Ubiquitin is a 76-amino-acid protein of approximately 8.5 kDa that plays critical roles in various biological processes in cells^[1]. Ubiquitination is an important and widespread post-translational modification (PTM) in eukaryotes^[2], whereby ubiquitin molecules are conjugated to target proteins through the coordinated action of multiple enzymes, thereby modulating target protein function, localization, stability, and intermolecular interactions^[3]. The ubiquitination process is typically executed sequentially by three main enzyme classes: ubiquitin-activating enzymes (E1), ubiquitin-conjugating enzymes (E2), and ubiquitin ligases (E3). First, the E1 enzyme uses adenosine triphosphate (ATP) to activate ubiquitin, forming a high-energy thioester bond and transferring it to the active site of an E2 enzyme. Subsequently, the E3 enzyme serves as a substrate-recognition factor, binding the target protein and facilitating the transfer of ubiquitin from the E2 enzyme to a lysine residue on the substrate, thereby forming an isopeptide bond^[4].

Ubiquitination is a dynamic and reversible process. Deubiquitinating enzymes (DUBs) remove ubiquitin or polyubiquitin chains from target proteins by hydrolyzing the isopeptide bonds between ubiquitin and substrate protein, thereby reversing the ubiquitination process^[5]. DUBs comprise multiple enzyme families, including the ubiquitin-specific protease family (USPs), the ovarian tumor protease family (OTUs), the JAB1/MPN/Mov34 metalloenzyme (JAMM) family, the ubiquitin C-terminal hydrolase family (UCHs), and the Machado-Joseph disease protein domain protease family (MJDs), among others^[6-8].

The DUB PSMD14 (26S proteasome non-ATPase regulatory subunit 14), also known as RPN11 or POH1, is a member of the JAMM family. By regulating protein ubiquitination levels, PSMD14 is involved in various physiological and pathological processes, including the cell cycle, apoptosis, autophagy, immune responses, and tumorigenesis. Its functions are largely mediated through the specific deubiquitination of diverse substrate proteins, thereby modulating their stability^[9]. In recent years, accumulating evidence has demonstrated that aberrant expression of PSMD14 is closely associated with multiple human diseases. In the context of malignant tumors, PSMD14 is upregulated in various cancers, including hepatocellular carcinoma (HCC)^[10], osteosarcoma^[11], head and neck squamous cell carcinoma^[12], breast cancer^[13], and lung adenocarcinoma^[14], where its high expression correlates with poor prognosis. Furthermore, PSMD14 plays important roles in non-tumor diseases. For example, in glucocorticoid-induced osteoporosis, PSMD14 stabilizes SLC7A11 to suppress osteocyte ferroptosis and ameliorate bone loss^[15]. In the prognostic assessment of brain injury following cardiac arrest, PSMD14 has been validated as a predictive biomarker for early neurological outcomes^[16] and in choroidal neovascularization, PSMD14 regulates the Keap1/Nrf2 antioxidant pathway by deubiquitinating and stabilizing Keap1^[17]. These findings suggest that the functions of PSMD14 are not limited to tumor regulation but extend broadly to the pathological processes of various diseases.

This review focuses on PSMD14-driven mechanisms of therapeutic resistance in cancer, while also considering its broader pathophysiological roles in certain non-malignant conditions that may contribute to target validation and safety assessment, particularly with respect to its deubiquitinating substrates and functions. Furthermore, we discuss the clinical value of PSMD14 as a prognostic biomarker and recent advances in therapeutic strategies targeting PSMD14, aiming to provide new insights for the development of precision therapeutic approaches.

PSMD14 deubiquitinates and stabilizes a broad range of substrate proteins, thereby regulating multiple biological processes as summarized in Figure 1.

In recent years, the oncogenic functions of the deubiquitinase PSMD14 have been documented in a wide range of cancer types. A systematic review and meta-analysis by Dong et al., encompassing 8 studies and 1,608 patients, demonstrated that high PSMD14 expression is significantly associated with reduced overall survival [OS; pooled hazard ratio (HR) = 1.92, 95% confidence interval (CI) 1.67-2.22, P < 0.00001] and disease-free survival (pooled HR = 1.84, 95%CI 1.33-2.55, P = 0.0003). Elevated PSMD14 expression was also correlated with larger tumor size, poorer differentiation, and a higher incidence of lymph node metastasis^[49]. Within this pan-cancer context, tumor type–specific studies have further elucidated the prognostic relevance of PSMD14.

To date, multiple studies have assessed the relationship between PSMD14 expression and patient outcomes using Kaplan–Meier survival analyses. In PDAC^[50] (P < 0.05), osteosarcoma^[11] (P = 0.013), non-small cell lung cancer^[51] (P < 0.001), lung adenocarcinoma^[40] (P = 0.026), and head and neck squamous cell carcinoma^[12] (P = 0.0132), high PSMD14 expression correlates with shorter OS, indicating an unfavorable prognosis.

Beyond oncology, PSMD14 has also been investigated as a predictor of early postoperative neurological outcomes in patients with cardiac arrest. In a prognostic model for brain injury following cardiac arrest, PSMD14 was identified as the most readily detectable peripheral blood biomarker candidate, with the highest area under the curve (AUC, 0.767). Its protein level showed a strong positive linear correlation with NDS neurological function scores (r ≈ 0.89, P < 0.001), and PSMD14 was validated for the first time as a predictor of early neurological outcome in patients who underwent out-of-hospital resuscitation^[16]. In addition, a bioinformatics study on non-syndromic Tetralogy of Fallot, a congenital heart disease, identified PSMD14 as a potential hub gene involved in proteostasis dysregulation during right ventricular outflow tract remodeling, further suggesting its broader pathophysiological relevance beyond cancer^[52].

Multivariate Cox regression analyses have further indicated that PSMD14 serves as an independent prognostic factor for OS in pancreatic cancer^[50] (HR = 3.849, 95%CI 2.118-6.995, P = 0.001) and non-small cell lung cancer^[51] (HR = 1.679, 95%CI 1.086-2.596, P = 0.020). In addition, multiple studies have constructed multigene prognostic risk score models incorporating PSMD14 expression using multivariate Cox regression analyses. The predictive performance of these models varied considerably across cancer types, as summarized in Table 1^[39,53-58]. Notably, the glioma model achieved relatively high AUC values (0.818-0.861), whereas models for hepatocellular carcinoma and cervical cancer showed moderate predictive performance, with AUC values of 0.694-0.901 and 0.676-0.730, respectively. In contrast, the breast cancer risk model exhibited more limited predictive performance, with 1- to 3-year AUC values ranging from 0.61 to 0.64, indicating that the predictive value of PSMD14-containing multigene signatures varies across cancer types and would require further optimization before broad clinical application.

In addition to OS, the predictive value of PSMD14 for early postoperative recurrence has been systematically evaluated in hepatocellular carcinoma. Xiong et al. conducted tissue microarray–based immunohistochemistry in 312 HCC patients who underwent R0 resection and found that high PSMD14 expression in tumor tissues was an independent risk factor for early postoperative recurrence (≤ 24 months) (HR = 2.17, 95%CI 1.42-3.32, P = 0.0003). Furthermore, a nomogram integrating clinicopathological parameters showed that inclusion of PSMD14 expression increased the C-index of the model for predicting 1-, 3-, and 5-year recurrence-free survival (RFS) from 0.68 to 0.75 (P < 0.01), indicating that PSMD14 markedly improves the accuracy of postoperative recurrence prediction^[59].

In recent years, THL has emerged as one of the most extensively studied PSMD14 inhibitors. THL inhibits PSMD14/Rpn11 deubiquitinase activity by chelating Zn²⁺ within the JAMM domain^[60]. Currently classified as a tool compound at the preclinical stage, this mechanism has been validated in multiple solid tumor models, including breast cancer^[18], head and neck squamous cell carcinoma^[43], ESCC^[61], and anaplastic thyroid cancer^[20], in which THL significantly inhibits tumor growth. In ER-positive breast cancer cell lines (MCF-7, T47D), THL inhibits the deubiquitinase activity of PSMD14, directly downregulates ERα protein (by shortening its half-life) and its downstream target genes such as GREB1, TFF1, and IL20, and markedly suppresses cell proliferation^[18]. In head and neck squamous cell carcinoma cell lines (SCC15, UM1), THL inhibits PSMD14 deubiquitinase activity, accelerates E2F1 degradation, disrupts the E2F1–Akt–SOX2 axis, reduces cancer stemness, and significantly enhances cisplatin sensitivity^[43]. In ESCC cell lines (KYSE 30, KYSE 150), THL disrupts the interaction between PSMD14 and SNAIL, promotes SNAIL ubiquitination and degradation, upregulates E-cadherin, downregulates N-cadherin and vimentin, and effectively reverses the EMT. Functionally, THL not only suppresses the migration, invasion, and stemness of ESCC cells but also induces apoptosis and synergistically enhances chemosensitivity to cisplatin^[61]. In anaplastic thyroid carcinoma cell lines (8505C, CAL62), THL similarly targets and inhibits the deubiquitinase activity of PSMD14, promotes E2F1 degradation, and diminishes the transcriptional activation of the PI3K/AKT/mTOR and Ras/Raf/MEK/ERK pathways (as reflected by reduced p-AKT and p-ERK levels). As a result, THL inhibits cell proliferation and migration and induces apoptosis. Concurrently, THL upregulates E-cadherin and downregulates mesenchymal markers such as N-cadherin and vimentin, thereby blocking EMT and reducing the invasive and metastatic capacities of ATC cells^[20].

Research on THL has progressed beyond its activity as a natural product and its single-target mechanism, with substantial advances in structural optimization, dual-target design, and improvement in synthetic routes. For ESCC, He et al. used the natural product THL as a lead compound, retained its core scaffold responsible for PSMD14 inhibition, and introduced histone deacetylase (HDAC) inhibitory moieties at the N7 position to generate a series of dual-target inhibitors. Among these, compound 8b, which contains an eight-carbon linker and a 1,2-phenylenediamine group, exhibited the most balanced dual-target inhibitory activities (PSMD14 IC₅₀ = 238.7 ± 27 nM; HDAC1 IC₅₀ = 141.2 ± 10.3 nM). In the ESCC cell line (KYSE 30), dual targeting of PSMD14 and HDAC by 8b destabilizes the oncogenic transcription factor Snail, restores histone H3 acetylation, promotes chromatin condensation and gene silencing, and consequently reverses the EMT phenotype^[62]. Whether applied as monotherapy or in combination strategies, compound 8b, as the first dual PSMD14/HDAC inhibitor, outperforms single-target PSMD14 or HDAC inhibitors, highlighting the therapeutic advantage of dual-target inhibition. Currently in the early discovery stage, compound 8b requires further pharmacokinetic optimization and comprehensive preclinical evaluation. These findings not only expand the pharmacological potential of THL but also provide a new paradigm for multi-target anticancer drug design. Future work should focus on further structural optimization of THL-based compounds, rational target-combination strategies, and comprehensive preclinical evaluation.

Beyond THL, other PSMD14 inhibitors include 8-thioquinoline (8TQ) and its derivative Capzimin, and 1,10-phenanthroline (OPA). Like THL, these agents function as metal-chelating inhibitors that target the catalytic Zn²⁺ ion within the JAMM domain. As a tool compound, 8TQ exerts potent inhibitory activity against Rpn11 (IC₅₀ = 2.8 ± 0.4 μM) by chelating the catalytic Zn²⁺ ion of Rpn11 via its endocyclic nitrogen and exocyclic sulfur donor atoms in a bidentate fashion, thereby blocking deubiquitinating activity and preventing the removal of ubiquitin chains from substrate proteins^[63]. Through structure–activity relationship (SAR) optimization of 8TQ, Capzimin was developed as a derivative with approximately sevenfold higher potency against Rpn11 (IC₅₀ = 0.34 μM). Similar to 8TQ, Capzimin is a noncompetitive inhibitor of the 26S proteasome that directly targets and inhibits the Rpn11•Rpn8 complex. By broadly suppressing protein degradation in diverse cancer cell lines, Capzimin ultimately inhibits cell proliferation and induces apoptosis^[64]. Currently classified as a validated chemical probe for PSMD14/Rpn11 with moderate cell-based activity (B-level probe), Capzimin has demonstrated anti-tumor efficacy in preclinical models of multiple malignancies. Notably, in lung adenocarcinoma cell lines (H1299, PC9, A549, H1975), Capzimin exhibited potent anti-tumor effects with IC₅₀ values ranging from 3.4 to 8.3 μM, suppressing proliferation, migration, and invasion while inducing mitochondrial-mediated apoptosis. In vivo, Capzimin significantly suppressed tumor growth in xenograft mouse models, and its combination with the TGF-β inhibitor galunisertib demonstrated enhanced efficacy^[65]. In prostate cancer cell lines (PC3, DU145, 22RV1, and LNCaP), Capzimin targets POH1, suppresses proliferation, induces cell cycle arrest, and significantly enhances the efficacy of chemotherapy and androgen deprivation therapy, suggesting its potential as a novel targeted therapeutic, particularly for refractory prostate cancer^[66].

OPA specifically binds to and inhibits the metalloprotease activity of Rpn11 without affecting other deubiquitinases. In multiple myeloma, high PSMD14 expression is strongly associated with poor prognosis and resistance to the proteasome inhibitor bortezomib. Currently classified as a tool compound at the preclinical stage, OPA targets the deubiquitinase activity of the 19S proteasome, blocks proteasome function, activates caspase-dependent apoptosis and endoplasmic reticulum stress signaling, effectively overcoming bortezomib resistance, and potentiates the synergistic antitumor effects of immunomodulatory drugs (such as lenalidomide and pomalidomide) and dexamethasone^[47]. Beyond hematological malignancies, OPA has also demonstrated efficacy in solid tumors: in glioblastoma, OPA treatment restores immune cell infiltration and enhances CAR-NK cell cytotoxicity by disrupting the PSMD14-BCKDK-BCAA metabolic axis^[42]. Recent advances in nanomedicine have expanded the therapeutic potential of OPA. An acid-responsive cobalt-doped ZIF-8 nanoplatform encapsulating OPA (CPP) has been developed to enable tumor-selective delivery and pH-triggered release. Upon exposure to the acidic tumor microenvironment, CPP releases both Co²⁺, which catalyzes Fenton-like reactions to generate reactive oxygen species (ROS), and OPA, which inhibits PSMD14 and promotes SLC7A11 ubiquitination and degradation, thereby depleting glutathione and sensitizing tumor cells to ferroptosis. In osteosarcoma xenograft models, CPP administration significantly suppressed tumor growth, demonstrating the therapeutic potential of combining PSMD14 inhibition with nanocarrier-based drug delivery^[67].

In addition to these metal-chelating inhibitors, researchers have also explored covalent inhibitors that irreversibly target PSMD14. Epidithiodiketopiperazines (ETPs) inhibit PSMD14 and thereby compromise proteasome function. However, ETPs can also inhibit other JAMM family proteins, including CSN5 and AMSH. Currently in the early discovery stage, SOP11 is a modified ETP with reduced nonspecific effects that stabilizes a subset of proteasome substrates, induces an unfolded protein response, and ultimately leads to tumor cell death^[68]. More recently, Eupalinolide B (EB), a natural sesquiterpene lactone, has been identified as another covalent PSMD14 inhibitor. EB covalently binds to PSMD14 at His183 via its reactive α-methylene-γ-lactone moiety, thereby inhibiting its deubiquitinase activity. In acute promyelocytic leukemia cell lines (HL-60), EB promotes AKT1 and CDK4 degradation, leading to G2/M cell cycle arrest and apoptosis. Currently in the early discovery stage, EB represents a novel covalent PSMD14 inhibitor and a promising lead compound for further development^[69].

Collectively, these inhibitor studies underscore that the therapeutic efficacy of PSMD14 blockade is fundamentally dictated by its context-dependent substrate networks. Table 2 provides a comprehensive catalog of the characterized PSMD14 substrates, their deubiquitination types, downstream biological effects, and associated drug resistance profiles across different malignancies, offering a molecular rationale for the inhibitor strategies discussed above.

As a core member of the JAMM family of deubiquitinases, PSMD14 has emerged as a focal point in cancer biology and preclinical research in recent years. This review systematically summarizes the multifaceted regulatory roles of PSMD14 in cell cycle progression, programmed cell death, invasion and metastasis, mitochondrial function, metabolic reprogramming, the immune microenvironment, and therapeutic resistance, while also evaluating its clinical value as a prognostic biomarker and the current landscape of inhibitor development. Taken together, the evidence indicates that PSMD14-targeted therapy has considerable promise in preclinical studies, although several key challenges remain.

For clinical application, PSMD14-targeted intervention faces four major obstacles: inhibitor selectivity and off-target toxicity, suboptimal pharmacokinetic properties, a lack of prospectively validated biomarkers, and the absence of established optimal combination regimens. Currently available inhibitors, including THL^[60], Capzimin^[64], and OPA^[47], chelate the catalytic zinc ion within the JAMM domain. However, because this zinc-binding site is highly conserved among other JAMM family members, the risk of off-target effects is difficult to eliminate completely. In addition, ETPs, although they inhibit PSMD14 through a covalent mechanism, can also target other JAMM family proteins, including CSN5 and AMSH^[68]. Moreover, most of these compounds exhibit limited bioavailability; only OPA has recently shown preliminary improvement via a nano-delivery system^[67]. More critically, while PSMD14 expression levels have been correlated with responsiveness to various chemotherapeutic agents, its utility as a predictive biomarker for guiding clinical treatment decisions remains unvalidated by prospective clinical trials, and clinical evidence for optimal combination strategies is largely lacking.

Beyond these challenges, the PSMD14-driven resistance mechanisms in ESCC^[61], head and neck squamous cell carcinoma^[43], glioblastoma^[45], breast cancer^[18], and multiple myeloma^[47,48] have been relatively well characterized, and pharmacological inhibitors have demonstrated the ability to reverse these resistance phenotypes. These findings suggest that these cancer types represent promising candidate indications for prioritized exploration of PSMD14-targeted therapy.

Based on current evidence, Capzimin appears to be the most promising inhibitor for further preclinical development, with an IC₅₀ of 0.34 μM and validated in vivo efficacy in solid tumor models including non-small cell lung cancer^[65] and prostate cancer^[66]. OPA has generated the most robust data for reversing bortezomib resistance in multiple myeloma^[47,48], and the recent development of a nano-delivery platform has opened new avenues for its therapeutic application^[67]. THL has demonstrated mechanistic validity and combination potential across multiple solid tumor types^{[18,20,43,45,61]}. However, as a natural product, its poor selectivity and moderate potency represent inherent limitations.

In terms of drug discovery strategy, dual-target inhibitor design warrants prioritization. Compound 8b, a PSMD14/HDAC dual inhibitor developed from the THL scaffold, has exhibited superior antitumor activity compared with single-target agents in ESCC^[62], suggesting that simultaneous targeting of PSMD14 and epigenetic regulators may offer unique advantages in aggressive solid tumors. Furthermore, based on the role of PSMD14 in stabilizing PD-L1 through deubiquitination, a recent study demonstrated that genetic knockdown of PSMD14 synergizes with anti-PD-1 therapy to exert antitumor effects in breast cancer models^[41]. Given that pharmacological inhibition can phenocopy the effects of genetic knockdown on PD-L1 stability, we propose the working hypothesis that PSMD14 inhibitors may similarly synergize with anti-PD-1/PD-L1 immune checkpoint blockade, particularly in tumors with high PSMD14 expression levels. This hypothesis warrants systematic evaluation in immunocompetent preclinical models, with priority given to tumors characterized by high PSMD14 expression and pre-existing T cell infiltration. Concurrently, the future clinical application of PSMD14-targeted therapies, if successful, would likely require biomarker-driven patient stratification strategies. We suggest that tumor PSMD14 expression levels be considered a primary stratification criterion for patient enrollment, with PD-L1 status or specific metabolic dependency features further guiding combination regimen selection. However, based on current evidence, the appropriateness of such exclusion criteria requires prospective validation in future clinical studies.

The pathological relevance of PSMD14 extends beyond oncology, as emerging evidence implicates its involvement in various non-tumor diseases. However, research in these areas remains in its early stages. Until substantial breakthroughs are achieved in the oncological setting, exploration of non-tumor indications remains a complementary, albeit worthwhile, direction for future investigation.

PSMD14, as a JAMM-family deubiquitinase, exerts multifaceted regulatory functions in tumorigenesis, progression, metastasis, drug resistance, and immune evasion by stabilizing diverse substrate proteins and has thus emerged as an attractive preclinical therapeutic target and prognostic biomarker. Although existing inhibitors such as THL, Capzimin, and OPA have shown promise in preclinical models, critical issues including selectivity, pharmacokinetic properties, and biomarker validation remain unresolved. Future research should prioritize the advancement of proof-of-concept clinical trials in cancer types such as ESCC, head and neck squamous cell carcinoma, glioblastoma, breast cancer, and multiple myeloma - in which inhibitor-mediated reversal of resistance has already been documented - and systematically explore combination strategies integrating PSMD14 inhibitors with immune checkpoint blockade. The future clinical applicability of PSMD14-targeted therapy ultimately depends on whether these prioritized directions can be validated through biomarker-guided clinical trials.