Lysine demethylases 6 A and 6B as epigenetic regulators in therapeutic resistance of cancer

Abstract

Histone 3 lysine 27 (H3K27) demethylation is a key post-translational modification of chromatin and plays a fundamental role in gene activation. Demethylation of H3K27 is mediated by Jumonji C domain-containing lysine demethylase 6 A (KDM6A) and its paralog, KDM6B, both of which are responsible for homeostasis, autoimmune response, infectious diseases, and cancers. To date, mounting studies dedicate the roles of KDM6A/B on tumor promotion or suppression, and there are many reviews systematically summarize the relevant mechanisms of KDM6A/B in tumor development and therapy. KDM6A and KDM6B also contribute to the regulation of therapeutic insensitivity to chemotherapy, targeted response, radiotherapy and immunotherapy. Herein, we outline the current knowledge of KDM6A/B in regulation of therapeutic resistance, and suggest that KDM6A/B holds immense potential in recovering therapeutic resistance.

1. Introduction

Epigenetics involves heritable changes in gene expression that rely on structural remodeling of chromatin and do not result from alterations in the DNA sequence [1]. Chromatin remodeling plays a pivotal role on determination of cell phenotype in homeostasis and pathological behaviors [2]. Chromatin is made up of nucleosomes, the basic structural units of chromatin. Each nucleosome consists of an octamer of histone proteins (made up of two copies each of H2A, H2B, H3, and H4) and about 146 base pairs of DNA wrapped around it in nearly two turns [3, 4]. Post-translational modification on N -terminal tails of histones is a vital way to impact on chromatin to be compacted, transcriptional silent state called heterochromatin or loose, transcriptional active state called euchromatin [5]. Theses modifications can be influenced by the interactions between environment cues and genome [6].

Currently, over 20 kinds of post-translational modifications of histones have been found that include phosphorylation, acetylation, SUMOylation, ubiquitination, crotonylation, citrullination, ADP-ribosylation, deamidation, formylation, O-GlcNAcylation, propionylation, crotonylation, proline isomerization, butyrylation, and methylation [7]. Methylation is one of the most common histone modifications, to date, histone 3 lysine 4 (H3K4), H3K27, H3K36, H3K79 and H4K20 are most extensively studied sites of methylation modification [8]. Dimethylated (me2) or trimethylated (me3) of H3K27 at gene promotor and transcriptional start sites usually refer to the condensed chromatin and repression of gene expression, and this process is reversible by lysine specific demethylase 6 A/B (KDM6A/B) [9, 10].

KDM6A, also called the ubiquitously transcribed tetratricopeptide repeat protein on X chromosome (UTX), and KDM6B also named the Jumonji domain containing protein 3 (JMJD3), belong to the family of JmjC domain-containing demethylases [11, 12]. KDM6A/B are not only the vital regulators in physiological environment, but also mutated, deleted, translocated or overexpressed in many cancers [13, 14]. Mounting studies propose that KDM6A/B are considered as oncogenes or suppressors regulating tumorigenesis in various types of cancers [15, 16]. In recent decades, KDM6A/B are reported to universally engage in regulation of therapeutic resistance in cancers, Which are key erasers of histone methylation, centrally govern therapeutic resistance in cancer by dynamically modulating H3K27me3/me2 levels [17]. Their epigenetic influence extends to critical processes such as drug transporter expression, cancer stemness, DNA damage repair, and immune microenvironment remodeling, thereby shaping tumor responsiveness to diverse treatment regimens—including chemotherapy, immunotherapy, targeted agents, a nd radiotherapy [13, 18–21]. Given their context-dependent roles, precise targeting of KDM6A/B holds promise for overcoming resistance, though therapeutic strategies must be carefully calibrated to tumor-specific molecular landscapes to maximize efficacy and minimize toxicity.

This review presents the structures and catalytic mechanisms of KDM6A/B, summarizes the functions of KDM6A/B in development and cancers, and fucuses on the sum up of resistance mechanisms of different treatments. Our work points out the potential opportunities of KDM6A/B for anti-resistance therapy and may provide insight for the innovation of therapeutic resistance strategies.

2. Structures and catalytic mechanisms of KDM6A/B

2.1 structures of KDM6A/B

Lysine-specific demethylases (KDMs) are divided into two main families based on their catalytic cofactors: the flavin adenine dinucleotide (FAD)-dependent LSD family, and the Fe(II)/α-ketoglutarate (also known as 2-oxoglutarate, 2OG)-dependent Jumonji C (JmjC) domain-containing family [22, 23]. Currently, only three members have been identified in the LSD family: LSD1, LSD1 + 8a, and LSD2. In contrast, more than 20 members of the JmjC family are known to catalyze the removal of mono-, di-, and tri-methyl groups from lysine residues [24]. The KDM6 demethylase family, which belongs to the JmjC family, is comprised by three distinct members called KDM6A (also known as UTX), KDM6B (also known as JMJD3) and KDM6C (also known as UTY) [25]. KDM6A and KDM6C are found on sex chromosome X and Y, respectively, while KDM6B is located at the chromosome 17 [25]. All three KDM6 protein members function as the demethylases to remove methyl groups from H3K27 dimethylation (H3K27me2) and H3K27 trimethylation (H3K27me3) dependent on JmjC domains at their C terminal [26]. Of note, KDM6C (also known as UTY) exhibits low demethylase activity due to subtle amino acid substitutions within the substrate-binding site of its JmjC domain [27]. There are four common domains within the structure of three KDM6 demethylase family members: TPR, helical, TPR, helical, zinc finger and JmjC domain (Fig. 1A). Among these, the TPR domain is present in both KDM6A and KDM6B but absent in KDM6C [28]. This domain facilitates protein-protein interactions and mediates the integration of multiprotein complexes, enabling coordinated chromatin regulation [17]. Notably, the zinc finger domain is unique to KDM6A and facilitates DNA binding to target specific genomic loci for demethylation [29].In contrast, the helical domain—conserved across all three members—mediates selective protein-protein interactions that promote chromatin-modifying complex assembly [30]. The C-terminal JmjC domain, central to all KDM6 enzymes, catalyzes histone lysine demethylation [25]. Its structure contains an eight-stranded β-sheet forming a catalytic pocket, where five conserved residues coordinate essential cofactors: Lys206 and His276 bind 2OG, while Thr185, His188, and Glu190 participate in Fe(II) binding [31]. Although KDM6A/B are well studied in cancer, all KDM6 members also exhibit non-enzymatic scaffolding roles in regulating macromolecular complexes [32].

Fig. 1.

A KDM6 demethylases structure. KDM6A/B/C are 1401/1641/1347amino acids residues in length. KDM6 demethylases have five domains: TPR, Helical, Linker, JmjC, and Zinc finger, KDM6A, All of the KDM6 demethylases have TPR, JmjC and Zinc finger domain. B The demethylase process of KDM6 demethylases

2.2 catalytic mechanisms of KDM6A/B

KDM6 is a JmjC domain-containing histone demethylase that specifically removes methyl groups from H3K27me3/me2, thereby regulating gene expression. Its catalytic mechanism depends on Fe(II), α-ketoglutarate (α-KG), and oxygen, and involves substrate recognition via the JmjC domain, cofactor binding, oxygen activation to generate an iron-oxo intermediate, demethylation of H3K27, and product release [17, 28]. The process initiates with Fe²⁺ coordination by His188, Glu190, and Thr185 residues in a monodentate manner, and bidentate binding of 2-OG (α-KG) to the metal center [10]. The H3K27me3 peptide is recognized by the JmjC domain, with a zinc finger domain specifically interacting with H3 residues 17–21, ensuring substrate specificity [10]. Oxidative decarboxylation of 2-OG produces CO₂, succinate, and a hydroxylated methyl group that is released as formaldehyde. This demethylates H3K27me3/me2 to H3K27me2/me1 or an unmodified state, facilitating transcriptional activation (Fig. 1B) [33–35].

Hypoxia, a common feature of the tumor microenvironment, reduces α-KG levels and increases the accumulation of 2-hydroxyglutarate (2-HG), leading to inhibition of KDM6A, H3K27 hypermethylation, and impaired cellular differentiation [29, 30]. In contrast, under hypoxic conditions, HIF-1α is stabilized due to impaired activity of oxygen-dependent prolyl hydroxylases (PHDs) and factor-inhibiting HIF-1 (FIH1). Stabilized HIF-1α translocates to the nucleus, dimerizes with HIF-1β (ARNT), and binds to hypoxia-response elements (HREs) in the KDM6B promoter, upregulating KDM6B expression to promote adaptation to low oxygen and activation of hypoxia-inducible genes [29, 30]. Thus, hypoxia negatively regulates KDM6A but positively regulates KDM6B.

KDM6A activity is also modulated by metabolic alterations and the tumor microenvironment. Mutations in IDH1/2, SDH, or FH disrupt Fe(II)/α-KG homeostasis; notably, oncogenic IDH1/2 mutations produce D-2HG, which inhibits KDM6A [28, 29]. KDM6B is regulated by multiple signaling pathways, including NF-κB, RAS-RAF, STAT3, and TGF-β/Smad, as well as by metabolic factors such as vitamin D, glutamine availability, and hypoxia [25, 31, 32, 36]. Additionally, TRAF6 and BMP4 activate KDM6B through specific mechanisms. KDM6 also engages in context-dependent interactions with proteins such as ER and RAR, influencing its catalytic activity and roles in differentiation, tumorigenesis, and metastasis [25, 31, 32, 36].

3. Roles of KDM6A/B in therapeutic resistance of cancers

KDM6A/B exhibits a dual - edged effect on the sensitivity to chemotherapeutic agents, which is contingent upon the specific tumor type and context. Not only that, but KDM6A/B has also demonstrated analogous dual - edged effects in other therapeutic modalities, including immunotherapy, targeted therapy, and radiotherapy. (Figures 2, 3 and 4)

Fig. 2.

The regulatory mechanisms of KDM6A that acts as a promotor in therapeutic resistance of 16 types of cancers

Fig. 3.

The regulatory mechanisms of KDM6A that acts as a suppressor in therapeutic resistance of 17 types of cancers

Fig. 4.

The regulatory mechanisms of KDM6B that acts as a promotor/suppressor in therapeutic resistance of 15 types of cancers

3.1 roles of KDM6A/B in chemoresistance

3.1.1 KDM6A

In a variety of malignancies, including acute myeloid leukemia (AML), bladder cancer, triple-negative breast cancer (TNBC), colorectal cancer, non-small cell lung cancer (NSCLC), and prostate cancer, KDM6A functions as a suppressor of chemoresistance. Its loss contributes to tumor cell insensitivity to chemotherapeutic agents through multiple mechanisms.

Despite being a cornerstone therapeutic approach for prolonging remission in acute myeloid leukemia (AML), conventional chemotherapy frequently culminates in the development of resistance, notably to agents such as cytarabine, in the majority of patients [33]. Genomic interrogation via exome sequencing in a cohort of 50 AML cases implicated loss of KDM6A function in conferring chemoresistance and disease relapse [34]. Consistent with this, AML cells lacking KDM6A exhibit diminished sensitivity to cytarabine and daunorubicin. Mechanistically, KDM6A deficiency reduces H3K27 acetylation at the promoter of the solute carrier family 29 member 1 (SLC29A1) gene, which encodes the nucleoside transporter ENT1. This epigenetic alteration suppresses ENT1 expression, impairing cellular uptake of nucleoside analogs and thereby fostering chemoresistance [35], thus leads to chemoresistance.

In muscle-invasive bladder cancer harboring inactivating mutations in KDM6A and components of the SWI/SNF complex, overexpression of Enhancer of Zeste Homolog 2 (EZH2) is linked to adverse prognosis and attenuated immune activity [36, 37]. EZH2 mediates resistance to cisplatin by repressing natural killer (NK) cell-related genes—MIP-1α, ICAM1, ICAM2, and CD86—thus dampening NK cell signaling. Concurrently, EZH2 upregulates pluripotency markers ALDH2, p63, and CK5, further promoting cisplatin resistance. These EZH2-driven resistance mechanisms are antagonized by KDM6A and SWI/SNF complex proteins [38].

In triple-negative breast cancer (TNBC), where cytotoxic chemotherapy remains a mainstay, KDM6A serves as a suppressor of chemoresistance. Depletion of KDM6A in TNBC models attenuates sensitivity to paclitaxel [39–41]. Chemoresistance and relapse in TNBC are often attributable to the enrichment of breast cancer stem cells (BCSCs), characterized by expression of pluripotency factors OCT4, NANOG, SOX2, and KLF4 [42, 43]. Extended chemotherapeutic exposure induces HIF-1-dependent S100A10 expression, which complexes with ANXA2, SPT6, and KDM6A, recruiting the latter to pluripotency gene promoters. KDM6A subsequently demethylates H3K27me3, activating stemness-associated transcription and promoting BCSC expansion and tumor recurrence [44]. This process is further modulated by p38 MAPK-driven nuclear translocation of SMARCD3, leading to transcriptional activation of stemness gene [45].

Genomic analyses in colorectal cancer correlate KDM6A loss-of-function mutations with recurrence, metastasis, and resistance to chemoradiotherapy [46]. Stemness maintaining of CSCs and tumor initiating cells (termed as CSCs sometimes) are considered as the important players for chemoresistance of colorectal cancer [47, 48]. KDM6A enhances stem-like properties through transcriptional activation of NOTCH, ID1, and TERT, conferring resistance to oxaliplatin and fluorouracil [49, 50]. Similarly, in non-small cell lung cancer (NSCLC) and prostate cancer, KDM6A collaborates with KMT2B to augment H3K4me3 levels and activate Wnt/β-catenin signaling, contributing to cisplatin and docetaxel resistance, respectively [51, 52]. KDM6A also drives enzalutamide resistance in metastatic castration-resistant prostate cancer (mCRPC) [53].

Collectively, these findings indicate that KDM6A loss-of-function indirectly facilitates chemoresistance across diverse malignancies—including AML, bladder cancer, TNBC, colorectal cancer, NSCLC, and prostate cancer—through mechanisms involving drug transporter downregulation, chromatin remodeling, and activation of stemness pathways. Nevertheless, the function of KDM6A is highly context-dependent, and it does not uniformly suppress drug resistance.

For instance, in esophageal squamous cell carcinoma (ESCC), KDM6A promotes genomic stability via a demethylase-independent mechanism. By interacting with SND1, it recruits RPA and Ku70 to nascent DNA, stabilizing replication forks and conferring resistance to camptothecin and hydroxyurea [54].

A notable exception is observed in testigenal germ cell tumors (TGCTs), where cisplatin-based chemotherapy often induces significant toxicity and resistance. Surprisingly, KDM6A inhibition sensitizes both cisplatin-sensitive and-resistant TGCT models to cisplatin. This effect involves p53 pathway activation and suppression of chromatin regulators such as BRD4, underscoring the context-dependent duality of KDM6A function [55–57].

Specially, alteration of KDM6A expression is highly associated with cisplatin resistance in osteosarcoma, and decreased KDM6A protein can resensitize osteosarcoma to cisplatin. Data dependent on RNA sequencing suggests that loss of KDM6A increase H3K27me3 level and decreased expression of PRKCA and MCL1, PRKCA, also knowns as protein kinase C alpha, has been propose to regulates various cellular functions including proliferation, apoptosis, and differentiation [58]. MCL1 is a member of SCL2 family that inhibits cellular apoptosis [59]. Inhibition of PRKCA and MCL1 inactivates RAF/ERK/MAPK cascades and decreases phosphorylation of BCL2 [60].

The deletion of KDM6A emerges as a shared mechanism driving chemoresistance across diverse cancer types. This effect is primarily mediated through epigenetic mechanisms that dysregulate drug transport, activate cancer stem cell programs, and alter key signaling pathways. However, the function of KDM6A is not invariant; it exhibits opposing or unique roles in specific malignancies, such as esophageal squamous cell carcinoma (ESCC), testicular germ cell tumors (TGCTs) and osteosarcoma, highlighting its complexity and context-dependency in tumor biology. Consequently, therapeutic strategies targeting KDM6A must be precisely tailored to the specific cancer type.

3.1.2 KDM6B

Inhibition of KDM6B sensitizes diffuse large B-cell lymphoma (DLBCL) cells to numerous chemotherapeutics including vincristine, doxorubicin, Bortezomib, carfilzomib, and panoinostat. Intriguingly, KDM6B positively regulates the activation of B-cell signaling and BCL6, both of which are critical for the survival of B cells [61].

In breast cancer and ovarian cancer cells, KDM6B engages in sensitivity regulation of alkylating agents. Alkylating agents induces poly (ADP-ribose) polymerase-1 (PARP-1) hyperactivation, which also known as PARP1-dependent cell death (PARthanatos), promotes cell death. Mechanistically, KDM6B enhances DNA damage by inhibiting MGMT expression and its direct DNA repair function, leading to PARthanatos and cell death. Moreover, KDM6B suppresses phosphorylation of CHK1 and activation of XRCC1-dependent DNA repair machinery to fix DNA damage escape from MGMT repair [62].

Although single inhibition of KDM6A has minor efficacy against temozolomide (TMZ) -resistant cells, dual inhibition of KDM6A and KDM5A strongly resensitizes TMZ-resistant cells to TMZ [63].

Same with KDM6A, in colorectal cancer, osteosarcoma and TGCTs, KDM6B promotes chemoresistance as described above [49, 50, 57, 60].

In total, KDM6A is associated with chemoresistance in multiple cancers through diverse mechanisms, We have made relevant summaries (Table 1).

Table 1.

The roles of KDM6A/B and regulatory genes and pathways related in chemoresistance

Demethylases	Cancer types	Roles in resistance	Drugs	Related genes and signal pathways	References
KDM6A	Triple-negative breast cancer	Promotor	Paclitaxel	A2BR, p38 MAPK, SMARCD3, FOXO3, OCT4, NANOG, SOX2, and KLF4	[44, 45]
	Colorectal cancer		Oxaliplatin	Notch	[49]
			Fluorouracil	ID1, TERT	[50]
	ESCC		Camptothecine, hydroxyurea	SND1, RPA, Ku70	[54]
	Testicular germ cell tumors		Cisplatin	p53, BRD4	[57]
	NSCLC			KMT2B, Wnt, β-catenin	[51]
	Osteosarcoma			PRKCA, MCL1, RAF/ERK/MAPK cascades, BCL2	[60]
	Prostate cancer		Docetaxel	Wnt, β-catenin, LEF1	[52]
	mCRPC		Enzalutamide	-	[53]
	Acute myeloid leukemia	Suppressor	Cytarabine	ENT1	[34, 35, 75]
	Muscle – invasive bladder cancer		Cisplatin	NK cell-associated genes MIP-1α, ICAM1, ICAM2, CD86 ALDH2, CK5	[38]
	Triple-negative breast cancer		Paclitaxel	-	[41]
	Rectal cancer		Fluorouracil, capecitabine	-	[46]
KDM6B	Colorectal cancer	Promotor	Oxaliplatin	Notch	[49]
			Fluorouracil	ID1, TERT	[50]
	Testicular germ cell tumors		Cisplatin	p53, BRD4	[57]
	Osteosarcoma			PRKCA, MCL1, RAF/ERK/MAPK cascades, BCL2	[60]
	Diffuse large B-cell lymphoma		Vincristine, doxorubicin, bortezomib, carfilzomib, panoinostat	BCL6, B-cell receptor signaling	[61]
	Glioblastoma		Temozolomide	KDM5B	[63]
	Breast cancer, ovarian cancer	Suppressor	MNNG	PARP-1, MGMT, CHK1, XRCC-1	[62]

3.2 roles of KDM6A/B in immunotherapy

3.2.1 KDM6A

The role of KDM6A in immunotherapy of urothelial carcinoma is evaluated based on bioinformatic analysis of open data base. This study proposes that hypoxic microenvironment may induce the formation of KDM6A deletion mutation, enhancing the differentiation of immunosuppressive cells, and ultimately results in immunological resistance and poor prognosis [64]. Bladder cancer is the most common type of urothelial carcinoma, and the impact of kdm6a on its immunotherapy has also been explored by relevant studies. A bioinformatic study conducted in bladder cancer proposes that loss of function mutation of KDM6A is associated with a lower number of tumor-infiltrating immune cells (TIICs) and immune escape [20]. Loss of KDM6A induces the expression of cytokines and chemokines such as CCL2, CXCL1, and IL6, M2 macrophage polarization, and proliferation of bladder cancer stem calls [65]. Importantly, KDM6A deficiency also downregulates antigen presentation-related genes, further impairing T-cell mediated anti-tumor immunity. Another study analyzing the role of KDMs in immune response on HCC illustrates that KDM6A significantly promotes the infiltration of B cells, CD8 + T cells, CD4 + T cells, macrophages, neutrophils, and dendritic cells (DCs) [66].

In breast cancer, KDM6A loss indues immune evasion by upregulating TGF-β extracellular secretion and suppressing expression of cytotoxic genes in CD8 + T cells in vitro [41]. Moreover, KDM6A loss leads to decreased expression of antigen presentation machinery components, thereby limiting T cell recognition of tumor cells. Loss of function mutation of KDM6A also promote immune escape of colorectal cancer, the main function of KDM6A is to promote inflammatory macrophage response and effector T cell response [19]. Specifically, KDM6A enhances the expression of MHC class II molecules, facilitating antigen presentation to CD4 + T cells. In medulloblastoma, KDM6A also enhances recruitment of CD8 + T cells to the tumor microenvironment by inducing Th1-type chemokines, which are responsible for T cell migration [67].

KDM6A also play a vital role of tumor immune regulation in pancreatic cancer. Pancreatic cancer progression is positively correlated with tumor-associated neutrophils (TAN) and neutrophil extracellular traps (NET) formation. Pancreatic cancer cells with KDM6A loss shows increased TAN and NET, high level of CXC motif chemokine ligand 1 (CXCL1) protein, which in turn recruits neutrophil, thus leas to tumor growth. CXCL1-CXCR2 axis blockade may be a promising targeted therapy in pancreatic cancer with KDM6A loss [68]. Concurrently, the impairment of antigen presentation and reduction in T cell infiltration contribute to the immunosuppressive milieu.

Daratumumab (anti-CD38 monoclonal antibody) resistance occurs in multiple myeloma (MM) is induced by loss of KDM6A dependent on its demethylase activity. CD48 is also downregulated upon KDM6A loss. KDM6A promote the expression of CD38/CD48 in the surface of MM cells, thus results in NK cell response [69]. KDM6A also upregulates the expression of CIITA and NLRC5 dependent on its acetylation activity in MM. CIITA and NLRC5 encode regulators of major histocompatibility complex (MHC) genes which increases MHC’s expression and T-cell infiltration [70].

3.2.2 KDM6B

In breast cancer, KDM6B is reported to influence macrophage polarization. Inhibition of KDM6B in tumor associated macrophages (TAMs) induces M2 polarization by activating β-catenin/c-Myc signaling [71]. Moreover, miR-138-5p is illustrated to inhibit the expression of KDM6B and promote the M2-like phenotype [71, 72].

Immune-suppressive myeloid cells surround with glioblastoma results the failure of immune checkpoint therapy and glioblastoma survival from anti-PD1 therapy. Mechanistic study shows that immune-suppressive myeloid cells appear with overexpression of KDM6B, which decreases antigen presentation, interferon response and phagocytosis via upregulating expression of immune suppressive factors including mafb, Socs3 and Sipra and limits anti-PD1 efficacy ultimately [73]. Recent report shows iron-loaded cancer-associated fibroblast (FerroCAFs) are associated with immunosuppression in prostate cancer, lung cancer, and ovarian cancer. Mechanistically, FerroCAFs are essential for recruitment of immune-suppressive myeloid cells by secreting myeloid cell-associated proteins such as CSF1, CCL2 and CXCL1. Aggregation of iron in FerroCAFs is induced by Hmox1-mediated iron release from heme degradation. Myeloid cell-associated proteins are expressed by KDM6B-depedent demethylase activity, which relies on enough iron substrate catalyzation. Repressing KDM6B activation in FerroCAFs incurs anti-tumor immunity and tumor suppression [74].

KDM6A/B impacts tumor - infiltrating immune cells, immune evasion, and immunotherapy response in various cancers, We have composed the relevant summaries (Table 2).

Table 2.

The roles of KDM6A/B and regulatory genes and pathways related in immune escape

Demethylases	Cancer types	Roles in resistance	Drugs	Related genes, signal pathways, and cells	References
KDM6A	Bladder cancer	Suppressor	-	Tumor infiltration immune cells, CCL2, IL6, CXCL1, M2 macrophage polarization, cancer stem cells	[20, 65]
	Triple-negative breast cancer			TGF-β	[41]
	Colorectal cancer			Inflammatory macrophage, effector T cell	[19]
	Medulloblastoma			Th1-type chemokines	[67]
	Multiple myeloma		Daratumumab	CD38, CD48, NK cells	[69]
			-	CIITA, NLRC5	[69]
	Hepatocellular carcinoma			B cells, CD8 + T cells, CD4 + T cells, macrophages, neutrophils, and DCs	[66]
	Pancreatic cancer			TAN, NET, CXCL1-CXCR2 axis	[68]
	Urothelial carcinoma			Hypoxia, infiltration T cells, B cells, Treg cells, and macrophages	[64]
KDM6B	Glioblastoma	Promotor	-	PD1, antigen presentation, interferon, phagocytosis, Mafb, Socs3 and Sirpa	[73]
	Prostate cancer, lung cancer, ovarian cancer			CAFs, HMOX1/iron/KDM6B signaling axis	[74]
	Breast cancer	Suppressor		miR-138-5p, M2 macrophage polarization β-catenin/cMyc	[71, 72]
	Hepatocellular carcinoma			B cells, CD8 + T cells, CD4 + T cells, macrophages, neutrophils, and DCs	[66]

3.3 roles of KDM6A/B in targeted therapy resistance

3.3.1 KDM6A

KDM6A, functioning as a histone demethylase, acts as a tumor suppressor in various cancers, with its activity or expression level directly influencing the sensitivity of tumor cells to targeted therapies. In some contexts, its loss sensitizes cells to treatment, whereas in others, its overexpression confers resistance.

Contrary to the function of repression of chemoresistance in AML, KDM6A is a promotor for resistance to poly ADP ribose polymerase (PARP) and BCL2 inhibition. KDM6A is indispensable for regulation of DNA damage repair (DDR) gene expression (mainly refer to BRCA and RAD), deficiency of KDM6A causes the impairment of DDR transcriptional activation, thus cells with loss of KDM6A can be more sensitive to olaparib (PARP inhibitor) and radiotherapy. In addition, overexpression of KDM6A is complicated in decreased expression BCL2 expression and upregulation of BCL2A1. AML cells with acquired KDM6A gain of function mutation is associated with venetoclax (BCL2 inhibitor) tolerance [75]. Similar with AML, head and neck squamous cell carcinoma can be sensitive to olaparib with the elevated H3K27me3 level of genes such as BRCA1 and FANCD2 induced by KDM6A loss [76]. RUNX3 is identified as a molecular driver of metastasis [77]. KDM6A deficient pancreatic cancer is sensitive to BET inhibitors JQ-1 or iBET-151, which downregulate Np63, MYC, and RUNX3 oncogenes, reverse squamous squamous differentiation, and confine tumor growth [78].

Resistance to trastuzumab is a significant challenge in the treatment of HER2-positive breast cancer. tRNA-derived fragment-27 (tRF-27) play crucial roles in trastuzumab resistance. Mechanistically, upregulated expression of tRF-27 is resulted from reduced inhibitory H3K27me3 modification at the promotor regions of tRF-27-related tRNA genes regulated by increased KDM6A activity [79].

Treatment with imatinib, a tyrosine kinase inhibitor targeting BCR-ABL fusion, though showed a complete response in chronic myelogenous leukemia (CML), many patients with CML eventually develop resistance to imatinib. KDM6A has been shown to overexpress in CML and promote imatinib resistance. Notably, KDM6A exert its demethylase activity to promote YY1-mediated transcriptional upregulation of TRKA, activation of TRKA is required for NGF– induced (NGF) expression of AKT and ERK, which invoke imatinib resistance and cell survival [80].

Proliferation of CSCs in glioblastoma potentiates resistance to receptor tyrosine kinases (RTKs). Accordingly, CSCs upregulation is dependent on activation of Notch signaling induced by demethylase activity of KDM6A [81].

In contrast to its role as a tumor suppressor, KDM6A, under specific cancer contexts and in response to targeted therapies, functions to promote and enhance treatment sensitivity—acting as a “targeted therapy–dependency factor.” Loss or functional impairment of KDM6A leads to resistance of tumor cells to the corresponding targeted agents. The presence of KDM6A generally sustains tumor cell dependence on inhibitors of specific signaling pathways, often by modulating the expression of key genes such as FGFRs, thereby rendering the cells more vulnerable to targeted drug–induced cell death. In the absence of KDM6A, this dependency is disrupted, resulting in therapy resistance.

Conversely, whole-exome sequencing applied in a pazopanib (tyrosine kinase inhibitor (TKI) for FGFR3-TACC3 fusion) resistance patient with bladder cancer shows the loss of function mutation of KDM6A [82].

In HCC, KDM6A positively influences the expression of fibroblast growth factor receptors (FGFRs), knockdown of KDM6A is corelated with decreased sensitivity of HCC cells to FGFRs inhibitor, lenvatinib [83].

Loss of KDM6A strongly increases H3K27 acetylation in super enhancer (SE) and activates SE regulating expression of Np63, MYC, and RUNX3 oncogenes. Np63 and MYC induce squamous-like pancreatic cancer and are correlated with poor prognosis [84].

In triple-negative breast cancer (TNBC), depletion of KDM6A reduces sensitivity to receptor tyrosine kinase inhibitors (e.g., dasatinib) and mTOR inhibitors (e.g., AZD2014) [41].

KDM6A functions as a therapeutic dependency factor, implying that its normal activity is required to maintain tumor sensitivity to specific targeted agents. Consequently, inactivation of KDM6A in these contexts serves as a driver of drug resistance. This stands in direct contrast to its role as a suppressor, where its loss enhances drug sensitivity, thereby underscoring the cancer type- and treatment context-dependent functionality of KDM6A.

3.3.2 KDM6B

In breast cancer, the most frequently mutated gene, PIK3CA, results in sustained activation of the PI3K/AKT signaling [85]. Although PI3K inhibitor GDC-0941 shows promising results from preclinical studies, intrinsic or acquired resistance limits the clinical efficacy. Aberrant activation of PI3K/AKT signaling not only directly promotes cell survival, but also mediates the phosphorylation inhibition of EZH2, thus leads to epigenetic switch from H3K27me3 to H3K27ac at the IFGBP5 gene promotor, activation of IGFBP5 stimulates expression of MAPK, which decreases BIM expression and promotes cell survival. With the occurrence of resistance, KDM6B is overexpressed, Activation of IGFBP5 is changed by KDM6B. KDM6B-IGFBP5 axis confers resistance to GDC-0941.

Glutamine metabolism engages in demethylase activity regulation of KDM6B and resistance to PLX4032 in melanoma with BRAFV600E mutation. Glutamine deficiency in core region of solid tumor gives rise to high H3K27me3 due to decreased α-ketoglutarate level, an indispensable substrate for demethylase activity of KDM6B. High level of H3K27me3 induced by low glutamine results in resistance to PLX4032 (BRAF inhibitor).

Same with KDM6A, in glioblastoma, KDM6B promotes resistance to RTKs inhibitors as described above [81]. In head and neck cancer, KDM6B also act as a promotor on olaparib-resistance, relevant resistant mechanism is same KDM6A as mentioned before [76].

KDM6A/B influence chemoresistance in various cancers via modulating gene expression and demethylase activity, We have drawn up the relevant outline (Table 3).

Table 3.

The roles of KDM6A/B and regulatory genes and pathways related in targeted therapy resistance

Demethylases	Cancer types	Roles in resistance	Drugs	Related genes and signal pathways	References
KDM6A	Acute myeloid leukemia	Promotor	Venetoclax	BCL2, BCL2A1	[75]
			Olaparib	PARP1, BRCA, RAD	[75]
	Chronic myelogenous leukemia		Imatinib	YY1, TRKA, NGF, AKT, ERK	[80]
	HER2-positive breast cancer		Trastuzumab	tRF-27	[79]
	Glioblastoma		Dasatinib, crenolanib, gefitinib	Notch	[81]
	Head and neck squamous cell carcinoma		Olaparib	PARP1/2, BRCA1, FANCD2	[76]
	Pancreatic cancer		JQ-1, iBET-151	BET, Np63, MYC, RUNX3	[78]
	Bladder cancer	Suppressor	Pazopanib	FGFR3-TACC3 fusion	[82]
	Hepatocellular carcinoma		Lenvatinib	FGFR4	[83]
	Triple-negative breast cancer		AZD2014	mTOR	[41]
			Dasatinib	RTK	[41]
KDM6B	Breast cancer	Promotor	GDC-0941	IGFBP5, PI3K/AKT, MAPK, BIM	[96]
	Melanoma	Suppressor	PLX4032	Glutamine	[97]

3.4 roles of KDM6A/B in radioresistance

3.4.1KDM6A

Machine leaning model study shows KDM6A is associated with radiation sensitivity in breast cancer [86]. Recent evidence proposes that inhibition of KDM6A sensitizes head and neck cancer cells to radiation with reduction of the stem-like potential of these cells [18].

3.4.2KDM6B

hypoxic microenvironment is also associated with resistance to radiotherapy in ESCC. Hypoxia increases expression of KDM6B, which promotes cell survival, migration, and DNA repair, furthermore, KDM6B also decreases cell apoptosis. KDM6B targeting, concomitant with conventional radiotherapy, is an effective strategy to overcome radioresistance [87].

Radiotherapy resistance of prostate cancer is aroused by high expression of p53. Mechanism dissection reveals that overexpression of p53 in radioresistant cells is caused by a decrease of H3K27me3 level at promoter of TP53 gene due to enhancement of KDM6B activity. In addition, p53 is responsible for DNA damage signaling response and cell recovery upon long-term stress of external beam radiotherapy. KDM6B is crucial epigenetic target for radiosensitization of prostate cancer [88].

KDM6A/B primarily contributes to radioresistance through overexpression, these findings collectively underscore the functional significance of KDM6A/B in mediating key resistance mechanisms and highlight the therapeutic promise of targeting these demethylases to improve radiotherapeutic outcomes. They acts as a promotor on resistance to radiotherapy [18]. We have compiled the related resume (Table 4).

Table 4.

The roles of KDM6A/B and regulatory genes and pathways related in radioresistance

Demethylases	Cancer types	Roles in resistance	Drugs	Related genes and signal pathways	References
KDM6A	Head and neck cancer	Promotor	-	CSCs	[18]
	Breast cancer	Suppressor	-	-	[46, 86]
	Rectal cancer
KDM6B	Head and neck cancer	Promotor	-	CSCs	[18]
	ESCC			HIF-1α, CAIX	[87]
KDM6B	Prostate cancer	Promotor	-	P53, DDR	[88]

3.5 roles of KDM6A/B in other types of adaptive resistance

Glucocorticoids (GC) is an effective avenue of T-lineage acute lymphoblastic leukemia (T-ALLs), enforced JDP2 overexpression can lead to adaptive GC resistance, whereas inactivation of KDM6A counteracts the GC resistance induced by JDP2 [89].

KDM6A and KDM6B, as critical histone demethylases, play a central role in therapeutic resistance across various cancers primarily through the regulation of epigenetic marks such as H3K27me3/me2. By modulating the expression of key genes, they are extensively involved in biological processes including drug transport, maintenance of cancer stem cell (CSC) properties, DNA damage repair, and remodeling of the immune microenvironment, thereby profoundly influencing tumor responses to chemotherapy, immunotherapy, targeted therapy, and radiotherapy. This complex functional landscape suggests that therapeutic strategies targeting KDM6A/B must be tailored to specific cancer types, molecular contexts, and treatment modalities to achieve personalized and precision-based intervention, aiming to reverse drug resistance while mitigating adverse effects.

4. Conclusion remarks and perspectives

KDM6A and KDM6B exhibit context-dependent dual roles in cancer biology, functioning as either oncogenic drivers or tumor suppressors. They achieve this by dynamically regulating H3K27 methylation states, thereby modulating chromatin accessibility and transcriptional programs critical for tumor progression and therapeutic resistance. Dysregulation of these enzymes contributes to multidrug resistance across diverse therapeutic modalities—including chemotherapy (e.g., in AML and TNBC), targeted therapies (e.g., PARP or BCL2 inhibitors), immunotherapy, and radiotherapy—via mechanisms such as stemness maintenance, DNA repair, immune evasion, and metabolic reprogramming. For instance, KDM6A loss promotes chemoresistance by repressing drug transporters like ENT1 or activating pluripotency factors (e.g., OCT4/NANOG), while KDM6B drives resistance via PI3K/AKT/IGFBP5 signaling in breast cancer or hypoxia-responsive pathways in esophageal cancers. In immunotherapy contexts, genetic alterations in KDM6A/B reshape the tumor microenvironment by modulating immune cell infiltration, cytokine signaling, and antigen presentation.

Substantial progress has been made in the research of KDM6 inhibitors, with development strategies primarily based on the catalytic mechanism of KDM6 as an Fe²⁺/α-ketoglutarate-dependent demethylase. Currently known KDM6 inhibitors can be broadly categorized into six classes, including biguanides, 2-OG analogs, quinoline derivatives, pyridone-based compounds, phenolic inhibitors, and other miscellaneous types [90–93]. Although a variety of inhibitors have been reported, none have yet progressed to clinical trials. Major challenges include poor selectivity (due to high conservation of the JmjC domain), low cellular permeability, insufficient stability, and variable efficacy resulting from cancer heterogeneity [94]. Future research directions involve the development of allosteric inhibitors, expansion of chemical diversity using natural products and metal complexes, and exploration of combination therapies [94, 95].

Future research priorities include unraveling their non-catalytic roles in chromatin remodeling complexes (e.g., SWI/SNF), developing context-specific inhibitors (e.g., for KDM6A-mutant germ cell tumors or KDM6B-overexpressing glioblastomas), validating predictive biomarkers, and exploring microenvironmental crosstalk (e.g., hypoxia or metabolic stress). Integrating mechanistic insights with clinically relevant models will be essential to harness their therapeutic potential while navigating their dualistic functions, ultimately improving outcomes in cancer treatment.

Author contributions

L.Z.Y and H.H.L : Writing - Original Draft, Conceptualization, Methodology. These authors made equal contributions.Z.C.Q and G.G.L: Investigation, Resources, Visualization, Validation. Y.P.Z: Visuliaztion.X.B.C and X.D.Y: Writing - Review & Editing, Supervision. Visualization, Resources, Validation.

Funding

This work was supported by Yunnan Basic Research Program - Joint Project of Kunming Medical University (202301AY070001-249). This work is also supported by Yunnan Provincial Department of Science and Technology Foundation for Youths (202301AU070195).

Data availability

No datasets were generated or analysed during the current study.

Data availability

Not applicable.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

Not applicable.