Mediator complex subunit 18: A novel breast cancer stem cell marker in tumor microenvironment

Abstract

Breast cancer remains the most frequently diagnosed malignancy among women, and treatment resistance and recurrence remain major clinical challenges, in part because of breast cancer stem cells (BCSCs). Established BCSC-associated phenotypes include CD44+/CD24-/low expression and high aldehyde dehydrogenase activity, although no single marker is universally specific. Guo et al recently reported that mediator complex subunit 18 (MED18) expression was associated with a prognostic gene signature, spatially colocalized with high-risk tumor regions, and correlated with inferred tumor, macrophage, fibroblast, and endothelial-cell activity. These bioinformatics, multi-omics, and spatial findings nominate MED18 as a candidate associated with BCSC-like and tumor-microenvironment programs; they do not establish biomarker accuracy, causal regulation of stemness, or therapeutic actionability. Functional perturbation studies and independent clinical validation are required.

Key Words: Breast cancer; Breast cancer stem cell; Mediator complex subunit 18; Tumor microenvironment; Multi-omics analysis

Core Tip: Breast cancer exhibits a tendency toward drug resistance and recurrence due to the presence of breast cancer stem cells (BCSCs). Therefore, the screening and validation of breast cancer biomarkers are crucial for diagnosis, treatment, and prognosis. Mediator complex subunit 18 (MED18) has emerged as a novel BCSC biomarker, positioned at the intersection of BCSC regulation and tumor microenvironment interactions. Although its identification demonstrates promising diagnostic and therapeutic potential, further mechanism validation through in vitro and in vivo studies is required to translate MED18 from a biomarker into a clinically actionable target.

This editorial refers to “Breast cancer stem cell activity driven by ME18D gene expression in the tumor microenvironment” by Guo et al, 2026; https://doi.org/10.4252/wjsc.v18.i1.111348.

---

INTRODUCTION

Breast cancer is one of the most common cancers in women and its incidence even briefly surpassed lung cancer in 2022 to become the most common cancer in the world[1]. However, the mortality rate of breast cancer has been reduced to some extent in recent years due to advances in prevention, diagnosis, and treatment technologies. The presence of breast cancer stem cells (BCSCs) remains a key factor contributing to poor treatment outcomes. They significantly promote tumor metastasis, recurrence, and treatment resistance. Similar to normal stem cells, BCSCs possess high plasticity, self-renew capacity, and multi-potent differentiation potential. However, unlike their normal stem cells, the regulatory mechanisms in BCSC are out of control, leading to sustained activation of proliferative signaling pathways that drive malignant tumor progression[2]. Therefore, identifying and eliminating BCSCs holds significant clinical importance for precision treatment of breast cancer and overcoming drug resistance.

This editorial reviews advances in BCSC-associated markers and evaluates the emerging evidence for mediator complex subunit 18 (MED18), with particular attention to the analytical strategy, prognostic observations, spatial evidence, and tumor-microenvironment associations reported by Guo et al[3]. It highlights the study’s contribution to candidate prioritization while distinguishing association with BCSC-like programs from biomarker performance, functional regulation of stemness, and therapeutic target validity. This balanced framework is intended to define a constructive roadmap for experimental and clinical development.

IDENTIFICATION AND EVALUATION OF BCSC-ASSOCIATED MARKERS

A molecule should not be classified as a BCSC marker solely because its expression correlates with prognosis or stemness-related signaling. A useful BCSC marker should ideally satisfy several complementary criteria: Enrichment in prospectively isolated tumor-initiating cells; reproducibility across cell lines, patient-derived samples, and in vivo models; a functional relationship with self-renewal or tumor initiation; prognostic or predictive relevance beyond established clinicopathological factors; and reliable measurement in clinical specimens. These criteria distinguish markers used to isolate or enrich BCSCs from molecules that influence stemness without serving as validated markers. Because BCSCs are heterogeneous and plastic, no single marker is expected to identify all BCSC states across every breast cancer subtype.

Among the better-supported BCSC-associated phenotypes, CD44+/CD24-/Low expression has been used to prospectively enrich tumorigenic breast cancer cells[4], and CD44 expression has also been associated with stem-like features in triple-negative breast cancer[5]. CD29 and CD49f have enriched cancer-initiating populations in BRCA1-mutant mouse mammary tumors[6], whereas sorted CD133+ breast cancer cells have shown tumorigenicity and resistance to DNA-damaging agents in selected models[7]. High aldehyde dehydrogenase activity is likewise associated with self-renewal, tumor initiation, drug resistance, and adverse prognosis[8]. Overall, combinations of phenotypic markers and functional assays are more informative than any single marker.

The positions of p21 and human epidermal growth factor receptor 2 (HER2) within this framework require particular caution. p21 (CDKN1A) is primarily an intracellular regulator of cell-cycle progression, senescence, and stress responses, and its expression has been associated with breast cancer prognosis[9-11]. Experimental alteration of p21 can influence epithelial-mesenchymal transition and stemness-related phenotypes[12]; however, this functional association does not establish p21 as a marker equivalent to CD44+/CD24-/Low expression or aldehyde dehydrogenase activity for BCSC isolation. It should therefore be described as a context-dependent regulator of stem-like properties rather than a validated BCSC marker. HER2 likewise has distinct roles that should not be conflated. ERBB2 amplification or HER2 overexpression defines a breast cancer subtype and provides an established therapeutic target. Although HER2 activity may be associated with tumor-initiating-cell enrichment in selected settings[13], HER2 should not be presented as a universal BCSC marker, nor should the clinical efficacy of HER2-directed therapy be assumed to result specifically from BCSC elimination.

BCSC marker performance and tumor-microenvironment interactions are also likely to differ among breast cancer subtypes. In hormone receptor-positive disease, endocrine signaling and treatment exposure may shape persistent stem-like populations. In HER2-positive disease, HER2 should be distinguished as a subtype-defining receptor from any context-dependent association with stem-cell-enriched phenotypes[13]. Triple-negative tumors may show greater enrichment of CD44+/CD24-/Low or aldehyde dehydrogenase-high populations, but substantial intertumoral and intratumoral heterogeneity remains[5,8]. Across these subtypes, residual or quiescent BCSCs may contribute to recurrence and treatment resistance[14]. Candidate markers and their tumor-microenvironment associations should therefore be evaluated within subtype-stratified cohorts rather than generalized across all breast cancers.

MED18: AN EMERGING CANDIDATE IN BCSC-LIKE PROGRAMS AND THE TUMOR MICROENVIRONMENT

BCSCs can occupy dynamic, interconvertible states that are commonly enriched using CD44+/CD24-/Low and ALDH1+ phenotypes[15]. Their heterogeneity and plasticity complicate both identification and therapeutic targeting. Transcription factors such as sex-determining region Y-box 2[16] and octamer-binding protein 4[17], together with Wnt[18] and Notch[19] signaling, contribute to stemness-related programs; however, these established pathway relationships should not be interpreted as evidence that MED18 directly regulates them in BCSCs.

World Journal of Stem Cells published an article titled “Breast cancer stem cell activity driven by ME18D gene expression in the tumor microenvironment” by Guo et al[3]. Although the published title uses “ME18D”, the gene examined in the main text and analyses is MED18. The study integrated public breast cancer transcriptomic data with differential-expression analysis, a least absolute shrinkage and selection operator-derived 10-gene prognostic signature, multi-omics analyses, and spatial transcriptomic assessment. The authors reported higher MED18 expression in malignant breast tissues than in normal controls, inclusion of MED18 in the prognostic signature, heterogeneous intratumoral expression, spatial overlap between MED18-high and high-risk-score regions, and positive correlations with inferred activities of tumor cells, macrophages, fibroblasts, and endothelial cells[3]. MED18 is a component of the conserved mediator complex, which links transcription factors with RNA polymerase II. MED18 has also been implicated in abnormal TP53 splicing in multiple myeloma[20], but that context-specific observation does not validate a role in BCSCs.

The integration of prognostic modeling, multi-omics analysis, and spatial transcriptomics is an important strength of the study by Guo et al[3], and the concordant findings provide a strong rationale for prioritizing MED18 for further investigation. The reported associations suggest that MED18 may participate in broader BCSC-related and tumor-microenvironment programs, although correlation and spatial overlap alone cannot determine whether MED18 directly coordinates macrophage, fibroblast, endothelial-cell, immune, angiogenic, or matrix-remodeling functions. A proposed model is that MED18 may participate in mediator-dependent transcriptional coordination of Wnt-, Notch-, and Hedgehog-related stemness programs.

The findings can be placed within four complementary levels of evidence. First, differential expression, participation in a prognostic risk signature, spatial colocalization, and correlations with inferred cellular activities provide meaningful association-level support for MED18 in BCSC-like programs. Second, independent evaluation of specificity, sensitivity, reproducibility, standardized measurement, and incremental prognostic value will clarify its biomarker performance. Third, perturbation and BCSC phenotypic assays can determine whether MED18 functionally contributes to stemness. Fourth, studies of selective dependency, therapeutic modulation, efficacy, and safety can establish whether MED18 has target potential. Within this framework, MED18 is best viewed as a promising, hypothesis-generating candidate that warrants systematic validation.

A practical validation roadmap can build directly on the study’s findings. MED18 knockdown and overexpression can be combined with mammosphere formation and serial-passage assays, ALDH activity measurements, assessment of CD44+/CD24-/Low or other relevant marker shifts, drug-resistance assays, and rescue experiments. In vivo limiting-dilution tumor-initiation assays and models of metastasis and recurrence would test biological relevance. ChIP-seq or CUT&Tag, co-immunoprecipitation, and promoter-activity assays could further evaluate MED18 occupancy, protein interactions, RNA polymerase II recruitment, and the proposed relationships with Wnt/β-catenin, Notch, or Hedgehog transcriptional programs.

When MED18 is assessed against the same marker framework, it meets important early-stage criteria through its association with a risk signature, BCSC-like programs, and spatially defined tumor-microenvironment features[3]. These observations provide preliminary prognostic and clinical-interest evidence. The next priorities are to examine enrichment in prospectively isolated tumor-initiating cells, reproducibility across independent models and breast cancer subtypes, functional relationships with self-renewal, and clinically applicable assays and cutoffs. MED18 can therefore be positioned as an emerging BCSC-associated candidate with a clearly defined path toward validation.

TRANSLATIONAL PROSPECTS OF BCSC BIOMAKERS AND THERAPEUTIC TARGETS

The therapeutic landscape for BCSC-associated markers provides useful context for considering the future development of MED18. CD44- or CD133-directed delivery systems and ALDH-related approaches have generated encouraging preclinical evidence[8,21-29]. Studies of CD105+/CD90+ and CD70+ populations describe candidate marker-associated phenotypes but do not evaluate corresponding therapeutic agents, so these entries were not retained in the therapeutic table[30,31]. Additional clinical and pathway-directed strategies range from exploratory trials to approved breast cancer therapies used for broader indications[32-48] (Table 1). To facilitate balanced comparison, Table 1 distinguishes the molecular feature, BCSC-related property, agent, mechanism, development setting, and current evidence status. Balixafortide-eribulin, reparixin-paclitaxel, and pembrolizumab-eribulin are correctly classified under C-X-C chemokine receptor 4 (CXCR4), CXCR1/2, and programmed cell death protein 1-related biology, whereas the clinical activity of HER2-directed agents is discussed separately from evidence for BCSC-specific targeting.

Table 1 Current evidence status for therapeutic strategies associated with breast cancer stem cell markers or pathways.

	BCSC-related property	Representative agent	Mechanism/modality	Setting/stage	Current evidence status/Limitation	Ref.
ALDH1A1	High enzymatic activity enriches self-renewing, tumor-initiating, and drug-resistant populations	Quercetin	Downregulation of ALDH1A1 and other stem-cell-associated proteins	Preclinical breast cancer models	No validated clinical BCSC-selective efficacy or standardized treatment threshold	[8,22]
CD44	CD44+/CD24-/Low phenotype and hyaluronan binding can enrich BCSC-like populations	HA-based nanocarriers; P245	CD44/HA-mediated delivery or anti-CD44 antibody targeting	Preclinical nanoparticle and xenograft studies	CD44 is heterogeneous and is also expressed by non-BCSC and normal cell populations	[21,23,26,27]
CD44/DLL4-Notch4	Associated with treatment-resistant and endocrine-resistant BCSC programs	FKBPL; AD-01	Modulation of the CD44 pathway and downregulation of DLL4/Notch4	Preclinical	Not a clinically validated CD44-directed therapy; pathway effects may not be BCSC-specific	[24,25]
CD133	Enriches tumorigenic and drug-resistant cells in selected breast cancer models	CD133-paclitaxel nanoparticles; CD133 mRNA dendritic-cell vaccine	Targeted drug delivery or immunization	Preclinical	Model-dependent expression and no validated clinical companion assay	[28,29]
CXCR4	Chemokine receptor associated with migration, homing, and stem-like behavior	Balixafortide plus eribulin	CXCR4 antagonism combined with microtubule inhibition	Phase I, HER2-negative metastatic breast cancer	Clinical benefit was not shown to result specifically from BCSC elimination	[32]
CXCR1/2	Inflammatory chemokine signaling has been linked to BCSC survival	Reparixin plus paclitaxel	CXCR1/2 inhibition combined with chemotherapy	Phase Ib, HER2-negative metastatic breast cancer	Exploratory clinical evidence without a validated BCSC-specific endpoint	[33]
PD-1/immune checkpoint	Immune modulation; not a BCSC marker	Pembrolizumab plus eribulin	PD-1 blockade combined with chemotherapy	Phase II, HR-positive/HER2-negative metastatic breast cancer	Not a BCSC-directed regimen and should not be classified as HER2-targeted therapy	[34]
HER2	Subtype-defining receptor; association with stem-cell-enriched phenotypes is context-dependent	T-DM1; T-DXd; A166; lapatinib	HER2-directed antibody-drug conjugates or tyrosine kinase inhibition	Approved therapies or phase I development	Clinical efficacy has not been demonstrated to be mediated specifically through BCSC targeting	[35-38]
HER2/Akt program	Stem-like properties in selected HER2-positive models	Disulfiram	Suppression of HER2/Akt signaling and stem-like traits	Preclinical	Not validated clinically as a HER2-directed BCSC treatment	[39]
CD49f/integrin α6	Associated with adhesion, tumor initiation, and radioresistance	LY294002	Indirect PI3K pathway inhibition downstream of integrin α6/Akt/ERK	Preclinical	Indirect pathway inhibition with limited marker and cell-type specificity	[40]
GD2	Identifies a tumorigenic BCSC-like subset in selected breast cancers	Dinutuximab	Anti-GD2 monoclonal antibody	Preclinical, including triple-negative breast cancer models	Antigen heterogeneity and no breast cancer approval	[41-43]
OAcGD2	BCSC-associated ganglioside in experimental models	mAb8B6	Anti-O-acetyl-GD2 monoclonal antibody	Preclinical	Early experimental evidence without clinical validation	[44]
ABCG2	Drug-efflux phenotype associated with treatment resistance but not specific to BCSCs	Ko143	ABCG2 transporter inhibition	In vitro and preclinical	Broad transporter expression raises selectivity and toxicity concerns	[45,46]
Notch pathway	Contributes to self-renewal and treatment resistance	MK-0752	Gamma-secretase/Notch pathway inhibition	Clinical investigation and reviewed studies	On-target pathway toxicity and lack of BCSC specificity	[47]
Hedgehog pathway	Stemness-associated developmental signaling	Vismodegib	Smoothened inhibition	Approved in another indication; investigated in breast cancer contexts	Not MED18-directed and no established BCSC-selective clinical benefit	[48]
MED18	Associated with a prognostic risk signature, BCSC-like programs, and tumor-microenvironment features	None identified	No validated therapeutic modality	Hypothesis-generating bioinformatics, multi-omics, and spatial analyses	Early-stage evidence; functional, biomarker-performance, druggability, efficacy, and safety validation remain priorities	[3]

BCSC: Breast cancer stem cell; ALDH1A1: Aldehyde dehydrogenase 1 family member A1; HA: Hyaluronic acid; ADC: Antibody-drug conjugate; CXCR4: C-X-C chemokine receptor 4; HER2: Human epidermal growth factor receptor 2; PD-1: Programmed cell death protein 1; HR: Hormone receptor; Akt: Protein kinase B; PI3K: Phosphoinositide 3-kinases; ERK: Extracellular signal-regulated kinase; OAcGD2: O-acetylated disialoganglioside GD2; MED18: Mediator complex subunit 18.

The intracellular position of MED18 within the mediator complex creates both scientific opportunities and practical development challenges. Early feasibility studies will need to address druggability, target specificity, the consequences of broad transcriptional interference, effects on normal stem and somatic cells, subtype heterogeneity, and the therapeutic window. Successful clinical translation would also require patient stratification, reproducible assay thresholds, and a validated companion diagnostic. Rather than diminishing the value of MED18 as a research candidate, these considerations define the evidence and development milestones needed to evaluate MED18-directed strategies responsibly.

ROLE OF MED18 IN OTHER MALIGNANT TUMORS

Findings in other malignancies further argue against assigning MED18 a universal oncogenic or tumor-suppressive role without context-specific evidence. The available studies differ in design and do not establish that MED18 has the same function across breast cancer, gastric cancer, uterine leiomyoma, bladder cancer, and multiple myeloma. They should therefore be interpreted as evidence of biological context dependence rather than as validation of the proposed BCSC mechanism.

In gastric cancer, Xuan and Wang[49] reported that the SNHG3/enhancer of zeste homolog 2 axis promoted methylation-associated suppression of MED18 and that reduced MED18 was associated with proliferation, migration, invasion, and metastasis. These findings support a potential tumor-suppressive, rather than oncogenic, role for MED18 in that setting. Consistently, Zhou et al[50] found that higher MED18 expression was associated with improved overall survival and lower prognostic risk in gastric adenocarcinoma. MED12 mutation-negative uterine leiomyomas may harbor deletions affecting mediator-complex genes, including MED18, but this observation does not by itself establish MED18 as a biomarker or therapeutic target[51]. In bladder cancer, MED18 was reported as a downregulated target of the oncogenic long non-coding RNA UCA1α, again suggesting a possible growth-inhibitory role[52]. Collectively, these studies show heterogeneous and largely indirect evidence; they do not resolve the function of MED18 in BCSCs and reinforce the need for tissue- and subtype-specific functional validation.

CONCLUSION

The study by Guo et al[3] provides a valuable bioinformatics, multi-omics, prognostic, and spatial foundation for prioritizing MED18 in BCSC research. Its principal contribution is the identification of a promising candidate that may connect BCSC-like programs with tumor-microenvironment features and that can now be examined through targeted functional studies. This interpretation preserves the exploratory importance of the work while aligning the strength of each conclusion with the available evidence.

Future studies can extend this foundation by combining the functional and mechanistic experiments described above with independent, adequately powered cohorts that assess analytical reproducibility, specificity, sensitivity, prognostic value beyond established clinicopathological factors, and performance across hormone receptor-positive, HER2-positive, and triple-negative breast cancers. Therapeutic evaluation should address druggability, target selectivity, efficacy, safety in normal stem and somatic cells, patient stratification, and companion diagnostic validity. The reported spatial clustering of MED18-high regions is particularly valuable for defining testable spatial hypotheses. If the required functional and clinical criteria are met, MED18 may ultimately contribute to biomarker stratification and the development of new BCSC-oriented therapeutic strategies.