Nature Communications2023ResearchNon-viral Gene Delivery

Engineering tumor-specific gene nanomedicine to recruit and activate T cells for enhanced immunotherapy

Yue Wang, Shi-Kun Zhou, Yan Wang, Zi-Dong Lu, Yue Zhang, Cong-Fei Xu, Jun Wang29 September 2026DOI 10.1038/s41467-023-37656-w

Read the original ↗

Summary

PD-1/PD-L1 blockade therapy is successful but often yields poor benefits due to insufficient T-cell infiltration and low intratumoral concentrations of PD-1/PD-L1 inhibitors. While strategies exist to increase either T-cell recruitment or inhibitor delivery, none actively recruit T cells while achieving tumour-specific delivery of PD-L1 inhibitors to specifically eliminate inhibition of tumour-infiltrating T cells. A strategy that addresses both. ### Nanoparticle Characterisation | Parameter | Value | |---------------|-----------| | Hydrodynamic diameter | 107.2 nm | | PDI | ~0.2 | | Zeta potential | +15.3 mV | | Stability in 10% FBS | ≥5 days (size/PDI stable).

Keywords

T cells Antigen presentation mRNA Nanoparticles PLGA Gene delivery Tumour microenvironment

Purpose: PD-1/PD-L1 blockade therapy is successful but often yields poor benefits due to insufficient T-cell infiltration and low intratumoral concentrations of PD-1/PD-L1 inhibitors. While strategies exist to increase either T-cell recruitment or inhibitor delivery, none actively recruit T cells while achieving tumour-specific delivery of PD-L1 inhibitors to specifically eliminate inhibition of tumour-infiltrating T cells. A strategy that addresses both hurdles simultaneously is needed.

Hypothesis: Tumour-specific gene nanomedicines utilising tumour-specific promoters (tyrosinase for melanoma; survivin for multiple tumours) to drive tumour cells to co-express CXCL9 (T-cell chemokine) and anti-PD-L1 scFv (αPD-L1) will: (1) establish a CXCL9 gradient for T-cell recruitment, (2) achieve high intratumoral αPD-L1 concentration for T-cell activation, and (3) result in enhanced antitumor efficacy across multiple tumour types with improved safety compared to systemic anti-PD-L1 antibody administration.

Aims: 1. Engineer and characterise melanoma-specific (NP_Tyr-C9AP) and universal (NP_Sur-C9AP) gene nanomedicines using PEG-PLGA/DOTAP nanoparticles encapsulating promoter-driven plasmids encoding CXCL9 and αPD-L1 2. Validate tumour-specific co-expression of CXCL9 and αPD-L1 in vitro and in vivo using tyrosinase (melanoma) and survivin (pan-tumour) promoters 3. Evaluate the synergistic effects of CXCL9 and αPD-L1 co-expression on T-cell recruitment, activation, and tumour cell killing in vitro using transwell migration and cytotoxicity assays 4. Assess therapeutic efficacy in multiple tumour models (B16-F10, Clone M-3, YUMM1.7 melanoma; CT26 colorectal; Panc02 pancreatic; 4T1 breast) and compare to systemic anti-PD-L1 antibody treatment 5. Characterise the tumour immune microenvironment after treatment, including T-cell infiltration, activation, and immunosuppressive cell populations

Delivery system:

Component: Nanoparticle Core; Description: PEG-PLGA (methoxy PEG5k-PLGA15k) + cationic lipid DOTAP (1,2-dioleoyl-3-trimethylammoniumpropane)

Component: Particle Properties; Description: • Hydrodynamic diameter: 107.2 nm<br>• PDI: ~0.2<br>• Zeta potential: +15.3 mV<br>• Spherical vesicle structure with PEG shell<br>• Stable in 10% FBS for ≥5 days

Component: Plasmid Payload (pTyr-C9AP); Description: Tyrosinase promoter-driven co-expression plasmid encoding CXCL9 and αPD-L1 (anti-PD-L1 scFv) with P2A self-cleaving peptide; includes EGFP reporter for promoter activity

Component: Plasmid Payload (pSur-C9AP); Description: Survivin promoter-driven co-expression plasmid (universal tumour-specific promoter)

Component: Control Plasmids; Description: pTyr-CXCL9 (CXCL9 only), pTyr-αPD-L1 (αPD-L1 only), pUC57 (empty vector)

Component: αPD-L1 Design; Description: Anti-PD-L1 scFv with 6×His tag for detection

Component: Gene Delivery; Description: Intravenous injection; 1 mg plasmid/kg body weight; every other day for 5 injections

Component: Targeting Mechanism; Description: Promoter-driven tumour-specific expression (not particle targeting); NPs deliver to both tumour and normal tissues but promoters restrict expression to tumour cells

Component: Mouse Models; Description: B16-F10, Clone M-3, YUMM1.7 (melanoma); CT26 (colorectal); Panc02 (pancreatic); 4T1 (breast)

Component: Mouse Strains; Description: C57BL/6 (B16-F10, Panc02, YUMM1.7); BALB/c (CT26, 4T1)

Approach:

Parameter: In Vitro Specificity; Details: 8 cell lines (B16-F10, CT26, Panc02, 4T1, C2C12, NIH/3T3, DC2.4, RAW264.7) transfected with NP_Tyr-C9AP; EGFP expression, CXCL9 and αPD-L1 mRNA/protein measured

Parameter: In Vivo Specificity (Tyr); Details: Bilateral tumour model: B16-F10 melanoma + Panc02 pancreatic tumour; i.v. NP_Tyr-C9AP; EGFP+ cells and CXCL9/αPD-L1 concentrations in tumours and organs

Parameter: In Vivo Specificity (Sur); Details: Four tumour models (B16-F10, CT26, Panc02, 4T1) + normal tissues; NP_Sur-C9AP; CXCL9/αPD-L1 concentrations measured

Parameter: T-Cell Recruitment Assay; Details: Transwell migration: supernatants from transfected B16-F10 cells in lower chamber; DiI-labelled CD8⁺ T cells in upper chamber; cell counting with Precision Count Beads

Parameter: PD-L1 Binding Assay; Details: IFN-γ-stimulated B16-F10 cells incubated with supernatants; αPD-L1 binding detected by anti-6×His tag antibody (flow cytometry)

Parameter: T-Cell Cytotoxicity Assay; Details: IFN-γ-stimulated B16-F10-OVA-EGFP cells + OVA-specific CD8⁺ T cells in culture supernatants; PI staining and CCK-8 for viability

Parameter: Transwell Migration + Cytotoxicity; Details: Combined assay: B16-F10-OVA-EGFP in lower chamber, OVA-specific CD8⁺ T cells in upper chamber; recruitment and killing measured

Parameter: Therapeutic Studies; Details: Tumour-bearing mice (n=6-10/group); 5 i.v. injections (every other day); tumour volume, weight, survival monitored

Parameter: Immune Profiling; Details: Flow cytometry: CD3⁺, CD4⁺, CD8⁺, CD69⁺, granzyme B⁺, perforin⁺, IFN-γ⁺ T cells; NK cells; MDSCs; Tregs; M1/M2 macrophage ratio

Parameter: Comparison to Anti-PD-L1; Details: Anti-PD-L1 antibody (2.5 mg/kg, i.v.) vs. NP_Tyr-αPD-L1 vs. NP_Tyr-C9AP vs. NP_Tyr-CXCL9 + anti-PD-L1

Parameter: Safety Assessment; Details: Body weight; H&E of major organs; ALT, AST, ALP (liver function)

Parameter: Sample Sizes; Details: n=4 (expression studies); n=6-10 (therapy studies); n=4 (toxicity)

Parameter: Statistical Tests; Details: Two-sided t-test; one-way ANOVA with Tukey's; two-way ANOVA with Greenhouse-Geisser; log-rank (Mantel-Cox)

Key methods:

Analysis Category: Nanoparticle Characterisation; Methods: DLS (size, PDI, zeta potential); TEM (morphology); stability in 10% FBS

Analysis Category: Gene Expression; Methods: qRT-PCR (CXCL9, αPD-L1 mRNA); ELISA (CXCL9, αPD-L1 protein); Western blot (αPD-L1)

Analysis Category: Promoter Specificity; Methods: EGFP reporter fluorescence (microscopy, flow cytometry)

Analysis Category: T-Cell Recruitment; Methods: Transwell migration assay; DiI or Hoechst 33342 labelling; Precision Count Beads for absolute cell counting; CLSM imaging

Analysis Category: PD-L1 Binding; Methods: Flow cytometry: Alexa Fluor 647 anti-6×His tag antibody

Analysis Category: T-Cell Cytotoxicity; Methods: PI staining (flow cytometry); CCK-8 viability assay

Analysis Category: Flow Cytometry (Immune Profiling); Methods: Antibodies: CD45, CD3, CD4, CD8, CD69, NK-1.1/NKp46, CD11b, Gr-1, F4/80, CD80, CD206, Foxp3, CD25, granzyme B, perforin, IFN-γ; BD FACSCelesta; FlowJo v10.0.7

Analysis Category: Histology; Methods: H&E staining; multiple immunofluorescence (CD3, CD8, CD69); LSM880 confocal microscopy

Analysis Category: Liver Toxicity; Methods: ALT, AST, ALP ELISA kits (Rayto)

Analysis Category: Statistical Analysis; Methods: GraphPad Prism v8.0

Key results: ### Nanoparticle Characterisation

Parameter: Hydrodynamic diameter; Value: 107.2 nm

Parameter: PDI; Value: ~0.2

Parameter: Zeta potential; Value: +15.3 mV

Parameter: Stability in 10% FBS; Value: ≥5 days (size/PDI stable)

NP_Tyr-C9AP Specificity (In Vitro):

Cell Type: B16-F10 (melanoma); EGFP+ Cells (%): 41.2%; CXCL9 (ng/mL): 0.47; αPD-L1 (ng/mL): 2.19

Cell Type: CT26 (colorectal); EGFP+ Cells (%): 3.2%; CXCL9 (ng/mL): Minimal; αPD-L1 (ng/mL): Minimal

Cell Type: Panc02 (pancreatic); EGFP+ Cells (%): <1%; CXCL9 (ng/mL): Minimal; αPD-L1 (ng/mL): Minimal

Cell Type: 4T1 (breast); EGFP+ Cells (%): <1%; CXCL9 (ng/mL): Minimal; αPD-L1 (ng/mL): Minimal

Cell Type: NIH/3T3 (fibroblast); EGFP+ Cells (%): 9.0%; CXCL9 (ng/mL): Minimal; αPD-L1 (ng/mL): Minimal

Cell Type: C2C12, DC2.4, RAW264.7; EGFP+ Cells (%): <1%; CXCL9 (ng/mL): Minimal; αPD-L1 (ng/mL): Minimal

Cell Type: Tyr promoter specificity; EGFP+ Cells (%): Melanoma-specific; CXCL9 (ng/mL): -; αPD-L1 (ng/mL): -

NP_Tyr-C9AP In Vivo Specificity (Bilateral Tumour Model):

Tissue: B16-F10 melanoma; EGFP+ Cells (%): 9.11%; CXCL9 (ng/g): 12.12; αPD-L1 (ng/g): 43.12

Tissue: Panc02 pancreatic tumour; EGFP+ Cells (%): Few cells; CXCL9 (ng/g): Not increased; αPD-L1 (ng/g): Not increased

Tissue: Heart, liver, spleen, lung, kidney; EGFP+ Cells (%): Few cells; CXCL9 (ng/g): Not increased; αPD-L1 (ng/g): Not increased

In Vitro T-Cell Recruitment (Transwell, 180 min):

Group: PBS; CD8⁺ T Cells Recruited (×10⁵): <0.1; CXCL9 (ng/mL): 0.02

Group: NP_Control; CD8⁺ T Cells Recruited (×10⁵): <0.1; CXCL9 (ng/mL): 0.02

Group: NP_Tyr-CXCL9; CD8⁺ T Cells Recruited (×10⁵): 1.26; CXCL9 (ng/mL): 0.44

Group: NP_Tyr-C9AP; CD8⁺ T Cells Recruited (×10⁵): 1.14; CXCL9 (ng/mL): 0.42

In Vitro T-Cell Cytotoxicity (PI-Positive B16-F10-OVA-EGFP):

Group: PBS; Dead Cells (%): 1.9%

Group: NP_Control; Dead Cells (%): 1.3%

Group: NP_Tyr-CXCL9 (recruitment only); Dead Cells (%): 12.7%

Group: NP_Tyr-αPD-L1 (activation only); Dead Cells (%): 7.6%

Group: NP_Tyr-C9AP (both); Dead Cells (%): 45.8%

In Vivo Antitumor Efficacy (B16-F10 Melanoma):

Group: PBS; Tumour Growth Inhibition Rate: -; Final Tumour Weight (% of PBS): 100%; Median Survival (days): ~27

Group: NP_Control; Tumour Growth Inhibition Rate: -; Final Tumour Weight (% of PBS): -; Median Survival (days): ~27

Group: NP_Tyr-CXCL9; Tumour Growth Inhibition Rate: 59.0%; Final Tumour Weight (% of PBS): 36.9%; Median Survival (days): 31

Group: NP_Tyr-αPD-L1; Tumour Growth Inhibition Rate: 69.5%; Final Tumour Weight (% of PBS): 29.5%; Median Survival (days): 30

Group: NP_Tyr-C9AP; Tumour Growth Inhibition Rate: 84.3%; Final Tumour Weight (% of PBS): 16.7%; Median Survival (days): 34

Tumour Immune Microenvironment (B16-F10, Post-Treatment):

Parameter: CD3⁺ T cells (% of CD45⁺); PBS: ~5%; NP_Tyr-CXCL9: 10.4%; NP_Tyr-αPD-L1: 8.9%; NP_Tyr-C9AP: 21.4%

Parameter: CD3⁺ T cells (×10⁶/g tumour); PBS: ~0.4; NP_Tyr-CXCL9: ~0.8; NP_Tyr-αPD-L1: ~0.7; NP_Tyr-C9AP: 1.87

Parameter: CD8⁺ T cells (×10⁶/g tumour); PBS: ~0.2; NP_Tyr-CXCL9: ~0.4; NP_Tyr-αPD-L1: ~0.3; NP_Tyr-C9AP: 0.94

Parameter: CD69⁺ CD8⁺ T cells (×10⁵/g); PBS: ~0.5; NP_Tyr-CXCL9: ~0.5; NP_Tyr-αPD-L1: ~1.1; NP_Tyr-C9AP: 2.33

Parameter: Granzyme B⁺ CD8⁺ T cells (%); PBS: ~15%; NP_Tyr-CXCL9: ~25%; NP_Tyr-αPD-L1: ~35%; NP_Tyr-C9AP: 68.4%

Parameter: Perforin⁺ CD8⁺ T cells (%); PBS: ~10%; NP_Tyr-CXCL9: ~20%; NP_Tyr-αPD-L1: ~25%; NP_Tyr-C9AP: 53.2%

Parameter: IFN-γ⁺ CD8⁺ T cells (%); PBS: ~5%; NP_Tyr-CXCL9: ~10%; NP_Tyr-αPD-L1: ~15%; NP_Tyr-C9AP: 31.1%

Parameter: CD8⁺/Treg ratio; PBS: ~3; NP_Tyr-CXCL9: ~5; NP_Tyr-αPD-L1: ~6; NP_Tyr-C9AP: 13.37

Parameter: MDSCs (% of CD45⁺); PBS: High; NP_Tyr-CXCL9: Reduced; NP_Tyr-αPD-L1: Reduced; NP_Tyr-C9AP: Reduced

Parameter: M1/M2 macrophage ratio; PBS: Low; NP_Tyr-CXCL9: Increased; NP_Tyr-αPD-L1: Increased; NP_Tyr-C9AP: Increased

NP_Tyr-C9AP vs. Anti-PD-L1 Antibody (B16-F10):

Group: Anti-PD-L1 antibody; Tumour Growth Inhibition Rate: 58.0%; Body Weight Change: Significant decrease

Group: NP_Tyr-αPD-L1; Tumour Growth Inhibition Rate: 64.2%; Body Weight Change: No significant change

Group: NP_Tyr-CXCL9 + anti-PD-L1; Tumour Growth Inhibition Rate: ~72%; Body Weight Change: Significant decrease

Group: NP_Tyr-C9AP; Tumour Growth Inhibition Rate: 78.2%; Body Weight Change: No significant change

NP_Tyr-C9AP in Other Melanoma Models:

Model: Clone M-3 melanoma; Tumour Growth Inhibition Rate: 69.5%

Model: YUMM1.7 melanoma; Tumour Growth Inhibition Rate: 56.0%

NP_Sur-C9AP (Universal) Efficacy in Multiple Tumours:

Tumour Model: B16-F10 melanoma; Tumour Growth Inhibition Rate: 84.4%; CD3⁺ T Cell Increase (vs. control): 16.8% of CD45⁺

Tumour Model: CT26 colorectal; Tumour Growth Inhibition Rate: 81.0%; CD3⁺ T Cell Increase (vs. control): 1.65×

Tumour Model: Panc02 pancreatic; Tumour Growth Inhibition Rate: 56.3%; CD3⁺ T Cell Increase (vs. control): 1.50×

Tumour Model: 4T1 breast; Tumour Growth Inhibition Rate: 66.5%; CD3⁺ T Cell Increase (vs. control): 3.67×

Safety:

Parameter: Body weight loss; Result: None observed

Parameter: Major organ damage (H&E); Result: None detected

Parameter: Liver toxicity (ALT, AST, ALP); Result: No significant increase

Interpretation: The authors conclude that "we have developed tumour-specific gene nanomedicines to specifically drive different tumour cells to coexpress CXCL9 and αPD-L1, which provide a solution for addressing both the lack of tumour-infiltrating T cells and the insufficient tumour enrichment of αPD-L1 inhibitors in ICB therapy." The strategy "skillfully addressed the two hurdles of αPD-L1 therapy." NP_Tyr-C9AP achieved 84.3% tumour growth inhibition in melanoma, and NP_Sur-C9AP demonstrated efficacy across multiple tumour types (colorectal, pancreatic, breast). The authors emphasise that "our nanomedicines-mediated local expression of αPD-L1 showed superior efficacy and safety when compared to systemic injection of anti-PD-L1 antibody." This approach "provides a strategy to maximise the immunotherapy outcome regardless of the heterogeneous tumour microenvironment."

10. Limitations (Explicitly Stated or Evident):

1. Plasmid safety concerns: The authors acknowledge that "the safety risks of plasmids, including immunogenicity and potential gene insertion, need to be investigated before clinical use."

2. No tumour antigen presentation enhancement: The authors note that "the lack of tumour antigen presentation is not a major hurdle of αPD-L1 therapy," but also state that "some strategies targeting tumour antigen presentation can be combined to further enhance the efficacy of our strategy."

3. Nanoparticle liver accumulation: The authors acknowledge that "nanoparticles are unable to selectively deliver drugs into one specific type of organ or cell, and preferentially accumulate in liver," though promoter specificity mitigates this for gene expression.

4. αPD-L1 scFv lacks ADCC/ADCP: The authors note that "αPD-L1 scFv lacks antibody-dependent cellular cytotoxicity or phagocytosis (ADCC/ADCP) functions which may reduce its antitumor efficacy in comparison to αPD-L1 antibody," but counter that atezolizumab (which also lacks ADCC/ADCP) has shown good clinical efficacy.

5. In vivo imaging of expression: The study did not include real-time imaging of CXCL9/αPD-L1 expression dynamics or T-cell trafficking in vivo.

6. No depletion studies: The study did not include T-cell depletion experiments to formally prove that therapeutic efficacy is CD8⁺ T-cell-dependent (though strongly suggested by immune profiling).

7. Survivin promoter specificity: While survivin is overexpressed in most cancers, some normal tissues (e.g., proliferating cells) may have low-level expression, potentially leading to off-target expression in certain normal tissues.

8. Therapeutic models are subcutaneous: All tumour models were subcutaneous; efficacy in orthotopic or metastatic models was not demonstrated.

9. Species specificity: Mouse-specific promoters and αPD-L1 scFv were used; translation to human would require human promoters (e.g., hTERT) and humanized scFv.

10. Dosing optimisation not explored: The study used a fixed dosing regimen (1 mg/kg, 5 injections); dose-response and schedule optimisation were not performed.

11. No antigen-specific T-cell tracking: While T-cell infiltration and activation were characterised, the study did not track antigen-specific T-cell responses (e.g., tetramer staining).

12. Long-term safety: The study assessed acute toxicity but did not evaluate long-term safety, potential autoimmunity, or the durability of therapeutic responses beyond the treatment period.

Report prepared based on the published Nature Communications article. For full experimental details, supplementary figures, and complete references, please refer to the original publication.

RSC pharmaceutics

Orthogonal ensemble and single-particle characterisation of antisense oligonucleotide lipid nanoparticles

Discover nano

Lipid nanoparticle-based mRNA delivery system for cancer management

Molecular cancer

siRNA-mediated mismatch repair disruption synergizes with oxaliplatin to restore immune surveillance in microsatellite-stable colorectal cancer

← Back to all articles

Let's engineer the next delivery breakthrough together

We co-develop nanocarrier and biosensing programmes with pharma, biotech and academic groups — from target selection through scale-up.

ResearchersRequest a technical consultationTalk formulation, characterisation or assay design with one of our scientists.Industry partnersExplore co-developmentFeasibility studies, co-development programmes and licensing.InvestorsExplore the market opportunitySourced market sizes, pipeline, investment and deal data for genetic medicine delivery.