Cuproptosis begins with the disruption of intracellular copper homeostasis (36). Dietary Cu²⁺ enters cells primarily through the high-affinity copper transporter SLC31A1 (CTR1) (37). Upregulation of SLC31A1 in hepatocellular carcinoma has been linked to oncogenic signaling pathways, suggesting a potential vulnerability of HCC cells to cuproptosis (38). Copper ionophores (e.g., Elesclomol and disulfiram) facilitate efficient copper delivery by forming complexes with copper ions and promoting their specific mitochondrial accumulation (39).Within the mitochondrial matrix, ferredoxin-1 (FDX1) serves as a central player. FDX1 acts not only as the principal reductase converting Cu²⁺ to the more toxic Cu⁺ but also as a key upstream regulator of the entire cuproptosis pathway (40). Genome-wide CRISPR screens confirm that FDX1 loss confers robust resistance to various copper ionophores, highlighting its essential role (41).

The reduced Cu⁺ does not indiscriminately damage cellular components but displays high specificity, primarily targeting key lipoylated enzymes within the tricarboxylic acid (TCA) cycle (42). This precision arises from the exceptionally high affinity of Cu⁺ for protein lipoyl groups, with dissociation constants on the order of 10^-17M (41). Among these, the E2 component of the pyruvate dehydrogenase complex—dihydropyrrolidine thioamide acyltransferase (DLAT) (35)—has been identified as a core target for Cu⁺ action (43). Notably, FDX1 plays a dual central role in this process: it serves not only as the key reductase catalyzing the reduction of Cu²⁺ to highly reactive Cu⁺, but also as a positive regulator of protein lipoylation (15). Studies indicate that FDX1 promotes the lipoylation modification of mitochondrial proteins such as DLAT by directly interacting with lipoic acid synthase (LIAS) (39). Consistently, FDX1 knockout ablates cellular protein lipoylation and leads to significant accumulation of TCA cycle intermediates, including pyruvate and α-ketoglutarate (25).These findings position FDX1 as the central node connecting copper toxicity to mitochondrial metabolic enzyme dysfunction.

The binding of Cu⁺ to lipoylated DLAT induces abnormal, disulfide-dependent oligomerization and insoluble aggregation of the protein, rather than simply inhibiting its activity. These aggregates appear as distinct protein foci under electron microscopy, marking the physical collapse of the central TCA cycle (15).This process coincides with the destabilization and substantial loss of mitochondrial iron-sulfur (Fe-S) cluster proteins. As essential cofactors for enzymes involved in oxidative phosphorylation and DNA repair, the loss of Fe-S clusters severely exacerbates mitochondrial dysfunction (38). Ultimately, the coalescence of lipoylated protein aggregates and Fe-S cluster degradation triggers intense mitochondrial proteotoxic stress. This is marked by the induction of heat shock proteins (e.g., HSP70), but the cell fails to recover and undergoes death (42).A key feature of cuproptosis is its strong dependence on active mitochondrial respiration. Consequently, cells reliant on glycolysis or located in hypoxic niches exhibit significant resistance, providing a rationale for selectively targeting tumor cells with robust mitochondrial respiration and TCA cycle turnover.

Inducing cuproptosis may reverse tumor-induced immunosuppression, potentially converting immunologically “cold” tumors into “hot” tumors with enhanced immune cell infiltration, thereby improving responses to immunotherapy (42).Preclinical studies demonstrate that cuproptosis induced by Elesclomol-Cu or Disulfiram-Cu effectively activates the cGAS-STING innate immune pathway in dendritic cells (DCs). This activation promotes DC maturation and antigen presentation capacity and stimulates the secretion of proinflammatory factors like type I interferons, collectively initiating robust tumor antigen-specific CD8⁺ T cell responses. This process lays a critical immunological foundation for the efficacy of immune checkpoint inhibitors (38).

Cuproptosis directly disrupts tumor cell metabolic reprogramming by targeting the TCA cycle, thereby altering the TME’s metabolite composition (26).For example, TCA cycle inhibition reduces tumor cell lactate production. As a key immunosuppressive metabolite in the TME, lactate directly impairs the function and proliferation of cytotoxic T cells and NK cells while promoting the activation and recruitment of immunosuppressive cells, such as M2 macrophages and regulatory T cells (Tregs) (15). Therefore, by reducing lactate levels in the TME, cuproptosis-induced metabolic reprogramming can indirectly alleviate the suppression of effector immune cells, helping to reverse the overall immunosuppressive state.

A close but complex relationship exists between copper metabolism and immune checkpoint molecule expression. Substantial evidence indicates that elevated intracellular copper levels upregulate PD-L1 expression, influencing both its transcription and protein stability, potentially through copper-mediated activation of STAT3 and EGFR signaling pathways (45). This suggests that cuproptosis induction as a monotherapy could paradoxically increase PD-L1 expression in surviving tumor cells, potentially enhancing their immune evasion and fostering adaptive resistance (45).

However, this elevated PD-L1 expression simultaneously provides an increased density of therapeutic targets, potentially rendering surviving tumor cells more susceptible to anti-PD-L1 monoclonal antibodies (mAbs). This dynamic interplay underscores the synergistic potential of combining cuproptosis induction with immune checkpoint blockade to effectively eliminate drug-tolerant persister cells (45, 46). Conversely, this interplay supports combination therapy: using copper chelators (e.g., tetrasulfomolybdate) or modulating cuproptosis can promote PD-L1 ubiquitination and degradation, thereby enhancing the infiltration of CD8⁺ T cells and NK cells into tumors (45). Therefore, precise modulation of cuproptosis and its associated pathways represents a promising strategy to concurrently eliminate tumor cells and reduce immune checkpoint barriers, achieving synergistic antitumor effects.

The mechanism is illustrated in three sequential phases (1): Initiation - Copper Uptake (Left): Copper ionophores, represented by purple transmembrane channels (e.g., elesclomol (ES) and disulfiram (DSF)), facilitate the influx of extracellular Cu²⁺ (blue spheres) into the cytoplasm and subsequently into the mitochondria (2). Execution - Mitochondrial Toxicity (Center): Within the mitochondrion, FDX1 reduces Cu²⁺ to Cu⁺ (orange spheres). The highly reactive Cu⁺ selectively binds to lipoylated TCA cycle enzymes (depicted as green spheres with lipid tails, e.g., DLAT). This binding triggers the formation of insoluble protein aggregates (shown as green tangled structures) and the loss of Fe-S clusters (yellow squares), leading to proteotoxic stress and the physical collapse of the TCA cycle (3). Consequence - Immune Modulation (Right): The rupture of the dying tumor cell releases DAMPs (red spiky shapes), which are engulfed by Dendritic Cells (DCs) to activate the cGAS-STING pathway, promoting Type I IFN secretion and CD8⁺ T cell infiltration (Path A). Concurrently, the collapse of the TCA cycle results in reduced lactate secretion (↓ Lactate), thereby attenuating lactate-driven M2 polarization/inhibiting M2 polarization (Path B).

These oncogenic CRLs are characteristically upregulated in HCC and are intimately linked to poor patient prognosis, the establishment of an immunosuppressive microenvironment, and resistance to therapy. Their primary oncogenic mechanisms include:

Suppression of Cuproptosis Sensitivity: Certain CRLs function as negative regulators of key cuproptosis genes via transcriptional or post-transcriptional mechanisms, thereby dampening the susceptibility of tumor cells to copper-induced cell death. For instance, MKLN1-AS and KDM4A-AS1—key components of the prognostic model CRDELSig—are significantly upregulated in HCC. Knockdown of these molecules has been shown to restore FDX1 expression and inhibit cell proliferation, suggesting their roles as negative regulators of the cuproptosis machinery (52) (54).

Activation of Oncogenic Signaling via ceRNA Mechanisms: A subset of CRLs functions as competitive endogenous RNAs (ceRNAs), specifically sequestering miRNAs to relieve the repression of downstream oncogenes. Specifically, MKLN1-AS competitively binds to miR-654-3p, thereby upregulating the expression of hepatoma-derived growth factor (HDGF) to drive the malignant evolution of HCC (54).

Remodeling the Immune Microenvironment Landscape: Multiple studies have confirmed that high-risk scores based on CRL signatures are positively correlated with the infiltration abundance of immunosuppressive populations, such as M2 macrophages and regulatory T cells (Tregs), within the tumor microenvironment. Concurrently, these scores are associated with the downregulation of CD8⁺ T cell and natural killer (NK) cell activity (53, 57). In particular, the overexpression of AL133243.2 exhibits a significant negative correlation with functional markers of NK cells, indicating its involvement in the functional suppression of effector immune cells (53).

In contrast to the aforementioned resistance-driving oncogenic molecules, a subset of CRLs is significantly downregulated in HCC, and their elevated expression correlates positively with favorable prognosis and therapeutic sensitivity. This observation highlights the marked heterogeneity of CRLs in modulating the direction of cuproptosis.

LINC02362 exemplifies this category of molecules. Research confirms that LINC02362 is significantly downregulated in HCC tissues, with its low expression linked to adverse clinicopathological characteristics. Mechanistically, LINC02362 functions as a competitive endogenous RNA (ceRNA) for miR-18a-5p, thereby relieving the post-transcriptional repression of the core cuproptosis gene FDX1. The overexpression of LINC02362 restores FDX1 levels, thereby sensitizing tumor cells to cuproptosis induced by copper ionophores (e.g., Elesclomol) and synergistically enhancing the antitumor efficacy of oxaliplatin (51). Crucially, this discovery demonstrates that specific CRLs can function as “accelerators” rather than “brakes” of the cuproptosis program, providing direct evidence for developing CRL-based sensitization strategies.

A critical, yet previously underexplored paradigm regarding CRL biology is the extent to which these molecules uniformly drive immunotherapy resistance. Accumulating evidence challenges the monolithic view of CRLs as solely immunosuppressive factors, revealing instead a functional dichotomy: while the majority of identified CRLs in HCC indeed facilitate immune evasion, a distinct subset possesses the capacity to remodel the tumor immune microenvironment (TIME) toward an immune-permissive state.

Achieving a profound understanding of CRL functional heterogeneity faces substantial challenges, primarily stemming from the inherent complexity of tumor biology:

Intratumoral Heterogeneity: Within a single HCC tumor, distinct spatial regions or cellular subpopulations (e.g., tumor core vs. invasive margin, cancer stem cells vs. differentiated cells) may exhibit vastly different CRL expression profiles. The functions of these CRLs may differ fundamentally across malignant cells, immune cells (e.g., tumor-associated macrophages), and stromal cells (60).

Dynamic Evolution under Therapeutic Pressure: The expression profiles and functions of CRLs may undergo adaptive alterations under the selective pressure of ICI therapy. For instance, a CRL initially exerting an immune-sensitizing effect may shift to support resistance during treatment due to functional or regulatory network rewiring (7).

Network Superposition and Counteraction: A single CRL may simultaneously participate in multiple, potentially opposing signaling pathways. Its ultimate net biological effect is the result of the interplay between these pathways and is highly contingent upon the specific intracellular and extracellular context.

Summary

The mechanisms by which CRLs drive immunotherapy resistance in HCC constitute a stereoscopic network encompassing the regulation of intrinsic cell death sensitivity, long-distance intercellular communication, and the systemic modulation of immune checkpoint networks. However, the simplistic categorization of CRLs as mere “resistance drivers” no longer captures their full biological complexity. The field is currently at a pivotal transition point, moving from “correlative description” to “functional classification.” Future core mandates include leveraging cutting-edge technologies—such as single-cell multi-omics, spatial transcriptomics, CRISPR screens, and conditional cell-specific knockout models—to precisely elucidate the function of individual CRLs within specific cell types and microenvironmental contexts (41, 60). Establishing a causal logical chain linking “regulatory direction (pro-/anti-cuproptosis) — immune phenotype (pro-/anti-tumor immunity) — therapeutic response” is paramount. Only by achieving this precise functional decoding can we realize a paradigm shift from “broadly targeting CRLs” to “precisely targeting specific functional CRL subsets, “ thereby designing truly effective and personalized combination strategies to overcome HCC immunotherapy resistance.

Homogeneity in Immune Microenvironment Prediction: In various solid tumors, including HCC, NSCLC, and gastric cancer, multi-gene signatures based on CRLs robustly stratify patient prognosis. More importantly, high-risk scores consistently map to immunosuppressive microenvironmental features, such as M2 macrophage polarization, elevated Treg infiltration, and functional exhaustion of CD8⁺ T cells (63). This indicates that the function of CRLs as biomarkers indicating an “immune-cold” status possesses pan-cancer applicability.

The Central Regulatory Role of ceRNA Networks: A shared mechanistic principle across these cancers is the competitive endogenous RNA (ceRNA) network. Whether via the MKLN1-AS/miR-654-3p/HDGF axis in HCC (54) or the LINC01128/miR-576-5p/CDKN3 axis in NSCLC (61), CRLs consistently exert oncogenic functions by sequestering specific miRNAs to disrupt post-transcriptional regulatory networks. This suggests that while the specific lncRNAs differ by tissue, the “ceRNA-mediated immune evasion” represents a universal strategy utilized by CRLs to drive tumor progression.

Despite the functional convergence on immune suppression, distinct cancer types possess specific “core” CRLs. For example, AL133243.2 is a focal point of immune microenvironment regulation in HCC (52) but is rarely reported in other cancers; conversely, USP2-AS1, which is critical in breast cancer, is not a primary driver in HCC (62). This reflects the spatiotemporal specificity of lncRNA expression and the differential weighting of their roles across distinct tumor initiation and progression trajectories.

The comparative analysis reveals that CRL research relies on specific tumor histological contexts. Future studies should focus on the following strategic directions:

Systematically Charting a Pan-Cancer Regulatory Atlas of CRLs: Utilize multi-omics big data to laterally compare the expression regulatory networks and downstream signaling pathways across different cancer types, distinguishing between “housekeeping CRLs” and “tissue-specific CRLs.”

Exploring Personalized Therapeutic Strategies: For CRLs with high tissue specificity, develop precise companion diagnostic tools to realize personalized interventions based on molecular subtyping. By targeting the unique “core” CRLs of each cancer type (e.g., AL133243.2 for HCC), therapies can achieve maximum efficacy with minimal off-target effects.

Antisense Oligonucleotides (ASOs) and Chemical Modification Technologies: ASOs specifically silence target RNA via RNase H-mediated degradation mechanisms. For HCC, the utilization of N-acetylgalactosamine (GalNAc) conjugation technology is key to breaking through delivery bottlenecks. GalNAc ligands can bind with high affinity to the asialoglycoprotein receptor (ASGPR) on the surface of hepatocytes, enabling the liver-specific uptake of ASOs. This not only drastically improves gene silencing efficiency but also effectively avoids off-target toxicity in the kidney and nervous system (66).

Small Molecule Inhibitors Interfering with Protein-Protein Interactions (PPI): Many CRLs function by recruiting specific proteins (e.g., FDX1). Developing small molecule compounds capable of occupying specific binding pockets in the secondary structure of lncRNAs to block the lncRNA-protein interaction interface represents a highly promising non-nucleic acid drug strategy (67).

CRISPR/Cas13 Editing Technology: Although still in the preclinical stage, RNA editing systems based on CRISPR/Cas13 offer higher specificity and efficiency than RNAi for targeting nuclear lncRNAs. Utilizing the Cas13d effector protein to specifically degrade oncogenic CRLs may become a disruptive technology for conquering “undruggable” lncRNA targets in the future (68).

“Prime and Kill” Strategy: Utilizing copper ionophores (e.g., Elesclomol) as “primers.” They not only induce cuproptosis in tumor cells but, more critically, trigger immunogenic cell death (ICD), releasing danger signals such as ATP and HMGB1 (38, 41).This immunogenic burst converts the originally immunosuppressed “cold tumor” into an inflamed “hot tumor, “ thereby significantly elevating the response rate of subsequent PD-1/PD-L1 inhibitors (“killers”) (15, 30).

Copper Chelators for Improving Hypoxia and Vascular Normalization: For HCC subtypes with high copper burden and active angiogenesis, the use of copper chelators (e.g., tetrathiomolybdate, TTM) can reduce bioavailable copper levels. The dual benefit of this strategy lies in: on one hand, reducing extracellular matrix cross-linking by inhibiting LOX enzyme activity to soften the tumor stroma; and on the other hand, inhibiting the VEGF pathway to promote vascular normalization, thereby breaking physical barriers and facilitating the deep infiltration of effector T cells into the tumor core (45, 69).

Balance between Stimuli-Responsiveness and Safety: Designing intelligent nanocarriers sensitive to tumor microenvironmental characteristics (acidic pH, high ROS, high GSH) is key to reducing systemic copper toxicity (Table 4). For example, ROS-responsive linkers ensure that copper ionophores are released only within tumor cells with high oxidative stress levels, while remaining stable in normal liver tissue (25, 46).

Dual Immune-Metabolic Reprogramming: Next-generation nanomedicines should pursue a “killing two birds with one stone” effect. For example, designing nanoplatforms that co-deliver cuproptosis inducers and STING pathway agonists. This design leverages mitochondrial DNA leakage generated by cuproptosis to synergistically activate the cGAS-STING pathway, triggering a potent Type I interferon response, thereby achieving synchronous activation of innate and adaptive immunity (70).

On one hand, cuproptosis functions as a potent immunogenic driver. Unlike apoptosis, which is often immunologically silent, cuproptosis triggered by copper ionophores (e.g., elesclomol) induces profound mitochondrial proteotoxic stress and membrane rupture (14, 41). This catastrophe triggers the release of Damage-Associated Molecular Patterns (DAMPs), including ATP, HMGB1, and mitochondrial DNA (mtDNA) (38). These danger signals are sensed by dendritic cells (DCs) via the cGAS-STING pathway or Toll-like receptors, facilitating DC maturation and the cross-priming of cytotoxic CD8+ T cells (17, 38). In this context, inducing cuproptosis effectively converts the “cold” immunosuppressive tumor microenvironment (TME) of HCC into an inflamed, “hot” state, overcoming the antigen-presentation defects typical of liver cancer.

Conversely, a growing body of evidence suggests that copper metabolism intrinsically supports immune evasion mechanisms. A seminal study by Voli et al. demonstrated that intratumoral copper levels act as a metabolic rheostat for PD-L1 expression (71). Mechanistically, copper ions bind to the CTR1 transporter and facilitate the phosphorylation of EGFR and STAT3, which transcriptionally upregulate PD-L1; simultaneously, copper promotes the ubiquitination inhibition of PD-L1, enhancing its protein stability. Consequently, sub-lethal copper accumulation—or the copper-rich environment characteristic of HCC—may paradoxically protect tumor cells from T cell killing by fortifying the PD-1/PD-L1 inhibitory axis (45). Furthermore, copper acts as a cofactor for enzymes that facilitate angiogenesis (e.g., lysyl oxidase), potentially fostering a hypoxic barrier that excludes T cell infiltration (14, 44).

This dichotomy presents a strategic dilemma: while acute, lethal doses of copper induce immunogenic death, chronic, sub-lethal copper elevation may drive adaptive immune resistance. We propose a “Threshold Hypothesis”: the immunological outcome of copper targeting depends critically on crossing a lethality threshold that overwhelms the tumor’s adaptive PD-L1 upregulation. In the specific context of HCC, where the liver is the central organ for copper homeostasis and tumors often exhibit baseline copper overload, the risk of copper-induced immune suppression is particularly high. Therefore, relying solely on cuproptosis induction is likely insufficient and may even select for resistant clones with high PD-L1 expression. This mechanistic insight provides a compelling rationale for the obligatory combination of cuproptosis inducers with immune checkpoint inhibitors (ICIs). By using ICIs to block the “brake” (PD-L1) reinforced by copper, and copper ionophores to press the “accelerator” (ICD), clinicians may achieve a synergistic therapeutic efficacy that neither monotherapy can provide (15, 16).

Effective induction of cuproptosis relies heavily on intact mitochondrial respiratory function. This dependency, however, faces a major obstacle in the pronounced intra- and inter-tumor heterogeneity characteristic of HCC (60).HCC cells of different molecular subtypes exhibit markedly divergent sensitivities to cuproptosis inducers. For instance, ARID1A-deficient HCC cells, constrained by glycolysis, depend more heavily on mitochondrial respiration and thus show heightened sensitivity to copper ionophores. Conversely, cells with TCA cycle dysfunction or electron transport chain impairments readily develop resistance (14). This metabolic heterogeneity underscores the need for future research to leverage single-cell multi-omics technologies. These approaches can precisely map CRLs and cuproptosis pathway-specific regulatory networks at single-cell resolution across tumor, immune, and stromal cells. Building on this foundation, establishing a precision classification system based on molecular profiles and mitochondrial functional states is a prerequisite for implementing effective targeted therapies.

A central concern for clinical translation is the risk of off-target toxicity caused by systemic perturbation of copper homeostasis. Since copper is an essential trace element, significant fluctuations in systemic levels can induce severe adverse effects, including hepatotoxicity, nephrotoxicity, and neurological impairment (15).To overcome this bottleneck, emerging nanodelivery systems offer distinct advantages. For instance, platelet-mimetic PTC systems achieve tumor enrichment through natural targeting, while macrophage-coated SonoCu systems utilize integrin-mediated active targeting (46).A promising future direction involves developing next-generation smart nanocarriers that respond to specific tumor microenvironment signals (e.g., pH, glutathione levels, or enzyme activity). Such systems would enable precise, controlled drug release, enhancing therapeutic efficacy while minimizing systemic toxicity.

A significant gap exists in the availability of validated clinical biomarkers for predicting tumor susceptibility to cuproptosis (72). Although prognostic models based on cuproptosis-related genes and CRLs show promise in retrospective analyses, their generalizability across diverse populations and HCC subtypes requires large-scale validation (15). Progress in this area hinges on the systematic validation and optimization of these predictive models using large prospective cohorts, followed by their translation into clinically applicable decision-making tools. Concurrently, non-invasive dynamic monitoring technologies should be actively developed (73), Examples include employing ⁶⁴Cu-PET imaging to visualize in vivo copper distribution (74, 75)and developing liquid biopsy assays for detecting extracellular CRLs in blood (64), These approaches could provide objective data for real-time treatment adjustment.

It must be acknowledged that most cuproptosis-related research remains confined to the preclinical stage, with no clinical trials conducted for HCC to date. Early clinical trials in other cancer types have yielded inconsistent results, underscoring the challenges in translation. For instance, a Phase III trial of Elesclomol combined with paclitaxel in melanoma was prematurely terminated due to a lack of survival benefit and increased toxicity; similarly, trials in relapsed acute myeloid leukemia failed to demonstrate significant efficacy (15).These setbacks highlight deficiencies in drug delivery, pharmacokinetics, and, crucially, patient selection strategies. Consequently, future clinical translation efforts must be strategically designed. Initial Phase I/II trials should be conducted in biomarker-enriched populations, prioritizing rational combination regimens. Given the close interaction between cuproptosis and the immune microenvironment, a triple therapy approach combining cuproptosis inducers, immune checkpoint inhibitors, and anti-angiogenic drugs holds particular synergistic potential (69). Additionally, exploring optimization strategies like pulsed or sequential dosing is crucial for preventing or overcoming adaptive resistance (76).