The molecular classification of breast cancer is determined by the status of hormone receptors. The prognostic outcomes for each subtype exhibit substantial variability [36], making it imperative to establish differentiated follow-up monitoring strategies on the basis of molecular typing, particularly for HER2-positive and triple-negative subtypes, which require enhanced posttreatment imaging surveillance and circulating tumour DNA detection. Future research should concentrate on analysing intrasubtype molecular heterogeneity and developing novel targeted therapies [37], especially therapeutic breakthroughs in triple-negative breast cancer [38]. Table 1 summarises the regulatory mechanisms of noncoding RNAs in the PCD of breast cancer.
PANoptosis represents an integrated form of cell death encompassing apoptosis, pyroptosis, and necrosis [39], which is orchestrated by the PANoptosome complex. To date, four primary types of PANoptosomes have been identified: the ZBP1-PANoptosome (comprising ZBP1, NLRP3, ASC, caspase-1, caspase-6, caspase-8, RIPK1 and RIPK3) [39], AIM2-PANoptosome (comprising AIM2, Pyrin, ZBP1, ASC, caspase-1, caspase-8, FADD, RIPK1 and RIPK3) [40], RIPK1-PANoptosome (comprising RIPK1, RIPK3, NLRP3, ASC, caspase-1 and caspase-8) [41] and NLRP12-PANoptosome (comprising NLRP12, ASC, caspase-8 and RIPK3) [42]. These complexes facilitate the activation of caspase-3/7, cleavage of GSDMD and GSDME, and phosphorylation of MLKL, leading to membrane pore formation and the progression of PANoptosis [39]. Given the high heterogeneity of breast cancer, continuous refinement of existing molecular typing systems and prognostic assessment tools is needed. A robust prognostic prediction model, validated across multiple dimensions, has been developed [43]. This model can effectively guide the prediction of chemotherapy sensitivity and optimisation of targeted-immune combination therapies, thereby providing molecular tools to advance precise medicine paradigms in breast cancer management [44]. PANoptosis cannot be suppressed by pyroptosis, apoptosis or necroptosis. The formation and activation of the PANoptosome within a single cell also support PANoptosis. The molecular and regulatory mechanisms of this process are shown in Fig. 2.
As the central regulatory hub of apoptosis, mitochondria mediate signal transduction through both intrinsic and extrinsic pathways. The extrinsic pathway is initiated by death receptor-mediated or granzyme-dependent mechanisms [45]. Specifically, the extrinsic pathway begins with the binding of death receptors, members of the TNFR superfamily such as Fas and DR4/5, to their respective ligands, such as FasL and TRAIL. This binding subsequently recruits and activates initiator caspases-8/10 via adaptor proteins such as FADD/TRADD, leading to downstream activation of the effector caspases-3/7 and culminating in apoptosis [43]. The intrinsic pathway involves mitochondria-related and endoplasmic reticulum stress pathways, where the Bcl-2 protein family plays a pivotal role by modulating mitochondrial outer membrane permeabilisation (MOMP) [46]. This family includes antiapoptotic members (e.g., Bcl-2 and Bcl-xL) and proapoptotic members (e.g., Bax and Bak). Upon activation, proapoptotic proteins form pore complexes on the mitochondrial membrane, resulting in the loss of the mitochondrial membrane potential and the release of apoptotic factors such as cytochrome C, thereby activating the caspase-9/3 cascade [46]. Dysregulated apoptosis signalling pathways in breast cancer cells can be reprogrammed to re-enter the apoptotic cycle, representing a critical therapeutic strategy. In breast cancer, overexpression of the antiapoptotic protein MCL1 impedes apoptosis by inhibiting mitochondrial NOX4 function. However, BH3 mimetics targeting MCL1 have demonstrated promising therapeutic potential [47]. Moreover, p53, a key tumour suppressor, exerts dual regulatory effects on ROS-induced apoptosis. Wild-type p53 promotes apoptosis through the transcriptional activation of proapoptotic genes such as Bax and PUMA [48]. In triple-negative breast cancer (TNBC), mutant p53 induces nonclassical apoptosis via aberrant interaction with MDM2, independent of wild-type p53 [49]. Compound G613 promotes the apoptosis of MCF-7 cells by inhibiting the formation of the p53-MDM2 complex and upregulating p53 expression [50]. Additionally, increasing the bioavailability of bile acid promotes the apoptosis of breast cancer cells by activating the caspase-8/Bid/ROS pathway, suggesting a novel therapeutic target for hormone receptor-negative breast cancer [51].
Apoptotic pathways interactively regulate other cell death modalities. In ferroptosis-apoptosis cross-regulation, long noncoding RNA LINC00618 regulates cell death through a dual mechanism: downregulating SLC7A11 to inhibit ferroptosis and upregulating BAX expression to promote apoptosis [52]. In autophagy-dependent apoptosis, high ROS levels in tumour cells enhance the sensitivity to apoptosis by inducing mitochondrial hyperpolarisation, curcumin or DNA damage [53].
Necroptosis is a caspase-independent form of cell death that significantly regulates the initiation and progression of breast cancer [54]. Necroptosis is a precisely regulated form of programmed necrosis, and its core mechanism hinges on the assembly of the RIPK1/RIPK3 kinase complex and subsequent phosphorylation-oligomerisation of MLKL. In triple-negative breast cancer, RIPK1 facilitates vascular mimicry by activating the AKT/eIF4E signalling pathway [55]. Moreover, RIPK3 expression exhibits spatiotemporal heterogeneity [56]. Owing to hypermethylation of the promoter, RIPK3 is silenced in primary tumours but significantly upregulated in recurrent lesions, with an enhanced dependence on cysteine metabolism, suggesting that targeting RIPK3 may inhibit tumour recurrence [56].
Notably, the DNA damage response protein MRE11 activates the cGAS‒STING pathway, thereby driving necroptosis mediated by the ZBP1‒RIPK3‒MLKL axis [57]. Low ZBP1 expression in TNBC is significantly associated with genomic instability, an immunosuppressive microenvironment, and a dismal prognosis, indicating its potential as a biomarker for functional defects in this pathway [58]. Necroptosis plays a dual role in tumour immune regulation, as it can activate antigen-specific immune responses by releasing damage-associated molecular patterns, while RIPK3/MLKL-mediated IL-1α release can suppress T-cell function and promote tumour immune escape [59]. In light of this dichotomy, novel targeted nanocomplexes have been developed to induce immunogenic cell death by specifically increasing RIPK3 phosphorylation and MLKL tetramerisation, thereby enhancing the cytotoxic effect of T cells on triple-negative breast cancer [60]. This paradoxical immune regulation underscores the need for precise modulation of necroptosis according to the treatment stage: leveraging its immune-stimulatory properties in the early stages while inhibiting its metastasis-promoting effects in the later stages [61]. For example, lung metastasis can be significantly reduced by inhibiting MLKL [62]. Epigenetic regulation also plays a crucial role in necroptosis. Z-DNA binding protein 1 has emerged as a potential therapeutic target for modulating the necrosis process in advanced tumours [63]. Moreover, necroptosis-related lncRNAs, such as LINC00472, which regulates RIPK1, have demonstrated their predictive value for the prognosis of patients with breast cancer and their guiding value for personalised treatment strategies [63].
Pyroptosis is a form of PCD, and its classical activation pathway is mediated by caspase-1 and executed through the formation of membrane pores by proteins from the gasdermin (GSDM) family [64]. The canonical pyroptosis pathway is initiated by the activation of the NLRP3 inflammasome complex, which includes NLRP3, the adaptor protein ASC containing the CARD domain, and pro-caspase-1 [65]. Notably, inflammatory responses play a dual role in tumorigenesis: they can both promote malignant transformation and metastasis and exert an antitumour effect via pyroptosis induction [66]. The nonclassical pyroptosis pathway is mediated by caspase-4/5/11. Although these caspases cannot directly cleave the precursors of IL-1β/IL-18, they can still trigger pyroptosis by cleaving GSDMD-NT, a process specifically activated by bacterial lipopolysaccharides [67, 68]. Additionally, caspase-8 regulates pyroptosis under specific conditions: nucleus-localised PD-L1 induces pyroptosis via caspase-8-mediated cleavage of GSDMC in the hypoxic tumour microenvironment [69]. Docetaxel cleaves GSDME via the ROS/JNK/caspase-3 pathway, and its pyroptotic effect can be enhanced by demethylation of the DFNA5 gene, highlighting the role of epigenetic regulation in chemosensitisation [70]. In the context of triple-negative immunotherapy for breast cancer, the combination of pyroptosis induction with immune checkpoint inhibitors has significant synergistic potential [62]. For example, TAT3 inhibitor nanoparticles (MPNPs) in conjunction with oncolytic viruses can activate GSDME-dependent pyroptosis and markedly enhance the efficacy of anti-PD-1 therapy [62]. The dopamine receptor DRD2 has been identified as a novel therapeutic target to remodel the immunosuppressive tumour microenvironment by promoting M1 macrophage polarisation, inhibiting the NF-κB signalling pathway, and inducing pyroptosis in breast cancer cells [62], which provides a promising new treatment direction for breast cancer. These findings establish a robust theoretical foundation and translational strategy for precision medicine approaches targeting pyroptosis regulation in breast cancer. Epigenetic modifications also play crucial roles in mediating tumour growth and metastasis by regulating pyroptosis-related pathways [71].
Ferroptosis is a form of iron-dependent PCD characterised by the aberrant accumulation of lipid peroxides and specific adjustments, as shown in Fig. 3 [72]. This process is specifically triggered by the peroxidation of polyunsaturated fatty acids (PUFAs), rather than a generalised increase in ROS [72]. Its regulatory network includes three core axes: the Xc/GSH/GPX4 antioxidant system, ACSL4/LPCAT3/15-LOX pathway, and FSP1/CoQ10/NAD(P)H axis, which counteract ferroptosis by inhibiting lipid peroxidation [73]. Moreover, pathways mediating ferroptosis in breast cancer include T-cell inhibition via GCH1-BH4 [74] and ferroptosis triggered by DHODH-CoQH2 following GPX4 inactivation [75]. To date, 8 glutathione peroxidases have been identified, among which GPX4 directly reduces lipid peroxides [27], making it a potential therapeutic target for breast, ovarian, liver, and prostate cancers [76]. In breast cancer, the Xc/GSH/GPX4 system is essential for maintaining redox homeostasis [77]. Cystine enters cells through Xc and is reduced to cysteine via the cystine reduction pathway dependent on GSH or thioredoxin reductase 1 (TXNRD1), thereby promoting GSH synthesis [78]. GSH, as a cofactor of GPX4, facilitates the reduction of phospholipid hydroperoxides (PLOOHs) to their corresponding alcohols (PLOHs), thereby maintaining redox balance [33].
Targeted inhibition of the Xc/GSH/GPX4 system axis effectively induces ferroptosis [79]. For example, DMOCPTL regulates EGR1 in TNBC cells and induces GPX4 ubiquitination to reduce EGR4 protein levels, thereby regulating mitochondria-mediated apoptosis [80]. In the FSP1/CoQ10/NAD(P)H axis, FSP1 reduces CoQ10 to generate the antioxidant CoQH2 independently of GPX4, inhibits lipid peroxidation [81], and cooperates with vitamin K metabolism to protect cells [82]. Ferroptosis in breast cancer involves subtype-specific regulation. In triple-negative breast cancer, the expression level of GPX4 is significantly associated with prognosis. Studies have demonstrated that GPX4 is markedly upregulated in the luminal androgen receptor subtype of TNBC, thereby conferring resistance to ferroptosis [83]. Conversely, the basal-like subtype is more susceptible to ferroptosis induction due to elevated expression levels of ACSL4 and FADS1/2 [84]. Notably, the sensitivity of BRCA1-deficient breast cancer cells to PARP inhibitors can be increased by inhibiting GPX4, thereby providing a novel strategy to overcome drug resistance [75].
In luminal breast cancer, ER/PR positivity is positively correlated with GPX4 expression, while m6A demethylation activates SLC7A11 by inhibiting the FGFR4/GSK-3β pathway, thereby increasing sensitivity to ferroptosis [85]. In HER2-positive breast cancer, overactivation of HER2 upregulates SLC7A11 via the PI3K/AKT/mTOR pathway to increase the antioxidant capacity [85]. The combination of chemotherapeutic drugs and ferroptosis inducers can reverse drug resistance in HER2+ breast cancer [86]. For clinical strategies targeting ferroptosis, existing interventions include GPX4 inhibitors, such as red ginseng polysaccharides, which induce ferroptosis by downregulating GPX4. Their combination with immunotherapy can enhance therapeutic efficacy [87]. Epigenetic regulation is prominent in breast cancer ferroptosis. DNA methylation, microRNAs (miRNAs), and m6A modifications differentially regulate the expression of ferroptosis-related genes [88]. High methylation of the SLC7A11 promoter suppresses its expression and increases cellular resistance to ferroptosis. m6A demethylation activates the SLC7A11/FPN1 axis by inhibiting the FGFR4/GSK-3β/β-catenin pathway, suggesting a novel strategy to overcome resistance to HER2 treatment [85]. The METTL3 inhibitor STM2457 promotes GPX4 degradation by reducing m6A modification, inducing ferroptosis and inhibiting metastasis in TNBC [89]. Noncoding RNAs and ferroptosis have complex relationships in cancer progression. Different ncRNAs can either promote or inhibit ferroptosis in cancer cells. ncRNA-mediated regulation of ferroptosis may represent a new therapeutic direction for treating breast cancer. Long noncoding RNAs (lncRNAs) contribute to improving prognosis and identifying potential therapeutic targets in breast cancer [90]. LncFASA promotes lipid droplet formation and inhibits lipid peroxidation by activating the SLC7A11‒GPX4 axis [91]. The mechanism of tumour suppression involves multiple pathways. LINC00152 confers tamoxifen resistance by inhibiting ferroptosis through BACH1, and sensitivity to endocrine therapy can be restored by silencing LINC00152 [92]. Additionally, the lncRNA H19 regulates chemotherapy resistance in breast cancer via the DNA damage response pathway [93].
Cuproptosis is a recently identified form of copper-dependent PCD that differs from apoptosis, necroptosis, and ferroptosis. Specifically, the accumulated copper ions within cells directly bind to lipoylated proteins in the TCA cycle, leading to their inactivation and subsequent cell death (Fig. 3) [94]. Notably, breast cancer cells exhibit elevated aerobic respiration due to their highly active mitochondrial metabolism, which may potentiate tumour angiogenesis via the cuproptosis pathway, thereby fostering a tumour-promoting microenvironment [95]. Copper homeostasis has a dual nature in that it enhances the capacity of copper ions as cofactors for SOD to scavenge reactive oxygen species [96]. Conversely, excessive Cu⁺ can generate ·OH via the Fenton reaction, inducing oxidative stress and cellular damage [97]. In this context, GSH can remarkably maintain copper homeostasis by neutralising ROS toxicity through reduction reactions [98]. During the cuproptosis process, the key target FDX1 interacts with components of the tumour microenvironment. The mitochondrial protein FDX1 has been validated as a target of the anticancer drug Elesclomol, which induces specific copper-mediated cell death by facilitating copper ion delivery [99]. Notably, the hypoxic microenvironment characteristic of TNBC markedly suppresses FDX1 expression. Under these conditions, lipoic acid synthase sustains TCA cycle function through Fe‒S cluster-mediated acylations [100].
In epigenetics, lncRNAs associated with cuproptosis significantly regulate the proliferation and metastasis of breast cancer cells and potentially predict prognosis and sensitivity to various therapies [101]. The copper transporter SLC31A1 is aberrantly overexpressed in breast cancer cells and is significantly positively correlated with tumour immune infiltration and the expression of immune checkpoint molecules such as PD-L1 [102]. SLC31A1 is significantly correlated with diverse immune cell infiltrations, immune cell biomarkers, and immune checkpoints in breast cancer and is regulated by the LINC01640/miR-204-5p/SLC31A1 axis, which may be central to copper-induced cell death in breast cancer, suggesting that SLC31A1 may serve as a potential target for adjuvant immunotherapy [103].
Oxeiptosis is a noninflammatory form of cell death initiated by oxidative stress and is sensitive to ROS [104]; this form of cell death is carried out primarily through the KEAP1/PGAM5/AIFM1 signalling pathway, thereby protecting the organism from ROS-induced damage and inflammation [105]. As shown in Fig. 4a, excessive accumulation of ROS triggers KEAP1 activation and enables KEAP1 to interact with PGAM5. This interaction facilitates the dephosphorylation of the intermembrane space protein AIFM1 by PGAM5, which in turn triggers the release and nuclear translocation of AIFM1, ultimately leading to large-scale DNA fragmentation [106]. Alloimperatorin enhances the expression of KEAP1 without altering the expression levels of PGAM5 or AIFM1, thereby inhibiting the proliferation and invasion of breast cancer cells [106].
When glucose is depleted, the cellular redox state becomes insufficient. Mesencephalic astrocyte-derived neurotrophic factor (MANF) mitigates protein oxidation and restores phagocytosis mediated by E3 ligase activity, thereby promoting the survival of BC cells under glucose starvation [106]. Cystine transported by SLC7A11 induces the response of actin cytoskeleton proteins to disulfide stress, leading to disulfide-induced apoptosis [107]. Central genes associated with disulfide-related immune checkpoints can predict the prognosis of breast cancer [108]. Breast cancer subtypes can be predicted by lncRNAs involved in disulfide metabolism. Specifically, in the basal subtype, LINC02188 is highly expressed, whereas LINC01488 and GATA3-AS1 exhibit the lowest expression levels [109]. LINC00511 is expressed at the highest level in the Her2 subtype, and GATA3-AS1 is expressed at the highest level in the LumA, LumB, and normal-like subtypes [108]. These findings provide a novel direction for identifying clinical therapeutic targets for breast cancer (Fig. 4b) [109].
Autophagy eliminates damaged organelles and abnormal proteins via the lysosomal degradation pathway to maintain intracellular homeostasis and suppress genomic instability [110]. Under physiological conditions, moderate levels of autophagy effectively mitigate both endogenous and exogenous inflammatory responses by inhibiting NLRP3 inflammasome activation [111]. Notably, breast cancer cells achieve metabolic adaptation by hijacking the autophagy mechanism: under glucose-deprivation conditions, MANF-mediated mitophagy eliminates dysfunctional mitochondria through the PRKN/Parkin-dependent pathway, maintains the ATP supply and promotes tumour survival [112], suggesting that targeting the MANF‒PRKN interaction can break the drug resistance barrier of breast cancer under metabolic stress.
Autophagy and ferroptosis involve complex metabolic interactions: autophagy-dependent degradation of ferritin leads to the accumulation of free iron and promotes lipid peroxidation through the Fenton reaction. Meanwhile, mitochondrial damage mediated by PCD protein 2 can increase the generation of mitochondrial reactive oxygen species, synergistically inducing autophagy and ferroptosis [113]. Notably, the combination of autophagy inhibitors (such as chloroquine) and ferroptosis inducers (such as erastin) can significantly inhibit breast cancer proliferation, suggesting the potential of synergistic therapy for dual death pathways [114]. Noncoding RNAs exert regulatory functions within the autophagy network [115]. LncRNAs significantly modulate autophagy plasticity via epigenetic and posttranscriptional mechanisms. For example, DDIT4-AS1 promotes protective autophagy by inhibiting mTORC1 complex activity, thereby facilitating the progression of triple-negative breast cancer. Its silencing enhances sensitivity to cisplatin-induced DNA damage [116]. H19 activates autophagy through the H19/SAHH/DNMT3B axis, potentially contributing to tamoxifen resistance in breast cancer (Fig. 4c) [117].
Paraptosis represents an atypical form of PCD [118]. The key features that distinguish paraptosis from classical apoptosis include its independence from caspase-9 activation, the absence of typical apoptotic markers such as chromatin condensation or DNA laddering, and cytoplasmic vacuolisation as the predominant morphological feature positively modulated by the mitogen-activated protein kinase (MAPK) signalling pathway [119]. Additionally, paraptosis can be specifically inhibited by the apoptosis inhibitory protein AIP-1/Alix, indicating a distinct regulatory network separate from the apoptotic pathway [120]. The occurrence of paraptosis is closely associated with dysfunction in the ER‒mitochondrial axis [121]. The accumulation of misfolded proteins within the ER lumen leads to osmotic imbalance, resulting in water efflux and expansion of the ER compartment. Persistent ER stress promotes the extensive release of Ca into the cytoplasm via the IP3R channel. Ca subsequently enters the mitochondrial matrix through the mitochondrial calcium uniporter to cause mitochondrial swelling and membrane potential collapse, ultimately leading to cell death (Fig. 4d) [122, 123]. Research has demonstrated that celastrol induces paraptosis in breast cancer cells via the IP3R-dependent Ca signalling pathway, highlighting the critical role of Ca homeostasis disruption in this death pattern [124]. Increasing attention has been given to emerging therapeutic strategies based on the paraptosis mechanism [125].