LMAN2 promotes breast cancer progression and deduces cell sensitivity to doxorubicin through stearoyl-CoA desaturase

Abstract

Background

Dysregulated lipid metabolism and chemoresistance are key drivers of breast cancer progression. Lectin, mannose-binding 2 (LMAN2) is frequently overexpressed in human breast tumors and functions as an oncogenic driver. However, whether LMAN2 contributes to chemoresistance remains unknown.

Methods

We integrated multi-omics data from 1,085 primary tumors and matched normal tissues (from GEPIA and UALCAN) with functional studies in breast cancer cell lines and a doxorubicin (ADM)-treated nude mouse xenograft model. LMAN2 expression was modulated via siRNA/shRNA-mediated silencing or lentivirus-driven overexpression. Cellular phenotypes-including proliferation, migration, apoptosis, and response to ADM were systematically assessed. RNA-sequencing, untargeted lipidomics, and rescue experiments identified stearoyl-CoA desaturase (SCD) as a critical downstream effector. IC50 shifts and epistasis analysis further validated the role of the LMAN2/SCD axis in chemoresistance.

Results

LMAN2 mRNA was elevated across all molecular subtypes (luminal > HER2 > triple-negative) and predicted poorer overall survival (P = 5 × 10⁻⁴) and progression-free survival (P = 0.018). Silencing LMAN2 reduced clonogenicity by ~ 45% and migration by 37–63%, whereas overexpression increased cell viability by 1.4–1.7-fold and doubled motility. Knockdown of LMAN2 decreased the ADM IC50 by 4–5 fold, abolished macroscopic colony formation, and elevated apoptosis rates from 15 to 18% to 39–41%; these effects were reversed upon LMAN2 overexpression. In vivo, shLMAN2 combined with ADM reduced tumor volume and weight by 72% and 75%, respectively, compared to ADM alone (P < 0.001). Mechanistically, LMAN2 loss downregulated genes involved in “cholesterol homeostasis” and reduced total cellular cholesterol by 24%. SCD emerged as the most significantly downregulated enzyme and fully rescued the phenotypic and chemoresistance effects resulting from LMAN2 modulation. Epistasis experiments confirmed that LMAN2-mediated chemoresistance strictly depends on SCD function.

Conclusions

LMAN2 is a robust prognostic biomarker that promotes breast tumor growth and anthracycline resistance by enabling SCD-dependent lipid desaturation. Therapeutic targeting of the LMAN2/SCD axis represents a promising strategy to overcome chemoresistance in breast cancer.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13058-026-02343-3.

Introduction

Breast cancer stands as the foremost malignancy among women. An approximate 2.3 million diagnoses and 670,000 deaths were recorded globally for 2022; by 2050, these figures are projected to climb to roughly 3.2 million diagnoses and 1.1 million fatalities [1, 2]. Chemotherapy is a mainstay in the therapeutic arsenal against breast cancer [3]. Doxorubicin (ADM) (Adriamycin^®; Pfizer, New York, NY, USA) serves as the cornerstone adjuvant therapy for a wide range of solid tumors, most notably early-stage and metastatic breast cancer, the most common malignancy afflicting women worldwide [4]. ADM resistance continues to cripple breast cancer chemotherapy, arising from a tangled web of processes—the drug being pumped back out of cells, programmed death being switched off, and pro-survival autophagy being switched on [5]. Hence, a rational strategy to amplify ADM’s cytotoxic punch and tilt the clinical balance in favor of breast cancer patients is urgently needed.

LMAN2 (Lectin, mannose-binding 2), a protein‐coding gene, produces a type-I transmembrane lectin that continuously traffics among the cell surface, Golgi stack, and endoplasmic reticulum [6, 7]. Bioinformatics analysis suggests that LMAN2 intersects the stem cell program, metastatic drive, differentiation trajectory, and DNA-damage repair circuitry specific to HER2-positive breast tumors [8]. Feng et al. subsequently demonstrated that LMAN2 operates as an oncoprotein: it fuels proliferation, accelerates cell-cycle transit, and drives invasion while shielding breast cancer cells from apoptosis; silencing the gene in vivo resensitizes DDP-resistant tumors to cisplatin [9]. Xiao et al. showed that depleting LMAN2 dampens tumor growth and inflammation in HER2 + breast cancer, an effect tied to its interaction with HEATR3 and the consequent shutdown of the Akt/ERK/NF-κB signaling axis [10].

Here, we interrogated LMAN2’s role in ADM resistance and uncovered a previously unrecognized pathway. LMAN2 is markedly up-regulated in breast cancer specimens versus matched normal tissue, and its high expression forecasts worse outcome. Mechanistically, we provide the first evidence that LMAN2 tilts the cholesterol-balance program by transactivating SCD, thereby hard-wiring ADM tolerance in breast cancer cells. Consequently, the LMAN2/SCD axis emerges as both a prognostic biomarker and a druggable vulnerability for patients failing anthracycline therapy.

Materials and methods

Specimens from the Cancer Genome Atlas

LMAN2 abundance and predictive value were interrogated through GEPIA (http://gepia.cancer-pku.cn), comparing breast tumors with normal breast tissues. Subclass-specific expression across luminal, HER2-positive, and triple-negative subtypes was mapped with UALCAN (http://ualcan.path.uab.edu). The survival impact of LMAN2 mRNA was subsequently gauged in breast cancer cases via the Kaplan–Meier Plotter resource.

Cell growth and transfection

MCF-7 and MDA-MB-231 strains were provided by Procell (Wuhan, China) and cultured in DMEM plus 10% FBS and 1% penicillin–streptomycin at 37 °C under 5% CO2. ADM (Selleck, Shanghai, China) was administered at 50 nM. Gene-specific siRNAs against LMAN2 (cluster ID 10960; siLMAN2-1, 5′-CGGAUAUAACUGACGGCAACA-3′; siLMAN2-2, 5′-CCAGGAGCCAAUGUGAACTTT-3′) and SCD (cluster ID 6319; siSCD, 5′-CGUCCUUAUGACAAGAACATT-3′), together with a non-targeting control (siCtrl), were synthesized by SyngenTech (Beijing, China). Transfection was carried out with Lipofectamine 3000 (2319757; Invitrogen) per the supplier’s protocol. LMAN2 and SCD overexpression lentivectors (Cat# BVC-LW308) were produced by Genechem (Shanghai, China).

RNA isolation and real-time polymerase chain reaction (PCR)

Total RNA was extracted from treated breast cancer cell lines with TRIzol regeant (Invitrogen). Five micrograms of RNA were converted to first-strand cDNA with the Hifair II 1st Strand cDNA Synthesis Kit (Yeasen, Shanghai, China). Quantitative PCR was then run on a LightCycler 480 (Roche Holding AG, Basel, Switzerland) with the BeyoFast SYBR Green One-Step qRT-PCR Kit (Beyotime, Shanghai, China): 40 cycles of 95 °C for 15 s, 60 °C for 20 s, and 72 °C for 15 s. Expression levels were quantified using the ΔΔCt method, with β-actin as the endogenous reference; all primer sequences are provided in Table 1.

Table 1.

Oligonucleotide primer sets used for human qRT-PCR

No.	Gene	Forward (5’-3’)	Reverse (5’-3’)
1	LMAN2	CTGGCCCCACTACACCTCT	GCTCCCGCTTGAGATGTTCA
2	SCD	TCTAGCTCCTATACCACCACCA	TCGTCTCCAACTTATCTCCTCC
3	ACSL4	CATCCCTGGAGCAGATACTCT	TCACTTAGGATTTCCCTGGTCC
4	FASN	AAGGACCTGTCTAGGTTTGATGC	TGGCTTCATAGGTGACTTCCA
5	ACSL3	GCCGAGTGGATGATAGCTGC	ATGGCTGGACCTCCTAGAGTG
6	ACAT2	GCGGACCATCATAGGTTCCTT	ACTGGCTTGTCTAACAGGATTCT
7	FADS2	TGACCGCAAGGTTTACAACAT	AGGCATCCGTTGCATCTTCTC
8	ELOVL5	TAACAGGAGTATGGGAAGGCA	ACCAGAGGACACGGATAATCTT
9	ACADSB	GATGGCAAATGTAGACCCTACC	AAGGCCCGGAGTATCACGA
10	FADS1	CTACCCCGCGCTACTTCAC	CGGTCGATCACTAGCCACC
11	ACACA	ATGTCTGGCTTGCACCTAGTA	CCCCAAAGCGAGTAACAAATTCT
12	β-actin	GTCTGCCTTGGTAGTGGATAATG	TCGAGGACGCCCTATCATGG
13	GAPDH	ACAACTTTGGTATCGTGGAAGG	GCCATCACGCCACAGTTTC

Western blot

Whole-cell lysates were solubilized in RIPA buffer (Thermo Fisher Scientific). Proteins were separated on 8% SDS-PAGE gels, transferred to PVDF membranes, blocked in 5% non-fat milk, and incubated with anti-LMAN2 (12455-MM04; Sino Biological, Beijing, China) followed by horseradish perioxidase–conjugated mouse secondary antibody (SA00001-1; Proteintech). Signals were visualized with ECL reagent (Millipore) and quantified in ImageJ (NIH, USA) relative to a GAPDH loading control (60004-1-Ig; Proteintech).

CCK-8 assay

Cell viability was tracked with the CCK-8 assay. In brief, 3 × 10³ cells per well were seeded overnight in 96-well plates; at 0, 24, 48, 72 and 96 h, 10 µL of CCK-8 reagent was added, and the plates were returned to 37 °C for 1 h. Absorbance was then read at 492 nm on a microplate reader (Bio-Rad Laboratories, Hercules, CA, USA). For IC₅₀ determination, transfected cells were plated at 5 × 10³ per well in 96-well plates. After attachment, increasing ADM concentrations (0–2,500 nM) were added for 48 h, and viability was quantified using CCK-8 assay. Absorbance at 450 nm was recorded on a microplate reader (Thermo Fisher Scientific), and IC₅₀ values were interpolated from the resulting dose–response curves.

Cell Colony formation assay

Briefly, 1 × 10³ cells were plated per well of six-well plates in 2 mL of medium and incubated at 37 °C and 5% CO₂ for 14 days, with medium renewal every third day. Colonies were then rinsed with PBS, fixed in 4% paraformaldehyde for 15 min, stained with 1% crystal violet (V5265; Sigma) for 5 min, and counted manually under a light microscope.

Transwell assay

Cell migration was gauged in 24-well Transwell inserts (8-µm pores; Corning). Then, 2 × 10⁵ serum-starved cells were seeded in the upper chamber (200 µL), while the lower chamber received 600 µL of DMEM–10% FBS as chemoattractant. After 24 h at 37 °C / 5% CO₂, trans-migrated cells were fixed, stained, and imaged under a 200× light microscope.

Apoptosis assay

Apoptosis was quantified with an apoptosis detection kit (#C1052; Beyotime) according to the manufacturer’s instructions; the fraction of apoptotic cells was analyzed on a FACSCalibur flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA).

In vivo tumorigenesis assay

Four-week-old female BALB/c nude mice (Vital River Laboratories, Beijing, China) were maintained in accordance with U.S. National Institute of Health guidelines and under protocols approved by the ethics committee of the First Affiliated Hospital of The Fourth Military Medical University. After randomization into four groups (n = 5 each)-shCtrl+vehicle, shLMAN2 + vehicle, shCtrl + ADM, and shLMAN2 + ADM—mice received a subcutaneous injection of 6 × 10⁶ MCF-7 cells (200 µL) stably transduced with either lenti-sh-LMAN2 or control shRNA lentivirus (Hanbio, Shanghai, China). When tumors reached ~ 100 mm³, ADM (5 mg kg⁻¹) or vehicle was administered intraperitoneally. Tumor dimensions were recorded weekly, and volume was estimated as 0.5 × length × width²; mice were sacrificed on day 25, and xenografts were excised and weighed. The maximal tumor size permitted by ethics committee and was not exceeded.

RNA sequencing

At 48 h post-transfection with siLMAN2 or non-targeting control siRNA, total RNA was isolated from MCF-7 cells. Libraries were constructed and sequenced by Shanghai Genechem. Gene-set enrichment analysis (GSEA) was then performed on normalized counts to uncover the pathways selectively modulated by LMAN2 depletion.

Lipidomics analysis

Untargeted lipidomics was executed by Shanghai Genechem (Shanghai, China). In brief, 1 × 10⁵ freshly isolated tumor cells were snap-frozen in liquid nitrogen, then lysed at 4 °C in ice-cold methanol (2 × 10⁶ cells ml⁻¹). After 10 min of vortexing and 20 min at − 20 °C, debris was cleared (18 213 g, 10 min, 4 °C), and the supernatant was injected into a Vanquish UHPLC system coupled to a Q Exactive mass spectrometer (Thermo Fisher Scientific). Chromatographic separation was achieved on a Waters ACQUITY HSS T3 column using 15-min (positive) and 17-min (negative) ESI gradients with data-dependent Top15 ddMS² acquisition. Features were identified and quantified with LipidSearch (Thermo Fisher Scientific), and pathway enrichment was computed with the LIPEA software [11].

Statistical analysis

Every assay was run in triplicate. Data are expressed as mean ± SD from a minimum of three independent runs and were processed using GraphPad Prism 8.0. Comparisons between two groups used unpaired Student’s t tests; multiple groups were evaluated by one-way ANOVA. P-values < 0.05 were considered statistically significant.

Results

LMAN2 is high expressed in breast tumors and indicates poor outcomes

GEPIA-based analysis showed markedly elevated LMAN2 transcript abundance in primary breast tumors compared to matched normal breast tissues (Fig. 1A). UALCAN subtype stratified data further demonstrated that this up-regulation spanned all major molecular sub-classes, with the highest median transcript abundance observed in luminal tumors, followed by HER2-positive and triple-negative cases; each subclass displayed a statistically significant increase relative to normal controls (one-way analysis of variance with Dunnett’s post-test, all P < 0.001) (Fig. 1B). Kaplan–Meier survival curves indicated that patients with LMAN2 expression above the median exhibited markedly reduced overall survival (P =5E-04) and shorter progression-free survival (P = 0.018) (Fig. 1C, D). Collectively, these results revealed LMAN2 to be a frequently over-expressed gene in breast cancer whose elevated expression is strongly linked to unfavorable clinical prognosis.

Fig. 1.

High LMAN2 expression is observed in breast tumors and indicates poor outcomes. (A) LMAN2 levels in normal and primary breast cancer tissues were interrogated using the GEPIA database. (B) The LMAN2 mRNA expression in breast cancer subclasses (Normal, Luminal, HER2-positive, and Triple-negative) by UALCAN. (C) Association of LMAN2 transcript levels with overall survival among breast cancer patients. (D) Correlation between LMAN2 gene expression and progression-free survival in breast cancer patients

LMAN2 promotes breast cancer cell proliferation and migration

Then, we investigated how LMAN2 expression modulates proliferation and migration in breast cancer cells using functional studies. Two independent siRNAs (siLMAN2#1 and siLMAN2#2) reduced LMAN2 mRNA by > 80% and virtually abolished the corresponding protein in two BC cells (Fig. 2A–B). This knock-down significantly impaired cell viability at 96 h (Fig. 2C) and reduced clonogenic survival by 43% (MCF-7) and 45% (MDA-MB-231) relative to siCtrl (Fig. 2D). Transwell assays revealed an even more pronounced defect: the number of cells traversing the membrane dropped by 37% and 63%, respectively (Fig. 2E). Conversely, lentiviral over-expression of LMAN2 raised transcript levels by five- to seven-fold and increased protein abundance accordingly (Fig. 2F–G). Ectopic LMAN2 enhanced cell viability by 40% in MCF-7 and 66% in MDA-MB-231 at 96 h (Fig. 2H), augmented colony formation by 1.4- and 1.5-fold (Fig. 2I), and doubled the migratory output in Transwell assays (Fig. 2J). Collectively, these complementary gain- and loss-of-function experiments establish LMAN2 as a potent driver of breast cancer cell proliferation and migration.

Fig. 2.

LMAN2 promotes breast cancer cell proliferation and migration. (A), (B) The mRNA and protein levels of LMAN2 in MCF7 and MDA-MB-231 cells transfected with siCtrl, siLMAN2#1, and siLMAN2#2. (C) Cell viability in MCF7 and MDA-MB-231 cells after LMAN2 knockdown was detected by CCK-8 assay. (D) Cell colony numbers in BC cells after LMAN2 knockdown. (E) Cell-migration ability in BC cells after LMAN2 knockdown. (F), (G) LMAN2 mRNA and protein abundance in BC cells transfected with Ctrl and LMAN2. (H) Cell viability in BC cells following LMAN2 overexpression was detected by CCK-8 assay. (I) Cell colony numbers in BC cells following LMAN2 overexpression. (J) Cell-migration ability in BC cells following LMAN2 overexpression. ***P<0.001

LMAN2 silencing sensitized breast cancer cells to ADM in vitro and in vivo

To determine whether LMAN2 modulates the ADM response, MCF-7 and MDA-MB-231 cells were transfected with LMAN2-targeting siRNA (si-LMAN2) or non-targeting control (siCtrl) and exposed to 50nM of ADM. CCK-8 assays revealed that both LMAN2 knockdown and ADM treatment reduced cell viability in both two BC cells, whereas the combination of siLMAN2 and ADM further decreased the cell viability of breast cancer cells (Fig. 3A). Clonogenic survival was next evaluated. ADM alone suppressed colony formation by ~ 50%; however, combining si-LMAN2 with ADM almost completely abolished macroscopic colonies (Fig. 3B–D). Flow-cytometric apoptosis analyses demonstrated that LMAN2 depletion increased the Annexin V-positive fraction from 18% to 41% in MCF-7 cells and from 15% to 39%, respectively, in MDA-MB-231 cells under ADM treatment (Fig. 3E, F). Transwell migration assays indicated that ADM modestly decreased cell motility, whereas si-LMAN2 plus ADM led to >80% reductions in migrated cells (Fig. 3G, H). Finally, the therapeutic relevance of LMAN2 loss was examined in vivo. Tumors in the shLMAN2 + ADM MCF-7 cohort were visibly smaller (Fig. 3I). Quantification revealed a 72% reduction in mean volume (Fig. 3J) and a 75% reduction in final tumor weight compared to the shCtrl + ADM MCF-7 cohort (Fig. 3K) (P < 0.001). Collectively, these results indicate that LMAN2 knockdown potentiates the efficacy of ADM by augmenting its cytotoxic and cytostatic effects, as observed in both in vitro and in vivo.

Fig. 3.

LMAN2 silencing sensitizes BC cells to doxorubicin in vitro and in vivo. (A) Cell viability of BC cells silencing LMAN2 on ADM (50 nM); (B)–(D) Cell colony numbers of BC cells silencing LMAN2 on ADM (50 nM). (E), (F) Cell apoptosis of BC cells silencing LMAN2 on ADM (50 nM). (G), (H) Cell-migration ability of BC cells silencing LMAN2 on ADM (50 nM). (I) Photographs of subcutaneous xenograft tumors (n = 5). (J), (K) Tumor volume and weight were recorded. *P < 0.05, **P < 0.01, ***P < 0.001

LMAN2 overexpression confers doxorubicin resistance in breast cancer

Then, MCF-7 and MDA-MB-231 cells were stably transduced with LMAN2-expressing lentivirus (LMAN2-OE) or empty vector (Ctrl), exposed to 50 nM of doxorubicin (ADM), and functionally interrogated. Cell viability assays showed that ectopic LMAN2 increased survival in two BC cells relative to the Ctrl + ADM cohort (Fig. 4A). Clonogenic assays corroborated these findings. Whereas ADM reduced colony numbers by ~ 40% in control cells, LMAN2-OE almost completely rescued this inhibition, restoring plating efficiency to 95% (MCF-7) or 78% (MDA-MB-231) of untreated controls (Fig. 4B–D). Apoptosis analyses revealed that LMAN2 overexpression decreased the cell apoptosis in two BC cells under ADM treatment (Fig. 4E, F). Transwell migration assays demonstrated that LMAN2-OE fully reversed ADM-induced motility suppression, restoring invaded cell counts to untreated levels (Fig. 4G, H). Taken together, these results demonstrate that forced expression of LMAN2 is sufficient to attenuate doxorubicin cytotoxicity, reduce apoptosis, and sustain the migratory capacity of breast cancer cells, functionally validating LMAN2 as a driver of chemoresistance.

Fig. 4.

LMAN2 overexpression confers doxorubicin resistance in BC cells. (A) Cell viability of MCF-7 and MDA-MB-231 cells overexpressing LMAN2 on ADM; (B)–(D) Cell colony-formation assay to detect the effect of LMAN2 overexpression on the clone-formation ability of MCF-7 and MDA-MB-231 cells after treatment with ADM. (E), (F) Cell apoptosis of BC cells overexpressing LMAN2 on ADM. (G)–(H) Cell-migration ability of BC cells overexpression LMAN2 on ADM

LMAN2 regulated lipid metabolism in breast cancer cells by modulating SCD expression

Furthermore, RNA sequencing of LMAN2-knockdown (siLMAN2) versus control (siCtrl) MCF-7 cells (n = 3) was performed to dissect how LMAN2 drives BC progression. The heatmap of differentially expressed genes that visualized in Fig. 5A. GSEA of these differentially expressed genes showed a highly significant negative enrichment of the “cholesterol homeostasis” hallmark (Fig. 5B), suggesting that LMAN2 sustains cholesterol biosynthesis. Volcano-plot analysis revealed 91 significantly altered species, 68% of which were down-regulated after LMAN2 knock-down (Fig. 5C). GSEA of the lipidome ranked “cholesterol metabolism” as the top enriched pathway (Fig. 5D). Consistently, total cellular cholesterol was reduced by 24% in si-LMAN2 MCF-7 cells (Fig. 5E). Pathway enrichment of the altered lipids further pinpointed “glycerophosphocholines” as the most impacted modules (Fig. 5F). Quantitative PCR validation confirmed that the top 10 cholesterol-related genes-including SCD, ACSL4, FASN, ACAT2, FADS2, ELOVL5, ACADSB, FADS1, and ACACA were uniformly suppressed upon LMAN2 silencing (Fig. 5G). Among then, stearoyl-CoA desaturase (SCD) was identified as the most down-regulated metabolic enzyme, and western blot verified a pronounced decrease of SCD protein levels in LMAN2 knock down MCF7 cells and increased in LMAN2 overexpression MCF7 cells (Fig. 5H). Collectively, these data position LMAN2 as an upstream regulator of cholesterol homeostasis in breast cancer cells, with SCD representing a key downstream node linking LMAN2 to lipid metabolism and tumor progression.

Fig. 5.

Cholesterol homeostasis is functionally coupled to LMAN2-driven breast cancer progression. (A) Heatmap of differentially expressed genes between control and LMAN2-silenced MCF-7 cells (n = 3). (B) Enrichment plots from GSEA revealed an inverse association between reduced LMAN2 expression and cholesterol homeostasis. (C) Volcano plots of the lipid metabolomics with significant changes in MCF7 cells transfected with siCtrl and siLMAN2. (D) GSEA indicated significant enrichment of DEGs in cholesterol metabolism pathways. (E) Relative total cholesterol in MCF7 cells transfected with siCtrl and siLMAN2. (F) Lipid pathway-enrichment analyses of the lipids identified in (D). (G) The relative mRNA expression of the top 10 decreased genes after LMAN2 knockdown in MCF7 cells. (H) SCD protein abundance in LMAN2 knock down and LMAN2 overexpression MCF7 cells

LMAN2 drives breast cancer progression via SCD-dependent rewiring of lipid metabolism

To determine whether SCD is the functionally relevant downstream target of LMAN2, we performed a series of rescue and loss-of-function experiments. First, cells received either siCtrl, siLMAN2, or siLMAN2 + SCD. Western blot analysis confirmed that siLMAN2 greatly reduced LMAN2 protein levels, and the SCD protein level was enhanced in siLMAN2 + SCD group than siLMAN2 (Fig. 6A). CCK-8 assays showed that SCD restoration rescued the proliferation defect induced by LMAN2 knock-down, significantly restoring cell viability to siCtrl levels (Fig. 6B). Colony-formation assays paralleled these data: siLMAN2 decreased colony numbers by ~ 70%, whereas concomitant SCD overexpression restored clonogenicity to baseline (Fig. 6C). Flow-cytometric apoptosis analyses revealed that SCD overexpression reduced the elevated Annexin V-positive fractions in two BC cells (Fig. 6D). Transwell migration assays demonstrated that SCD overexpression fully reversed the impaired motility caused by LMAN2 silencing (Fig. 6E). Second, stably overexpressing LMAN2 (LMAN2-OE) in breast cancer cell lines triggered an increase in SCD protein. This induction was abolished when SCD was subsequently silenced with siSCD (Fig. 6F). Functionally, cell proliferation by LMAN2-overexpression was completely neutralized by SCD knock-down, as demonstrated by both CCK-8 (Fig. 6G) and cell colony formation assays (Fig. 6H). Similarly, the anti-apoptotic effect of LMAN2-OE was nullified when SCD was concomitantly inhibited (Fig. 6I). Finally, the enhanced migratory capacity conferred by LMAN2 overexpression was reduced upon SCD silencing (Fig. 6J). Collectively, these reciprocal rescue experiments establish SCD as a requisite downstream effector through which LMAN2 promotes proliferation, suppresses apoptosis, and augments migration in breast cancer cells.

Fig. 6.

LMAN2 promotes breast cancer cell development through targeting SCD. (A) LMAN2 and SCD protein abundance was assessed in BC cells treatment with siCtrl, siLMAN2, and siLMAN2 + SCD. (B), (C) Proliferation ability was examined by CCK-8 and colony-formation assays in BC cells treatment with siCtrl, siLMAN2, and siLMAN2 + SCD. (D) Flow cytometry showing the apoptosis rate in BC cells treatment with siCtrl, siLMAN2, and siLMAN2 + SCD. (E) Transwell assays in BC cells treatment with siCtrl, siLMAN2, and siLMAN2 + SCD. (F) Western blot analysis was used to evaluate LMAN2 and SCD expression in BC cells treated with Ctrl, LMAN2-OE, and LMAN2-OE+siSCD. (G), (H) Proliferation ability was examined by CCK-8 and colony formation assays in BC cells treated with Ctrl, LMAN2-OE, and LMAN2-OE+siSCD. (I) Flow cytometry showing the apoptosis rate in BC cells treated with Ctrl, LMAN2-OE, and LMAN2-OE+siSCD. (J) Transwell assays were performed in BC cells treated with Ctrl, LMAN2-OE, and LMAN2-OE+siSCD

LMAN2 promotes chemoresistance in breast cancer though SCD

Finally, we quantitatively defined the LMAN2/SCD axis as a critical modulator of anthracycline sensitivity in breast cancer. In MCF-7 cells, LMAN2 knockdown lowered the ADM IC₅₀ from 1.5 µM to 0.34 µM; this sensitization was completely rescued by simultaneous lentiviral over-expression of SCD (IC₅₀ restored to 1.29 µM) (Fig. 7A). An identical epistatic pattern was observed in MDA-MB-231 triple-negative cells: LMAN2 silencing reduced the IC₅₀ from 2.39 µM to 0.89 µM, and co-expression of SCD reverted the value to 2.09 µM (Fig. 7B). Conversely, enforced LMAN2 over-expression increased the IC₅₀ of MCF-7 cells to 3.01 µM, an effect that was abolished when SCD was suppressed by siSCD (IC₅₀ returned to 0.72 µM; Fig. 7C). Similarly, in MDA-MB-231 cells, LMAN2 over-expression raised the IC₅₀ to 9.86 µM, whereas dual SCD knock-down reduced it to 2.76 µM (Fig. 7D). Collectively, these gain- and loss-of-function experiments establish that LMAN2 confers ADM chemoresistance through, and strictly depends on, functional SCD.

Fig. 7.

LMAN2 promotes chemoresistance in BC through SCD. (A) ADM IC₅₀ values in MCF-7 cells subjected to LMAN2 knockdown and /or SCD overexpressing on; (B) ADM IC₅₀ values in MDA-MB-231 cells subjected to LMAN2 knockdown and /or SCD overexpressing; (C) IC₅₀ values of MCF-7 cells overexpressing LMAN2 and /or silencing SCD on ADM; and (D) IC₅₀ values of MDA-MB-231 cells overexpressing LMAN2 and /or silencing SCD on ADM

Discussion

Our data establish the type-I membrane lectin LMAN2 as a frequently over-expressed gene that heralds poor outcomes across the major breast cancer subtypes. This is consistent with previous reports in breast cancer [8–10]. LMAN2 has since been advanced as a potential early-detection biomarker for oral squamous cell carcinoma [12]. Importantly, gain- and loss-of-function experiments demonstrate that LMAN2 is both necessary and sufficient for sustained proliferation, clonogenicity, and motility of luminal and triple-negative cells. We therefore propose that LMAN2 could be a prognostic and oncogenic driver in breast cancer.

Altered lipid homeostasis is linked to both higher breast cancer incidence rates and a worse clinical prognosis [13]. Cholesterol is essential for maintaining cellular homeostasis in tumor cells, including those of breast cancer [14]. There is abundant experimental evidence showing that disturbed cholesterol balance fuels cancer biology; indeed, cholesterol is an indispensable membrane component, and tumor cells must ramp up its synthesis or uptake to sustain relentless proliferation [15]. The core mechanism we identify is an LMAN2-orchestrated transcriptional circuit that preserves cholesterol balance through obligatory induction of SCD. SCD1 inserts the first double bond into saturated fatty acids, generating Δ9-monounsaturates [16], and is itself a canonical transcriptional target of sterol-regulatory element–binding proteins (SREBPs) [17]. We show that LMAN2 silencing suppresses the entire SREBP-regulated cholesterol biosynthetic module (FADS2, FASN, ELOVL5, ACACA, SCD, etc.) and lowers total cellular cholesterol by ~ 25%. LMAN2 is a lectin involved in glycoprotein trafficking, emerging evidence suggests proteins in the ER-Golgi intermediate compartment can influence ER stress signaling. We hypothesize that LMAN2 loss-of-function disrupts its role in protein maturation/secretion, causing an accumulation of client proteins or altering ER membrane composition. This disturbance triggers a low-grade or sustained Endoplasmic Reticulum (ER) Stress, leading to the activation of the Unfolded Protein Response (UPR). Key UPR branches, notably IRE1α-XBP1 and ATF6, result in the nuclear translocation of the transcription factors XBP1s (spliced form) and ATF6, which globally rewire metabolic gene expression. Crucially, both XBP1s and ATF6 have been shown to bind directly to the promoters of key lipid metabolism genes, including SCD, and repress their transcription). Therefore, we posit that LMAN2 knockdown → ER Stress/UPR activation → Nuclear translocation/activity of XBP1s/ATF6 → Transcriptional repression of the SCD gene.

Among the module components, fatty acid desaturases 1 and 2, the enzymes that craft polyunsaturated fatty acids, are markedly up-regulated in a TNBC subgroup that experiences the worst outcomes [18]. Dual blockade of SCD1 and FADS2 may offer a novel, metabolism-based therapeutic angle against cancer [19]. Fatty acid synthase, the gatekeeper of de novo lipogenesis, is markedly up-regulated in breast tumors [20] and represents a druggable metabolic vulnerability [21]. Fatty acid elongase 5 (ELOVL5) is transcriptionally activated by SREBP-1 and operates as a key negative-feedback regulator of de novo lipogenesis [22]. Research shows that ELOVL5 is markedly over-expressed in breast tumor samples relative to paired adjacent non-cancerous tissue [23]. Reducing ELOVL5 drives breast cancer metastasis by boosting lipid-droplet stores, which in turn amplifies TGF-β receptor signaling [24]. Acetyl-CoA carboxylase, a SREBP-1 downstream target, curbs antitumor immunity [25]; its suppression helps to establish an immunosuppressive lung pre-metastatic niche in breast cancer [26]. Acetyl-CoA acetyltransferase 2 maintains the equilibrium between cholesterol and triglyceride metabolism [27]. Finally, the long-chain fatty acyl-CoA synthetase (ACSLs) activate fatty acids by ligation to CoA and thus sit at a central hub of lipid metabolism. Among the five existing isoforms, ACSL3 and ACSL4 are recurrently dysregulated in cancer, fueling tumor initiation, progression, metastasis, and immune-evasion, and are emerging as actionable therapeutic targets [28].

We can infer that SCD-derived MUFAs are indispensable for the oncogenic and chemo-resistant phenotypes imposed by LMAN2. Pharmacological or genetic inhibition of SCD phenocopies LMAN2 loss, whereas enforced SCD expression fully rescues proliferation, migration, and ADM resistance in LMAN2-depleted cells. These findings align with emerging evidence that TNBC cells generate MUFAs via SCD and secrete lipid-bound MUFAs that protect them from pro-ferroptotic lipid peroxidation [29]. Blocking SCD not only curbs intrinsic tumor expansion but also re-engineers the brain immune milieu, offering a two-pronged approach against breast cancer brain metastases [30]. Our epistatic IC₅₀ analyses extend these observations by demonstrating that SCD re-activation is sufficient to revert the four- to five-fold ADM sensitization achieved by LMAN2 knock-down. Thus, the LMAN2/SCD axis constitutes a biochemical bottleneck whose disruption lowers the threshold for anthracycline cytotoxicity.

In this study, we showed that LMAN2 mRNA or protein abundance could be exploited as a companion diagnostic to stratify patients most likely to benefit from SCD-targeted therapy or dose-dense anthracycline regimens. LMAN2 is a surface-exposed lectin with a large luminal domain; it is amenable to antibody–drug conjugation or ligand-directed CAR-T-cell approaches. LMAN2 silencing sensitizes tumors without overt toxicity in nude mice, positioning the gene as an attractive synthetic lethal target when combined with standard-of-care anthracyclines.

The interplay between SCD activity, MUFA production, and cholesterol metabolism constitutes a critical axis in the development of cancer chemoresistance. SCD, by converting saturated fatty acids (SFAs) to monounsaturated fatty acids (MUFAs), fundamentally reshapes the cellular lipid landscape [31]. This altered MUFA/SFA ratio directly impacts membrane phospholipid composition, affecting membrane fluidity, permeability, and the stability of cholesterol-rich lipid rafts [32]. These biophysical changes can impair the influx or enhance the efflux of chemotherapeutic agents, reducing their intracellular accumulation. Furthermore, MUFAs are essential substrates for the esterification of cholesterol into cholesteryl esters, a process that regulates free cholesterol levels and facilitates its storage in lipid droplets. Dysregulation of this esterification can lead to toxic free cholesterol accumulation or, conversely, deplete cholesterol pools necessary for maintaining membrane integrity and signaling platforms. Cholesterol itself is a master regulator of lipid raft-mediated pro-survival signaling pathways, such as PI3K/AKT and EGFR, which are frequently hyperactivated in resistant cells. By modulating the composition and function of these signaling hubs, SCD-driven lipid remodeling can sustain proliferative and anti-apoptotic signals. Additionally, SCD inhibition has been linked to the induction of endoplasmic reticulum (ER) stress and the unfolded protein response (UPR), as well as sensitization to ferroptosis—a form of cell death driven by lipid peroxidation [33]. Cancer stem cells (CSCs), which are often drug-resistant, appear particularly dependent on SCD activity to maintain their stemness and viability, potentially through the regulation of pathways like β-catenin and YAP/TAZ [31]. Therefore, the SCD-MUFA-cholesterol axis promotes resistance through a convergent mechanism involving altered drug pharmacokinetics, enhanced survival signaling, and evasion of stress-induced cell death. Targeting this metabolic vulnerability, for instance by combining SCD inhibitors with standard chemotherapy, represents a promising strategy to overcome resistance.

However, there are still several limitations to overcome, as follows. (i) Our clinical analyses were retrospective and confined to transcript-level data; validation in prospective cohorts with quantitative IHC is essential. (ii) The mouse experiments employed subcutaneous xenografts, which lack the metastatic complexity and immune micro-environment of autochthonous mammary tumors. Orthotopic or MMTV-PyMT transgenic studies will provide a more stringent test of therapeutic index. (iii) Although SCD is the dominant metabolic effector identified, unbiased CRISPR-Cas9 suppressor screens may reveal additional nodes (e.g., PI3K/AKT, YAP/TAZ) that buffer LMAN2 loss. (iv) Finally, the possibility that LMAN2 participates in antigen cross-presentation or T-cell priming—processes critically dependent on Golgi trafficking—remains unexplored and could influence the design of combination regimens that include immune-checkpoint blockade. (v) Our data establish a strong correlation between SCD inhibition, cholesterol depletion, and chemosensitization, but do not yet provide direct mechanistic proof that the change in cholesterol is the primary driver of the functional phenotypes. The proposed rescue experiments are essential to move from correlation to causation.

In summary, we delineate a hitherto unrecognized Golgi–lipid nexus in which LMAN2 orchestrates SCD-dependent MUFA production to fuel breast cancer progression and anthracycline resistance. Targeting this axis offers a rational strategy to enhance the efficacy of cytotoxic chemotherapy while minimizing collateral toxicity.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ADM

Doxorubicin

LMAN2

Lectin, mannose-binding 2

SCD

stearoyl-CoA desaturase

PCR

polymerase chain reaction

Author contributions

YCJ conceived the study, designed and executed all cellular and animal experiments, prepared the figures and drafted the manuscript. CFQ performed bioinformatic and survival analyses. HML carried out qPCR, Western blotting, apoptosis and cholesterol assays. LCT conducted the in-vivo drug treatments and tumor measurements. LR coordinated the study. WT supervised the project and critically revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the CSCO Clinical Oncology Research Foundation (grant No. Y-HH202101-0092).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

No human subjects were involved; all animal experiments were approved by the Animal Ethics Committee of First Affiliated Hospital of The Fourth Military Medical University.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.