Celastrol induces ferroptosis in multiple myeloma associated with suppression of the squalene synthase-mediated lipid defense axis

AI Med

DOI: 10.71423/aimed.20261006

https://www.ai-press.org/journal/aimed

Article

Celastrol induces ferroptosis in multiple myeloma associated with

suppression of the squalene synthase-mediated lipid defense axis

Huizi Ye¹, Bin Peng^1,*

1 Shanghai Health Commission Key Lab of Artificial Intelligence (AI)-Based Management of Inflammation and Chronic Diseases, Gongli

Hospital of Shanghai Pudong New Area, Shanghai University of Medicine & Health Sciences, Shanghai, China

* Corresponding author(s):

Bin Peng, pb01312@glhospital.com

Received: 30 August 2026

Accepted: 18 September 2026

Published: 30 September 2026

© 2026 The Author(s). AI Med

published by AI Press.

This is an open access article

distributed under the terms of the

Creative Commons Attribution 4.0

International License (CC BY 4.0).

Abstract: Multiple myeloma (MM) remains incurable due to pervasive drug resistance,

motivating exploration of non-apoptotic cell death pathways. Squalene synthase (SQS,

encoded by FDFT1), the first committed enzyme in sterol biosynthesis within the mevalonate

(MVA) pathway, has been linked to ferroptosis defense, but its role in MM is poorly defined.

This study aimed to investigate whether SQS modulates celastrol-induced ferroptosis in MM.

Four MM cell lines (RPMI-8226, U266, KMS-11, SKO-007) were treated with celastrol.

Cell viability, intracellular iron, reactive oxygen species (ROS), malondialdehyde (MDA),

and glutathione (GSH) were measured. Ferroptosis inhibitors (ferrostatin-1, deferoxamine

mesylate) and lentiviral SQS overexpression were used to assess functional dependence.

Bulk RNA-seq data from the MMRF-COMMPASS cohort (n = 764 primary CD138⁺ plasma

cell samples) were analyzed for FDFT1-ferroptosis correlations and pathway activity using

GSVA and GSEA. Celastrol reduced viability (24-h IC50: 2.41-4.44 µmol/L) with iron

accumulation, ROS elevation, MDA increase, and GSH depletion in all four cell lines.

Ferrostatin-1 and deferoxamine partially rescued viability and restored SQS, GPX4,

SLC7A11, and FSP1 protein levels. SQS overexpression partially protected RPMI-8226 cells

and mitigated iron/MDA accumulation while preserving GSH. In MMRF-COMMPASS,

FDFT1-high tumors showed significantly elevated ferroptosis defense pathway activity

(GSVA P_adj = 1.37 × 10^-6; GSEA NES = 2.13, P_adj = 2.68 × 10^-3), although FDFT1 was

not an independent prognostic marker (univariate HR = 1.064, P = 0.441; multivariate HR =

1.059, P = 0.482). SQS is a functional ferroptosis defense node in MM. Celastrol-induced

ferroptosis

is

associated

with

suppression

of

the

MVA-SQS

axis,

supported

by

pharmacological, genetic, and clinical transcriptomic evidence. These findings provide a

preclinical rationale for targeting SQS-mediated metabolic vulnerability in MM.

Keywords: celastrol; squalene synthase; FDFT1; multiple myeloma; ferroptosis; mevalonate

pathway

Multiple myeloma (MM) is a plasma-cell malignancy

characterized by substantial biological heterogeneity and

recurrent treatment resistance [1,2]. Although proteasome

inhibitors, immunomodulatory drugs, monoclonal antibodies,

and

stem-cell

transplantation

have

improved

outcomes,

relapse remains nearly universal, and the disease is considered

incurable [3,4]. This persistent therapeutic failure underscores

the urgent need to identify non-apoptotic cell death pathways

and metabolic dependencies that can be exploited to overcome

resistance in refractory MM clones [5,6].

Ferroptosis is an iron-dependent form of regulated cell

death driven by catastrophic lipid peroxidation, distinct from

apoptosis, necroptosis, and autophagy [7,8]. The canonical

defense

axis

comprises

the

cystine/glutamate

antiporter

SLC7A11, glutathione (GSH), and glutathione peroxidase 4

(GPX4),

which

collectively

suppress

phospholipid

peroxidation

[9,10].

A

parallel

glutathione-independent

system centered on ferroptosis suppressor protein 1 (FSP1)

and coenzyme Q10 (CoQ10) provides additional protection at

the

plasma

membrane

[11,12],

while

mitochondrial

dihydroorotate dehydrogenase contributes a third layer of

defense [13]. In MM specifically, ferroptosis inducers have

shown promising preclinical activity, and iron metabolism is

increasingly recognized as a determinant of bortezomib

response

[14-17].

Furthermore,

the

bone

marrow

microenvironment

can

modulate

MM

cell

ferroptosis

sensitivity through stromal-cell-driven lanosterol biosynthesis

[18].

The

mevalonate

(MVA)

pathway

is

the

central

biosynthetic route for sterols, including cholesterol, CoQ10,

AI Med, 2026

2 / 10

and isoprenoid intermediates [19,20]. Squalene synthase (SQS,

encoded by FDFT1) catalyzes the first committed step in

sterol biosynthesis, converting two molecules of farnesyl

pyrophosphate to squalene [21,22]. MVA-derived metabolites,

particularly CoQ10 and squalene itself, have been implicated

in

ferroptosis

defense

[23-25].

Notably,

squalene

accumulation protects certain lymphomas and solid tumors

from oxidative cell death [23], whereas the consequences of

SQS

suppression

in

hematological

malignancies

remain

poorly defined.

Celastrol is a pentacyclic triterpenoid, itself an MVA

pathway product, isolated from Tripterygium wilfordii [26],

with broad anti-inflammatory and antitumor activities [27,28].

In MM, celastrol inhibits proliferation and induces apoptosis

through NF- κ B and proteasome inhibition [29,30]. Recent

studies have linked celastrol to ferroptosis in multiple cancer

types [31-33], but whether ferroptosis contributes to its anti-

MM activity and which metabolic nodes are involved remain

unknown.

In the present study, we investigated whether SQS

modulates celastrol-induced ferroptosis in MM. We assessed

ferroptosis-associated phenotypes in four human MM cell

lines, evaluated the functional contribution of SQS through

pharmacological rescue and genetic overexpression, and

interrogated the clinical transcriptomic context of FDFT1 in

the MMRF-COMMPASS cohort (n = 764). Our findings

identify SQS as a functional ferroptosis defense node whose

suppression is associated with celastrol-induced cell death in

MM.

2. Materials and Methods

2.1. Cell Culture and Reagents

RPMI-8226 (Chinese Academy of Sciences, Shanghai,

China), U266, KMS-11, and SKO-007 (gifts from Prof. Jian

Hou,

Changzheng

Hospital,

Shanghai,

China)

were

maintained in RPMI-1640 (Gibco, Grand Island, NY, USA)

supplemented

with

10%

fetal

bovine

serum

and

1%

penicillin/streptomycin at 37 °C in a humidified 5% CO2

incubator. All cell lines were confirmed mycoplasma-negative.

Celastrol (purity ≥98%, Sigma-Aldrich, St. Louis, MO, USA;

catalog no. C0869; molecular formula C29H38O4, molecular

weight 450.61 g/mol) was dissolved in dimethyl sulfoxide

(DMSO) as a 50 mmol/L stock; the final DMSO concentration

in culture medium did not exceed 0.1% (v/v). Ferrostatin-1

(Fer-1; SML0583) and

deferoxamine mesylate (DFOM;

D9533) were purchased from Sigma-Aldrich. Antibodies used

were: SQS (ab236666, Abcam, Cambridge, UK), GPX4

(52455, Cell Signaling Technology, Danvers, MA, USA),

SLC7A11 (HA721868, HuaAn Biotechnology, Hangzhou,

China), FSP1 (24972S, Proteintech, Rosemont, IL, USA),

SREBP2 (ab30682, Abcam), and GAPDH (D11005, Sangon

Biotech, Shanghai, China).

2.2. Cell Viability Assay

Cells were seeded in 96-well plates at a density of 1-2 ×

10⁴ cells per well and treated with celastrol at concentrations

of 0-10 µmol/L for 24, 48, or 72 h. For rescue experiments,

Fer-1 (5 µmol/L) or DFOM (10 µmol/L) was added 1 h before

celastrol treatment. Cell viability was assessed using the Cell

Counting

Kit-8

(CCK-8;

Dojindo,

Kumamoto,

Japan)

according to the manufacturer's instructions. Absorbance at

450 nm was measured using a BioTek microplate reader

(Agilent, Santa Clara, CA, USA). Half-maximal inhibitory

concentration (IC50) values were calculated by non-linear

regression using GraphPad Prism 10.0 (GraphPad Software,

San Diego, CA, USA).

2.3. Biochemical Assays and ROS Detection

After 24 h of celastrol treatment, intracellular iron

concentration was measured using a colorimetric iron assay

kit (NJJCBIO, Nanjing, China). Malondialdehyde (MDA)

levels were quantified using the thiobarbituric acid reactive

substances (TBARS) assay kit (Beyotime Biotechnology,

Shanghai, China). Reduced glutathione (GSH) was measured

using a GSH assay kit (Beyotime Biotechnology). All

measurements were normalized to total protein concentration

determined

by

the

bicinchoninic

acid

(BCA)

method.

Intracellular reactive oxygen species (ROS) were detected

using 10 µmol/L 2',7'-dichlorofluorescein diacetate (DCFH-

DA; Beyotime Biotechnology). Cells were incubated with

DCFH-DA for 30 min at 37 °C, washed twice with phosphate-

buffered

saline,

and

imaged

using

an

Olympus

IX73

fluorescence microscope (Olympus, Tokyo, Japan). Mean

fluorescence intensity was quantified using ImageJ software

(National Institutes of Health, Bethesda, MD, USA). All

assays were performed in triplicate and repeated at least three

times.

2.4. Western Blotting

Cells

were

lysed

in

radioimmunoprecipitation

assay

(RIPA) buffer (Beyotime Biotechnology) containing protease

and phosphatase inhibitor cocktails. Protein concentration was

determined by the BCA assay. Equal amounts of protein (20-

30 µg per lane) were separated by sodium dodecyl sulfate-

polyacrylamide

gel

electrophoresis

and

transferred

to

AI Med, 2026

3 / 10

polyvinylidene difluoride membranes (Millipore, Billerica,

MA, USA). Membranes were blocked with 5% non-fat milk

for 1 h at room temperature and incubated with primary

antibodies overnight at 4 °C. After washing, membranes were

incubated with horseradish peroxidase-conjugated secondary

antibodies (Cell Signaling Technology) for 1 h at room

temperature.

Signals

were

developed

using

enhanced

chemiluminescence

reagent

(Thermo

Fisher

Scientific,

Waltham, MA, USA) and imaged on a ChemiDoc MP system

(Bio-Rad, Hercules, CA, USA). Densitometric analysis was

performed using ImageJ with GAPDH as the loading control.

2.5. Lentiviral SQS Overexpression

A lentiviral vector expressing human FDFT1 (SQS) with a

GFP

reporter

and

puromycin

resistance

cassette

was

constructed by Obio Technology (Shanghai, China). RPMI-

8226 cells were transduced at a predetermined multiplicity of

infection in the presence of 6-8 µg/mL polybrene. Stable

transductants were selected with 2 µg/mL puromycin for 2

weeks, after which GFP-positive cells exceeded 90% as

confirmed by fluorescence microscopy. SQS overexpression

was

validated

by

western

blotting.

An

empty

vector-

transduced cell line served as the negative control.

2.6. Bioinformatic Analysis of MMRF-COMMPASS Data

Bulk

RNA-seq

data

and

corresponding

clinical

information

from

the

MMRF-COMMPASS

study

were

retrieved via TCGAbiolinks [34]. Primary bone marrow

CD138⁺ plasma cell samples (n = 764) were selected for

analysis. Gene-level raw counts were normalized using the

variance-stabilizing transformation implemented in DESeq2

[35].

Pearson

correlation

coefficients

between

FDFT1

expression

and

24

core

ferroptosis-related

genes

were

calculated, with Benjamini-Hochberg (BH) false discovery

rate (FDR) correction applied for multiple testing. Survival

analysis was performed using Kaplan-Meier estimation and

Cox proportional hazards regression via the survival R

package [36], with multivariate models adjusted for age, sex,

and International Staging

System (ISS) stage. Pathway

activity was estimated using gene set variation analysis

(GSVA) with single-sample gene set enrichment analysis

(ssGSEA) [37], utilizing custom ferroptosis-related gene sets

and Hallmark/KEGG gene sets from the Molecular Signatures

Database (msigdbr) [38]. Differential expression between

FDFT1-high and FDFT1-low groups (defined by median split)

was analyzed using limma [39], and gene set enrichment

analysis (GSEA) was conducted using fgsea [40]. A two-sided

P < 0.05 or FDR < 0.05 was considered statistically

significant.

2.7. Statistical Analysis

Wet-lab

experiments

were

performed

at

least

three

independent times. Data are presented as mean ± standard

deviation (SD). Comparisons between two groups were

performed using the two-tailed Student's t-test. Multiple group

comparisons were analyzed by one-way or two-way analysis

of variance (ANOVA) followed by Dunnett's post hoc test, as

appropriate. All statistical analyses for experimental data were

conducted using GraphPad Prism 10.0. For bioinformatic

analyses, statistical significance was defined as P < 0.05 or

FDR < 0.05 after BH correction.

3. Results

3.1. Celastrol Reduces MM Cell Viability and Induces

Ferroptosis-Associated Biochemical Changes

Celastrol reduced the viability of RPMI-8226, U266,

KMS-11, and SKO-007 cells in a concentration- and time-

dependent manner (Figure 1a). The 24-h IC50 values were

2.41, 2.66, 3.30, and 4.44 µmol/L, respectively, with RPMI-

8226 showing the greatest sensitivity. After 24 h of treatment,

celastrol significantly increased intracellular iron and MDA

levels, decreased GSH content (Figure 1b-d), and elevated

DCFH-DA fluorescence, indicating increased ROS production

(Figure 1e), in all four cell lines. These changes are consistent

with oxidative stress and a ferroptosis-associated phenotype.

AI Med, 2026

4 / 10

Figure 1. Celastrol reduces MM cell viability and induces

ferroptosis-associated biochemical changes. (a) Viability of

RPMI-8226,

U266,

KMS-11,

and

SKO-007

cells

after

treatment with celastrol for 24, 48, or 72 h, measured by

CCK-8

assay.

(b-d)

Intracellular

iron,

malondialdehyde

(MDA), and glutathione (GSH) levels after 24 h of celastrol

treatment. (e) Representative DCFH-DA fluorescence images

and quantification of mean fluorescence intensity after 24 h.

Data represent mean ± SD (n = 3). *P < 0.05, **P < 0.01,

***P < 0.001 vs. DMSO control.

3.2. Celastrol Suppresses SQS and Ferroptosis-Defense

Proteins; Inhibitors Rescue the Phenotype

Celastrol treatment for 24 h suppressed SQS protein

abundance in a concentration-dependent manner in all four

MM cell lines (Figure 2a). The ferroptosis-defense proteins

GPX4, SLC7A11, and FSP1, as well as the MVA pathway

master regulator SREBP2, were also reduced. To confirm

ferroptosis dependence, Fer-1 (a lipid peroxidation inhibitor)

and DFOM (an iron chelator) were employed. Both inhibitors

partially

attenuated

celastrol-induced

loss

of

viability

(Supplementary Figure S1), reduced celastrol-associated ROS

elevation, iron accumulation, and MDA production, and

partially

restored

GSH

levels

in

RPMI-8226

cells

(Supplementary

Figure

S2).

Western

blot

analysis

demonstrated

that

both

Fer-1

and

DFOM

significantly

reversed celastrol-mediated

suppression

of SQS, GPX4,

SLC7A11, and FSP1 in RPMI-8226 cells (Figure 2b). The

incomplete rescue indicates that ferroptosis contributes to

celastrol-induced cytotoxicity but may not be the sole cell

death mechanism involved.

Figure 2. Celastrol suppresses SQS and ferroptosis-defense

proteins; Fer-1 and DFOM rescue. (a) Representative western

blots

and

densitometric

quantification

of

SQS,

GPX4,

SLC7A11, FSP1, and SREBP2 upon celastrol treatment for 24

h in RPMI-8226, U266, KMS-11, and SKO-007 cells. (b)

Representative western blots for SQS, GPX4, SLC7A11, and

FSP1 in RPMI-8226 cells pre-treated with Fer-1 (5 µmol/L)

or DFOM (10 µmol/L) before celastrol exposure. Numbers

below

each

blot

indicate

densitometric

quantification

normalized to GAPDH. Data are presented as mean ± SD (n =

3). *P < 0.05, **P < 0.01, ***P < 0.001 vs. DMSO control;

&P < 0.05, &&P < 0.01, &&&P < 0.001 vs. respective

celastrol-treated group.

3.3. SQS Overexpression Partially Protects RPMI-8226

Cells from Celastrol

To provide genetic evidence for the functional role of SQS,

stable SQS-overexpressing RPMI-8226 cells were generated.

SQS protein abundance was increased by approximately 1.38-

fold relative to the empty vector control (Figure 3a, b).

Compared with vector control cells, SQS-overexpressing cells

showed significantly greater viability after celastrol exposure

(Figure 3c), reduced intracellular iron and MDA levels, and

partial preservation of GSH (Figure 3d-f). These results

provide genetic evidence that SQS expression levels modulate

celastrol sensitivity and ferroptosis-associated biochemical

changes in MM cells.

AI Med, 2026

5 / 10

Figure 3. SQS overexpression partially protects RPMI-8226

cells from celastrol-induced cytotoxicity. (a, b) Validation of

SQS overexpression by western blotting and densitometry. (c)

Cell viability after celastrol treatment. (d-f) Intracellular iron,

MDA, and GSH levels. Data represent mean ± SD (n = 3). *P

< 0.05, **P < 0.01, ***P < 0.001 vs. DMSO; &P < 0.05,

&&P < 0.01, &&&P < 0.001 vs. respective celastrol-treated

group.

3.4. Clinical Transcriptomic Context from MMRF-

COMMPASS

To place the in vitro findings in a clinical context, we

analyzed the MMRF-COMMPASS cohort (n = 764 primary

bone

marrow

CD138⁺

plasma

cell

samples).

FDFT1

expression did not differ significantly across ISS stages

(Kruskal-Wallis P = 0.068) and was not associated with

overall survival in univariate (HR = 1.064, 95% CI: 0.908-

1.247, P = 0.441) or multivariate Cox regression adjusted for

age, sex, and ISS stage (HR = 1.059, P = 0.482). These

findings indicate that FDFT1 is not an independent prognostic

biomarker in MM, consistent with its role as a metabolic

enzyme supporting ferroptosis defense rather than a driver of

disease progression.

Pearson

correlation

analysis

revealed

significant

associations between FDFT1 and ferroptosis core genes after

BH correction (Table 1; Figure 4a). Key findings include a

positive correlation with FSP1/AIFM2 (r = 0.135, FDR =

0.001), a negative correlation with IREB2/IRP2 (r = -0.240,

FDR < 0.001), the iron regulatory protein that promotes iron

uptake and ferroptosis sensitivity, and the strongest positive

correlation with GSR (glutathione reductase; r = 0.425, FDR <

0.001),

linking

MVA

pathway

activity

to

glutathione

homeostasis. This pattern — positive with FSP1 and GSR,

negative with IREB2 — is consistent with a model in which

elevated SQS expression supports a ferroptosis-resistant

cellular state.

GSVA/ssGSEA identified significant pathway activity

differences between FDFT1-high and FDFT1-low tumors

(Table 1; Figure 4b). Ferroptosis defense was elevated in

FDFT1-high tumors (P_adj = 1.37 × 10^-6), as was the

ferroptosis core gene set (P_adj = 1.7 × 10^-4), while the

ferroptosis induction pathway was not significantly different

(P_adj

=

0.364).

GSEA

confirmed

ferroptosis

defense

enrichment (NES = 2.13, P_adj = 2.68 × 10^-3) and identified

oxidative stress response (NES = 2.53), mTORC1 signaling

(NES = 2.88), and sterol biosynthesis (NES = 3.03) as top

enriched pathways in FDFT1-high tumors (Figure 4c, d). The

convergence

of

GSVA

and

GSEA

—

both

showing

ferroptosis defense but not induction elevated in FDFT1-high

tumors — provides cohort-level support for the hypothesis

that SQS supports ferroptosis defense in MM.

Figure 4. Clinical transcriptomic analysis of FDFT1 in

MMRF-COMMPASS (n = 764). (a) Correlation heatmap:

MVA pathway genes vs. ferroptosis core genes. (b) GSVA

pathway comparison: FDFT1-high vs. FDFT1-low (selected

pathways). (c) GSEA enrichment dotplot for top enriched

pathways. (d) GSEA enrichment plot for the ferroptosis

defense gene set.

AI Med, 2026

6 / 10

Table 1. FDFT1-ferroptosis gene correlations and pathway

activity in MMRF-COMMPASS (n = 764)

Gene / Pathway

r or NES

P value

FDR (BH)

Correlation (Pearson

r)

GSR (glutathione

reductase)

+0.425

6.83 × 10⁻³⁵

< 0.001

FSP1/AIFM2

+0.135

1.93 × 10⁻⁴

0.001

IREB2 (IRP2)

-0.240

1.93 × 10⁻¹¹

< 0.001

GPX4

+0.060

0.096

0.145 (NS)

SLC7A11

-0.051

0.161

0.228 (NS)

GSVA pathway

(P_adj)

Ferroptosis defense

-

-

1.37 × 10⁻⁶

Ferroptosis core

-

-

1.7 × 10⁻⁴

Ferroptosis induction

-

-

0.364 (NS)

GSEA (NES)

Ferroptosis defense

2.13

-

2.68 × 10⁻³

Oxidative stress

response

2.53

-

5.47 × 10⁻²¹

mTORC1 signaling

2.88

-

3.72 × 10⁻²⁰

Sterol biosynthesis

3.03

-

3.94 × 10⁻¹³

NS, not significant. Full correlation matrix and GSVA results are

provided in Supplementary Materials.

4. Discussion

In this study, we demonstrate that celastrol exerts anti-

MM efficacy by inducing a regulated cell death cascade with

a robust ferroptotic component, and that this process is

associated with suppression of the MVA-SQS lipid defense

axis. Through a combination of phenotypic characterization,

pharmacological rescue, genetic overexpression, and clinical

transcriptomic analysis, we identify SQS as a functional

modulator of celastrol responsiveness in MM.

Ferroptosis is characterized by catastrophic lipid ROS

accumulation resulting from failure of cellular antioxidant

defense systems, primarily governed by the System Xc^--GSH-

GPX4 axis and the parallel FSP1-CoQ10 system [7-12]. Our

results show that celastrol exposure in MM cells leads to iron

overload, GSH depletion, MDA accumulation, and ROS

elevation, all of which are attenuated by the ferroptosis

inhibitors Fer-1 and DFOM. Importantly, SQS protein levels

were

downregulated

upon

celastrol

exposure,

and

this

downregulation

was

partially

reversed

by

ferroptosis

inhibitors.

These

observations

place

SQS

within

the

ferroptotic

cascade,

although

whether

its

suppression

represents an upstream trigger or a secondary consequence of

oxidative injury warrants further investigation.

The functional relevance of SQS downregulation is

supported by genetic rescue experiments. Enforced SQS

overexpression in RPMI-8226 cells was sufficient to partially

attenuate celastrol-induced cytotoxicity, reduce intracellular

iron and MDA accumulation, and preserve GSH levels. SQS

catalyzes the first committed step in sterol biosynthesis within

the

MVA

pathway

[21,22].

MVA-derived

metabolites,

including CoQ10 and squalene, have been shown to reinforce

ferroptosis defense [23-25]. It is therefore plausible that SQS

suppression by celastrol compromises the cellular pool of

these protective intermediates, thereby lowering the threshold

for lipid peroxidation. The FSP1-CoQ10 axis [11,12] and

DHODH-mediated

mitochondrial

defense

[13]

represent

parallel ferroptosis-suppressing systems that may also be

affected when MVA pathway flux is disrupted. Interestingly,

the relationship between MVA pathway perturbation and

ferroptosis appears context-dependent: squalene accumulation

protects certain solid tumor cells and lymphomas from

oxidative cell death [23], whereas in our hematological model,

SQS downregulation is associated with enhanced ferroptotic

sensitivity.

The downregulation of SREBP2 observed in our western

blot data is notable, as SREBP2 is the master transcriptional

regulator of MVA pathway genes including FDFT1. Celastrol

has previously been reported to modulate SREBP signaling in

other cancer contexts [27,28]. It is therefore possible that

celastrol suppresses SQS expression at least in part through

SREBP2 downregulation, rather than through direct physical

interaction

with

the

SQS

protein.

This

transcriptional

mechanism would be consistent with the observed parallel

downregulation of GPX4, SLC7A11, and FSP1, which are

known targets of SREBP2 and related lipid metabolic

regulatory networks. We acknowledge that the present study

does not establish a direct physical interaction between

celastrol and SQS, and the causal chain linking celastrol,

SREBP2, SQS, and ferroptosis requires further mechanistic

dissection.

The clinical transcriptomic analysis of the MMRF-

COMMPASS cohort (n = 764) provides translational context.

FDFT1 expression was not independently associated with

overall survival, indicating that SQS is not a prognostic

biomarker but rather a functional metabolic target. This is not

surprising, as many validated drug targets are not themselves

prognostic

markers;

their

therapeutic

relevance

lies

in

druggability

and

functional

contribution

to

resistance

phenotypes.

Critically,

FDFT1-high

tumors

showed

significantly elevated ferroptosis defense pathway activity but

not ferroptosis induction pathway activity, providing cohort-

level evidence that SQS expression is associated with a

ferroptosis-resistant state in MM. The positive correlation

between FDFT1 and FSP1 (r = 0.135) and GSR (r = 0.425),

together with the negative correlation with IREB2 (r = -0.240),

connect MVA pathway activity to both the glutathione-

AI Med, 2026

7 / 10

dependent and glutathione-independent ferroptosis defense

arms, as well as to iron regulatory signaling.

Several limitations should be acknowledged. First, the

present study is restricted to in vitro models; rigorous in vivo

validation using human MM xenograft or immunocompetent

models will be essential to establish translational relevance.

Second, loss-of-function experiments (SQS knockdown or

knockout) and pharmacological controls with known SQS

inhibitors such as zaragozic acid A or lapaquistat would

further strengthen the causal evidence. Third, the downstream

metabolic

consequences

of

SQS

suppression-particularly

regarding carbon flux redistribution within the MVA pathway

and direct measurement of CoQ10, squalene, and cholesterol

depletion-remain

to

be

elucidated

using

stable-isotope-

resolved metabolomics. Fourth, the mechanism by which

celastrol downregulates SQS and SREBP2, whether through

direct transcriptional modulation, post-translational effects, or

upstream signaling interference, requires further investigation.

Fifth, while the MMRF-COMMPASS analysis

provides

correlative

clinical

support,

prospective

validation

in

independent cohorts is needed.

From a translational perspective, our findings position the

MVA-SQS axis as a candidate metabolic vulnerability in MM.

Given that statins are clinically used, and SQS inhibitors

represent

a

promising

repurposing

strategy

[41,42],

repurposing

these

agents

or

developing

SQS-selective

inhibitors in combination with established anti-MM therapies

represents a rational future direction. The observation that

celastrol, a natural product with known anti-inflammatory and

anti-tumor activities [27-30], can modulate this axis further

supports the exploration of MVA-targeting strategies for MM

treatment.

5. Conclusions

Our study demonstrates that celastrol induces ferroptosis-

associated cell death in multiple myeloma cells, and that this

process is functionally associated with suppression of the

MVA-SQS lipid defense axis. Pharmacological rescue with

ferroptosis inhibitors, genetic overexpression of SQS, and

clinical

transcriptomic

evidence

from

the

MMRF-

COMMPASS cohort collectively support SQS as a ferroptosis

defense

node

whose

modulation

influences

celastrol

sensitivity in MM. These findings provide a preclinical

rationale for further investigation of SQS-targeted strategies to

overcome drug resistance in MM.

Acknowledgments: Not applicable.

Ethics approval and consent to participate: Not applicable.

Clinical trial number: Not applicable.

Availability

of

data

and

materials:

The

MMRF-

COMMPASS dataset used in this study is publicly available

via the GDC Data Portal (https://portal.gdc.cancer.gov/). All

other data supporting the results of this study are included

within the article and its supplementary materials.

Competing interests: The authors declare no conflicts of

interest.

Funding: This research was funded by the Shanghai Pudong

New Area Science, Technology and Economy Commission,

grant number PKJ2024-Y68.

Authors'

contributions:

Methodology,

validation,

investigation, and writing—original draft preparation, H.Y.;

conceptualization, supervision, funding acquisition, writing—

review and editing, B.P. All authors have read and agreed to

the published version of the manuscript.

References

[1] Malard, F.; Neri, P.; Bahlis, N. J.; Terpos, E.; Moukalled,

N.; Hungria, V. T. M.; Manier, S.; Mohty, M. Multiple

myeloma. Nat Rev Dis Primers 2024, 10, 45. DOI:

10.1038/s41572-024-00529-7.

[2] Scalzulli, E.; Grammatico, S.; Vozella, F.; Petrucci, M. T.

Proteasome inhibitors for the treatment of multiple

myeloma. Expert Opin Pharmacother 2018, 19, 375-386.

DOI: 10.1080/14656566.2018.1441287.

[3] Garfall, A. L. New Biological Therapies for Multiple

Myeloma.

Annu

Rev

Med

2024,

75,

13-29.

DOI:

10.1146/annurev-med-050522-033815.

[4] Gertz, M. A. Is There Still a Role for Stem Cell

Transplantation in Multiple Myeloma?. Hematol Oncol

Clin

North

Am

2024,

38,

407-420.

DOI:

10.1016/j.hoc.2023.12.005.

[5] Forster, S.; Radpour, R.; Ochsenbein, A. F. Molecular and

immunological

mechanisms

of

clonal

evolution

in

multiple myeloma. Front Immunol 2023, 14, 1243997.

DOI: 10.3389/fimmu.2023.1243997.

[6] Neri, P.; Leblay, N.; Lee, H.; Gulla, A.; Bahlis, N. J.;

Anderson, K. C. Just scratching the surface: novel

treatment approaches for multiple myeloma targeting cell

membrane proteins. Nat Rev Clin Oncol 2024, 21, 590-

609. DOI: 10.1038/s41571-024-00913-y.

[7] Dixon, S. J.; Olzmann, J. A. The cell biology of ferroptosis.

Nat Rev Mol Cell Biol 2024, 25, 424-442. DOI:

10.1038/s41580-024-00703-5.

AI Med, 2026

8 / 10

[8] Stockwell, B. R.; Friedmann Angeli, J. P.; Bayir, H.; Bush,

A. I.; Conrad, M.; Dixon, S. J.; Fulda, S.; Gascón, S.;

Hatzios, S. K.; Kagan, V. E.; Noel, K.; Jiang, X.;

Linkermann, A.; Murphy, M. E.; Overholtzer, M.; Oyagi,

A.; Pagnussat, G. C.; Park, J.; Ran, Q.; Rosenfeld, C. S.;

Salnikow, K.; Tang, D.; Torti, F. M.; Torti, S. V.;

Toyokuni, S.; Woerpel, K.; Zhang, D. D. Ferroptosis: A

Regulated Cell Death Nexus Linking Metabolism, Redox

Biology, and Disease. Cell 2017, 171, 273-285. DOI:

10.1016/j.cell.2017.09.021.

[9] Lei, G.; Zhuang, L.; Gan, B. Targeting ferroptosis as a

vulnerability in cancer. Nat Rev Cancer 2022, 22, 381-

396. DOI: 10.1038/s41568-022-00459-0.

[10] Zheng, J.; Conrad, M. Ferroptosis: when metabolism

meets cell death. Physiol Rev 2025, 105, 651-706. DOI:

10.1152/physrev.00031.2024.

[11] Bersuker, K.; Hendricks, J. M.; Li, Z.; Magtanong, L.;

Ford, B.; Tang, P. H.; Roberts, M. A.; Tong, B.;

Maimone, T. J.; Zoncu, R.; Bassik, M. C.; Nomura, D. K.;

Dixon, S. J.; Olzmann, J. A. The CoQ oxidoreductase

FSP1 acts parallel to GPX4 to inhibit ferroptosis. Nature

2019, 575, 688-692. DOI: 10.1038/s41586-019-1705-2.

[12] Doll, S.; Freitas, F. P.; Shah, R.; Aldrovandi, M.; da Silva,

M. C.; Ingold, I.; Goya Grocin, A.; Xavier da Silva, T. N.;

Panzilius, E.; Scheel, C. H.; Mourão, A.; Buday, K.; Sato,

M.; Wanninger, J.; Vignane, T.; Mohana, V.; Rehberg,

M.; Flatley, A.; Schepers, A.; Kurz, A.; White, D.; Sauer,

M.; Sattler, M.; Tate, E. W.; Schmitz, W.; Schulze, A.;

O’Donnell, V.; Proneth, B.; Popowicz, G. M.; Pratt, D. A.;

Angeli, J. P. F.; Conrad, M. FSP1 is a glutathione-

independent ferroptosis suppressor. Nature 2019, 575,

693-698. DOI: 10.1038/s41586-019-1707-0.

[13] Mao, C.; Liu, X.; Zhang, Y.; Lei, G.; Yan, Y.; Lee, H.;

Koppula, P.; Wu, S.; Zhuang, L.; Fang, B.; Poyurovsky,

M. V.;

Olszewski,

K.;

Gan,

B. DHODH-mediated

ferroptosis defence is a targetable vulnerability in cancer.

Nature 2021, 593, 586-590. DOI: 10.1038/s41586-021-

03539-7.

[14]

Campanella,

A.;

Santambrogio,

P.;

Fontana,

F.;

Frenquelli, M.; Cenci, S.; Marcatti, M.; Sitia, R.; Tonon,

G.; Camaschella, C. Iron increases the susceptibility of

multiple myeloma cells to bortezomib. Haematologica

2013, 98, 971-979. DOI: 10.3324/haematol.2012.074872.

[15] Camiolo, G.; Barbato, A.; Giallongo, C.; Vicario, N.;

Romano, A.; Parrinello, N. L.; Parenti, R.; Sandoval, J. C.;

García-Moreno, D.; Lazzarino, G.; Avola, R.; Palumbo,

G. A.; Mulero, V.; Li Volti, G.; Tibullo, D.; Di Raimondo,

F. Iron regulates myeloma cell/macrophage interaction

and drives resistance to bortezomib. Redox Biol 2020, 36,

101611. DOI: 10.1016/j.redox.2020.101611.

[16] Xian, M.; Wang, Q.; Xiao, L.; Zhong, L.; Xiong, W.; Ye,

L.; Su, P.; Zhang, C.; Li, Y.; Orlowski, R. Z.; Zhan, F.;

Ganguly, S.;

Zu,

Y.; Qian,

J.; Yi, Q.

Leukocyte

immunoglobulin-like receptor B1 (LILRB1)

protects

human multiple myeloma cells from ferroptosis by

maintaining cholesterol homeostasis. Nat Commun 2024,

15, 5767. DOI: 10.1038/s41467-024-50073-x.

[17] Zhang, J.; Liu, Y.; Zuo, L.; Fan, F.; Yan, H.; Zhao, F.; Li,

J.; Ma, C.; Li, Q.; Xu, A.; Xu, J.; Zhang, B.; Hu, Y.; Sun,

C. Class II ferroptosis inducers are a novel therapeutic

approach for t(4;14)-positive multiple myeloma. Blood

Adv

2024,

8,

5022-5038.

DOI:

10.1182/bloodadvances.2023010335.

[18] Jiang, H.; Wang, L.; Zhang, Q.; Wang, S.; Jia, L.; Cheng,

H.; Wang, J.; Li, X.; Xie, Y.; Wang, Y.; Hu, M.; Guo, J.;

Li, Q.; Peng, Z.; Wang, M.; Xie, Y.; Li, T.; Wang, Y.;

Geng, B. D.; Swaminathan, S.; Bergsagel, P. L.; Liu, Z.

Bone marrow stromal cells dictate lanosterol biosynthesis

and ferroptosis of multiple myeloma. Oncogene 2024, 43,

1644-1653. DOI: 10.1038/s41388-024-03020-5.

[19] Guerra, B.; Recio, C.; Aranda-Tavío, H.; Guerra-

Rodríguez, M.; García-Castellano, J. M.; Fernández-

Pérez, L. The Mevalonate Pathway, a Metabolic Target in

Cancer Therapy. Front Oncol 2021, 11, 626971. DOI:

10.3389/fonc.2021.626971.

[20] Xue, L.; Qi, H.; Zhang, H.; Ding, L.; Huang, Q.; Zhao,

D.; Wu, B. J.; Li, X. Targeting SREBP-2-Regulated

Mevalonate Metabolism for Cancer Therapy. Front Oncol

2020, 10, 1510. DOI: 10.3389/fonc.2020.01510.

[21] Chang, H.; Cheng, T.; Wang, A. H. Structure, catalysis,

and inhibition mechanism of prenyltransferase. IUBMB

Life 2021, 73, 40-63. DOI: 10.1002/iub.2418.

[22] Mavridi, D.; Katavati, T.; Kourounakis, A. P. The

Contribution of Cholesterol and Squalene Synthase in

Cancer:

Molecular

Mechanisms,

Lipid

Rafts

and

Therapeutic Approaches. Med Res Rev 2025, 46, 352-381.

DOI: 10.1002/med.70012.

[23] Garcia-Bermudez, J.; Baudrier, L.; Bayraktar, E. C.; Shen,

Y.; La, K.; Guarecuco, R.; Yucel, B.; Fiore, D.; Tavora,

B.; Freinkman, E.; Chan, S. H.; Lewis, C.; Min, W.;

Inghirami, G.; Sabatini, D. M.; Birsoy, K. Squalene

accumulation

in

cholesterol

auxotrophic

lymphomas

prevents oxidative cell death. Nature 2019, 567, 118-122.

DOI: 10.1038/s41586-019-0945-5.

AI Med, 2026

9 / 10

[24] Sun, Q.; Liu, D.; Cui, W.; Cheng, H.; Huang, L.; Zhang,

R.; Gu, J.; Liu, S.; Zhuang, X.; Lu, Y.; Chu, B.; Li, J.

Cholesterol mediated

ferroptosis

suppression

reveals

essential roles of Coenzyme Q and squalene. Commun

Biol 2023, 6, 1108. DOI: 10.1038/s42003-023-05477-8.

[25] Tosi, G.; Paoli, A.; Zuccolotto, G.; Turco, E.; Simonato,

M.; Tosoni, D.; Tucci, F.; Lugato, P.; Giomo, M.;

Elvassore, N.; Rosato, A.; Cogo, P.; Pece, S.; Santoro, M.

M. Cancer cell stiffening via CoQ10 and UBIAD1

regulates ECM signaling and ferroptosis in breast cancer.

Nat Commun 2024, 15, 8214. DOI: 10.1038/s41467-024-

52523-y.

[26] Yang, H.; Wang, J.; Xie, J.; Yang, H.; Wu, X. Molecular

mechanisms and delivery strategies of celastrol targeting

inflammation-associated

autoimmune

diseases.

J

Ethnopharmacol

2026,

355,

120719.

DOI:

10.1016/j.jep.2025.120719.

[27] Jiang, Q.; Liu, Z.; Cheng, J.; Geng, Z.; Gu, Y.; Wei, Y.;

Yu, M.; Bi, Y. The Antitumor Potential of Celastrol:

Research

Progress

on

Antitumor

Mechanisms

and

Strategies

for

Toxicity

Reduction

with

Efficacy

Enhancement.

Biomolecules

2026,

16,

620.

DOI:

10.3390/biom16050620.

[28] Lu, Y.; Liu, Y.; Zhou, J.; Li, D.; Gao, W. Biosynthesis,

total synthesis, structural modifications, bioactivity, and

mechanism of action of the quinone‐methide triterpenoid

celastrol. Med Res Rev 2021, 41, 1022-1060. DOI:

10.1002/med.21751.

[29] Shanmugam, M. K.; Ahn, K. S.; Lee, J. H.; Kannaiyan,

R.; Mustafa, N.; Manu, K. A.; Siveen, K. S.; Sethi, G.;

Chng, W. J.; Kumar, A. P. Celastrol Attenuates the

Invasion and Migration and Augments the Anticancer

Effects of Bortezomib in a Xenograft Mouse Model of

Multiple Myeloma. Front Pharmacol 2018, 9, 365. DOI:

10.3389/fphar.2018.00365.

[30] Zhong, Y.; Xu, G.; Huang, S.; Zhao, L.; Zeng, Y.; Xiao,

X.; An, J.; Liu, J.; Yang, T. Celastrol induce apoptosis of

human multiple myeloma cells involving inhibition of

proteasome activity. Eur J Pharmacol 2019, 853, 184-

192. DOI: 10.1016/j.ejphar.2019.03.036.

[31] Ding, Y.; Bai, Y.; Chen, T.; Chen, S.; Feng, W.; Ma, S.;

Zhang, Q. Disruption of the sorcin‒PAX5 protein‒protein

interaction

induces

ferroptosis

by

promoting

the

FBXL12-mediated

ubiquitination

of

ALDH1A1

in

pancreatic cancer. J Hematol Oncol 2025, 18, 27. DOI:

10.1186/s13045-025-01680-8.

[32] Liu, M.; Fan, Y.; Li, D.; Han, B.; Meng, Y.; Chen, F.;

Liu, T.; Song, Z.; Han, Y.; Huang, L.; Chang, Y.; Cao, P.;

Nakai, A.; Tan, K. Ferroptosis inducer erastin sensitizes

NSCLC cells to celastrol through activation of the ROS–

mitochondrial fission–mitophagy axis. Mol Oncol 2021,

15, 2084-2105. DOI: 10.1002/1878-0261.12936.

[33] Luo, P.; Liu, D.; Zhang, Q.; Yang, F.; Wong, Y.-K.; Xia,

F.; Zhang, J.; Chen, J.; Tian, Y.; Yang, C.; Dai, L.; Shen,

H.-M.; Wang, J. Celastrol induces ferroptosis in activated

HSCs

to

ameliorate

hepatic

fibrosis

via

targeting

peroxiredoxins and HO-1. Acta Pharm Sin B 2022, 12,

2300-2314. DOI: 10.1016/j.apsb.2021.12.007.

[34] Colaprico, A.; Silva, T. C.; Olsen, C.; Garofano, L.; Cava,

C.; Garolini, D.; Sabedot, T. S.; Malta, T. M.; Pagnotta, S.

M.;

Castiglioni,

I.;

Ceccarelli,

M.;

Bontempi,

G.;

Noushmehr,

H.

TCGAbiolinks:

an

R/Bioconductor

package for integrative analysis of TCGA data. Nucleic

Acids Res 2016, 44, e71. DOI: 10.1093/nar/gkv1507.

[35] Love, M. I.; Huber, W.; Anders, S. Moderated estimation

of fold change and dispersion for RNA-seq data with

DESeq2.

Genome

Biol

2014,

15,

550.

DOI:

10.1186/s13059-014-0550-8.

[36] Therneau, T. M. survival: A Package for Survival

Analysis in R. R package version 3.5-5; 2023. Available

online: https://CRAN.R-project.org/package=survival.

[37] Hänzelmann, S.; Castelo, R.; Guinney, J. GSVA: gene set

variation analysis for microarray and RNA-Seq data.

BMC Bioinformatics 2013, 14, 7. DOI: 10.1186/1471-

2105-14-7.

[38] Dolgalev, I. msigdbr: MSigDB Gene Sets for Multiple

Organisms in a Tidy Data Format. R package version

7.5.1;

2022.

Available

online:

https://CRAN.R-

project.org/package=msigdbr.

[39] Ritchie, M. E.; Phipson, B.; Wu, D.; Hu, Y.; Law, C. W.;

Shi,

W.; Smyth,

G. K. limma

powers

differential

expression analyses for RNA-sequencing and microarray

studies.

Nucleic

Acids

Res

2015,

43,

e47.

DOI:

10.1093/nar/gkv007.

[40] Korotkevich, G.; Sukhov, V.; Budin, N.; Shpak, B.;

Artyomov, M. N.; Sergushichev, A. Fast gene set

enrichment analysis. bioRxiv 2016, DOI: 10.1101/060012.

[41] Longo, J.; van Leeuwen, J. E.; Elbaz, M.; Branchard, E.;

Penn, L. Z. Statins as Anticancer Agents in the Era of

Precision Medicine. Clin Cancer Res 2020, 26, 5791-

5800. DOI: 10.1158/1078-0432.CCR-20-1967.

[42] Kourounakis, A. P.; Bavavea, E. New applications of

squalene synthase inhibitors: Membrane cholesterol as a

AI Med, 2026

10 / 10

therapeutic target. Arch Pharm (Weinheim) 2020, 353,

2000085. DOI: 10.1002/ardp.202000085.