Progress of siRNA Nanomedicines in Modulating the Microenvironment of Triple-Negative Breast Cancer

Abstract

Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype, and its poor prognosis, chemoresistance, and limited immunotherapy response are closely linked to pathological remodeling of the tumor microenvironment (TME). Targeting the TME has emerged as a promising strategy, but conventional agents such as monoclonal antibodies and small-molecule inhibitors are limited by undruggable targets, off-target toxicity, and poor tumor penetration. Small interfering RNA (siRNA) enables sequence-specific silencing of multiple TME-related genes—including TGF-β, PD-L1, and HIF-1α—offering a flexible approach to reverse immunosuppression, fibrosis, and hypoxia. However, poor in vivo stability and inefficient intratumoral delivery have hindered clinical translation. Recent advances in intelligent nanocarriers, including pH/enzyme-responsive and biomimetic membrane-coated systems, together with spatial omics-guided target selection, have substantially improved siRNA delivery and therapeutic efficacy. This review summarizes the immunoregulatory mechanisms and key targets within the TNBC microenvironment, the design principles and technological bottlenecks of siRNA nanodelivery systems, and emerging combination strategies with immune checkpoint inhibitors. It further discusses subtype-specific targeting needs, long-term resistance risks, and translational challenges from preclinical models to clinical applications, providing a systematic framework for siRNA nanomedicine-based remodeling of the TNBC microenvironment and highlighting critical directions for future precision combination therapy.

Introduction

Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype and is characterized by poor prognosis, rapid acquisition of chemoresistance, high rates of recurrence and metastasis, and limited response to immunotherapy.^1,2 These clinical challenges reflect not only the intrinsic invasiveness and genomic heterogeneity of TNBC cells but also pathological remodeling of the tumor microenvironment (TME). The aberrant TME promotes tumor proliferation and malignant progression by driving immune escape, abnormal angiogenesis, and metabolic reprogramming, thereby serving as a critical extrinsic determinant of drug resistance and metastatic maintenance.^3–5 Accordingly, therapeutic strategies have shifted from tumor cell-centric targeting toward comprehensive remodeling of the TME, which has become a frontier in basic and translational TNBC research.

Unlike hormone receptor-positive and human epidermal growth factor receptor 2 (HER2)-positive breast cancers, TNBC lacks actionable molecular targets, limiting the clinical benefit of conventional therapies.^3,6,7 Within the TME, cancer-associated fibroblasts (CAFs), immunosuppressive myeloid cells, and regulatory T cells (Tregs) establish a pro-tumor inflammatory network that, together with physicochemical abnormalities such as hypoxia and acidosis, suppresses antitumor immunity and promotes drug resistance, recurrence, and progression.^8,9 Targeting key TME components and their signaling pathways may therefore reduce tumor survival advantages and complement existing antitumor strategies. However, current TME-targeted approaches, including monoclonal antibodies and small-molecule inhibitors, are often constrained by the poor druggability of key targets, off-target toxicity, and inadequate tumor penetration.^10,11 Small interfering RNA (siRNA) offers sequence-specific gene silencing and can simultaneously regulate multiple genes involved in tumorigenesis, immunosuppression, and drug resistance within the TME.^12–14 Compared with conventional drugs, siRNA can target a broad range of otherwise undruggable targets with high design flexibility, specificity, and manageable off-target effects.^15–17 Nevertheless, poor in vivo stability and inefficient intratumoral delivery have long constrained clinical translation.¹⁷

Over the past two years, advances in nanomaterials and spatial omics have begun to address these delivery barriers and have driven multiple breakthroughs in siRNA-based targeting of the TNBC TME.^18–20 Microenvironment-responsive nanocarriers, such as pH/enzyme dual-sensitive liposomes and biomimetic membrane-coated systems, have been optimized to enhance siRNA stability, tumor-targeted accumulation, and deep penetration, overcoming limitations of conventional delivery platforms.^21–24 In parallel, single-cell sequencing and spatial transcriptomics have delineated cellular heterogeneity, spatial organization, and key regulatory targets within the TNBC TME, offering high-resolution omics evidence for precise siRNA target selection. Leveraging these advances, preclinical studies have demonstrated that siRNA-mediated TME remodeling can relieve the immunosuppressive state of TNBC and, when combined with immune checkpoint inhibitors, exert synergistic antitumor effects.²⁵ Nevertheless, several critical questions remain unresolved: whether distinct TNBC molecular subtypes require tailored TME-targeting strategies, whether long-term siRNA-mediated TME remodeling selects for resistant clones, and whether delivery improvements observed in animal models translate to clinically relevant models.

In contrast to previous reviews that have focused on the optimisation of a single delivery vehicle or the modification of nanomaterials,^26–29 this review examines how siRNA-based nanomedicines can systematically alleviate immunosuppression and metabolic dysregulation by precisely modulating interactions among immune cells, CAFs, and the extracellular matrix (ECM) in the TNBC microenvironment, thereby helping overcome therapeutic resistance and recurrence. The success of Patisiran in the treatment of hereditary transthyretin‑mediated amyloidosis with polyneuropathy (hATTR‑PN) associated with polyneuropathy, together with the development of inclisiran for a broad patient population rather than for a rare disease, also highlights the therapeutic potential of siRNA technology.^30,31 We focus on the tumor immune microenvironment of TNBC and integrate the logic of target selection, design principles of intelligent nanocarriers, and mechanisms of siRNA-mediated immune synergy. We discuss key targets, emerging delivery systems, and combination strategies for siRNA-mediated remodeling of the TNBC microenvironment, with emphasis on current limitations and translational barriers, aiming to provide a reference for basic research and clinical translation of precision combination therapies for TNBC.

Immunoregulatory Mechanisms and Therapeutic Targets in the TNBC Tumor Microenvironment

The immune regulation mechanisms and therapeutic targets of the tumor microenvironment in TNBC. The occurrence and development of TNBC are closely related to its TME, a complex ecosystem composed of various immune cells, stromal cells, and extracellular components.^32,33 In the TME, the functional states of various immune cells and their interaction networks profoundly impact tumor immune evasion and treatment response. Among these, the balance of polarization of tumor-associated macrophages, functional differentiation of T cell subpopulations, antigen presentation capacity of dendritic cells, and cytotoxic activity of natural killer cells collectively determine the intensity of the anti-tumor immune response.^34,35 Additionally, the abnormal expression of immune checkpoint molecules and stromal remodeling mediated by cancer-associated fibroblasts further exacerbate the formation of an immunosuppressive microenvironment. In-depth analysis of the regulatory mechanisms of these key components in the breast cancer TME not only helps in understanding the molecular basis of tumor immune evasion but also provides important targets for the development of novel immune-targeted treatment strategies.

Key Immune Cells

TAMs, as the most abundant immune cell population in the TME, can account for over 50% in breast cancer. TAMs are classified into two types: M1 and M2. M1 macrophages secrete pro-inflammatory cytokines and stimulate cytotoxic T lymphocyte responses against cancer cells, exhibiting pro-inflammatory and anti-tumor activities. M2 tumor-associated macrophages utilize multiple mechanisms to create an immunosuppressive microenvironment in breast cancer.³⁶ On one hand, they secrete factors like IL-10 and TGF-β to induce Treg expansion, while high expression of PD-L1/CD80 directly leads to T cell exhaustion; on the other hand, they recruit suppressive cells such as MDSCs through CCL2/CCL5, collectively contributing to the formation of an immunosuppressive environment.^37–39 Additionally, TAMs promote angiogenesis and ECM remodeling through VEGF/MMP-9, enhancing tumor invasiveness and creating a physical barrier that hinders CD8⁺ T cell infiltration.⁴⁰ Clinical studies have shown that M2-type TAMs dominate the breast cancer microenvironment, and their infiltration level is significantly associated with increased malignancy and poor prognosis.^41,42

In the complex regulatory network of the tumor immune microenvironment, T cells exhibit significant functional heterogeneity, and their dynamic balance directly determines the efficacy of the anti-tumor immune response.^43,44 CD8⁺ cytotoxic T cells, as the primary effector population, can directly kill tumor cells by secreting perforin, granzymes, and other effector molecules after recognizing tumor antigens through TCR, and produce pro-inflammatory factors like IFN-γ to activate macrophages’ anti-tumor functions.⁴⁵ However, in an immunosuppressive microenvironment, these effector T cells often exhibit a state of functional exhaustion, characterized by sustained high expression of inhibitory receptors like PD-1, TIM-3, and others.^46,47 In contrast, CD4⁺ T cell subsets demonstrate a more complex regulatory network: Th1 cells support CD8⁺ T cell responses by secreting IL-2 and IFN-γ, while Th2 cell-produced IL-4 and IL-13 may promote tumor progression.⁴⁸ Tregs, as a critical immune suppressive population, inhibit effector T cell function by high expression of CTLA-4 and secretion of inhibitory cytokines like IL-10 and TGF-β, promoting the formation of an immunosuppressive microenvironment.⁴⁹ Recent studies show that by precisely regulating Tregs activity, reshaping T cell memory subpopulations, or targeting tumor microenvironment metabolic abnormalities, the durability and effectiveness of T cell therapy can be significantly enhanced.^50–52

Dendritic cells (DCs) serve a dual role in immune regulation in breast cancer as specialized antigen-presenting cells in the TME.⁵³ Although DCs should present tumor antigens via MHC molecules and activate T cell responses, tumor-associated DCs often exhibit functional defects leading to decreased antigen presentation capacity.^54,55 Mature dendritic cells play a central role in the anti-tumor immune response. These cells express high levels of MHC II molecules (such as HLA-DR) and co-stimulatory molecules (CD80/CD86/CD40), effectively uptake, process, and present tumor antigens, stimulating initial T cell activation and proliferation.⁵⁶ Mature DCs promote differentiation of CD8⁺ T cells into cytotoxic T lymphocytes (CTLs) by secreting Th1 cytokines like IL-12, while also enhancing NK cell cytotoxic activity.^57,58

Another important innate immune cell, natural killer (NK) cells, play a crucial role in tumor immune surveillance.⁵⁹ They recognize stress ligands through surface activating receptors (such as NKG2D, DNAM-1), release perforin and granzymes to induce tumor cell killing, and secrete cytokines like IFN-γ to activate adaptive immune responses.^60,61 However, the tumor microenvironment can inhibit NK cell function through various mechanisms: downregulation of NKG2D expression by inhibitory factors like TGF-β; reduced cytotoxic activity in hypoxic conditions; overexpression of ligands like HLA-E on tumor cells that bind to inhibitory receptors.^62,63 In summary, the dynamic interplay among TAMs, T cells, DCs, and NK cells within the TNBC microenvironment forms a complex immunosuppressive network. To facilitate a clearer understanding of the spatial distribution and functional interactions of these key immune cell populations, this cellular landscape is schematically depicted in Figure 1A.

Figure 1.

(A) The TNBC microenvironment consists of tumor cells, exhausted CD8⁺ T cells and immunosuppressive Tregs, B cells (often within tertiary lymphoid structures), M2‑like macrophages, dysfunctional dendritic cells, aberrant endothelial cells, and heterogeneous CAFs. Together, these components form an immunosuppressive and pro‑tumorigenic network that drives progression, metastasis, and therapeutic resistance. (B) The cellular interaction regulatory network of major immune checkpoints in the microenvironment of TNBC. APCs, tumor cells, Tregs and macrophages interact via multiple immune checkpoint pathways (PD-1/PD-L1/2, CTLA-4/CD80/86, LAG-3/MHC, TIGIT/PVR, TIM-3/CEACAM-1, B7-H3, Vista); simultaneously, macrophages can regulate the secretion of the chemokines CXCL9/10 via the IFNγ-JAK-STAT signaling axis, thereby participating in tumor immune regulation.

Immune Checkpoint Molecules

Programmed death-ligand 1 (PD-L1) is a crucial immune checkpoint molecule on the surface of tumor cells that, upon binding to the PD-1 receptor on T cells, transmits inhibitory signals, leading to T cell exhaustion and apoptosis.^64,65 In breast cancer, the expression of PD-L1 shows significant subtype variability, with expression rates in TNBC reaching up to 20–30%.⁶⁶ Upregulation of PD-L1 can be induced by cytokines like IFN-γ or driven by genomic alterations such as PTEN loss.^67,68 Its high expression is closely associated with tumor immune evasion but also suggests potential responsiveness to PD-1/PD-L1 inhibitor therapy.

CD47 is a “don’t eat me” signal molecule widely expressed on the surface of tumor cells, which inhibits the phagocytic activity of macrophages by binding to the SIRPα receptor on the macrophage surface.^69,70 In breast cancer, CD47 is often overexpressed (especially in TNBC and triple-negative subtypes), aiding tumors in evading immune clearance. Its expression levels are associated with disease progression and poor prognosis.⁷¹

An increased proportion of CTLA-4⁺ regulatory T cells has been observed in the peripheral blood and tumour microenvironment of patients with TNBC, and their expression levels correlate positively with higher tumour grade and lymph node metastasis.⁷² The persistent overexpression of CTLA-4 on Treg cells has been identified as a key mechanism of immune evasion. CTLA-4 expression is primarily driven by FoxP3 and is further enhanced by TGF-β in the tumour microenvironment.⁷³ High levels of CTLA-4 expression are associated with a poor prognosis; however, it has also been developed as a therapeutic target for anti-CTLA-4 antibodies (such as ipilimumab). Notably, when used in combination with PD-1 inhibitors, it has been shown to significantly enhance the anti-tumour immune response.

In addition to CTLA-4, another inhibitory receptor of considerable interest, LAG-3, is also recognised as being expressed on the surface of activated T cells, NK cells and regulatory T cells; its primary ligand has been identified as MHC class II molecules.^74,75 LAG-3 inhibits the proliferation of effector T cells and the secretion of cytokines such as IFN-γ by transmitting inhibitory signals, thereby contributing to T-cell exhaustion.⁷⁶ In TNBC, high expression of LAG-3 has been shown to contribute to T-cell exhaustion and is closely associated with more advanced disease stages, distant metastases and poor clinical outcomes.⁷⁷ The upregulation of LAG-3 is primarily driven by sustained T-cell receptor stimulation and is further maintained by inflammatory cytokines such as IFN-γ in the tumour microenvironment.⁷⁸

TIM-3 is widely recognised as an inhibitory receptor expressed on the surface of exhausted CD8⁺ T cells, NK cells and myeloid immune cells; by binding to ligands such as galectin-9 and HMGB1, it suppresses anti-tumour immune responses and promotes the formation of an immunosuppressive microenvironment.^79,80 In TNBC, it has been observed that the upregulation of TIM-3 promotes M2-type macrophage polarisation and tumour invasion and metastasis. Given its crucial role in mediating immune evasion, TIM-3 has been identified as one of the key candidate targets for reversing resistance to immunotherapy.

B7-H3 (CD276) and B7-H4 (VTCN1) both belong to the B7-CD28 superfamily of co-inhibitory molecules and have been observed to be abnormally overexpressed in breast cancer.⁸¹ B7-H3 promotes immune evasion by inhibiting T-cell activation; its overexpression is associated with high tumour grade, metastasis and poor prognosis, and targeted drugs have demonstrated antitumour activity in clinical trials.⁸² B7-H4, on the other hand, negatively regulates the immune response by inhibiting T-cell proliferation;^83,84 its overexpression is associated with high histological grade, lymph node metastasis and shorter overall survival. Both are potential targets for immunotherapy, although B7-H4-targeted drugs are still at an early stage of development. There are complex interrelationships between immune checkpoint signalling pathways. Figure 1B illustrates the regulatory network of interactions in triple-negative breast cancer, where multiple immune checkpoint signalling pathways are formed between various immune cells.

Extracellular Matrix and CAFs

In the breast cancer TME, the extracellular matrix (ECM) and CAFs collaborate to construct an immunosuppressive barrier. CAFs excessively secrete ECM components such as type I/III collagen, fibronectin, and hyaluronic acid, forming a dense physical barrier that significantly hinders the infiltration of CD8⁺ T cells.⁸⁵ Additionally, CAFs remodel the ECM through the secretion of cell factors like TGF-β, IL-6, and MMPs, which not only activate immunosuppressive Tregs and M2-type TAMs but also produce degradation products (such as endothelin) with direct T cell inhibitory functions.^86,87 This ECM-CAFs interaction network further excludes T cells from the tumor edge through a CXCL12/SDF-1 chemotactic gradient, leading to a typical “immune exclusion” phenotype.⁸⁸ Studies have shown that CAF-induced ECM stiffening can upregulate PD-L1 expression 2.3 times and is significantly associated with immune therapy resistance.⁸⁹ Targeting the CAFs-ECM axis (such as FAP inhibitors or LOXL2 antibodies) can improve T cell infiltration depth by up to 3.2 times, providing new strategies to overcome immune resistance.^90,91

RNAi Nanodelivery Technology

RNAi technology, with its efficient and specific gene silencing capabilities, has emerged as one of the most promising breakthroughs in cancer therapy. SiRNA, as the core molecule of RNAi, can precisely target the mRNA of disease-causing genes and induce their degradation, providing a novel intervention strategy for cancers that are challenging to treat with traditional therapies, such as triple-negative breast cancer.⁹² However, the clinical application of siRNA still faces multiple challenges—from stability issues at the molecular level to delivery efficiency within cells, and to targeted accumulation in tumor tissues—each step impacts the ultimate therapeutic outcome.⁹³ In recent years, the rapid advancement of nanotechnology has provided innovative solutions to these bottlenecks: intelligent responsive lipid nanoparticles, multifunctional polymer carriers, inorganic nanoparticles with photothermal control, and exosome delivery systems with inherent targeting capabilities are continuously driving RNAi therapy towards higher efficiency and lower toxicity.^94,95

Principles and Advances in Gene Interference Based on RNAi Technology

SiRNA, as the core effector molecule of RNAi technology, achieves gene silencing by guiding the degradation of target mRNA in a sequence-specific manner. The mechanism begins with artificially synthesized double-stranded siRNA entering the cytoplasm through delivery systems, integrating into the RNA-induced silencing complex, and, mediated by the Argonaute 2 protein, the antisense strand precisely recognizes complementary mRNA and guides its cleavage, thereby disrupting gene expression.^96–98 Since the discovery of RNAi by Andrew Fire and Craig Mello in 1998, siRNA technology has experienced a breakthrough development from basic research to clinical translation.⁹⁹

Early studies established its gene silencing capability in mammalian cells and addressed stability and immunogenicity issues through chemical modifications. With breakthroughs in delivery technologies like lipid nanoparticles, the approval of the first siRNA drug, Patisiran, in 2018 marked the clinical maturity of this technology.¹⁰⁰ In the field of cancer therapy, siRNA has demonstrated significant efficacy in difficult-to-treat tumors such as TNBC by silencing oncogenes or immune modulatory molecules using targeted delivery systems (such as antibody-modified nanoparticles).^101,102

Early-stage studies have shown that combining siRNA with immunotherapy can significantly enhance anti-tumor effects. Current research focuses on improving delivery efficiency, developing intelligent responsive carriers, and exploring multi-targeted combination therapy approaches. Despite facing challenges such as delivery to solid tumors and large-scale production, siRNA technology has evolved into an important platform for gene therapy, providing a novel intervention strategy for precision cancer treatment.

Key Technological Bottlenecks Faced by siRNA Nanodelivery Systems

To enhance the therapeutic efficacy of siRNA nanodelivery systems and overcome the biological barriers they face is crucial. The clinical translation of siRNA nanodelivery systems encounters multi-level biological challenges, spanning the entire delivery process from the molecular level to tissue targeting. At the molecular level, the inherent instability and negative charge of siRNA molecules pose primary obstacles.¹⁰³ The phosphodiester backbone structure of siRNA is highly sensitive to nucleases commonly found in serum, and its strong hydrophilic and negative charge characteristics severely hinder transmembrane transport. Additionally, siRNA molecules are rapidly cleared by the kidneys. During the cellular delivery process, siRNA faces three critical barriers: firstly, its transmembrane transport efficiency is significantly limited; secondly, once entering cells via endocytic pathways, most siRNA molecules become trapped in endosomes; finally, siRNA molecules not released in a timely manner are at risk of lysosomal degradation. Furthermore, in terms of tissue targeting, in TNBC as a stroma-vascular tumor, the dense extracellular matrix significantly restricts the penetration depth of nanoparticles, abnormal vascular structures, and interstitial fluid pressures as high as 20–30 mmHg hinder the tumor accumulation of nanoparticles; tumor heterogeneity results in significantly different delivery efficiencies in various regions.¹⁰⁴

Due to the inherent deficiencies of naked siRNA, such as easy degradation and low transmembrane efficiency, clinical application typically relies on delivery vehicles for administration.^105,106 However, the introduction of carrier systems also brings new safety considerations: carrier materials may activate the complement system, leading to immunogenic reactions; after intravenous administration, the significant dose-limiting toxicity often results from first-pass effects in the liver; in addition, the long-term accumulation of carrier components may lead to organ-specific toxicity, all of which need to be systematically evaluated during drug development. The ultimate goal of overcoming these delivery barriers is to enable siRNA to reach the cytoplasm and assemble into the RNA-induced silencing complex (RISC), where the antisense strand guides sequence-specific cleavage of target mRNA. This RISC-mediated gene silencing pathway is schematically illustrated in Figure 2.

Figure 2.

Nanocarriers overcome the poor cellular permeability and nuclease susceptibility of free siRNA, shuttling it across biological barriers to reach target mRNAs for silencing. Nanocarrier‑mediated siRNA reprograms the tumor microenvironment by ablating checkpoint signals that sustain immune evasion, shifting the balance from immunosuppression to immune activation and enabling effective T‑cell‑dependent tumor regression.

Precise Therapeutic Advantages of siRNA

With its unique RNA interference mechanism, siRNA demonstrates significant advantages in precision therapy, precisely recognizing and degrading target mRNA through base-pairing, thereby disrupting the synthesis of pathogenic proteins at the source, effectively intervening with “undruggable” targets such as intracellular proteins and non-coding RNAs.¹⁰⁷ Compared to traditional monoclonal antibody therapy, siRNA has a broader range of targets and a deeper mechanism of action: monoclonal antibodies mainly act by binding specifically to cell surface receptors or soluble proteins in circulation, with their effects limited to extracellular and membrane-related targets; whereas siRNA can enter cells through effective delivery systems, specifically degrade target gene mRNA, thereby regulating protein expression at the post-transcriptional level.¹⁰⁸

SiRNA delivery systems enable precise treatment, acting through base-pairing with mRNA to silence the target gene—designing targeted siRNA allows for the silencing of the mRNA of interest, targeting intracellular targets that are difficult for traditional drugs to intervene with, including transcription factors, mutant proteins, and other “undruggable” targets.¹⁰⁹ In addition, siRNA can achieve comprehensive intervention of complex disease pathways through multi-gene silencing. In terms of therapeutic characteristics, siRNA demonstrates significant long-lasting effects; a single dose can maintain therapeutic effects for weeks to months, providing a clear advantage over antibody therapies that require frequent dosing.¹¹⁰ Additionally, the chemical production process of siRNA is relatively straightforward, cost-effective, and suitable for the treatment management of chronic diseases. However, monoclonal antibodies still play an irreplaceable role in mediating immune effector functions such as antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC).¹¹¹ The unmodified siRNA also has many disadvantages that need to be overcome. It is worth noting that in immune checkpoint regulation, siRNA, by directly degrading target mRNA to block protein synthesis, can not only inhibit the expression of membrane proteins but also reduce the generation of intracellular proteins. This strategy of source blockade may lead to more thorough and long-lasting immune modulation effects. In the future, combined treatment strategies that combine siRNA’s gene silencing capabilities with the targeted specificity of antibodies, such as novel treatment modalities like antibody-siRNA conjugates, are expected to open up new avenues for tumor immunotherapy, achieving more precise and efficient control of the TME.¹¹²

Research Progress of Common Nanocarriers in RNAi Delivery Systems

Liposome

Liposomes are widely used for siRNA delivery. siRNA can be encapsulated within an aqueous core or bound to the lipid bilayer, relying on the lipid shell to protect against nuclease degradation and utilizing fusion-type or ionizable lipids to achieve efficient endosomal escape. These carriers exhibit good biocompatibility and are fully biodegradable; they are also easily surface-functionalized to enable active targeting and prolonged circulation. The representative drug Patisiran was approved by the FDA in 2018 and is the world’s first siRNA-liposomal nanoparticle drug, marking the transition of lipid carriers from preclinical research to clinical application.^113,114 DLin-MC3-DMA is a classic ionizable lipid that is nearly electrically neutral at physiological pH to reduce systemic toxicity; it becomes protonated in the acidic environment of endosomes, promoting efficient endosomal escape.¹¹⁴ It is the core lipid component of Patisiran, an FDA-approved siRNA drug. Owing to the strong negative charge of siRNA molecules, their cellular transmembrane delivery is greatly restricted. Cationic liposomes (eg, DOTAP/DOPE) can efficiently encapsulate siRNA via electrostatic interactions to form stable nucleic acid‑lipid complexes. However, the net positive charge on such carriers’ surfaces may induce toxic side effects like cell membrane damage and inflammatory reactions.¹¹⁵ By introducing polyethylene glycol (PEG) modifications, stealth liposomes can be constructed, significantly improving pharmacokinetic properties: prolonging blood circulation time (3–5 times half-life extension) and reducing reticuloendothelial system clearance. Further functional design includes integrating pH-sensitive lipid components (such as DSPE-Hydrazone-PEG), enabling specific degradation in the characteristic acidic microenvironment of tumor tissues (pH 6.5–6.8), achieving spatiotemporal control release. These engineered modification strategies systematically optimize the carrier’s biological distribution, cellular uptake, and endosomal escape efficiency, resulting in a 10–20 times increase in therapeutic index.¹¹⁶

Although several liposome-based (including LNP) siRNA delivery systems have entered clinical trials, the use of liposomes for siRNA delivery still faces significant limitations: following intravenous administration, liposomes readily bind to plasma proteins and aggregate, leading to premature dissociation of the siRNA and its degradation by nucleases;¹¹⁷ simultaneously, the formulations are easily cleared by the reticuloendothelial system and naturally tend to accumulate in the liver, making it difficult to achieve extrahepatic targeted delivery. Although PEG modification can prolong the circulation half-life, it reduces cellular uptake efficiency and induces the antibody-mediated accelerated clearance (ABC) effect; cationic lipids possess inherent cytotoxicity, and both siRNA and carrier components can easily activate innate immune pathways, triggering inflammatory responses.¹¹⁸ Novel liposomes hold promise for overcoming these limitations and achieving more efficient delivery. Ismail et al developed a poly(phosphorylcholine)-modified liposome (PMPC-Lipo) siRNA delivery system. Its core innovation lies in the amphoteric PMPC modification, which enables penetration of the blood-brain barrier via nAChRs and ChTs receptors without the need for additional targeting ligands, while avoiding the ABC effect. Compared to conventional PEGylated liposomes, this carrier avoids the rapid hepatic clearance caused by anti-PEG antibodies and complement activation, has a longer circulation time, and can penetrate deep into gliomas to achieve highly efficient siRNA delivery.¹¹⁹ Gang Guo’s team developed dopamine‑modified lipid‑polymer hybrid nanoparticles (lip@PSD‑siP) for co‑delivery of siPD‑L1 and photothermal agent IR780. The core innovation is dopamine‑mediated chemical modification of polyethyleneimine (PEI), which greatly reduces PEI’s intrinsic high cytotoxicity. Its distinctive lipid‑polymer hybrid structure allows lysosomal escape after cellular uptake and glutathione (GSH)‑triggered siRNA release. Compared with classic transfection reagent Lipo3000‑siP and unmodified lip@PEI‑siP, this nanocarrier achieves more potent PD‑L1 knockdown. Coupled with mild photothermal effects, it induces tumor immunogenic cell death (ICD), facilitates dendritic cell maturation, boosts tumor‑infiltrating CD4⁺/CD8⁺ T cells, downregulates immunosuppressive regulatory T cells (Tregs), and elicits long‑lasting antitumor immune memory.¹²⁰ The key challenges of liposomes—including protein adsorption, RES clearance, the ABC effect, and cationic lipid cytotoxicity—along with those of other nanocarrier platforms, are summarized in Table 1.

Table 1.

Key Bottlenecks in Mainstream siRNA Delivery Vectors

Carrier Type	Key Challenges
LNPs	Prone to aggregation due to adsorption by plasma proteins and premature degradation of siRNA; prone to clearance by the reticuloendothelial system; PEG modification tends to induce the ABC effect; cationic lipids exhibit cytotoxicity;
Polymer NPs	Poor biodegradability and long-term toxicity (due to positive charge + polymer); Weak cellular uptake and insufficient targeting; Insufficient in vivo stability
Inorganic Nanoparticles	Poor biocompatibility; most are non‑degradable, raising long‑term toxicity concerns Low cellular uptake efficiency Some materials (eg, silver, quantum dots) show high toxicity
Exosomes	Complex sourcing and preparation; difficult to standardize Low drug loading capacity and poor in vivo stability Low siRNA encapsulation efficiency
Dendrimers (eg, PAMAM)	Cytotoxicity increases with generation Complex synthesis routes Limited clinical translation

Polymer NPs

Currently, advanced polymer nanocarriers used for siRNA delivery mainly include the following systems: FDA-approved PLGA biodegradable materials achieve efficient siRNA encapsulation and controlled release through process optimization; chitosan derivatives exhibit excellent siRNA compaction ability and tumor microenvironment responsive properties; branched polyethyleneimine (such as 25kDa bPEI) demonstrate outstanding endosomal escape performance; and the new generation dendritic polymers (such as 4th generation PAMAM) achieve efficient siRNA loading at the single-molecule level due to precise 3–5nm nanostructure and abundant surface functional groups.¹²¹ The latest research integrates various carrier advantages through the construction of intelligent hybrid systems (such as PLGA-PAMAM core-shell structures, chitosan-dendrimer composite materials) and develops multi-stimuli-responsive carriers (such as ROS/pH/enzyme triple-responsive systems), enhancing siRNA tumor tissue accumulation efficiency by 3–5 times.¹²² Davis et al developed cyclodextrin-based cationic polymer nanoparticles (CDP-NPs) for systemic siRNA delivery, enhancing tumor enrichment through transferrin-mediated targeting and PEG shielding. Compared to unmodified cationic polymers, this system reduces nonspecific interactions and, for the first time, achieves RNA interference at the tumor site in humans; however, it still suffers from polymer-intrinsic cytotoxicity and defects to serum protein corona formation.¹²³ Zhang et al developed a self-assembled dendritic molecular nano system decorated with PD-L1 antibodies for siRNA delivery, targeting phosphoinositide-dependent protein kinase 1 (PDK1), an enzyme involved in cancer metabolism and metastasis.¹²⁴ Intended to target PD-L1, this system aims to deliver siRNA specifically for cancer, regulating tumor metabolism, while promoting immunotherapy based on the programmed cell death 1 (PD-1)/PD-L1 pathway.

Inorganic Nanomaterials

Inorganic nanomaterials combined with siRNA demonstrate unique delivery advantages: Mesoporous silica nanoparticles leverage their regular pore structure (pore size of 2–10nm) and ultra-high surface area (>1000m²/g) to achieve high drug loading (>200μg/mg) and controlled release of siRNA; Their surface-rich silanol groups are easily modified with targeting ligands (such as folic acid, RGD peptides).¹²⁵ However, the clinical translation of inorganic nanocarriers is constrained by poor biodegradability, which raises long-term toxicity concerns. Table 2 provides a broader comparison of the advantages, limitations, and representative outcomes of various siRNA nanocarrier platforms. Gold nanoparticles (10–50nm) achieve stable siRNA loading through thiol chemistry, and their unique photothermal conversion performance (photothermal efficiency >90%) not only promotes endosomal escape but also achieves near-infrared light-triggered spatiotemporal release, enhancing tumor inhibition rates by 2–3 times.¹²⁶ Rebolledo et al developed a phosphonate-modified modular mesoporous silica nanoparticle (MSN-PEI) delivery system for combined chemotherapy and gene therapy in triple-negative breast cancer. This platform can co-load two chemotherapeutic agents as well as a nucleic acid drug targeting anti-apoptotic genes. Compared to other cationic polymer nanocarriers, this carrier enables efficient nucleic acid binding and protection against nucleases while reducing the cytotoxicity associated with the polymer.¹²⁷ Yu et al developed a biomimetic nano-co-delivery system called RB/siPEG10@PMSNs-CM, designed to address CDK4/6 inhibitor resistance in breast cancer. This system co-loads ribociclib (RB) and PEG10 small interfering RNA (siPEG10)—the latter specifically reverses drug resistance and inhibits tumor metastasis by targeting and silencing the PEG10 gene. As the carrier, mesoporous silica nanoparticles (MSNs) were used as the core, which feature high drug-loading capacity and pH-responsive release; they were modified with PEG-β-cyclodextrin to enhance stability and biocompatibility; and they were coated with breast cancer cell membrane (CM), endowing the nanoparticles with homogeneous targeting capability and enabling tumor-specific accumulation.¹²⁸ Inorganic nanomaterials used for siRNA delivery are endowed with excellent physicochemical stability and tunable pore structures, enabling high drug loading and convenient surface modification to achieve targeted and stimulus-responsive release; siRNA is physically protected from nuclease degradation by inorganic scaffold, and this type of non-viral delivery platform has also been shown to possess low intrinsic immunogenicity and the potential for integrated diagnosis and therapy.

Table 2.

Comparison of Nanocarrier Packaging Modes and Properties for siRNA Delivery Systems

Nanocarrier	Dominance	Outcome
liposome	High encapsulation rate, flexible target modification, higher biocompatibility	A sialic acid-lipid derivative co-loaded liposome targets neutrophils and suppresses neutrophil-mediated pro-inflammatory and pro-metastatic processes.129 Natural peptide ligands of lipoproteins cross the blood-brain barrier.130
Polymer nanoparticles	Controlled degradation with flexible functional modifications	Polymeric nanoparticles constructed through metal and pH-responsive phenylborate bonding for combined potentiation of metal immunotherapy and chemotherapy.131
Inorganic nanomaterials	High drug carrying capacity, photothermal synergistic therapy	The use of gold nanoparticles with a multifunctional carrier core for electrostatic adsorption of Bcl-2 siRNA and intercalated binding of DOX enabled highly efficient gene-chemotherapy synergistic therapy.132
Exosome/cell membrane	Natural targeting, low immunogenicity	Significant inhibition of breast cancer progression using exosomes of M1 TAMs as carriers encapsulating siSIRPα and siSTAT6 in combination with PD-L1 inhibitors.133
dendrimer	High single-molecule loading for precise dimensional control	After loading siRNA, the MUC1 aptamer-modified dendritic multilayer nanoconjugates showed the most pronounced inhibition of target gene expression by the targeted dendritic polymers compared to the non-targeted dendritic polymers and liposome-2000.134
Multi-functional composite nanosystems	Integration of the strengths of each carrier	Cationic polymer PEI and lipid nanoparticles were co-formulated with siRNA to form lipid polyplexes (LPPs) for siRNA delivery.135

Exosomes/Cell Membrane-Coated NPs

Exosomes/cell membrane carriers combined with siRNA exhibit unique biological advantages: Naturally sourced exosomes retain targeting molecules from the parent cells’ surface (such as integrins, adhesion proteins), enabling specific recognition of homologous tumor tissues (increasing targeting efficiency by 2–3 times);¹³⁶ Their phospholipid bilayer structure provides excellent biocompatibility and low immunogenicity, especially engineered exosomes effectively penetrate the blood-brain barrier, achieving significant gene silencing effects in breast cancer brain metastasis models. Zhao et al developed a biomimetic nanoparticle CBSA/siS100A4@Exosome, which couples siS100A4 with cationic bovine serum albumin (CBSA) and envelops exosome membranes to form uniform particles.¹³⁷ Experimental results demonstrate that this nanoparticle exhibits higher targeting to the premetastatic niche (PMN) for lung metastases compared to traditional liposomes, effectively inhibiting postoperative metastasis in TNBC (Figure 3 and Table 1). One study developed a cholesterol-rich bovine milk-derived exosome-based siRNA delivery vector. The core innovation lies in increasing the cholesterol content of the exosome membrane, which triggers cell membrane fusion to enable direct delivery of siRNA into the cytoplasm. Compared to native exosomes, cholesterol-depleted exosomes, and conventional transfection reagents, this carrier induces stronger gene silencing and pro-apoptotic effects in vitro; following oral or intravenous administration to tumor-bearing mice, its antitumor activity outperforms that of lipid nanoparticles, providing a safe and convenient bio-nanocarrier for siRNA delivery.¹²³ The major siRNA nanocarrier classes, including exosome-based systems, are illustrated in Figure 3, and the key challenges of exosomes—including complex sourcing and preparation, low drug loading capacity, and suboptimal in vivo stability—along with those of other nanocarrier platforms, are summarized in Table 1. Representative examples of the application of these nanocarrier platforms in siRNA delivery are compiled in Table 2.

Figure 3.

(A) siRNA offers exceptional sequence‑specificity, potent silencing of previously undruggable genes, rapid programmability based on target sequences, and synergistic compatibility with various therapeutics. Nanocarriers further provide active targeting, protection from enzymatic degradation, stimuli‑responsive controlled release, and co‑loading capacity for combination therapy. When siRNA is encapsulated, the system enables co‑delivery with other agents, precision targeting to desired cells, environment‑triggered liberation at lesions, and efficient endosomal escape via proton‑sponge or membrane‑fusion effects—collectively overcoming biological barriers and maximizing silencing efficacy. (B) Major siRNA nanocarriers include lipid nanoparticles (LNPs, eg, patisiran), cationic polymers (PEI, PLGA, dendrimers), inorganic materials (mesoporous silica, gold nanoparticles), and exosomes/EVs. These systems protect siRNA from degradation, enable active targeting, facilitate endosomal escape, and allow co‑delivery with other therapeutics.

Challenges and Potential Solutions for siRNA Delivery Systems

As outlined in Key Technological Bottlenecks Faced by siRNA Nanodelivery Systems, the sequential barriers of nuclease degradation, renal clearance, innate immune activation, inefficient endosomal escape, and poor tumor penetration in TNBC cannot be addressed by any single carrier. A logical strategy therefore begins with molecular stabilization, since serum stability is a prerequisite for all downstream delivery steps. Chemical modifications such as 2′-OMe, 2′-F, and phosphorothioate backbones reduce nuclease degradation and TLR3/7/8 activation, while direct ligand–siRNA conjugates provide a carrier-free route for selective tumor uptake.^138,139

Once stable siRNA reaches the target tissue, intracellular delivery becomes the limiting step. Ionizable lipids with tuned pKa address this by promoting endosomal escape, and their combination with targeting ligands and tumor-penetrating peptides such as iRGD and LyP-1 further improves tumor accumulation and stromal penetration.¹⁴⁰ Microenvironment-responsive elements triggered by pH, glutathione, MMP-2/9, or hypoxia extend this logic by restricting release to the tumor site and reducing systemic toxicity, a feature particularly well suited to the protease-rich, hypoxic TNBC microenvironment.^141,142

At the systemic level, biomimetic platforms such as cell membrane-coated nanoparticles and engineered exosomes offer complementary advantages in immune evasion and homologous targeting,¹⁴³ while local delivery through injectable hydrogels or microneedle patches bypasses systemic barriers altogether. Combination with chemotherapy, immune checkpoint blockade, or photothermal/photodynamic therapy provides additional synergistic antitumor effects.¹⁴⁴

Together, these approaches form an integrated strategy: molecular modification ensures stability, carrier engineering controls intracellular fate, and biomimetic or local administration overcomes systemic and tissue-level barriers. Rational integration of these complementary strategies holds promise to enhance the efficacy, specificity, and safety of siRNA delivery for TNBC microenvironment modulation.

Research Progress on Precise Regulation of Breast Cancer TME Based on RNAi Delivery Systems

RNA interference technology regulates protein expression by selectively degrading target mRNA. Specific targeting of the TME with siRNA molecules can effectively inhibit the abnormal expression of genes related to proliferation, drug resistance, and metastasis, achieving an anti-tumor effect.^5,12,145 Notably, this technology can not only directly act on tumor cells but also reshape the secretion profile of stromal cells, disrupting the crosstalk between the tumor and the microenvironment, which is crucial in inhibiting tumor progression.^146,147 This section will systematically elucidate the latest research breakthroughsand therapeutic potential of siRNA technology in regulating the breast cancer microenvironment.

RNAi Technology Targeting TAMs

TAMs are one of the most abundant populations of immune cells in the tumor microenvironment. Among them, M2-type TAMs play a key role in tumor progression and immune evasion by secreting immunosuppressive cytokines, promoting angiogenesis, and suppressing T-cell function.^148,149 Currently, immunotherapy strategies targeting TAMs primarily focus on the following areas: (1) inhibiting TAM recruitment and survival; (2) blocking the immunosuppressive function of M2-type TAMs; (3) reprogramming M2-type TAMs into the antitumor M1 phenotype; (4) utilizing TAMs as carriers for the delivery of therapeutic drugs or genes.^150,151 Based on this, strategies using RNAi technology to achieve M2-to-M1 polarization are gradually gaining attention, with the core approach being to reverse the immunosuppressive microenvironment by silencing specific genes. M2-type pro-tumor macrophages (TAMs) overexpress IL4R. In Byungheon Lee’s study, M1 macrophage-derived exosomes (IL4R-Exo(si/mi)) were utilized to target the reprogramming of TAMs.¹⁵² These exosomes were modified with an IL4R-binding peptide to target M2-type TAMs and loaded with NF-κB p50 siRNA and miR-511-3p, effectively internalizing into TAMs, downregulating M2 markers, and promoting M1 polarization.¹⁵³ In vivo experiments showed that IL4R-Exo(si/mi) selectively enriched in tumor tissues significantly inhibited tumor growth, reduced immunosuppressive factors, and enhanced anti-tumor immune responses. VEGF and PlGF are overexpressed in breast cancer cells and M2-TAMs, collectively promoting tumor progression and immunosuppression.¹⁵⁴ To address this, mannose and PEG dual-modified pH-responsive nanoparticles (PEG-MT/PC NPs) were developed for targeted delivery of siVEGF/siPIGF to M2-TAMs and tumor cells. These NPs, aided by PC’s charge reversal, achieved endosomal escape and siRNA release, significantly inhibiting tumor growth and metastasis. Leslie M. Shaw’s team developed a chemically modified siRNA conjugated to an albumin-binding dendrimer to target insulin receptor substrate 2 (IRS2) in triple-negative breast cancer. In an immunocompetent in situ tumor model, this conjugate enabled efficient delivery of siRNA to tumor cells, the stroma, and immune cells, leading to IRS2 silencing, which significantly inhibited tumor growth and modulated the microenvironment. However, it remains unclear how different cell populations in the TME differ in their uptake of siRNA and how this contributes to the therapeutic response; furthermore, the effects of long-term disruption of IRS2’s physiological functions in normal tissues have not been fully evaluated.¹⁵⁵ Generally, the effects of nanoparticles on the TME are not unidimensional but rather produce broad regulatory effects through multiple targets and pathways. For example, the lipid nanoparticle (LNP@siFmr1) developed by Bai et al, which carries small interfering RNA (siFmr1) targeting the Fmr1 gene, effectively silenced the Fmr1 gene and downregulated the expression of the fragile X intellectual disability protein (FMRP) in a mouse model of in situ breast cancer, systematically reshaped the TME: on the one hand, by recruiting and activating CD8⁺ T cells, and on the other hand, by promoting the polarization of immunostimulatory macrophages and the maturation of DCs, synergistically reversing the immunosuppressive state and thereby significantly inhibiting tumor growth.²⁵

RNAi Technology Targeting CAFs Cells

Although ligand-modified active targeting strategies are designed to enhance the tumour enrichment of siRNA, the high interstitial pressure and dense extracellular matrix within solid tumours often significantly undermine their actual synergistic effects; whilst the targeting advantages reported in many studies are significant in animal models, they are difficult to replicate in more complex pathological environments. CAFs are not only major drivers of stromal remodeling but also profoundly influence chemoresistance and immune evasion through various paracrine signaling and metabolic interactions.^156,157 In recent years, RNAi strategies targeting CAFs have gradually garnered attention; their core principle lies in silencing specific functional genes in CAFs to remodel the microenvironment and restore treatment sensitivity. However, the high heterogeneity and phenotypic plasticity of CAFs often make it difficult for single-gene silencing strategies to achieve lasting therapeutic effects. The chemotherapy drug etoposide induces ROS in the tumor microenvironment’s CAFs, activating autophagy, which in turn promotes exosome secretion.¹⁵⁸ These exosomes protect CAFs from ROS damage and transmit CXCR4 and c-Myc signals to breast cancer cells, enhancing their resistance to EPI. However, siRNA targeting ATG5 can inhibit autophagy initiation, thereby reducing exosome release from CAFs. TGF-β secreted by CAFs and tumor cells promotes vascular abnormalities, stromal fibrosis, and immunosuppression. Yang et al developed a dual-targeting nano-platform (HA-LSL/siTGF-β) with HAase/GSH-responsive release characteristics for delivering siTGF-β. This platform mediated by HAase-triggered endosomal escape and GSH-mediated siRNA release effectively silenced the TGF-β signaling pathway in stromal and tumor cells, significantly reducing matrix density, enhancing nano-drug penetration, improving tumor oxygenation, and T cell infiltration.¹⁵⁹ As illustrated in Figure 4A, the HA-LSL/siTGF-β platform achieves siRNA delivery through a two-step responsive mechanism: HAase triggers deshielding of the HA shell to enhance cellular uptake, while the high GSH concentration in the cytoplasm promotes disulfide bond cleavage to release siTGF-β. This dual-responsive design ensures efficient silencing in both tumor cells and CAFs, thereby disrupting TGF-β signaling and converting the immunosuppressive TME into an immune-permissive state. The combination with PD-1 inhibitors further amplifies T cell-mediated antitumor immunity (Figure 4B–D). Targeting CAFs as an anti-stromal therapeutic strategy has been extensively explored, particularly in tumors characterized by connective tissue proliferation, such as hepatocellular carcinoma and pancreatic cancer. In these malignancies, the dense fibrotic stroma generated by CAFs not only forms a physical barrier that hinders drug penetration into deep tumor regions but also creates a pro-survival microenvironment, thereby driving treatment resistance.^160,161 Professor Leonard Kaps’s team has made a series of advances in the treatment of stromal-driven hepatocellular carcinoma. First, the team developed cationic nanogel (NG) particles based on biodegradable polycarbonate as a siRNA delivery system targeting MFAP-5, a protein expressed in CAFs. In a clinically relevant HCC mouse model, following systemic administration, this system demonstrated significant antitumor efficacy by inhibiting tumor angiogenesis, providing preliminary evidence of the mechanistic value and translational potential of MFAP-5 RNAi in anti-stromal therapy.¹⁶² On the other hand, the team also engineered multi-compartment polyionic complex micelles (PICMs) based on a triblock peptide (pSar-b-pGlu(OBn)-b-pLys) for the co-delivery of anti-MFAP-5 siRNA and the endosome escape enhancer desloratadine (DES). In the same HCC model, this PICMs platform achieved approximately 60% knockdown efficiency of MFAP-5 in CAFs and significantly reduced liver tumor burden (by approximately 80%), with no significant toxic reactions observed.¹⁶³

Figure 4.

Anti-tumour efficacy of the HA-LSL/siTGF‑β dual-targeted delivery system in a mouse breast cancer model. (A) Schematic representation of HA-LSL/siTGF‑β nanoparticles: these nanoparticles target tumour cells overexpressing CD44 via hyaluronan, release siRNA in response to tumour-derived HA-lyase and GSH, and promote vascular normalisation, T-cell infiltration, alleviation of hypoxia, and reversal of the immunosuppressive tumour microenvironment. (B) Tumour growth curves in mice treated with saline, HA‑LSL/siScr, HA‑LCL/siTGF‑β, HA‑LSL/siTGF‑β combined with anti-PD‑L1, and HA‑LCL/siTGF‑β combined with anti-PD‑L1. (C) Tumour suppression rate; (D) Kaplan–Meier survival curves. Compared with the control groups in each set, the combination therapy of HA‑LSL/siTGF‑β and anti-PD‑L1 demonstrated the strongest antitumour effect and prolonged survival.¹⁵⁹ The down arrow indicates a decrease. One-way ANOVA was used for statistical analysis. ***p < 0.001. Copyright 2023 Elsevier.

Silencing Immune Checkpoint Gene CD47

CD47, a key signaling molecule for tumor immune evasion due to its “don’t eat me” function, has become a hot topic in immune surveillance research. Regulating CD47 gene expression can effectively inhibit a tumor’s immune evasion capabilities.¹⁶⁴ Yu et al developed a PEG-PLGA nanoparticle decorated with iron-polyphenol (NP_siCD47@Fe-TA), capable of delivering siCD47 while inducing iron-mediated cell death in tumor cells.¹⁶⁵ This nano system significantly enhanced macrophage tumor phagocytosis by downregulating CD47 expression on tumor cells and exposing the “eat me” signal, effectively inhibiting tumor growth and metastasis. However, this system relies on passive accumulation mediated by the enhanced permeability and retention effect, without active tumor-homing capability, which poses substantial challenges for its clinical translation. Li et al developed a cancer cell membrane-camouflaged ultra-small iron oxide nanoparticle-loaded polyethyleneimine nanogel system (CM-USIO-NG) for co-delivering docetaxel and siCD47. This system effectively downregulated the tumor cell “don’t eat me” signal and promoted M1 polarization of macrophages, combining immunogenic cell death induced by chemotherapy and immune checkpoint blockade to achieve a significant anti-tumor effect and joint functionality of magnetic resonance imaging.¹⁶⁶ This novel composite nanoparticle delivery system, by loading siCD47, overcomes the functional limitations of single siCD47 and significantly enhances anti-tumor efficacy (Figure 5). Furthermore, CD47 expression typically exhibits significant heterogeneity between tumors and even within individual tumors. At the same time, a growing body of evidence suggests that blocking CD47 may, through compensatory feedback loops (such as the secretion of IFN-γ by activated macrophages), produce a seemingly paradoxical phenomenon in which alternative immune checkpoint molecules, such as PD-L1, are upregulated.¹⁶⁷ This mechanistic interaction (crosstalk) suggests that siRNA-mediated CD47 silencing alone may be insufficient to achieve durable tumor control, as residual tumor cells can readily evade immune surveillance through this complementary pathway.

Figure 5.

Synergistic anti-tumour and anti-metastatic effects of Fe-TA-modified PEG-PLGA nanoparticles (NP_siCD47@Fe-TA) loaded with siCD47. (A) Schematic representation of the nanoparticle formulation, comprising a PEG-PLGA polymer embedded with DOTAP, MC3 and siCD47, surface-modified with a trivalent iron–tannic acid (Fe-TA) coordination complex to combine ferrocide-induced cell death with CD47-blockade-mediated immunotherapy. (B) Tumour weight measurements at the end of the study in tumour-bearing mice across treatment groups (n = 5–6 per group). Groups: G1, PBS (control); G2, NP_siCD47 (without Fe‑TA); G3, NP_siNC@Fe‑TA (random-sequence siRNA control group containing Fe‑TA); G4, NP_siCD47@Fe‑TA (complete formulation). (C) Representative images of lung metastatic nodules in each group. (D) Haematoxylin and eosin (HE) staining images showing the number of lung metastatic nodules. The NP_siCD47@Fe‑TA group exhibited the lowest tumour burden, the fewest metastatic lesions, and the longest overall survival.¹⁶⁵ The black downward-pointing arrow indicates the location of the lung nodule. One-way ANOVA was used for statistical analysis. *p < 0.05, ****p < 0.0001. Copyright 2025 Wiley-VCH Verlag.

Silencing Immune Checkpoint Gene PD-L1

Since the wide clinical application of PD-L1 inhibitors, significant progress has been made in targeting PD-L1 treatment strategies, including the development of monoclonal antibody drugs and gene silencing technologies based on RNAi. Current research focuses primarily on two areas: first, modifying nanocarriers to improve the tumor-targeted delivery of PD-L1 siRNA; and second, developing combination therapies that enhance antitumor activity while reducing systemic adverse reactions. Nevertheless, significant challenges remain regarding the biocompatibility of the carriers, the stability of siRNA in vivo, and the consistency of efficacy across different tumor subtypes.

In response to these challenges, several researchers have conducted representative studies: Du et al developed a novel polyethyleneimine-modified neutrophil elastase (PEI-elastase) nanoparticle delivery system for co-delivering PD-L1 siRNA and elastase.¹⁶⁸ By encapsulating PEI-elastase/PD-L1 siRNA complexes in pH-responsive anionic liposomes, an intelligent delivery system P-E/S Lip was constructed. Experimental results showed that compared to single therapy, P-E/S Lip rapidly released its payload in the tumor acidic microenvironment: on one hand, PEI-elastase exerted strong tumor-killing effects and induced immunogenic cell death; on the other hand, PEI-mediated siRNA delivery effectively reduced tumor cell PD-L1 expression. In a 4T1 tumor model, P-E/S Lip not only significantly inhibited primary tumor growth but also suppressed distant metastases by activating a systemic immune response. Although the in vivo results are encouraging, the potential immunotoxicity associated with high doses of exogenous elastase continues to limit its further clinical translation. To overcome photodynamic therapy (PDT)-mediated compensatory PD-L1 upregulation, Fang et al developed a macrophage membrane-camouflaged spermine-disulfide bond nano-carrier, co-loaded with siPD-L1 and a photosensitizer.¹⁶⁹ The study found that while photodynamic therapy induced PD-L1 upregulation, this nano system accelerated siRNA release through ROS, effectively counteracting the PDT-induced PD-L1 elevation. In a 4T1 breast cancer model, this “self-synergistic” strategy significantly suppressed PD-L1 expression and enhanced anti-tumor immune responses. However, the variable photopenetration depth in solid tumors remains a key limitation to the broad applicability of this combinatorial phototherapy–siRNA approach. With the ongoing advances in photodynamic therapy, its integration with siRNA-based treatment holds promise for meaningful breakthroughs. Song et al developed a pH/reduction dual-responsive co-delivery system based on chitosan-coated hyaluronic acid micelles (R/C/D@HAssOA) for simultaneously loading doxorubicin (DOX) and siPD-L1.¹⁷⁰ This nano system exhibited excellent drug loading performance, colloidal stability, and biocompatibility, responding to the TME acidic pH and high glutathione (GSH) levels to achieve controlled drug release. The results showed that through CD44 receptor targeting mediated by hyaluronic acid (HA), R/C/D@HAssOA significantly promoted the uptake of DOX and siPD-L1 in tumor cells. SiPD-L1 effectively silenced PD-L1 expression, promoted CD4⁺/CD8⁺ T cell tumor infiltration, while DOX exerted direct cytotoxic effects. In vivo anti-tumor evaluations demonstrated that compared to free DOX and single DOX-loaded micelles, R/C/D@HAssOA exhibited significantly enhanced therapeutic effects in a breast cancer model, showing stronger tumor growth inhibition and more pronounced immune cell activation. These findings confirm that this dual responsive micelle system, through the synergistic action of combining chemotherapy drugs and immunotherapy siRNA, provides new insights for developing highly effective combined anti-tumor strategies (Figure 6 and Table 3). Gong et al developed siRNA PD-L1@HA-ZIF-8 nanoparticles, which encapsulate PD-L1 siRNA via electrostatic interactions and are modified with hyaluronic acid. After accumulating at the tumor site, the nanoparticles degrade to release Zn²⁺ and PD-L1 siRNA: Zn²⁺ induces ROS production, triggers pyroptosis, induces ICD, releases DAMPs, and promotes DC maturation, thereby converting a “cold” tumor into a “hot” tumor; PD-L1 siRNA downregulates PD-L1 expression in tumor cells, reverses immune evasion, and restores T-cell cytotoxicity.¹⁷¹ This nanosystem is simple to prepare; by using MOFs to load siRNA and incorporating HA for targeting, it has yielded satisfactory results in both in vitro and in vivo evaluations, clearly demonstrating the significant feasibility and development potential of the siRNA gene therapy strategy. The molecular targets and regulatory mechanisms of these siRNA-mediated interventions across different TME components are schematically summarized in Table 3.

Figure 6.

Schematic illustration of siRNA-based therapeutic targets in the breast cancer TME. This figure categorizes key cell types and their associated siRNA-targetable molecules: in TAMs, siRNAs targeting Siglec‑G, ATP6, Redd1, VEGF/PIGF, and NF‑κB are shown; in tumor cells, siCD47 blocks the CD47 immune checkpoint; in CAFs, siRNAs against ATG5, TGF‑β, and Twist are indicated; and in endothelial cells, siRNAs targeting VEGF and Notch1 are illustrated in the context of PD‑L1 regulation. This schematic summarizes a multimodal siRNA strategy for concurrently relieving immunosuppression, normalizing vasculature, and remodeling the stromal barrier to enhance antitumor immunity.

Table 3.

Therapeutic Targets of siRNAs in TME of Breast Cancer

siRNA Target	Nanocarrier	TME Component	Outcome
Siglec-G	Liposome	TAMs	Blockade of the CD24/Siglec-G-mediated ‘don’t eat me’ signaling pathway activates macrophage phagocytosis and promotes antigen presentation.172
ATP6	PEG-PDPA	TAMs	Silencing of the ATP6 gene inhibits mitochondrial ATP synthesis and mtDNA stability, driving the polarization of TAMs towards an anti-tumor M1 phenotype.173
Redd1	Outer membrane vesicles	TAMs	Redd1 gene silencing inhibits tumor growth by modulating the tumor metabolic microenvironment.174
VEGF/PIGF	Mannose-modified pH-responsive nanoparticles	TAMs	Overexpression of VEGF and PIGF in breast cancer cells and M2 TAMs synergistically promotes tumor progression and immunosuppression, and silencing of VEGF and PIGF reduces the expression of M2 TAMs.154
Twist	HES-CH	CAFs	Silencing the Twist gene inhibits the EMT process, remodels the interstitial tumor barrier to facilitate drug penetration, and significantly reduces the in situ tumor load and the number of lung metastases in TNBC.175
TGF-β	LPEI-SS-LPEI	CAFs	Promoting vascular normalisation and inhibiting epithelial-to-mesenchymal transition (EMT) facilitated nanomedicine penetration.159
Notch1	Lipopeptide	Endothelial cells	Notch1 gene silencing synergizes with the mTOR inhibitor metformin to overcome treatment resistance and chemotherapy resistance in TNBC.28
VEGF	Extracellular vesicles	Endothelial cells	Extracellular vesicles derived from dendritic cells loaded with VEGF-A siRNA and doxorubicin reduce glioma angiogenesis in vitro.176
IDO1	Liposome	T-cells	Nanoliposomal system targeting delivery of IDO1 siRNA and doxorubicin synergistically induces immunogenic cell death (ICD) and blocks IDO1-mediated immune escape to remodel the tumor microenvironment.177
TGF-β	LPEI	CAR-T cells	Targeted silencing of TGF-β can effectively degrade ECM and significantly enhance the infiltration and activity of CAR-T cells in solid tumors.178
CD47	Tecto dendrimers	DCs	Silencing of the CD47 gene enhances macrophage recognition and promotes DC maturation and subsequent cytotoxic T cell activation.179
YTHDF 1	G5 PAMAM	DCs	Antitumor immunity is regulated by m6A methylation, a process mediated by YTHDF1 in DCs; inhibition of YTHDF1 enhances maturation of DCs.180
STAT3	Heterocyclic lipidoid and BR647	DCs	STAT3 silencing promotes type I interferon production and enhances DC maturation through activation of the STING pathway.181
S100A4	Exosome	DCs	Silencing S100A4 prevents it from driving breast cancer progression by promoting tumor metastasis and shaping the pre-metastatic microenvironment (PMN).137
HIF-1α/ TIGIT	Chitosan lactate	T cells	Inhibition of TIGIT and HIF-1 alpha molecules inhibits tumor cell expansion.182
MCT4	Dendrimer/copper(II)	T cells	Delivery of siMCT4 blocks tumor lactate metabolism.183
AXL	Dendritic polylysine material	B cells	Co-delivery of PTX, siAXL inhibits AXL gene and reverses PTX resistance in tumor cells.184
Bcl-2	AuNPs	B cells	Silencing the Bcl-2 gene in conjunction with chemotherapy reduces colony formation and cell migration in breast cancer cells.132

RNAi Technology Targeting DC Cells

However, inhibitory signals within the tumour microenvironment frequently lead to impaired DC maturation and compromised cross-presentation, thereby weakening their ability to support T cells and B cells.^185,186 Silencing negative regulatory factors in DCs, such as SOCS1 or PD-L1, via siRNA can restore their antigen-presenting and co-stimulatory functions, thereby indirectly enhancing T-cell activation and B-cell responses.¹⁸⁷ Furthermore, targeting the transcription factor NR4A1, which regulates metabolic adaptation in DCs, or BATF3, a key factor in maintaining cross-presentation function, helps DCs to maintain their essential functions within an inhibitory microenvironment.^188,189 However, the restoration of DC function may be accompanied by a risk of excessive inflammation, and their regulatory effects on T and B cells are influenced by the dynamic regulation of other immune subsets within the microenvironment. Song et al developed a dual-module nanoplatform based on core-shell dendrimeric macromolecules (CSTD) for synergistic chemo-immunotherapy in breast cancer. The first module consists of a G5-G3-D complex loaded with doxorubicin DOX, which induces ICD via chemotherapy to promote DC maturation; the second is a CBAA-G5/G3-Man/YTHDF1 siRNA aggregate, which actively targets DCs via mannose and delivers YTHDF1 siRNA to silence YTHDF1 expression, thereby enhancing antigen cross-presentation by DCs. The combined use of these two modules promotes DC maturation via the dual pathways of chemotherapy-induced ICD and DC gene editing, thereby activating the anti-tumour immune response of CD4⁺/CD8⁺ T cells. This approach significantly inhibits tumour growth in an in situ breast cancer model whilst demonstrating good safety.¹⁸⁰ Liu et al developed a dual-targeted nanovaccine based on a heterocyclic lipid (A18) and a polyester (BR647), which encapsulated both the OVA antigen and STAT3 siRNA, and was modified with oligohyaluronan (HA) and mannose to achieve dual active targeting of tumours and draining lymph nodes. On the one hand, this nanovaccine targets DCs via mannose, promoting antigen cross-presentation and DC maturation, whilst A18 activates the STING pathway to enhance DC activation; on the other hand, it targets tumour cells via HA, delivering STAT3 siRNA to silence STAT3, thereby downregulating MDSCs and Tregs and inducing the polarisation of TANs from N2 to N1, thus reversing the immunosuppressive microenvironment. Through the synergistic action of multiple mechanisms, this platform enhances the anti-tumour immune response of CD8⁺/CD4⁺ T cells, significantly inhibiting tumour growth in a melanoma model; furthermore, when used in combination with anti-PD-L1 antibodies, it further prolongs survival.¹⁸¹

RNAi Technology Targeting T Cells

T cells are the key effector cells responsible for mediating the killing of tumour cells; their level of activation depends directly on the efficiency of antigen presentation by DCs and is bidirectionally regulated by various immune cell subsets within the tumour microenvironment.^190,191 Tregs can significantly suppress the function of effector T cells by secreting IL-10 and TGF-β. Silencing endogenous T cell immune checkpoint molecules (such as PD-1) or the negative signalling regulator PTPN2 using siRNA can effectively reverse T cell exhaustion and restore their ability to kill tumour cells.^192,193 However, T cells themselves have a low efficiency of siRNA uptake; more importantly, if one merely enhances the effector function of T cells without improving the antigen-presenting capacity of DCs and simultaneously alleviating Treg-mediated immunosuppression, it is generally difficult to induce a durable anti-tumour immune response. This suggests that siRNA intervention strategies targeting T cells must take into account both the upstream antigen-presenting role of DCs and the immunosuppression caused by Tregs, thereby achieving synergistic regulation across multiple cellular levels. Kim et al developed an aptamer-conjugated nanoliposome co-delivery system for the combination of immunogenic chemotherapy and the reversal of immunosuppression. The surface of these liposomes is covalently conjugated with anti-CD44 and anti-PD-L1 DNA aptamers, enabling dual active targeting of breast cancer cells, whilst utilising a pH gradient method to encapsulate DOX (and IDO1 siRNA). Following intravenous injection, this nanosystem undergoes tumour-specific accumulation and endocytosis via aptamer-mediated receptor recognition; on the one hand, DOX induces ICD in tumour cells, promoting CRT exposure, ATP release and HMGB-1 secretion, thereby activating DCs and recruiting CD8⁺ T cells; on the other hand, IDO1 siRNA silences IDO1 gene expression, decreases Treg infiltration and lifts IDO1-mediated immunosuppression, whilst the PD-L1 aptamer blocks PD-1/PD-L1 interactions to further restore T-cell activity. Through the synergistic effects of ICD induction and IDO1 silencing, this dual-targeted co-delivery system reshapes the immunosuppressive TME into an immunostimulatory TME, significantly inhibiting tumour growth and lung metastasis in a subcutaneous breast cancer model, whilst increasing CD8⁺ T cell infiltration and TNF-α secretion.¹⁷⁷

RNAi Technology Targeting Endothelial Cells

Tumour endothelial cells constitute a key barrier affecting nanomedicine delivery and the infiltration of immune cells. In the tumour microenvironment, pro-angiogenic factors such as VEGF drive abnormal endothelial cell proliferation and vascular structural disorganisation, thereby limiting the intratumoural distribution of siRNA and the infiltration of effector T cells.^194,195 Silencing pro-angiogenic signalling molecules such as VEGFR2 or ANG-2 in endothelial cells via siRNA promotes vascular normalisation, improves permeability and perfusion, and enhances the delivery efficiency of siRNA and immune cells; siRNA interventions targeting adhesion molecules such as E-selectin or ICAM-1,¹⁹⁶ on the other hand, help to restore the endothelial cells’ support for the transendothelial migration of CD8⁺ T cells.^197,198 However, this strategy faces challenges such as endothelial cell heterogeneity and a short window for vascular normalisation, and requires the integration of time-controlled release and microenvironment-responsive designs to achieve precise regulation. Shamshiripour et al developed a co-delivery system based on dendritic cell-derived extracellular vesicles (DC-EVs), loaded with VEGF-A siRNA and DOX, to inhibit glioma angiogenesis and progression. In this study, microvesicles (MVs) and exosomes were isolated from the culture supernatant of mature DCs, and dual drug loading was achieved via electroporation and co-incubation, with MVs demonstrating a significantly higher loading efficiency than exosomes. Following intranasal administration, DC-EVs effectively crossed the blood-brain barrier to target gliomas within the brain. Mechanistically, VEGF-A siRNA silences VEGF-A expression, inhibits the formation of tubular structures in endothelial cells and impairs the migratory capacity of apical cells; DOX induces necrotic apoptosis in tumour cells. This dual-loading system blocks angiogenesis whilst moderately activating the Ang2/Tie2 compensatory pathway to prevent excessive hypoxia and stromal transformation, and reduces the formation of pseudo-vessels. In an in vivo glioma model, this strategy significantly reduced microvascular density, normalised vascular morphology, and promoted the intratumoural infiltration of CD8⁺/CD4⁺ T cells, effectively inhibiting the growth and invasion of in situ gliomas whilst demonstrating good safety.¹⁷⁶

Clinical Translation and Prospects of Cancer Immunotherapy Based on siRNA

The advances and limitations of siRNA delivery systems discussed above are directly reflected in the clinical translation of tumour-targeted RNAi therapies; early-stage trials have provided both cautionary lessons and validation of their feasibility. Since the approval of the first siRNA drug, patisiran,³⁰ in 2018, RNAi therapies have achieved clinical success in the treatment of liver-targeted hereditary diseases, while GalNAc-siRNA conjugates (such as vutrisiran) have further extended the dosing interval to three months.¹⁹⁹ These results demonstrate that RNAi can deliver clinical efficacy when delivery challenges are adequately addressed. However, the advantages in hepatic uptake underpinning these successes cannot be directly extrapolated to extrahepatic solid tumours, such as triple-negative breast cancer. As the LNP delivery system naturally targets the liver, approximately 65% of RNAi trials focus on liver diseases. Excluding liver diseases, trials relating to cancer account for only 9.7%. Furthermore, the literature states that 45.8% of oncology RNAi programmes were terminated prior to Phase II.²⁰⁰ These early-stage failures help to clarify the delivery requirements for next-generation tumour-targeting systems. Several siRNA-based anti-tumour drugs have now progressed to Phase I–II clinical trials. Among these, KRAS-G12D is an oncogene that is frequently mutated in pancreatic cancer and drives malignant tumour proliferation; MYC, as a core proto-oncogene, is widely involved in the progression of various solid tumours, including hepatocellular carcinoma, through its abnormal amplification; PLK1, meanwhile, is a key kinase regulating cell mitosis, and high expression of this gene can induce unlimited proliferation of tumour cells. As shown in Table 4, the aforementioned candidate siRNA drugs, utilising various delivery vehicles such as peptides, lipid nanoparticles and gold nanoparticles, are currently undergoing clinical trials for malignant tumours including skin cancer, pancreatic cancer, glioblastoma and ovarian cancer.

Table 4.

Representative Clinical Trials of siRNA-Based Cancer Therapies

Drug	Target	Indication	Delivery System	Phase	Source
STP705	TGF-β1/COX-2	Skin Cancer	Peptide	Phase II	NCT04844931
STP707	TGF-β1/COX-2	Solid Tumors	Peptide	Phase I	NCT05037149
DCR-MYC	MYC	Hepatocellular Carcinoma	LNP	Phase I	NCT02110563
siG12D-LODER	KRAS G12D	Pancreatic cancer	LODER	Phase II	NCT01676259
NU-0129	Bcl-2L12	Glioblastoma	Gold nanoparticles	Phase I	NCT03091337
siRNA-EphA2	EphA2	Advanced/Recurrent Ovarian Cancer	Liposomes	Phase I	NCT01591356
Atu027	PKN3	Solid Tumors	LNP	Phase II	NCT00938574
TKM-080301	PLK1	Hepatocellular Carcinoma	LNP	Phase II	NCT01437007

CALAA-01 was the first to demonstrate, via biopsy, RNAi-mediated cleavage of target mRNAs and dose-dependent tumour accumulation; however, its termination was not due to a lack of tumour targeting, but rather to vector-related immunotoxicity.¹²³ DCR-MYC and TKM-080301, meanwhile, demonstrated different failure patterns associated with early-stage lipid nanoparticle formulations. TKM-080301 completed a Phase I/II clinical trial in 2019 but failed to demonstrate an improvement in overall survival in patients with hepatocellular carcinoma, with a median overall survival of only 7.5 months.²⁰¹ DCR-MYC was discontinued in 2016 due to a narrow therapeutic window and toxicity to normal tissues. Some projects have already generated proof-of-concept data. ALN-VSP02 is a lipid nanoparticle (LNP) that co-delivers KSP and VEGF-A siRNA; it was well tolerated in a Phase I study; one patient with liver metastases from endometrial cancer achieved complete remission, whilst several other patients experienced long-term disease stabilisation.²⁰² Atu027 targets PKN3 in endothelial cells and has completed a Phase I/II evaluation in combination with gemcitabine, supporting the feasibility of vascular-targeted RNAi.²⁰³ siG12D-LODER is a local sustained-release matrix implanted directly into pancreatic tumours.²⁰⁴ STP707 is an intravenous formulation that dual-targets TGF-β1 and COX-2. In a Phase I clinical trial, 74% of evaluable patients achieved disease stabilisation, and the treatment demonstrated good safety, indicating that multi-target immunomodulatory siRNA therapies are well tolerated clinically.²⁰⁵

The overall trend across these programmes is that intravenously administered candidate drugs have generally failed to achieve sufficient tumour accumulation and therapeutic thresholds, whilst locally administered or endothelial-targeted therapies have produced more consistent signals of efficacy. Applying these clinical insights to the treatment of TNBC further highlights the challenges of translating these therapies into clinical practice. In addition to the general delivery barriers discussed in Key Technological Bottlenecks Faced by siRNA Nanodelivery Systems, TNBC presents three further obstacles. Firstly, TNBC cells overexpress various endosomal and lysosomal regulators, including members of the Rab family, leading to hyperactive endocytic cycling and reduced endosomal escape efficiency. Overcoming this issue requires higher cationic carrier loading, which may approach the cytotoxic threshold.^206,207 Secondly, the significant molecular heterogeneity of TNBC means that single-target silencing often triggers compensatory signalling. For example, BCL-2 silencing can upregulate MCL-1, whilst VEGF silencing can elevate PLGF.²⁰⁸ Multi-target co-delivery may resolve this issue, but it also presents its own engineering challenges, including payload competition and batch-to-batch consistency. Thirdly, although TP53 mutations and homologous recombination defects are common in TNBC, current siRNA delivery systems have not yet been designed to address these characteristics. Mechanisms for triggering the DNA damage response and synthetic lethal strategies targeting HRD-associated genes remain largely at the preclinical stage. Furthermore, a dense extracellular matrix and elevated interstitial fluid pressure—which have been recognised as ubiquitous barriers—are more pronounced in the fibro-proliferative subtype of TNBC, further limiting the penetration of nanoparticles beyond the perivascular regions.²⁰⁹ These obstacles do not preclude the possibility of treatment. In TNBC, a more realistic role for siRNA is as a precision-regulating component within combination therapy regimens. This may include silencing ABCB1 to enhance chemotherapy sensitivity, downregulating TGF-β or STAT3 to improve response to immune checkpoint inhibitors, or disrupting DNA repair pathways to achieve synthetic lethality in tumours with homologous recombination defects.^210,211 Antibody-siRNA conjugates targeting TNBC surface antigens (such as Trop-2, EGFR and PD-L1) offer a method to circumvent the bottlenecks associated with systemic administration.²¹² Localised sustained-release platforms for post-surgical tumour bed therapy represent another option. Patient selection based on biomarkers such as the degree of immune infiltration, HRD score and specific driver gene mutations should improve the efficiency of clinical trials. With continuous improvements in chemical modifications, scalable production and combination strategies, RNAi therapies may progress from proof of concept to delivering meaningful clinical benefits in TNBC.

Conclusion

The convergence of RNA interference technology, smart nanocarriers and the immunology of the tumour microenvironment has brought about a revolution in the therapeutic approach to TNBC: treatment strategies are no longer limited to the direct destruction of tumour cells, but instead focus on targeted intervention in the pathological signalling networks within the tumour microenvironment that mediate immune suppression, matrix fibrosis, hypoxia and treatment resistance. siRNA nanomedicines are capable of targeting intracellular “undruggable” targets that are difficult for monoclonal antibodies and small-molecule drugs to access, thereby enabling sequence-specific regulation of multiple genes.^17,213–215 However, this field still faces significant translational challenges: whilst numerous studies have demonstrated excellent efficacy in preclinical models, very few protocols have progressed to late-stage clinical trials, highlighting a significant objective gap between mouse animal models and the actual tumour microenvironment in humans.

siRNA therapy targeting the tumour microenvironment is not merely a simple single-gene knockdown; it essentially constitutes a systematic disruption of the tumour’s multicellular interactive network. Single-target silencing is highly prone to inducing compensatory feedback pathways in the body. Taking TGFB1 as an example, this gene mediates fibrosis and tumour invasion whilst suppressing anti-tumour immunity; however, when TGFB1 is silenced solely within tumour cells, CAFs compensatorily secrete effector molecules such as CTGF and PDGF, continuing to drive collagen deposition and ultimately leading to treatment resistance.²¹⁶ Consequently, research efforts should shift towards rationally designed multi-target siRNA co-delivery strategies, whilst combining RNA interference technology with other therapeutic approaches in a synergistic manner. At the same time, spatial omics opens up a entirely new paradigm for target screening. Compared with traditional single-cell sequencing, spatial transcriptomics and spatial proteomics technologies can precisely localise functional targets within key pathological regions—such as those characterised by immune exclusion, hypoxia and matrix barriers—and eliminate ineffective candidate genes that are merely highly expressed but do not participate in the pathological processes of the microenvironment; this represents a highly promising direction for development in this field.^217,218 With regard to siRNA delivery to the tumour microenvironment, short-term efforts should focus on improving the penetration efficiency of delivery vehicles through the high interstitial pressure and hypoxic regions of solid tumours, with an emphasis on developing microenvironment-responsive materials that adapt their size and charge. At the same time, therapeutic strategies should shift from simple gene silencing to the remodelling of physical barriers within the microenvironment; In the medium term, the focus should shift to the design of carriers capable of multi-pressure, time-controlled release, in combination with low-dose radiotherapy and chemotherapy or immune checkpoint inhibitors, to convert “cold” tumours into “hot” tumours by inducing immunogenic cell death, whilst utilising organ-targeted sorting technology to address the challenges of extrahepatic delivery and the immunogenicity associated with repeated dosing. In terms of the R&D timeline, we must ensure that GLP-compliant toxicology studies are completed by month 24 and that an IND clinical approval is obtained by month 42. Clinical endpoints should prospectively adopt functional imaging parameters or changes in circulating biomarkers, rather than relying solely on traditional tumour volume reduction. Should you require a detailed plan for the specific technical approach, I would be happy to elaborate.

Multiple technical bottlenecks continue to hinder clinical translation; however, several viable technical pathways have now been established to specifically address these challenges. Delivery to solid tumours is the most critical rate-limiting step: the dense extracellular matrix, elevated tumour interstitial pressure and vascular structural heterogeneity characteristic of human TNBC severely limit the intratumoural penetration and uniform distribution of nanoparticles, whilst commonly used subcutaneous tumour models struggle to replicate these pathological features.^219–221 To address this, tumour-penetrating peptides such as iRGD and LyP-1 can be introduced to construct tumour microenvironment-responsive dissociable shells, or in situ injectable sustained-release hydrogels can be employed for local administration, thereby circumventing the barriers to intratumoural penetration associated with systemic delivery.^222–224 At the intracellular level, inefficient endosomal escape results in the majority of siRNA being retained in endolysosomes, significantly reducing the effective loading of the RISC complex; By optimising the carrier’s pKa value, introducing structures that utilise the proton sponge effect, and developing membrane-fusion-type biomimetic modifications, the efficiency of endosomal escape can be effectively enhanced, thereby reducing the lysosomal degradation of siRNA. On the other hand, safety risks such as carrier-induced complement activation, siRNA-activated innate immunity, non-specific accumulation in the liver, and long-term accumulation of carrier materials in vivo must not be overlooked. Chemical modifications such as 2′‑OMe, 2′‑F and thio-phosphate backbones can be applied to siRNA to reduce its immunogenicity; the use of amphoteric ionic modifications in place of traditional PEG modifications helps to circumvent the antibody-mediated accelerated clearance (ABC) effect. Although biomimetic carriers, such as cell-membrane-coated nanoparticles and engineered exosomes, can improve biocompatibility, they still face industrial challenges such as low siRNA encapsulation efficiency, poor batch reproducibility and difficulties in clinical quality control; there is an urgent need to establish standardised preparation processes and quality evaluation systems. Furthermore, the tumour microenvironment of TNBC varies greatly across different molecular subtypes, meaning there is no one-size-fits-all nanomedicine; spatial multi-omics can be utilised to develop companion diagnostic biomarkers to enable patient stratification, and to match individualised siRNA target combinations and nanocarrier strategies based on tumour fibrosis levels and immune cell infiltration phenotypes.

Future key research directions will primarily include: firstly, promoting the deep integration of spatial multi-omics and nanomedicine; utilising biomarkers to identify targets and stratify patients, thereby achieving mechanism-driven, precision-based remodelling of the tumour microenvironment; secondly, developing carrier platforms suitable for GMP-compliant large-scale production, with a focus on technological systems that balance therapeutic efficacy and manufacturing feasibility, such as ionisable lipid nanoparticles, antibody–siRNA conjugates and locally injectable hydrogels; thirdly, elucidating the optimal dosing schedules, dosages and mechanisms of synergy for combination therapies involving siRNA with immune checkpoint inhibitors, radiotherapy, chemotherapy and phototherapy. Concurrently, in the early stages of drug evaluation, greater use should be made of human organoids and in situ PDX models that preserve the intact matrix structure, thereby minimising discrepancies between preclinical research and clinical trial results.

In summary, strategies for tumour microenvironment remodelling based on siRNA nanomedicines provide powerful technical tools for addressing core challenges such as TNBC immune evasion and treatment resistance. To truly achieve clinical translation, not only are technological innovations in materials science required, but also in-depth interdisciplinary collaboration across fields such as tumour immunology, spatio-omic biomarkers and translational pharmacology. By relying on the aforementioned targeted solutions to overcome key bottlenecks—such as delivery, biosafety and personalised therapy—one by one, it is hoped that a completely new pathway for precision combination therapy for refractory triple-negative breast cancer will be opened up.

Funding Statement

This work was supported by the National Natural Science Foundation of China (Grant Nos. 81672380) and the Nanjing Medical Science and Technology Development Program (Grant Nos. ZKX23012).

Data Sharing Statement

No new data were generated or analyzed in this review. Data sharing is not applicable.

Ethics Approval

This is a review article. No human or animal studies were conducted; therefore, ethics approval is not applicable.

Consent for Publication

Not applicable. This review contains no individual person’s data.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.