FMRP silencing via siRNA lipid nanoparticles to reprogram the tumor microenvironment and enhance anti-PD-1 efficacy in triple-negative breast cancer

Abstract

Fragile X mental retardation protein (FMRP), encoded by the Fmr1 gene, is notably upregulated in various malignancies, including triple-negative breast cancer, pancreatic ductal adenocarcinoma, and colon carcinoma, and potentially mediates immune evasion through multiple mechanisms. Here a small interfering RNA (siRNA) targeting Fmr1 is designed for tumor site delivery to suppress FMRP expression and enhance immune activation. Lipid nanoparticles loaded with siFmr1 (LNP@siFmr1) are prepared using a microfluidic-based electrospray method, achieving optimal particle size, encapsulation efficiency, and stability. In an orthotopic breast cancer mouse model, LNP@siFmr1 demonstrates effective passive targeting of tumor sites, leading to significant downregulation of FMRP expression. This suppression coincides with the recruitment and activation of CD8⁺ T cells and immunostimulatory macrophages, effectively inhibiting tumor growth. Furthermore, the combination with PD-1 antibodies achieves approximately 80 % suppression of tumor growth. These findings provide new insights into tumor-targeted siRNA delivery and immunotherapy strategies involving FMRP targeting and PD-1 antibodies.

Graphical abstract

Highlights

• •A microfluidics-based electrospray method was innovatively applied for preparation of siRNA-loaded lipid nanoparticles.
• •LNP@siFmr1 successfully delivered siFmr1 to the tumor, knocked down the FMRP expression, and activated the immune response.
• •The combination therapy of LNP@siFmr1 and PD-1 antibodies achieved efficient suppression of tumor growth.

1. Introduction

Triple-negative breast cancer (TNBC), characterized by the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), represents a clinically aggressive subtype with high metastatic potential, poor prognosis, and limited treatment options [[1], [2], [3]]. While immunotherapy has emerged as a promising strategy for TNBC, the efficacy of immune checkpoint inhibitors (ICIs) remains constrained by primary resistance and the immunosuppressive tumor microenvironment (TME) [[4], [5], [6]].

Fragile X mental retardation protein (FMRP), an RNA-binding translational regulator encoded by FMR1, plays a critical role in tumor immune evasion by orchestrating immunosuppressive networks within the TME [[7], [8], [9], [10], [11], [12], [13]]. Mechanistically, FMRP-expressing cancer cells secrete interleukin-33 (IL-33) to recruit regulatory T cells (Tregs) and protein S (PROS1) to polarize tumor-associated macrophages (TAMs) toward an immune-suppressive M2 phenotype. These coordinated actions generate "immune deserts" by excluding cytotoxic CD8⁺ T cells and dampening their antitumor activity [13]. Targeting FMRP thus presents a novel therapeutic avenue to reprogram the TME and enhance ICI responsiveness.

RNA interference (RNAi)-mediated FMRP silencing using small interfering RNA (siRNA) offers a precision strategy to disrupt this immunosuppressive axis [14]. Unlike traditional small-molecule drugs or biologics, siRNA therapeutics require unique development paradigms where pharmacodynamic effects are sequence-dependent, while pharmacokinetic properties hinge on delivery systems [15]. Systemic siRNA delivery to extrahepatic tumors remains a translational challenge. Lipid nanoparticles (LNPs), clinically validated carriers for nucleic acid therapeutics, provide an optimal solution [[16], [17], [18]]. The lipid composition and formulation methods critically determine LNP characteristics (e.g., size, surface charge, and targeting capacity), which govern their passive accumulation in tumors [19]. Our LNP design employs the ionizable lipid ALC-0315, which demonstrates superior siRNA encapsulation and endosomal escape compared to conventional formulations like DLin-MC3-DMA [20,21]. Utilizing an innovative coaxial electrospray microfluidic platform [22,23], we achieve rapid single-step synthesis of uniform LNPs with enhanced reproducibility and scalability over traditional methods (e.g., preformed vesicles or complex microfluidic mixers) [[24], [25], [26]].

We further propose that FMRP silencing synergizes with programmed cell death protein 1 (PD-1) blockade. While PD-1/PD-L1 inhibitors have revolutionized cancer treatment, their efficacy in "immune-cold" tumors like TNBC is limited by inadequate T-cell infiltration [27,28]. By disrupting FMRP-driven immunosuppression, our strategy may potentiate PD-1 antibody therapy through dual mechanisms: (1) enhancing CD8⁺ T-cell recruitment via TME remodeling, and (2) reversing T-cell exhaustion. This combinatorial approach addresses a critical unmet need in TNBC management, offering a blueprint for overcoming ICI resistance through targeted TME modulation.

Therefore, we employed a microfluidic-based coaxial electrostatic spray system to prepare lipid nanoparticles encapsulating siFmr1, enabling passive targeting and accumulation within tumor tissues to facilitate siRNA delivery into tumor cells (Scheme 1). The selected siFmr1 sequence silenced the Fmr1 gene in tumor cells and inhibited the expression of FMRP, thereby recruiting and activating T cells. Additionally, it downregulated the expression of PROS and IL33, enhancing the levels of immunostimulatory macrophages and dendritic cells (DC) within the tumor, collectively improving the tumor immune microenvironment. On this basis, PD-1 antibodies are further utilized to relieve the suppression of tumor cells on T cells. Ultimately, the combination therapy of LNP@siFmr1 and PD-1 antibodies achieved effective inhibition of tumor growth in a murine breast cancer model while demonstrating good safety.

Scheme 1.

Lipid nanoparticles loaded with FMRP siRNA were prepared using a microfluidic-based electrospray method to activate tumor immunity in TNBC and synergistically enhance the therapeutic efficacy with PD-1 antibodies. Created with Biorender.com.

2. Materials and methods

2.1. Materials

((4-hydroxybutyl)azanediyl)bis(hexane-6,1-diyl)bis(2-hexyldecanoate) (ALC-0315), 1,2-Dioctadecanoyl-sn-glycero-3-phophocholine (DSPC), 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000 (DMG-PEG2000), and cholesterol were purchased from AVT Pharmaceutical Tech Co., Ltd. (Shanghai, China). siFmr1 RNA, siFmr1-FAM RNA, siFmr1-Cy5.5 RNA and qPCR primers were purchased from Sangon Biotech Co., Ltd. (Shanghai, China). The siRNA and qPCR primer sequences are shown in Table 1. FMRP Recombinant Rabbit Monoclonal Antibody [JM91-41] (# ET1703-70), HRP Conjugated Goat anti-Rabbit IgG polyclonal Antibody (# HA1001) and GAPDH Recombinant Rabbit Monoclonal Antibody [SA30-01] (# ET1601-4) were purchased from HUABIO (Zhejiang, China). PROS Polyclonal Antibody and IL-33 Polyclonal Antibody were purchased from Baijia Biotechnology Co., Ltd. (Taizhou, China). InVivoMAb anti-mouse PD-1 (CD279) (#BE0146) was purchased from Bio X Cell. RIPA Lysis Buffer and Protease Inhibitors Cocktail were purchased from Seven Biotech (Beijing, China). BCA protein quantification kit, PAGE gel kit and Total RNA extraction reagent were purchased from Epizyme Biotech (Shanghai, China). Cell Counting Kit-8 (CCK-8) was purchased from APExBIO Technology. Hoechst33342 was obtained from Solarbio (Beijing, China). Lyso-Tracker Red was obtained from Beyotime Biotechnology (Shanghai, China). Rapid Blocking Buffer (TBS-T), Color Prestained Protein Marker (10–180 kDa) and Protein Loading Buffer (SDS, 4 × ) were purchased from BestEnzymes Biotech Co., Ltd. (Lianyungang, China). Ethanol and other chemical reagents were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). The antibodies used for flow cytometry were sourced from Thermo Fisher Scientific (USA) and included Fixable Viability Dye eFluor 506, anti-CD16/CD32, anti-CD45-APC-eFluor 780, anti-CD3e-FITC, anti-IA/IE-PerCP-eFluor 710, anti-CD8-PE-Cy7, anti-CD80-PE, and anti-CD86-PE-Cy7. TNF-α and IFN-γ ELISA kits were obtained from Biolegend (USA).

Table 1.

siRNA and qPCR primer sequences.

Name	Strand	Sequence (5′–3′)
siFmr1#1 RNA	sense	GAGGAUGAUAAAGGGUGAGUUTT
	anti-sense	AACUCACCCUUUAUCAUCCUCTT
siFmr1#2 RNA	sense	CGCACCAAGUUGUCUCUUAUATT
	anti-sense	UAUAAGAGACAACUUGGUGCGTT
siFmr1#3 RNA	sense	GCGUUUGGAGAGAUUACAAAUTT
	anti-sense	AUUUGUAAUCUCUCCAAACGCTT
siFmr1-FAM RNA	sense	FAM-GCGUUUGGAGAGAUUACAAAUTT
	anti-sense	AUUUGUAAUCUCUCCAAACGCTT
siFmr1-Cy5.5 RNA	sense	Cy5.5-GCGUUUGGAGAGAUUACAAAUTT
	anti-sense	AUUUGUAAUCUCUCCAAACGCTT
primer of βactin	forward	GTGCTATGTTGCTCTAGACTTCG
	reverse	ATGCCACAGGATTCCATACC
primer of Fmr1	forward	GGTCAAGGAATGGGTCGAGG
	reverse	AGTTCGTCTCTGTGGTCAGAT

2.2. The construction of a coaxial electrostatic spray system based on microfluidics

The method for chip preparation follows previous literature [22], utilizing soft lithography to fabricate a two-phase flow focusing chip. The final PDMS (polydimethylsiloxane) microfluidic chip obtained has a microchannel width and depth of 100 μm, and the outlet consists of a glass capillary tube with an inner diameter of 250 μm. Additionally, a polytetrafluoroethylene tube of suitable length was connected to the sample inlet, with its opposite end affixed to a syringe needle for sampling.

The syringe was connected to the inlet tube of the microfluidic chip, with a syringe pump (LSP01-3A, Longer, China) for sample injection. The positive terminal of the DC high-voltage power supply (TCM6000i, Teslaman, China) was connected to a copper coil placed at the sample collection site, while the negative terminal was connected to the needle of the external syringe. This setup generates an electric field between the microfluidic chip and the sample collection tube, enabling electrostatic spraying.

2.3. Preparation of lipid nanoparticles

ALC-0315 (5 mg/ml), cholesterol (2 mg/ml), DSPC (1 mg/ml), and DMG-PEG2000 (0.5 mg/ml), were dissolved in ethanol to serve as the inner phase solution. siRNA was dissolved in a pH 4 acidic buffer solution at a concentration of 5 μM to serve as the outer phase solution. The flow rate of the outer phase was 120 μl/min, while the inner phase was 40 μl/min, resulting in a flow rate ratio of 3:1. When the liquid reaching the capillary outlet, a high-voltage direct current power supply was activated and adjusted to form a stable Taylor cone. The lipid nanoparticles were collected and the acidic buffer was replaced with pH 7.4 PBS buffer using ultrafiltration centrifugation method, while the ethanol content was reduced to below 0.5 % to meet intravenous injection requirements. The final lipid nanoparticles were stored at 4 °C for subsequent experiments.

2.4. Characterization of lipid nanoparticles

The Taylor cone image and the electrospray process were captured by a high-speed camera (Phantom VEO 410L, Vision Research Inc., USA). The particle size and zeta potential were measured using a Malvern particle size analyzer (Zetasizer Nano ZS 90, Malvern Co., Ltd., UK). The morphology of the nanoparticles was identified using a transmission electron microscope (Titan G2 60–300, FEI Company, USA). The encapsulation rate of lipid nanoparticles for siRNA was detected using agarose gel electrophoresis, imaged with a chemiluminescence imaging system (BIO-RAD, ChemiDoc XRS+, USA), and quantified using Image J software. The release of siRNA at different pH levels was detected using the dialysis bag method, and the concentration of siFmr1-FAM in the release medium was measured using a fluorescence spectrophotometer (F-7100, Hitachi, Ltd., Japan).

2.5. Cell culture

4T1 cells (mouse breast cancer cells, purchased from Procell company in China, have been correctly identified by STR analysis) were cultured at 37 °C, 5 % CO₂, and the medium consisted of RPMI-1640 supplemented with 10 % fetal bovine serum (FBS) and 1 % Penicillin-Streptomycin Solution.

2.6. Cell uptake

In a confocal dish, 20,000 cells were seeded and cultured overnight. LNP@siFmr1-FAM was administered at a concentration of 100 nM. After 2h, 4h, 8h and 24h, the culture medium was aspirated, and the cells were washed twice with PBS. Subsequently, 1 ml of 4 % paraformaldehyde was added for fixation for 10 min, and the cells were washed twice with PBS. Hoechst (10 μg/ml) was added for 5 min for nuclei staining, and the cells were washed twice. The images were captured using a laser confocal microscope (Zeiss LSM900, Germany).

For the 6-well plate, each well was seeded with 200,000 cells and incubated to allow adherence. After 24 h incubation with the drugs, the cells were digested, centrifuged, and washed three times with PBS, then resuspended in 200 μl PBS for detection by flow cytometry (BD, LSRFortessa, USA).

2.7. Lysosomal escape

Add 20,000 cells to a confocal dish and allow them to adhere overnight. Introduce LNP@siFrm1-FAM at a concentration of 100 nM and incubate for 2, 4, 8, and 24 h. Remove the culture medium and stain with Lyso-Tracker at a concentration of 50 nM, incubating at 37 °C for 20 min. Remove the staining solution and wash twice with PBS, then capture images using a fluorescence microscope (Leica Thunder Dmi8, Germany).

2.8. Western blot (WB)

In a 6-well plate, 200,000 cells were seeded in each well and allowed to adhere overnight. LNP was administered at a concentration of 100 nM. After incubated for 48 h, the cells were washed twice with PBS. The cells were scraped and collected, and 120 μl RIPA lysis buffer (containing 1 % cocktail protease inhibitor) was added for lysis on ice for 20 min. The samples were centrifuged at 12,000 rpm for 15 min, and the supernatant was collected. Protein quantification was performed using a BCA kit, following the manufacturer's instructions. The protein samples were mixed with loading buffer and heated at 100 °C for 10 min to denature the proteins. A 7.5 % PAGE gel was prepared according to the gel preparation kit instructions, with 20 μg of protein loaded per well. The gel was electrophoresed at 250 V for 30 min in fast electrophoresis buffer. The proteins are transferred to a PVDF membrane at 400 mA for 25 min using fast transfer buffer. The membrane was blocked with rapid blocking solution for 10 min and washed three times with TBST. Primary antibody (diluted according to the manufacturer's ratio) was added to the membrane and incubated overnight at 4 °C. The membrane was washed with TBST three times. Secondary antibody was added and incubated at room temperature for 1 h, followed by washing the membrane with TBST three times. After adding the detection solution to the membrane, imaging was performed using a chemiluminescence imaging system (BIO-RAD, ChemiDoc XRS+, USA).

2.9. Real-time reverse transcription-polymerase chain reaction (qPCR)

200,000 cells were seeded per well in 6-well plates and allowed to adhere overnight. LNP was administered at a concentration of 100 nM. After 48 h of incubation, the medium was discarded and 1 ml of total RNA extraction reagent was added to each well. The cells were lysed on ice, and the lysate were collected. 0.2 ml of chloroform was added, vigorously shaken and mixed for 15 s. The mixture was allowed to stand at room temperature for 3–5 min before centrifugation at 12,000×g for 15 min. The supernatant was collected and an equal volume of isopropanol was added, mixed thoroughly, and the samples were incubated at room temperature for 10 min to precipitate RNA. The samples were then centrifuged at 12,000×g and 4 °C for 10 min to collect the RNA pellet. The RNA pellet was washed with 75 % ethanol, air-dried at room temperature, and dissolved in DEPC water. RNA concentration was measure. 1000 ng of RNA was used for reverse transcription by combining 4 μl of reverse transcription mix, 1 μl of DNase, and 14 μl of water, bringing the total volume to 20 μl. The reaction was carried out in a PCR thermal cycler at 37 °C for 2 min, 55 °C for 15 min, and 85 °C for 5 min. The concentration of DNA obtained from reverse transcription was measured and diluted to 100 ng/μl 10 μl of qPCR mix, 0.4 μl each of forward and reverse primers, 1 μl of cDNA template, and 8.2 μl of water were added to the PCR tube. Assays were performed in a qPCR instrument (BIO-RAD, CFX CONNECT, USA) using a 2-step procedure, pre-denaturing at 95 °C for 5 min, 95 °C for 10 s, 60 °C for 30 s, and 40 cycles.

2.10. Cell viability assay

The 96-well plate was seeded with 5000 cells per well and allowed to adhere overnight. Different concentrations of LNP were administered and incubated for 24 h, the culture medium was removed, and 100 μl of medium containing 10 % CCK8 was added. After incubating for 2 h, the absorbance was measured at 450 nm.

2.11. Hemolysis test

Anticoagulant sheep blood was diluted with PBS. LNPs were added at varying concentrations, with PBS as the negative control and water as the positive control. The samples were incubated at 37 °C for 30 min, followed by centrifugation at 3000 rpm for 5 min. Photographs were taken, and the supernatant was collected for absorbance measurement at 541 nm.

2.12. Construction of a mouse orthotopic breast cancer model

All animal experiments approved by the Institutional Animal Careand Use Committee (IACUC), The Second Xiangya Hospital, Central South University, China (20250577).

Five million 4T1 cells were suspended in 100 μl of PBS and injected into the fourth left mammary fat pad of 6-8-week-old female BALB/c mice. Subsequent experiments were initiated when the tumor volume reaches 100 mm³.

2.13. In vivo distribution experiment

siFmr1-Cy5.5 and LNP@siFmr1-Cy5.5 were administered intravenously via the tail vein (n = 3) at a dosage of 100 nmol/kg, based on the siFmr1-Cy5.5 concentration. In vivo imaging was performed at 4h, 12h, and 24h post-administration using the IVIS Spectrum Imaging system (PerkinElmer, USA). After the in vivo imaging, the tumors and heart, liver, spleen, lungs, and kidneys were dissected for ex vivo imaging. The fluorescence intensity was quantified using IVIS Spectrum Imaging software (PerkinElmer, USA).

Furthermore, tumors from mice injected with LNP@siFmr1-Cy5.5 for 12 h were fixed with 4 % paraformaldehyde, embedded in OCT, and prepared as frozen sections. The nuclei were stained with DAPI, and the distribution of red fluorescence within the tumors was observed using laser confocal microscopy.

2.14. In vivo anti-tumor experiment

A total of 24 tumor-bearing mice were randomly divided into 4 groups. Group G1 served as the control group, receiving a tail vein injection of 200 μl of PBS every 2 days for a total of 7 times. Group G2 received a tail vein injection of 200 μl of siFmr1 every 2 days, with a dosage of 100 nmol/kg, for a total of 7 times. Groups G3 and G4 received a tail vein injection of 200 μl of LNP@ siFmr1 every 2 days, with a dosage of 100 nmol/kg based on siFmr1 concentration, for a total of 7 times. Group G4 received an intraperitoneal injection of 250 μg of PD-1 antibody every 4 days, for a total of 4 times. After the administration was completed, observation continued until the tumor volume in the control group reached 1000 mm³, at which point the experiment was terminated. Tumor tissues were collected, photographed, weighed, and subjected to H&E staining analysis.

To assess the immune activation, tumors and inguinal lymph nodes were excised and processed into single-cell suspensions. Red blood cells were lysed using the lysis buffer for 2 min, followed by centrifugation at 300g for 5 min. DPBS (containing 5 % FBS) was used as staining buffer, and Fixable Viability Dye eFluor 506 was added for staining at room temperature for 10 min to distinguish live and dead cells. After the staining and washing process, anti-CD16/CD32 antibodies were added and incubated at 4 °C for 5 min to block nonspecific binding. T cells were identified using anti-CD45, CD3, and CD8 antibodies. DCs were identified with anti-CD45, CD3, CD11c, IA/IE, CD80, and CD86 antibodies. Macrophages were identified using anti-CD45, CD11b, F4/80, and CD86 antibodies. Staining was performed at 4 °C for 30 min. The cells were centrifuged and washed twice, and the proportion of activated cells was then analyzed using flow cytometry.

2.15. Biological safety assessment

During the anti-tumor experiment, the body weight of the mice was measured every two days. After the experiment, blood was collected from the mice, and the levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (CR) in the plasma were measured to evaluate liver and kidney function. H&E staining analysis was performed to evaluate histopathological changes of the major organs, including the heart, liver, spleen, lungs, and kidneys.

2.16. Pharmacokinetics

Six healthy mice were divided into two groups and received tail vein injections of either free siFmr1-Cy5.5 or LNP@siFmr1-Cy5.5 at a dosage of 100 nmol/kg siFmr1-Cy5.5. Blood samples (50 μl) were collected from the orbital vein at 5 min, 30 min, 1 h, 2 h, 4 h, 6 h, and 8 h post-injection to measure Cy5.5 fluorescence intensity.

2.17. Statistical methods

The data are presented as mean ± standard deviation. The statistical significance of the differences between the two groups was assessed using a two-tailed Student's t-test, while the differences among three or more groups were evaluated using one-way analysis of variance (ANOVA). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001.

3. Results and discussion

3.1. Preparation and characterization of lipid nanoparticles

We prepared lipid nanoparticles loaded with siRNA using a microfluidic-based coaxial electrospray platform for the first time. Fig. 1A shows the Taylor cone at the capillary nozzle during the electrospray process, where the liquid is drawn into a conical shape under the influence of the electric field [29], gradually narrowing in width, and forming fine droplets at the tip of the cone. The video of the electro-spray process is shown in Video S1. In this process, the aqueous phase of the exterior phase is mixed in a controllable and sufficient manner with the organic phase of the interior phase, allowing the ionizable lipids to acquire a positive charge under acidic conditions and bind with negatively charged siRNA, thereby generating lipid nanoparticles.

Fig. 1.

The characterization of LNP. A) Image of the Taylor cone during the electro-spray process. B) Detection of LNP particle size using the DLS method. C) The transmission electron microscope image of LNP. D) Images and E) quantitative results of the encapsulation efficiency of different formulations of LNP detected by gel electrophoresis. F) Changes in particle size and PDI of LNP under storage conditions at 4 °C. G) The changes in particle size, PDI, and H) zeta potential of LNP under simulated in vivo conditions. I) Release of siRNA under simulated physiological conditions. Results are presented as mean ± SD (n = 3).

By leveraging the precise control of liquid flow rates afforded by the microfluidic system and the accurate regulation of fluid mixing enabled by the electric field, we successfully obtained LNPs with uniform size distribution. As illustrated in Fig. 1B, the dynamic light scattering (DLS) analysis revealed a particle size of 160.3 ± 3.7 nm with a polydispersity index (PDI) of 0.051 ± 0.018. A particle size less than 200 nm facilitates passive targeting of tumor sites via the enhanced permeability and retention (EPR) effect, enabling the nanoparticles to exert their function. Moreover, the zeta potential of the LNPs was measured at −0.723 ± 0.905 mV, indicating that the LNPs exhibited electrical neutrality in PBS buffer, which is beneficial for reducing adsorption to proteins in plasma and enhancing nanoparticle stability.

The following are the Supplementary data related to this article:

Transmission electron microscopy (TEM) images revealed that the LNP exhibited a uniform spherical morphology, with multiple internal cavities that were likely hydrophilic regions accommodating nucleic acid molecules (Fig. 1C). The encapsulation of siRNA by the lipid nanoparticles was assessed under various N/P ratios. N/P refers to the molar ratio of positively charged nitrogen atoms in ionizable lipids to negatively charged phosphorus atoms in nucleic acids. Free siRNA migrated into the gel, bound with dye, and emitted fluorescence, while siRNA encapsulated within the nanoparticles remained in the sample wells (Fig. 1D). Quantification using ImageJ software showed that the encapsulation efficiency reached 81.5 ± 3.3 % at an N/P ratio of 9.47 (Fig. 1E).

We examined the stability of the nanoparticles under two different conditions. The first condition was preservation at 4 °C (Fig. 1F), focusing on the stability from the preparation of the nanoparticles to their use in cell or animal experiments. Under this condition, the LNPs showed no significant changes in particle size and PDI over a period of 7 days, indicating good stability. The second condition involved 37 °C with 10 % fetal bovine serum, while being shaken at 100 rpm in a shaking incubator to simulate in vivo conditions (Fig. 1G and H). Under this condition, the LNPs exhibited no significant changes in particle size, PDI, or zeta potential over 96 h, preliminarily demonstrating that stability can be maintained in vivo. Furthermore, siRNA leakage was examined under various pH conditions (Fig. 1I), with minimal leakage observed at pH 4, 5, and 6, while a leakage of 29 % was noted at pH 7.4 over 48 h. When delivering siRNA using lipid nanoparticles, it is first necessary for the cells to uptake the lipid nanoparticles, after which the ionizable lipids interact with the negatively charged lipids on the endosomal membrane and disrupt the endosomal membrane structure, allowing siRNA to escape into the cytoplasm to exert its effects [[30], [31], [32]]. Given that free siRNA has difficulty penetrating cell membranes, we aim to minimize leakage of siRNA during circulation in vivo, allowing it to enter cells in nanoparticle form to exert its effects.

3.2. In vitro evaluation of LNP@siFmr1

The cellular uptake efficiency of LNP@siFmr1-FAM was evaluated after co-culturing with 4T1 cells. Laser confocal microscopy and flow cytometry were employed for detection. As shown in Fig. 2A, blue fluorescence indicates the cell nuclei, while green fluorescence represents LNP@siFmr1-FAM. After 4 h of culture, only a small amount of uptake was observed in some cells, whereas obvious green fluorescence was evident in all cells after 24 h. Quantitative analysis of flow cytometry demonstrated that the LNP@siFmr1-FAM treatment group exhibited a markedly higher cellular uptake compared to the control group after 24 h of culture (Fig. 2B). At the same time, we conducted lysosomal escape experiments, demonstrating that after being internalized by the cells, LNP@siFmr1 can deliver siFmr1 into the cytoplasm, as shown in Fig. S1.

Fig. 2.

In vitro evaluation of LNP@siFmr1. A) Images of cell uptake captured by laser confocal microscopy, with siRNA labeled using green fluorescent FAM. Scale bar = 10 μm. B) Quantification of cell uptake at 24 h through flow cytometry. C) The results of the knockdown of Fmr1 gene mRNA by LNP carrying different sequences of siRNA detected through qPCR. D) Images and E) quantitative results of FMRP protein knockdown using LNPs carrying different sequences of siRNA. F) The images and quantitative analysis of G) PROS and H) IL33 knockdown treating with LNP@siFmr1. I) The effect of LNP at different concentrations on cell viability. J) Hemolytic conditions of LNP at different concentrations. Results are presented as mean ± SD (n = 3 or 5).

In order to achieve optimal knockdown efficiency of the Fmr1 gene, we designed three different sequences of siRNA: siFmr1#1, siFmr1#2 and siFmr1#3, and prepared LNPs accordingly. After co-culturing the LNPs with 4T1 cells for 48 h, RNA was extracted for qPCR experiments and proteins were extracted for WB experiments. LNP@siFmr1#3 reduced Fmr1 mRNA levels to 39.9 ± 0.8 % of the control group (Fig. 2C). Furthermore, it decreased the FMRP protein level to 52.9 ± 1.6 % of the control group (Fig. 2D and E). Therefore, we selected siFmr1#3 as the optimal sequence, and in the subsequent sections, LNP@siFmr1 will refer to LNP@siFmr1#3.

We further investigated the effects of LNP@siFmr1 on PROS and IL33 through WB experiments. As shown in Fig. 2F, G, and H, the PROS level in the LNP@siFmr1 treatment group was 73.3 ± 0.8 % of that in the control group, while the IL33 level was 49.2 ± 3.6 % of that in the control group. The reduction of PROS and IL33 helps to relieve the suppression of tumor cells on the immune system, enhancing the anti-tumor effect of the immune system [33,34]. We assessed the toxicity of LNP@siFmr1 on cells using CCK8 assays, and there are no significant changes in cell viability across the concentration range from 6.25 nM to 200 nM (Fig. 2I), indicating that LNP@siFmr1 exhibits good safety and that the knockdown of the Fmr1 gene does not directly harm the cells. Furthermore, hemolysis experiments were further conducted to evaluate the biocompatibility of LNP@siFmr1. The hemolysis rate remained below 5 % across the concentration range of 12.5 nM–400 nM (Fig. 2J), confirming the safety of LNP@siFmr1 for intravenous injection.

3.3. In vivo distribution of LNP@siFmr1

The in vivo distribution of siFmr1-Cy5.5 and LNP@siFmr1-Cy5.5 was evaluated on an orthotopic breast cancer mouse model using the IVIS imaging system. Following administration via tail vein injection, free siFmr1-Cy5.5 was rapidly cleared and exhibited negligible fluorescence accumulation at the tumor site (Fig. 3A and B). In contrast, LNP@siFmr1-Cy5.5 demonstrated significant tumor-site accumulation at 4 h, 12 h, and 24 h post-administration, indicating the effective delivery of siRNA to the tumor through enhanced passive targeting mechanisms (Fig. 3A and B). Fluorescence imaging and quantification of excised tumor tissues further confirmed the markedly improved tumor-targeting performance of LNP@siFmr1-Cy5.5 over free siFmr1-Cy5.5 (Fig. 3C and D).

Fig. 3.

Tumor-targeting ability assessment of LNP@siFmr1. A) IVIS fluorescence imaging of tumor-bearing mice after administration of siFmr1-Cy5.5 and LNP@siFmr1-Cy5.5 at 4, 12, and 24 h. The red circle indicates the tumor site. B) Quantification of fluorescence at the tumor site. C) Imaging and D) quantification of fluorescence in excise tumors. E) Fluorescence imaging of major organs in vitro. F) Fluorescence quantification of major organs at 24 h. G) Fluorescence images of tumor sections after injection of LNP@siFmr1-Cy5.5 at 12 h. Blue: DAPI; red: Cy5.5. Scale bar = 10 μm. Results are presented as mean ± SD (n = 3).

The major organs (heart, liver, spleen, lung, kidney) and tumors were harvested for fluorescence imaging and quantification. The results further corroborate the enhanced targeting capability of LNP@siFmr1-Cy5.5. Both siFmr1-Cy5.5 and LNP@siFmr1-Cy5.5 demonstrated substantial accumulation in the liver and spleen, which can be attributed to the high vascularization of these organs and liver's central role in metabolic processes (Fig. 3E and F).

The tumors were excised for frozen sectioning at 12 h after administration of LNP@siFmr1-Cy5.5, and the fluorescence distribution was analyzed using a fluorescence microscope. The blue fluorescence indicated DAPI-stained nuclei, and the red fluorescence indicated siFmr1-Cy5.5. Prominent red signals were observed in the tumor tissues, within the tumor tissues, demonstrating an extensive and widespread distribution (Fig. 3G). These findings highlight the effective tumor-targeting capability of LNP@siFmr1-Cy5.5, demonstrating its potential to efficiently deliver siRNA to tumor site.

We also studied the pharmacokinetics of siFmr1- Cy5.5 and LNP@siFmr1- Cy5.5 in healthy mice, and the results are shown in Fig. S2. It can be observed that the area under the curve and blood half-life of LNP@siFmr1-Cy5.5 have significantly improved compared to siFmr1- Cy5.5. This indicates that the lipid nanoparticles successfully enhanced the stability and in vivo retention time of siRNA, laying the foundation for achieving good therapeutic effects.

3.4. Investigation of therapeutic efficacy

The therapeutic efficacy was evaluated using the mouse orthotopic breast cancer model. Treatment began when the tumor volume reached approximately 100 mm³. Tumor-bearing mice were randomly assigned to four groups: the PBS control group (G1), the siFmr1 group (G2), the LNP@siFmr1 group (G3), and the combination treatment group comprising LNP@siFmr1 and PD-1 antibody (G4). Among these, PBS, siFmr1, and LNP@siFmr1 were administered via tail vein injection every two days for a total of seven times; PD-1 antibody was administered via intraperitoneal injection every four days for a total of four times. The mice were monitored continuously and euthanized at the experimental endpoint for data collection. The overall process is illustrated in Fig. 4A.

Fig. 4.

Investigation of therapeutic efficacy. A) Schematic diagram of dosing regimen. B-C) Tumor volume curves of PBS, siFmr1, LNP@siFmr1 and LNP@siFmr1 + anti-PD-1 group. D) Excised tumors and E) tumor weight of different groups. F) Body weight monitoring of tumor-bearing mice in different groups. G) H&E staining, TUNEL staining and Ki67 staining of tumor tissues. Scale bar = 50 μm. P values were analyzed by one-way ANOVA test. ∗∗∗∗P < 0.0001.

The tumor volume of different groups was monitored, and the individual tumor volume trajectories for each mouse were depicted (Fig. 4B and C). The siFmr1 group exhibited minimal therapeutic efficacy compared to PBS group, which was attributed to the rapid metabolism of siFmr1 in the body and its limited accumulation at the tumor site. In contrast, LNP@siFmr1 demonstrated a significant inhibitory effect on tumor growth, as the lipid nanoparticles effectively delivered siFmr1 into tumor cells, knocked down Fmr1 gene expression, and activated cytotoxic immune responses against the tumor. WB analysis confirmed the significant inhibition of FMRP expression in tumor tissues (Fig. S3). Moreover, the tumor weight of the combination therapy of LNP@siFmr1 with PD-1 antibodies group was approximately 80 % of that in PBS group, significantly suppressed tumor growth. The addition of PD-1 antibodies further alleviated the immunosuppressive effects of tumor cells on T cells, thereby effectively activating the anti-tumor immune response. The image and weight of excise tumors further confirmed the therapeutic efficacy (Fig. 4D and E).

Additionally, H&E staining was performed on tumor tissues to evaluate histopathological changes (Fig. 4G). In the PBS control group, the majority of tumor cells were viable, exhibiting a dense arrangement and intact nuclei. In contrast, the siFmr1 group demonstrated a slight increase in cell death compared to the PBS control group. Notably, the proportion of dead tumor cells was significantly elevated in the LNP@siFmr1 group and the PD-1 combination treatment group. These groups displayed prominent nuclear fragmentation and progressively enlarged intercellular spaces, indicative of an enhanced tumor-killing effect. Among all treatments, the combination therapy of LNP@siFmr1 and PD-1 antibodies demonstrated the most pronounced therapeutic efficacy, consistent with the observed inhibition of tumor growth. We further performed TUNEL staining on tumor tissues to identify dead cells, indicated by green fluorescence. The results revealed that the PD-1 combination treatment group exhibited the highest proportion of green fluorescence, confirming its superior tumor-killing efficacy (Fig. 4G). Additionally, immunohistochemical staining was conducted to detect Ki67-positive cells in tumor tissues, a key marker for evaluating tumor cell proliferation. The findings showed that the PD-1 combination treatment group had the lowest proportion of Ki67-positive cells, demonstrating the strongest ability to inhibit tumor cell proliferation (Fig. 4G). These findings collectively demonstrate that the delivery of siFmr1 effectively induces tumor cell death while significantly enhancing the therapeutic efficacy of PD-1 treatment. The synergistic efficacy of siFmr1 delivery and PD-1 therapy underscores their potential as a powerful combinatorial approach for robust tumor suppression.

The body weight of mice was monitored, revealing no significant weight loss of treatment groups (Fig. 4F). Liver and kidney function were assessed by measuring plasma levels of ALT, AST, BUN, and CR, which remained within the normal physiological range (Fig. S4). Histopathological examination through H&E staining of major organs, including the heart, liver, spleen, lungs, and kidneys, revealed no evident pathological changes or tissue damage (Fig. S5). These findings collectively indicated that LNP@siFmr1 exhibited minimal toxicity at the experimental concentration, underscoring its favorable safety profile.

3.5. Immune activation assessment

We measured the levels of the TNF-α and IFN-γ within the tumor microenvironment as indicators of T cell activation and antitumor potential. Results demonstrated that the PD-1 antibody combination therapy group exhibited significantly elevated levels of TNF-α and IFN-γ, which were 3.4-fold and 2.8-fold higher compared to the PBS group, respectively (Fig. S6).

To further evaluate immune activation, the levels of CD8⁺ T cells, CD86⁺ macrophages, and CD80⁺CD86⁺ DCs in tumor tissues and lymph nodes (LNs) were quantified using flow cytometry (Figs. S7–9). The proportion of immunostimulatory macrophages in tumor tissues significantly increased in the LNP@siFmr1 group (40.07 ± 5.22 %) compared to the PBS group (27.33 ± 2.72 %) and siFmr1 group (30.10 ± 3.46 %). This finding suggests that the effective delivery of siFmr1 to tumor tissues facilitated FMRP knockdown, which subsequently promoted the conversion of macrophages into immunostimulatory phenotypes. Moreover, combined treatment with a PD-1 antibody further elevated the proportion of CD86⁺ macrophages (46.40 ± 7.31 %), indicating enhanced activation within the tumor microenvironment (Fig. 5A and B).

Fig. 5.

Immune activation assessment. A-B) Percentage of CD86⁺ macrophage in tumor. C-D) Percentage of CD8⁺ T cells in tumor. E-F) Percentage of CD8⁺ T cells in lymph nodes. G-H) Percentage of CD80⁺CD86⁺ DCs in lymph nodes. Results are presented as mean ± SD (n = 3). P values were analyzed by one-way ANOVA test.

Furthermore, the proportion of CD8⁺ T cells in tumor tissues showed a notable increase in the LNP@siFmr1 group (55.93 ± 4.48 %) and PD-1-combination therapy group (70.60 ± 4.83 %), which were significantly higher than those observed in the PBS group (32.93 ± 9.52 %) (Fig. 5C and D). These findings demonstrate that FMRP knockdown improved immune activation in the tumor microenvironment and the combined administration of PD-1 antibody amplified CD8⁺ T cell activation, further augmenting the cytotoxic functions.

Given that that lymph nodes (LNs) are critical sites of immune activation and harbor abundant immune cells, we assessed the immune activation status within the tumor-draining lymph nodes. The combination treatment group exhibited significantly elevated proportions of CD8⁺ T cells (43.83 ± 3.10 %) and CD80⁺CD86⁺ DCs (37.50 ± 2.05 %) compared to the PBS group (28.47 ± 5.45 % and 18.97 ± 2.11 %, respectively). These findings highlight the ability of the combination treatment to stimulate immune cell activation within the LNs, thereby amplifying anti-tumor immune responses (Fig. 5E–H).

3.6. Limitations and prospects

Although this study presents very appealing results, there are still some limitations. LNP@siFmr1 can effectively target tumor sites; however, there is also a certain amount of distribution in the liver, spleen, and lungs, where potential off-target effects and cumulative toxicity during prolonged administration require further investigation. Meanwhile, although the activation of the immune system can effectively inhibit tumor growth, it does not completely eradicate the tumor, necessitating further combination with other drugs or therapies to achieve complete tumor cure [35,36].

In future research, we aim to further enhance the targeting efficiency and biosafety of lipid nanoparticles through methods such as formulation optimization [37,38] and biomimetic modification of membrane proteins [39,40]. Additionally, various types of nucleic acid molecules, such as mRNA and plasmids [41,42], can be delivered to exert different therapeutic effects. Furthermore, this approach may be applicable to the treatment of a wider range of diseases, including stroke [43], kidney injury [44], and glioma [45,46].

4. Conclusion

The immunosuppressive tumor microenvironment poses a significant challenge in TNBC treatment. Given the critical role of FMRP in tumor progression, we designed an siRNA (siFmr1) for inhibiting the expression and innovatively applied a microfluidic-based electrospray method to fabricate siRNA-loaded LNP for the first time. LNP@siFmr1 demonstrated superior particle size, encapsulation efficiency, and stability, and exhibited efficient tumor-targeting capacity in an orthotopic mouse breast cancer model. siFmr1 was successfully delivered to tumor tissues, silenced FMRP expression, and facilitated the infiltration and activation of immune cells, leading to significant inhibition of tumor growth while exhibiting excellent biosafety. Moreover, combining LNP@siFmr1 with PD-1 antibody treatment achieved complete tumor growth suppression, confirming the improvement of the tumor immune microenvironment and the activation of anti-tumor immunity. This novel strategy provides promising insights for advancing tumor immunotherapy, particularly in overcoming resistance to PD-1 antibody therapy in clinical settings.

CRediT authorship contribution statement

Letian Bai: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Xinyan Hao: Writing – review & editing, Methodology, Data curation. Wenjie Xu: Methodology. Hai Huang: Methodology. Mengen Guo: Software. Yanwen Zhang: Writing – review & editing. Daxiong Xiang: Conceptualization. Guoping Yang: Supervision, Funding acquisition. Junyong Wu: Methodology, Funding acquisition. Chuanpin Chen: Supervision, Funding acquisition.

Declaration of competing interest

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.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

Data will be made available on request.