HF@SLN remodels the tumor immune microenvironment and suppresses pancreatic tumor progression

We further explored whether the HF@SLN can reprogram M2 BMDMs in vivo. First, we investigated the biodistribution of HF@SLN by intravenously injecting cyanine 7-labeled HF@SLN into subcutaneous Pan02-bearing mice. The accumulation of HF@SLN reached the peak value in the tumor 4 h after injection; in addition, there was no apparent HF@SLN accumulation in the tumor 24 h after injection, which is proved by both qualitative and quantitative analysis (Fig. 3a and b). Moreover, 4 h after injection, most HF@SLN located in the liver (46.4 ± 8.6%), followed by the tumor (10.8 ± 8.1%), kidney (7.9 ± 5.1%), spleen (3.9 ± 2.9%), lung (3.2 ± 2.2%), and heart (2.9 ± 2.0%) (Fig. 3c and d), which was consistent with the expected biodistribution behavior of lipid-based nanomaterials [29].

Fig. 3: Biodistribution, biosafety, and therapeutic effects of HF@SLN NPs in the subcutaneous pancreatic cancer model. (A) Biodistribution of cyanine 7-labeled HF@SLN and (B) quantitative analysis in tumor tissues (n = 5), after intravenous injection. (C) Biodistribution of cyanine 7-labeled HF@SLN and (D) quantitative analysis in main organs (n = 5), after 4 h intravenous injection. (E) IVIS images of mice from all groups after surgery and on day 30 (n = 10). (F) Tumor weight of each mouse (n = 10). (G) Body weight of mice (n = 10). Serum concentration of (H) ALT and (I) AST in mice (n = 10). Serum concentration of (J) TGF-β and (K) TNF-α in mice (n = 10). The percentages of (L) CD80+ macrophages, (M) CD206+ macrophages, and (N) the ratio of CD80+/CD206+ macrophages in mice (n = 10). The percentages of (O) CD4+ T cells, (P) CD8+ T cells, and (Q) the ratio of CD8+/CD4+ T cells in mice (n = 10). Data are presented as means ± SD. The error bars are based on standard errors of the individual animals. *Indicates the statistically significant differences comparing HW and HF@SLN with saline. #Statistically significant differences of HW and HF@SLN. ***P < 0.001, **P < 0.01, *P < 0.05 *, ###P < 0.001, and #P < 0.05. ANOVA with Tukey's post-test.

The therapeutic efficacy of HF@SLN was first evaluated in a postoperative recurrence model of subcutaneous pancreatic cancer. Partial tumor resection was performed to mimic clinically relevant residual disease. Two days after surgery, tumors were visualized, and the mice were randomly divided into three groups with different treatments (Saline, HW+50 mg/kg gemcitabine, and HF@SLN+50 mg/kg gemcitabine). After 30 days of treatments, the tumors were visualized again, and inhibited tumor growth was observed in both HW and HF@SLN groups (Fig. 3e). In addition, the HF@SLN groups showed no statistically significant difference in tumor weight (0.14 ± 0.07g) compared to the HW groups (0.25 ± 0.11g) and saline groups (0.69 ± 0.54g) (Fig. 3f). During the whole therapeutic process, the average body weight steadily increased in all groups without statistically significant differences (Fig. 3g). Additionally, the histology of main organs, such as the heart, liver, spleen, lung, and kidney, was organized in all groups (Fig. S5). The concentration of alanine transaminase (ALT) and aspartate transaminase (AST) were similar among the groups without statistically significant differences (Fig. 3h and i), indicating that liver function was normal in all groups. Increasing body weights, organized histology of main organs, and normal liver functions proved the biosafety of HW and HF@SLN during the therapeutic process.

We next examined whether HF@SLN-mediated tumor suppression was associated with remodeling of the tumor immune microenvironment (TME). Cytokine analysis revealed that HF@SLN statistically significantly reduced the immunosuppressive cytokine (transforming growth factor beta, TGF-β, 368.6 ± 43.9 pg/ml) while simultaneously increasing pro-inflammatory (tumor necrosis factor-alpha, TNF-α, 441.1 ± 30.4 pg/mL) levels compared to the HW (TGF-β: 418.2 ± 47.8 pg/ml and TNF-α: 376.5 ± 59.7 pg/mL) and saline (TGF-β: 487.4 ± 27.4 pg/mL and TNF-α: 309.9 ± 50.9 pg/mL) groups (Fig. 3j and k), suggesting conversion of the TME from an immune-suppressive to immune-activated state.

Flow cytometric analysis further demonstrated pronounced macrophage repolarization within tumors following HF@SLN treatment. Specifically, tumors from each mouse were prepared for sing-cells suspensions for the following flow cytometry analysis (gating strategy was presented in Fig. S6–9). Mice treated with HF@SLN, and HW, respectively, exhibited increased percentages of CD80+ macrophages (HF@SLN: 47.1 ± 3.3%, HW:34.5 ± 9.1%, and Saline:20.3 ± 9.0%) and decreased percentages of CD206+ macrophages (HF@SLN: 18.0 ± 2.8%, HW:30.0 ± 5.9%, and saline: 45.3 ± 11.0%) compared to the saline groups (Fig. 3l–n). Additionally, there was a decreased percentage of CD4+ T cells in mice treated with HF@SLN (Fig. 3o), but there were no statistically significant differences in CD8+ T cells in all groups (Fig. 3p). However, mice treated with HF@SLN and HW presented a higher CD8+/CD4+ ratio than the Saline group (HF@SLN: 2.5 ± 0.6, HW:1.7 ± 0.4, and saline: 1.2 ± 0.4, (Fig. 3q), indicating a more cytotoxic immune landscape.

To further elucidate the molecular basis underlying HF@SLN-mediated immune remodeling, transcriptomic analysis was performed on tumors collected from saline- and HF@SLN-treated mice. Gene ontology enrichment analysis revealed substantial activation of immune-associated biological processes following HF@SLN treatment, including leukocyte activation, lymphocyte activation, T-cell activation, adaptive immune responses, and immune system processes (Fig. 4a). These transcriptomic findings strongly support the ability of HF@SLN to convert immunologically cold pancreatic tumors into more immune-responsive microenvironments.

Fig. 4: The therapeutic effects and potent immunotherapeutic machines of HF@SLN NPs in the orthotopic pancreatic cancer model. (A) The top 25 enhanced biological processes on mice treated with HF@SLN compared to the saline in the transcriptome analysis. (B) IVIS images of mice from all groups on day 7, 21, and 30 (n = 10). (C) Survival rate of Pan02 orthotopic tumor-bearing mice treated with saline, HW, and HF@SLN (n = 10). (D) Tumor weight of mice treated with saline (n = 6), HW and HF@SLN (n = 10). Percentages of (E) CD206+ macrophages, (F) CD80+ macrophages, and (G) the ratio of CD80+/CD206+ macrophages in mice (n = 10). Percentages of (H) CD8+ T cells, (I) CD4+ T cells, and (J) the ratio of CD8+/CD4+ T cells in mice (n = 10). Immunofluorescence staining results of (K) human tumour treated HF@SLN in the expression of TFEB and (L) the local magnification. Data are presented as means ± SD. The error bars are based on standard errors of the individual animals. *Indicates the statistically significant differences comparing HW and HF@SLN with Saline. #Statistically significant differences of HW and HF@SLN. ***P < 0.001, **P < 0.01, ###P < 0.001, ##P < 0.01, and #P < 0.05. ANOVA with Tukey's post-test.

Encouraged by these results, we further evaluated HF@SLN in an orthotopic pancreatic tumor model that more faithfully recapitulates clinical pancreatic cancer progression. As shown in Fig. 4b, most mice developed orthotopic pancreatic cancer after seven days. Then the mice were respectively treated with saline, HW+50 mg/kg gemcitabine, and HF@SLN+50 mg/kg gemcitabine, and the tumor situation was continuously recorded after 21 and 28 days. All mice treated with HW or HF@SLN survived throughout the treatment period, whereas four mice in the saline group died during therapy (Fig. 4c). The gradually increased body weight was observed in all groups (Fig. S10). After 30 days of treatments, HF@SLN treatment also achieved the strongest tumor growth inhibition, reducing tumor weight to 0.37 ± 0.12 g compared to saline group (1.16 ± 0.44g) and HW group (0.69 ± 0.17g) (Fig. 4d).

Consistent with observations in the subcutaneous model, HF@SLN profoundly remodeled the orthotopic pancreatic TME by decreasing CD206+ macrophages, increasing CD80+ macrophages, and enhancing CD8+ T cell infiltration. Notably, the HF@SLN group exhibited the highest CD8+/CD4+ ratio together with the lowest M2/M1 macrophage ratio, further supporting effective immune activation within tumors. Specifically, after processing the tumor tissues into the single cell suspension, the saline group had the highest percentage of CD206+ macrophages (43.78 ± 7.47%) compared to the HW group (23.56 ± 6.35%) and HF@SLN (13.87 ± 4.15%) (Fig. 4e). The lowest percentage of CD80+ macrophages (13.05 ± 5.97%) was observed in the saline group compared to the HW (26.39 ± 7.21%) and HF@SLN (39.91 ± 4.63%) groups (Fig. 4f). Interestingly, the HF@SLN group presented the highest CD8+ percentage of T cells (32.43 ± 3.17%) followed by the HW (29.50 ± 1.88%) and saline (22.23 ± 1.86%) groups. However, there were no statistically significant differences in all groups CD4+ percentage of T cells (Fig. 4h and i). HW and HF@SLN groups exhibited an increased M2/M1 ratio and CD8+/CD4+ compared to the saline groups, indicating the activated TME (Fig. 4g–j). Finally, ex vivo treatment of human pancreatic tumor tissues with HF@SLN induced robust TFEB activation compared with free HW or saline controls (Fig. 4k, l and Fig. S11), highlighting the translational potential of this strategy in clinically relevant human tumor tissues.

Overall, these findings demonstrate that coordination-driven assembly of bioactive therapeutics into aMDFs represents a versatile strategy for simultaneously improving drug delivery and preserving intrinsic pharmacological activity. By integrating PIKfyve inhibition with nanostructure-mediated tumor delivery, HF@SLN effectively activates TFEB signaling, reprograms immunosuppressive macrophages, remodels the pancreatic tumor microenvironment, and potentiates antitumor immunity, thereby providing a promising nanomedicine platform for enhancing pancreatic cancer immunotherapy.