myeloma and Fatty Acid
Metabolism. JBMR Plus 3: e10173.
Mollinedo F, de la Iglesia-Vicente J, Gajate C, Estella-Hermoso de Mendoza A, Villa-Pulgarin JA,
Campanero MA, et al. (2010). Lipid raft-targeted therapy in multiple myeloma. Oncogene 29:
3748-3757.
Nickels JT, Jr. (2018). New links between lipid accumulation and cancer progression. J Biol
Chem 293: 6635-6636.
Nieman KM, Kenny HA, Penicka CV, Ladanyi A, Buell-Gutbrod R, Zillhardt MR, et al. (2011).
Adipocytes promote ovarian cancer metastasis and provide energy for rapid tumor growth.
Nat Med 17: 1498-1503.
Record M, Silvente-Poirot S, Poirot M, & Wakelam MJO (2018). Extracellular vesicles: lipids as
key components of their biogenesis and functions. J Lipid Res 59: 1316-1324.
Sanfilippo KM, Keller J, Gage BF, Luo SH, Wang TF, Moskowitz G, et al. (2016). Statins Are
Associated With Reduced Mortality in Multiple Myeloma. Journal of Clinical Oncology 34:
4008-U4115.
Schlaepfer IR, Nambiar DK, Ramteke A, Kumar R, Dhar D, Agarwal C, et al. (2015). Hypoxia
induces triglycerides accumulation in prostate cancer cells and extracellular vesicles
supporting growth and invasiveness following reoxygenation. Oncotarget 6: 22836-22856.
Schmeel LC, Schmeel FC, & Schmidt-Wolf IGH (2017). In Vitro Apoptosis Induction by
Fenofibrate in Lymphoma and Multiple Myeloma. Anticancer Res 37: 3513-3520.
Sellick CA, Hansen R, Stephens GM, Goodacre R, & Dickson AJ (2011). Metabolite extraction
from suspension-cultured mammalian cells for global metabolite profiling. Nat Protoc 6: 1241-
1249.
Shimano H, & Sato R (2017). SREBP-regulated lipid metabolism: convergent physiology -
divergent pathophysiology. Nat Rev Endocrinol 13: 710-730.
Stalinska J, Zimolag E, Pianovich NA, Zapata A, Lassak A, Rak M, et al. (2019). Chemically
Modified Variants of Fenofibrate with Antiglioblastoma Potential. Transl Oncol 12: 895-907.
Tabas I, & Ron D (2011). Integrating the mechanisms of apoptosis induced by endoplasmic
reticulum stress. Nat Cell Biol 13: 184-190.
Tadros S, Shukla SK, King RJ, Gunda V, Vernucci E, Abrego J, et al. (2017). De Novo Lipid
Synthesis Facilitates Gemcitabine Resistance through Endoplasmic Reticulum Stress in
This article is protected by copyright. All rights reserved.

Pancreatic Cancer. Cancer Research 77: 5503-5517.
Tirado-Velez JM, Joumady I, Saez-Benito A, Cozar-Castellano I, & Perdomo G (2012). Inhibition
of Fatty Acid Metabolism Reduces Human Myeloma Cells Proliferation. Plos One 7.
Tirinato L, Pagliari F, Limongi T, Marini M, Falqui A, Seco J, et al. (2017). An Overview of Lipid
Droplets in Cancer and Cancer Stem Cells. Stem Cells International.
Tonkin A, Hunt D, Voysey M, Kesaniemi A, Hamer A, Waites J, et al. (2012). Effects of
fenofibrate on cardiovascular events in patients with diabetes, with and without prior
cardiovascular disease: The Fenofibrate Intervention and Event Lowering in Diabetes (FIELD)
study. Am Heart J 163: 508-514.
Tziomalos K, & Athyros VG (2006). Fenofibrate: a novel formulation (Triglide) in the treatment
of lipid disorders: a review. Int J Nanomedicine 1: 129-147.
Wallington-Beddoe CT, & Pitson SM (2017). Enhancing ER stress in myeloma. Aging (Albany
NY) 9: 1645-1646.
Wong WW, Clendening JW, Martirosyan A, Boutros PC, Bros C, Khosravi F, et al. (2007).
Determinants of sensitivity to lovastatin-induced apoptosis in multiple myeloma. Mol Cancer
Ther 6: 1886-1897.
Xiao G, Zhang T, Yu S, Lee S, Calabuig-Navarro V, Yamauchi J, et al. (2013). ATF4 protein
deficiency protects against high fructose-induced hypertriglyceridemia in mice. J Biol Chem
288: 25350-25361.
Yeh KY, Lai CY, Lin CY, Hsu CC, Lo CP, & Her GM (2017). ATF4 overexpression induces early onset
of hyperlipidaemia and hepatic steatosis and enhances adipogenesis in zebrafish. Sci Rep 7:
16362.
You BJ, Hour MJ, Chen LY, Luo SC, Hsu PH, & Lee HZ (2019). Fenofibrate induces human
hepatoma Hep3B cells apoptosis and necroptosis through inhibition of thioesterase domain
of fatty acid synthase. Sci Rep-Uk 9.
Zak Z, Gelebart P, & Lai R (2010). Fenofibrate induces effective apoptosis in mantle cell
lymphoma by inhibiting the TNFalpha/NF-kappaB signaling axis. Leukemia 24: 1476-1486.
This article is protected by copyright. All rights reserved.

6. Bullet point summary
What is already known:
Reprogramming of lipid metabolism promotes the adaptive survival of cancer cells under
harmful conditions.
What this study adds:
The present study describes the abnormal accumulation of lipids in MM cells upon proteasome
inhibition.
A series of lipid-regulating drugs exerts significant synergistic killing effects with PIs on MM
cells.
Clinical significance:
The combination of lipid-regulating drugs and PIs has implications as a new therapeutic
strategy for MM.
This article is protected by copyright. All rights reserved.

Fig. 1. PIs induce abnormal lipid accumulation in MM cells. For this experiment, 1×107
MM.1S cells (1×106 cells mL-1) were treated with 100 nM MG-132 (or an equal volume of
solvent) for approximately 24 h. A, Unsupervised PCA of the negative control groups (NC
groups) and the MG-treated groups (MG groups). The model parameters are R2X=0.677. B,
PLS-DA of the NC and MG groups. The model parameters are R2Y=0.997 and Q2=0.939. C,
Heat maps showing the 106 differentially abundant lipid metabolites in MM.1S cells between
the NC and MG groups. The color of each section is proportional to the significance of changes
in metabolites (red, upregulated; blue, downregulated). Rows: lipid metabolites; columns:
MM.1S cell samples. D, Volcano plot of lipid metabolites detected in MM.1S cells from the
NC and MG groups. The differentially abundant lipid metabolites shown in red are present at
levels above the threshold (FC > 2, and P < 0.05). E, The increased levels of lipid metabolites
of MM.1S cells in the MG-treated group (compared with the NC groups) are classified in the
pie-doughnut chart. The percentage represents the proportion of these lipids among the 101
elevated lipids (lipids with a low percentage are not shown in the figure).
This article is protected by copyright. All rights reserved.

Fig. 2. Synergistic inhibition of growth and induction of apoptosis by MG-132 and
lovastatin in MM cells. A, Lovastatin reduced MM cell viability in a dose-dependent manner.
The 50% inhibitory concentrations (IC50 values) of lovastatin in MM.1S and RPMI8226 cells
were 35.2 and 27.1 μM, respectively, when cells were exposed to various concentrations of
lovastatin (from 210 to 50000 nM) for 48 h. B, MG-132 reduced MM cell viability in a dose-
dependent manner. The IC50 values of MG-132 in MM.1S and RPMI 8226 cells were 54.0 and
165.0 nM, respectively, when cells were exposed to various concentrations of MG-132 (40 to
10000 nM) for 48 h. C, The effects of MG-132 and lovastatin alone and in combination on the
viability of MM.1S and RPMI8226 cells that were exposed to various concentrations of MG-
132 and lovastatin for 48 h were assessed using a CCK-8 assay (n=5). A CI analysis was
performed using CalcuSyn software to determine the CI values of the combination of lovastatin
and MG-132 for MM.1S and RPMI8226 cells. Synergy was present when the CI is less than
1.0.
This article is protected by copyright. All rights reserved.

Fig. 3. Changes in the lipid profile induced by lovastatin. For this experiment, 1×107 MM.1S
cells (1×106 cells mL-1) were treated with 100 nM MG-132 (or 100 nM MG-132+15 μM
lovastatin, or 100 nM MG-132, 15 μM lovastatin and 8 μM FEN simultaneously) for
approximately 24 h. A, Unsupervised PCA of the MG groups and the MG+LOVA groups. The
model parameters are R2X=0.645. B, Partial PLS-DA of the MG and MG+LOVA groups. The
model parameters are R2Y=0.997 and Q2=0.939. C, Heat maps of the 249 differentially
abundant lipid metabolites in MM.1S cells between the MG and MG+LOVA groups. The color
of each section is proportional to the significance of changes in metabolites (red, upregulated;
blue, downregulated). Rows: lipid metabolites; columns: MM.1S cell samples. D, Volcano plot
of lipid metabolites detected in MM.1S cells from the MG and MG+LOVA groups. The
differentially abundant lipid metabolites shown in red are present at levels above the threshold
(FC > 2, and P < 0.05). E, The downregulated and upregulated lipid metabolites in MM.1S
cells from the MG+LOVA groups (compared with the MG groups) are classified in the left and
right pie-doughnut charts, respectively. F, Heat maps of the differentially abundant lipid
metabolites in MM.1S cells between the MG+LOVA groups and the group subjected to the
combination treatment with MG-132, lovastatin and FEN (MLF groups). The color of each
section is proportional to the significance of changes in metabolites (red, upregulated; blue,
downregulated). Rows: MM.1S cell samples; columns: lipid metabolites.
This article is protected by copyright. All rights reserved.

Fig. 4. Synergistic killing of MM cells by a novel FEN derivative. A, Effects of MG-132 and
FEN alone and in combination on the viability of MM.1S cells exposed to various
concentrations of MG-132 and FEN for 48 h. Treatment with FEN and MG-132 led to
significantly lower cell viability than treatment with either FEN or MG-132 alone (n=5). A CI
analysis was performed using CalcuSyn software to determine the CI values of the combination
of lovastatin and MG-132 for MM.1S and RPMI8226 cells. Synergy was present when the CI
was less than 1.0. A synergistic interaction between MG-132 and FEN was observed in MM
cells. B, FEC is the rapidly hydrolyzed product of FEN and did not show significant synergistic
activity with PIs (n=5). C, FCE acted synergistically (CI < 1.0) with MG-132 to inhibit growth
at almost all combination doses tested (n=5). D, At 10 μM, a concentration that does not induce
apoptosis of MM.1S cells, FCE showed synergy with MG-132 at concentrations of 100 nM
and 150 nM. The cells were stained with both annexin V and PI after treatment with drugs for
8 h and analyzed using flow cytometry. E, The combination of MG-132 and FCE inhibited the
growth of human MM xenograft tumors in SCID mice. The right flank of a CB-17 SCID mouse
This article is protected by copyright. All rights reserved.

was subcutaneously injected with 2 × 107 MM.1S cells in 100 μL of PBS and 100 µL of
Matrigel basement membrane matrix, and tumor growth was monitored using calipers. SCID
mice were not treated with the compounds until xenografts exceeded 1000 mm3 in size
(n = 8 for each group). The tumor volumes of mice treated with Ixa and FCE [Ixa: 7 mg/kg,
i.g., twice a week; FCE: 300 mg/kg, i.g., once a day] differed significantly from mice treated
with Ixa (7 mg/kg, i.g., twice a week, **P < 0.01) and mice treated with Ixa and FEN (Ixa: 7
mg/kg, i.g., twice a week; FEN: 300 mg/kg, i.g., once a day, ##P < 0.01). F, The tumor growth
is presented as the means ± S.D. (MM.1S, n = 8). The average tumor weights of the Ixa group
and FEN and Ixa group were 0.87 ± 0.57 g and 1.09 ± 0.75 g, respectively, whereas the tumor
weight in the combined FCE and Ixa-treated group was 0.36 ± 0.35 g (compared with the Ixa
group or FEN and Ixa group, *P < 0.05). G, The combination treatment with FCE and Ixa did
not significantly decrease the body weight of mice compared with the treatment with Ixa alone
or the combined FEN and Ixa treatment (P > 0.05). H, Tumors isolated from MM.1S-bearing
mice are shown.
This article is protected by copyright. All rights reserved.

Fig. 5. PIs induce abnormal lipid metabolism through a mechanism involving the
ATF4/SREBP pathway. A, The expression of the ATF4 mRNA in MM.1S and RPMI 8226
cells (cell density, 1×106 cells mL-1) was increased after a 1-h treatment with MG-132 (*P <
0.05, 100 nM and 400 nM, respectively). The data were acquired using real-time PCR (n=5).
B, The protein expression level of theATF4 in MM.1S and RPMI8226 cells (cell density, 1×106
cells mL-1) after treatment with 400 nM or 800 nM MG-132 for different durations (2, 4, 6, or
8 h) was detected by western blot. C and D, An exhaustive analysis of genes regulated byATF4
in MM.1S (cell density, 1×106 cells mL-1) cells after treatment with MG-132 (800 nM) for 6 h
was acquired in a series of ChIP-seq experiments. The expression of the SREBP1 (C) and
SREBP2 (D) genes related to lipid synthesis was directly regulated by ATF4. Compared with
the controls (NS-treated, without MG-132 treatment), the SREBP1 (C) and SREBP2 (D) genes
were predominantly located in the ±5-kb region surrounding transcriptional start sites, and the
significantly enriched peaks were identified in the genome-wide analysis. E, The expression of
the SRBEP1/2 mRNAs in MM.1S and RPMI8226 cells (cell density, 1×106 cells mL-1) treated
with MG-132 (100 nM and 400 nM, respectively) for 1 h was determined using real-time PCR
quantification (n = 5, *P < 0.05). F, The protein expression level of both precursor (~120 kDa)
and mature forms (60-78 kDa) of SREBP in MM.1S and RPMI8226 cells treated with various
concentrations of MG-132 for 4 h were detected by western blot.
This article is protected by copyright. All rights reserved.

Table 1. The average number of peaks detected in the MG-132 group and negative control group, respectively
Group
Peak Number (X±SD )
104 ± 62
Negative control
MG-132 treated
10177 ± 878
Table 2.The ATF4 binding sites for SERBP1 or SREBP2 based on ChIP-seq analysis data
Name
Chromosome
Chr17
Start
End
Annotation
SREBP1
SREBP2
17740481
42227575
17741066
42227707
Promoter (<=1kb)
Promoter (1-2kb)
Chr22
This article is protected by copyright. All rights reserved.

Graphical Abstract
This article is protected by copyright. All rights reserved.