Breast cancer (BC) is one of the most common malignancies worldwide, responsible for 670,000 deaths reported globally and 2.3 million diagnoses. This tumor may interest both women and men, but the incidence is massively tilted for the female gender, representing 99% of the whole cases, against 0.5%–1% of the male group. Breast cancer is a type of tumor with different outcomes and presentations; it is a very heterogeneous disease in which both environmental and genetic factors are involved (World Health Organization - www.who.int/news-room/fact-sheets/detail/breast-cancer). Human breasts are paired mammary glands that develop in females during puberty under the influence of various pubertal hormones; it has an inner structure made of epithelial components consisting of lobules and ducts (distributed throughout the fibrous and the adipose tissues), leading out to the nipple (1).

Most breast cancers are adenocarcinomas, and they can be invasive or non-invasive, according to their tendency to be circumscribed in the lobules and/or in the ducts (i.e., lobular carcinoma in situ and ductal carcinoma in situ) or to metastasize and “invade” other organs or tissues (i.e., Paget’s disease, Triple Negative Breast Cancer (TNBC), medullary carcinoma, inflammatory breast cancer, mucinous carcinoma, tubular carcinoma, phyllodes tumor and infiltrating lobular/ductal carcinoma) (2–10). The possibility of developing breast cancer can be increased by various genetic or non-genetic factors, like the presence of mutations in breast cancer susceptibility one and two genes (BRCA1 and BRCA2), smoke, obesity, or prolonged exposure to estrogen and progesterone hormones (11–13). Based on molecular and histological evidence, breast cancer comprises several histological and biological/molecular subtypes with distinct behaviors and responses to therapy: Luminal A, Luminal B, HER2 positive, and TNBC (Figure 1). Luminal A breast cancers are positive for the estrogen receptor (ER) and/or for the progesterone receptor (PR) but not for the Human Epidermal Growth Factor Receptor 2 (HER2); luminal B are ER+ and/or PR+ and HER2^+/−, with a general higher proliferation rate; HER2 breast cancers are only HER2+, while TNBCs are ER-, PR- and HER2-. This last subtype represents the most aggressive form of the tumor, a feature determined by its ability to metastasize and by the lack of targeted treatment strategies. In fact, according to the molecular characteristics of each BC type, different approaches are adopted to reduce/suppress the tumor (14).

In luminal-like breast cancers, the standard procedure is represented by an endocrine therapy (15–17), while in HER2+ BCs an anti-HER2 humanized monoclonal antibody currently represents the most effective treatment (18). These compounds can be used alone or in combination with other procedures like surgery, radiotherapy, or chemotherapy, but a consistent portion of patients develop a resistance to the deputed drug (15). For triple-negative breast cancers there is no standardized treatment regimen and chemotherapy still represents the primary systemic treatment, but the efficacy of conventional post-operative adjuvant chemo-radiotherapy is poor (19). One exception is represented by a specific TNBC subtype named “LAR” (Luminal AR+) characterized by an enrichment in androgen response, fatty acid metabolism, and oxidative phosphorylation.

Interestingly, the LAR-subtype is closely related to the L2 (ER + luminal BC) hormone-responsive cells, and its cell lines were uniquely sensitive to the AR antagonist Bicalutamide; therefore, the next-generation AR antagonist Enzalutamide is currently being evaluated in AR-positive (AR+) TNBC in combination with Paclitaxel (20).

In this very heterogeneous and partially uncovered context, a pivotal role could be represented by epigenetics: understanding the intricate interplay of the epigenetic modifiers could be critical for unraveling the complexities of BC progression and developing targeted therapeutic interventions (21, 22). Non-coding RNAs are essential components of the complex epigenetic regulation machinery (Figure 2), and they play crucial roles in the post-transcriptional regulation of gene expression. Indeed, the dysregulation of their functions can result in unfavorable outcomes across various disease pathways (23).

Non-coding RNAs (ncRNAs) are RNA sequences that do not translate into proteins and can be generally categorized into three main classes (24): long non-coding RNAs (lncRNAs), small non-coding RNAs (sncRNAs) and circular RNAs (circRNAs). CircRNAs are a class of endogenous non-coding RNA, characterized by their covalently closed-loop structures without a 5′cap or a 3′poly(A) tail (25), while sncRNAs and lncRNAs are regulatory RNAs that differ in size (26).

This review article aims to highlight the roles of non-coding RNAs in a very heterogeneous disease such as breast cancer. Over the years, many therapeutic approaches have been established and employed for BC, although they are inadequate or outdone in severe conditions. Given the avenue of new technologies and the growing body of research showing non-coding RNA involvement in BC tumorigenesis, progression, and invasion, this review aims to draw researchers’ and clinicians’ attention to the ncRNA’s potential use as biomarkers and therapeutic targets, examining their mechanisms and regulatory functions. We curated relevant studies primarily through a PubMed search using keywords tailored to each section of this manuscript, for example, “Epigenetics and Breast Cancer”, “non-coding RNAs in cancer”, “small non-coding RNAs and Breast Cancer”, “long non-coding RNAs and Breast Cancer” and “circRNAs and Breast Cancer”, deepening the research by focusing on peer-reviewed articles published in English, with a preference for those published during the last decade (2015–2025). Additional references were included based on cross-referencing and input from expert contributors to ensure thematic completeness and scientific accuracy.

We carefully evaluated the information extracted from these studies to ensure an accurate representation of the original research findings. The articles mentioned in this review were chosen prioritizing our purpose to give the reader an overview of every non-coding element in relationship with BC by following the same path: biogenesis, general biological role, eventual implication in tumorigenesis, invasiveness and migration and drug resistance, with a final focus on the consequent possible application as biomarkers or target for new advanced treatment. While every effort was made to include the most pertinent evidence, due to the volume of research in this area and the limitations inherent to a narrative format, we could not include every relevant article, and we apologize to the authors whose work could not be incorporated into this review.

This manuscript offers an overview of current knowledge in this field; despite the richness of the literature available, our work emerges as a comprehensive and didactic paper of particular interest for those who approach to the field for the first time and need a brief overlook to the theme. At the same time, the constant progress of the research highlights the need for adjourned reports of the state of the art.

Long non-coding RNAs (lncRNAs) are transcripts longer than 200 nucleotides that are not translated into proteins and exhibit limited evolutionary conservation (27). Classification of lncRNAs is often based on their genomic positioning relative to protein-coding genes. Sense lncRNAs are transcribed from the strand as adjacent/overlapping protein-coding genes, while anti-sense lncRNAs are generated from the opposite strand. Bidirectional lncRNAs originate from the protein-coding gene’s promoter but are transcribed in the opposite direction. Intergenic lncRNAs are generated from regions between protein-coding sequences, and intronic lncRNAs are transcribed within the introns of coding genes (28). In addition, lncRNAs situated between two encoding protein genes can be classified into two main groups: enhancer-associated (elncRNA), which often regulate the expression of nearby genes on the same chromosome, and promoter-associated lncRNAs, which regulate chromosomal status and epigenetic inheritance (29–32).

The functional roles of lncRNAs are closely linked to their subcellular localization (Figure 3): in the nucleus, they participate in the modulation of epigenetic regulators and influence transcriptional programs through chromatin remodelling/interactions, and through the spatial organization of the nuclear compartment via scaffolding (33). Within the chromatin, lncRNAs can act as molecular scaffolds or guides for proteins, facilitating or inhibiting their recruitment and activity at specific genomic loci (34). In the cytoplasm, lncRNAs can regulate mRNA post-transcriptional processes, affecting mRNA stability and translation and modulating signaling pathways. Importantly, in the cytoplasm, they also act as miRNA “sponges”: they can bind miRNAs, thus reducing their availability to interact with target mRNAs, thereby indirectly regulating gene expression (35).

Furthermore, lncRNAs can be also transported into mitochondria, where they are implicated in the regulation of mitochondrial metabolism, apoptosis, and in their crosstalk with the nuclei (36). LncRNAs can also be packaged into exosomes, which are then released into the extracellular environment; next, exosome-localized lncRNAs can reach recipient cells, where they contribute to epigenetic regulation, cell-type reprogramming, and genomic instability (37).

Due to their extensive gene regulatory capabilities, lncRNAs influence a wide range of physiological processes, including cell differentiation, growth, and responses to diverse stresses and stimuli. Moreover, they play key roles in the nervous, cardiovascular, hematopoietic and immune systems and their associated pathologies. The involvement of lncRNAs in oncogenesis, specifically in cancer initiation and progression, is increasingly recognized. LncRNA exert their effects on cancer cell proliferation and survival, often by modulating key oncogenic or tumor-suppressive transcription factors, such as p53 (38–40).

In breast cancer, a growing body of evidence highlights the aberrant expression of specific lncRNAs across different BC subtypes, with strong correlations with tumor initiation, progression, and clinical outcomes (Table 1). Furthermore, lncRNAs are particularly attractive as therapeutic targets due to several advantageous properties: high tissue-specificity, regulation of specific elements of key cellular networks, limited toxic effects associated with their targeting, the often fast-turnover and their low expression levels, which could facilitate quicker effects with lower doses (41–43).

LncRNAs can intervene in the regulation of breast cancer stem cells (BCSCs) related pathways; examples are the lncRNA LINC01133 (which induces Kruppel-like factor 4 (KLF4) gene) (44, 45) and the long intergenic non-coding RNA 00511 (LINC00511) which functions as a miR-185-3p “sponge”, indirectly activating (via the E2F1 protein targeting) the transcription of Nanog, a promoter of regeneration and prolonged proliferative potential of stem-like cancer cells, able to mediate oncogenic reprogramming. Thus, the LINC00511/miR-185-3p/E2F1/Nanog axis may also have therapeutic potential, by regulating breast cancer stemness and tumorigenesis (46). Another important example is represented by LINC00617, which can also impact on the BCSCs self-renewal capacity through the activation of SOX2 transcription mediated by hnRNP-K recruiting (47). Several long non-coding RNAs can also regulate breast cancer stem cells through epigenetic modifications, like the repression of the tumor suppressor long non-coding RNA (TSLNRs) EPB41L4A-AS2 through the enrichment of H3K27me3 (48).

A long non-coding RNA that actively intervenes in the development and maintenance of BC is HOX transcript antisense RNA (HOTAIR) that inhibits miR-7 expression, leading to increased SETDB1 expression in BCSCs, inducing the epithelial-mesenchymal transition (EMT).

HOTAIR can additionally foster H3K27 acetylation and E-cadherin promoter methylation, this mechanism inhibits E-cadherin production furtherly promoting EMT (49).

Non-coding RNAs are also implied in breast cancer metastatic progression; as depicted in Figure 3, they can act by different mechanisms, such as the degradation or silencing of specific mRNAs, the target of enzymes and microprocessor subunits involved in miRNA biogenesis, and the sponging of miRNAs, thus altering the expression of several genes and modulating different cell signaling pathways. Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) has been correlated with an increased tumor size and stage, and a consequent poor prognosis in human patients: it undergoes tight transcriptional control in tumor cells by several transcription factors, both positively and negatively (50). For example, hypoxia-inducible factor 1α (HIF-1α) upregulates MALAT1 with the mediation of AMP-activated protein kinase (AMPK) (51); the induced lncRNA then acts as a miRNA sponge of miR-3064-5p, a mechanism that promotes tumor growth and migration in breast cancer cells (52).

Conversely, the depletion of MALAT1 triggers the arrest of the cell cycle followed by a reduced cellular proliferation rate. It activates p53 – a tumor suppressor that participates in apoptosis and senescence processes–and its target genes (53).

According to several studies, this long non-coding could be essential in developing and metastasizing TNBC and HER2-positive BC because the presence of metastatic lymph nodes is correlated with MALAT1 expression in breast cancer patients. In addition, MYC and its downstream immune regulatory genes (CD47 and PD-L1) are related to metastasis and relapse in these subtypes of BCs and are positively regulated by MALAT1 (54). Huang and colleagues demonstrated that the knockdown of MALAT1 in MCF-7 cells reduced EGF expression, suggesting that it might initiate angiogenesis in BC, through modification of miR-145 (55). According to multiple lines of evidence, MALAT1 is also implicated in regulating signaling pathways associated with cancer progression, such as the Wnt signalling (56), but it is still uncertain how it affects these pathways.

Another important example of the contribution of the lncRNAs in breast cancer metastasis formation is represented by neuroblastoma-associated transcript 1 (NBAT1) (57), and by the aforementioned HOTAIR. The former induces BC cells invasiveness by reducing H3K27me3 levels, while the latter induces migration and invasion by recruiting the polycomb repressive complex 2 (PRC2) which leads to the variation of H3K27 methylation levels and global gene expression alterations (58).

LncRNAs can often interfere in protein translation; it was in fact shown that abnormally expressed lncRNAs can be also related to multidrug resistance in breast cancer. In endocrine therapies, the resistance to Tamoxifen can be mediated by the induction of an autophagy mechanism; this mechanism can be triggered by long non-coding RNAs–such as H19 – acting on key mediators of the process (59). In HER2+ BCs, an augmented expression of the lncRNA AFAP1-AS1 can induce resistance to Trastuzumab by binding to AUF1, thus promoting ERBB2 translation (60). Furthermore, several lncRNAs can also be involved positively or negatively in the resistance to Doxorubicin/Adriamycin: lncRNA BORG, for example, can activate the NF-kB signaling decreasing the genomic damage, whilst LINC00668 targets staphylococcal nuclease domain-containing 1 (SND1) and improves the resistance to DOX (61).

An altered immune response in the tumor microenvironment can also markedly affect cancer occurrence and development. In this context, lncRNAs can regulate the function of immune cells impacting the antigen presentation ability of dendritic cells (DCs): for instance, the lncRNA XIST (62) down-modulation in M1-type macrophages (M1) leads to the transformation in anti-inflammatory M2 macrophage (M2) to promote tumor cell proliferation and migration. Recent studies have shown that lncRNAs can intervene in immunosuppression and may be a potential target for cancer immunotherapy, but the mechanism of tumor immune escape is highly complex and needs to be extensively investigated (62). Undoubtedly, they are promising predictive biomarkers and therapeutic targets for breast cancer immunotherapy, although further research is still required.

All these correlations make long non-coding RNAs good candidates as biomarkers for tumor diagnosis and prognosis and for predicting disease progression, but also as therapeutic targets in the shape of small molecule inhibitors, siRNAs, antisense oligonucleotides (ASOs), and CRISPR-Cas9. Vaidya and Collogues, for instance, studied the Differentiation Antagonizing Non-Coding RNA (DANCR), which is a non-coding RNA involved in the regulation of different oncogenic mechanisms and undruggable by conventional molecules; it was demonstrated that the delivery of siRNA against DANCR and its subsequent inhibition epigenetically represses the expression of cancer-driven pathways, such as Wnt signaling, EMT, and phosphorylation of several kinases: siDANCR-NP effectively inhibits migration and invasion of cancer cells in vitro and tumor growth in vivo (63). A small molecular inhibitor, AC1Q3QWB, has been implemented in breast cancer-rived xenografts, resulting in efficient disruption of PRC2 recruitment by HOTAIR without notable off-target effects (64).

On the other hand, these compounds are not easy to design, and although lncRNAs are opening a new door for clinical diagnosis and treatment of breast cancer, there are still many difficulties that must be faced and overcome.

Circular RNAs (circRNAs) represent a subclass of single-stranded RNAs derived from precursor mRNAs or lncRNAs and characterized by a covalently closed loop structure. CircRNAs are generated by a process known as “RNA back-splicing” in which the 3′-end of an exon is joined to the 5′-end of the same or an upstream exon, via a 3′, 5′-phosphodiester bond. This event creates a closed circular structure containing a characteristic back-splicing junction (Figure 4).

CircRNAs are broadly categorized based on their composition. Most of them are composed only by exonic sequences and are referred to as exonic circRNAs (ecircRNAs). Less frequently, circRNAs may be formed entirely from intronic regions (IcirRNAs) or may contain both exonic and intronic sequences (EIciRNAs). EcircRNAs are mostly localized in the cytoplasm, although the precise mechanism governing their nuclear export remains insufficiently understood (65); interestingly, some ecircRNAs are found within the nucleus where they increase the nuclear retention of specific proteins or recruit proteins to chromatin (66). Conversely, most intron-containing circRNAs are retained in the nucleus, where they may regulate their parental gene expression (67).

Recent investigations have extended our understanding of circRNA biology by identifying a subset of circRNAs localized in mitochondria, extending the complexity of the mitochondrial transcriptome (68–70). However, the presence and functional relevance of circRNAs in other organelles and subcellular compartments remain largely unexplored and warrant further study.

The unique structure of these RNAs makes them more resistant to exonucleases than their linear counterparts, providing them with a longer half-life; in fact, circRNAs are often stable and accumulate in most cell types, with an especially high abundance in neural tissues. These features make circRNAs attractive candidates as diagnostic biomarkers and therapeutic targets.

A growing body of research has documented the distinct expression profiles and functional significance of circRNAs in a range of pathological conditions, including cancer (70–72), cardiovascular disease (73), neurological disorder (74), and autoimmune disease (75). Despite these advancements, the mechanisms underlying the abnormal landscape of circRNAs and how circRNAs exert their physiopathological roles remain poorly understood. Table 2 summarizes key circular RNAs involved in the BC pathological processes.

Of note, one of the main features of circular RNAs is that they can act as miRNA “sponges,” or competitive inhibitors. By circRNAs interaction/sequestration, miRNAs are prevented from binding to their mRNA targets, inhibiting miRNA-mediated gene silencing and protecting target mRNAs from degradation (76). Some circRNAs can be classified as oncogenic sponges, as they facilitate multiple malignant behaviors, including tumor proliferation, distance metastatization, and drug resistance. An example is Has_circRPPH_015, which is upregulated in BC tissues. At the same time, its knockdown restrains aggressive behaviors of BC cell line MCF-7: in fact, this circRNA can bind to miR-326 and negatively regulates ELK; on the contrary, an elevated expression of miR-326 inhibits cell proliferation, colony formation, and cell invasion in this BC line (77).

CircRNAs bind often to transcription factors promoters, regulating their expression. At the same time, they can also function as scaffold in the modulation of protein-protein interactions, or even have translational potential.

CircRNAs participate both positively and negatively in breast cancer development and progress, acting as either oncogenes or tumor suppressors and their aberrant expression can be associated with tumoral cell proliferation, apoptosis, autophagy, invasion, migration, and treatment resistance.

A differential expression of circRNAs has been recently associated with diverse breast tumor status, drawing attention to the possibility to outline a circRNA “signature” in different tumor biopsies or cell lines. Cancer cells can in fact release these ncRNAs in urine, plasma, and saliva, opening the possibility of using circular RNAs as potential biomarkers in diagnosis and prognosis, as for the case of circSMARCA5, Hsa_circ_0104824 that were shown to be decreased in BC patients’ blood compared to controls (78, 79). In other studies, circPSMA1 appeared upregulated in serum/plasma of BC patients compared with those of healthy controls (80–82). More specifically, the overexpression of circPSMA1 promoted TNBC cell proliferation, migration, and metastasis both in vitro and in vivo (81), together with another circRNA - circ-Amotl1 - which is involved in tumorigenesis, enhancing the stability of c-MYC and the expression of its targets (66). CircPVT1 can work through both ceRNA and protein scaffolding mechanisms: it sponges miR-181a-2-3p to modulate ESR1 mRNA stability and downstream estrogen/ERα-target genes, while it represses type I IFNs and ISGs by binding MAVS to disturb RIGI–MAVS complex formation. This dual function contributes to ERα-positive BC development (83). CircRNAs are thus becoming more clinically relevant in breast cancer diagnosis, particularly for their early detection and stratification into different subtypes; which ameliorates the disease prognosis; however, they have been poorly explored in HER2-related BC subtypes, and more investigations are needed.

Beyond tumorigenesis, these non-coding RNAs also play a role in progression and metastasis: in a recent work, Peng and colleagues showed that an overexpression of circEZH2 impacted on the vitality and the invasion of breast cancer cells, while its knockdown led to the opposite effects (84). The same molecular mechanism is furtherly used by circROBO1, another important actor in the migration and invasiveness of BC cancer cell, especially in liver metastasis (85).

Resistance to treatments represents still a challenging issue in breast cancer therapy and survival; in this scenario, a more personalized approach could represent a significant improvement and circRNAs might be promising predictive biomarkers, according to their ability to regulate BC cell sensibility to drugs/treatments (86). For example, the circular RNA circCDYL2 confers Trastuzumab resistance in BC patients by stabilizing GRB7 and preventing its ubiquitination degradation; this enhances its interaction with FAK, which thus sustains the activities of downstream AKT and ERK1/2 (87). A recent work demonstrated that circRNA-CREIT is aberrantly downregulated in Doxorubicin-resistant TNBC cells: the RNA binding protein DHX9 is responsible for its reduction by interacting with the flanking inverted repeat Alu (IRAlu) sequences and inhibiting back-splicing. Mechanistically, circRNA-CREIT acts as a scaffold for proteins interaction, affecting the PKR/eIF2α signaling axis - related to stress granules (SGs) assembly - and the RACK1/MTK1 apoptosis signaling pathway. Further investigations revealed that a combination of the SG inhibitor ISRIB and Doxorubicin synergistically inhibits TNBC tumor growth. Besides, circRNA-CREIT could be packaged into exosomes and disseminate Doxorubicin sensitivity among TNBC cells (88).

In hormone therapies, Zheng and colleagues observed that circRNA-SFMBT2 appeared to be directly related to cell proliferation and Tamoxifen resistance in vitro (89), whereas circ_0025202 has been reported as a potential predictive biomarker of BC resistance to Tamoxifen (90), but more research is needed to confirm the data.

In the end, circular RNAs can be potential biomarkers also in chemotherapy resistance, as well as radiotherapy and immunotherapy resistance. The inhibition of cirCDR1as, for example, increases the sensitivity to 5-fluorouracil and Cisplatin of initially resistant BC cells, while an augmented expression of circSMARCA5 improves the chemosensitivity to Cisplatin of BC cells and tumors (78). CircKDM4C is strongly associated with Doxorubicin resistance cells both in vivo and in vitro being a potential biomarker for a Doxorubicin response prediction (91). Similarly, circUBE2D2 is involved in Doxorubicin resistance in TNBC cells, acting at the cellular level as a sponge of miR-512-3p, resulting in the upregulation of CDCA3 expression (92). Several studies have also investigated the role of circRNAs in ADM-resistance (93) and the resistance to taxanes (94, 95). Lastly, a recent work by Li and colleagues reported that the circular HER2 RNA (circHER2) encodes for a novel protein–HER2-103 – that is expressed in a marked percentage of TNBC cases, with a worse overall prognosis than circ-HER2/HER2–103 negative patients; in their work, HER2-103 enhanced both homo and hetero dimerization of EGFR/HER3, AKT phosphorylation and malignant phenotypes.

Furthermore, Pertuzumab, an antibody employed in HER2+ tumor treatment, could represent a potential antagonist due to the congruence of the amino acid sequence of HER2-103 and HER2 CR1 domain. This antibody in fact decreased the in vivo tumorigenicity only of the triple-negative tumoral cells expressing the circ-HER2/HER2–103 (96). Altogether, these studies show that circRNAs may predict responsiveness to chemo-, radio-, immuno-, and hormone-therapies, and further clinical investigations are encouraged for validation.

Small non-coding RNAs (sncRNAs) are a class of highly abundant ncRNAs that are typically <100 nucleotides (nt) long, transcribed from noncoding genomic regions with the ability to regulate various aspects of gene expression during normal animal physiology and development. sncRNAs control gene expression by regulating chromatin architecture, transcription, RNA splicing, editing, translation, and turnover. They are further divided into different subtypes: micro RNAs (miRNAs), PIWI-interacting RNAs (piRNAs), small-interfering RNAs (siRNAs), small nuclear RNAs (snRNAs) and small nucleolar RNAs (snoRNAs) (100).

As thoroughly discussed, non-coding RNAs represent a vast resource in the comprehension and treatment of breast cancer; they act at different levels and with different mechanisms both in the development and in the inhibition of these tumors. All the direct and indirect correlations described in this manuscript suggest non-coding RNAs as perfect candidates as biomarkers for diagnosing tumors, judging patient prognosis, and predicting disease progression. Moreover, multiple studies have proved that many of these ncRNAs are stably expressed in BC patients’ blood, plasma, urine, and other body fluids even at early stages, providing evidence for a novel class of non-invasive biomarkers for BC.

Among them, an emerging and promising diagnostic tool is represented by circRNAs. Their unique structure, as mentioned above, provides them with stability and specificity, making them interesting candidates for therapeutic strategies. Furthermore, many studies show circRNAs ability to regulate BC cell sensibility to treatment, predicting responsiveness to chemo-, radio-, immuno-, and hormone-therapies, highlighting their potential to overcome the resistance issue. All these various implications highlight the potential diagnostic value of these non-coding RNAs.

Furthermore, with the continuous discovery of ncRNAs structural information and regulatory functions, small molecule inhibitors against ncRNAs have been developed with broad prospects for clinical diagnosis and treatment of tumors. RNA interference (RNAi) can be harnessed to inhibit the expression of cognate mRNA by exogenous RNA-based molecules that can be synthetically designed against any target RNA. The new anti-tumor drugs against ncRNAs have become a new promising trend in cancer treatment. At present, the research of new molecules targeting ncRNAs has made some progress.

a. small (or short) interfering RNAs, siRNAs, which target transcripts via RNA-induced silencing complex (RISC), downregulating mRNA levels;

There are many examples of RNAi, such as siRNAs, that can target all kinds of proteins, including traditionally undruggable proteins, and there is also evidence that these molecules could be superior to antibodies or small molecule inhibitors when inhibiting the same pathway, demonstrating a high therapeutic potential (200).

Although ncRNAs are opening a new door for clinical diagnosis and treatment of breast cancer and these nucleic acid-based approaches have great potential for clinical application, their limitations should be considered, and further investigations are needed to.

• increase their stability,

In this perspective, several strategies are under investigation to efficiently deliver these molecules against BC-related ncRNA targets, such as synthetic ionizable lipids (LNP), including ASOs targeting MALAT1 and HOTAIR lncRNAs (201) or ZIF-90 (202) nanoparticles enveloping dual antisense oligonucleotide targeting miR-21/miR-155 to treat TNBC and inhibit metastasis.

Undoubtedly, research improvements in this field will provide more action strategies in the understanding of BC cancer biology, in its correct diagnosis, and in the development of personalized, targeted therapies that may also be helpful in the more aggressive forms of the disease, opening new avenues for precision medicine in a heterogeneous disease such as breast cancer.