Improving cancer immunotherapy via co-delivering checkpoint blockade and thrombospondin-1 downregulator

Abstract

The use of checkpoint-blockade antibodies is still restricted in several malignancies due to the modest efficacy, despite considerable success in anti-tumor immunotherapy. The poor response of cancer cells to immune destruction is an essential contributor to the failure of checkpoint therapy. We hypothesized that combining checkpoint therapy with natural-product chemosensitizer could enhance immune response. Herein, a targeted diterpenoid derivative was integrated with the checkpoint blockade (anti-CTLA-4) to improve immunotherapy using thermosensitive liposomes as carriers. In vivo, the liposomes enabled the co-delivery of the two drug payloads into the tumor. Consequently, the regulatory T cell proliferation was restrained, the cytotoxic T cell infiltration was enhanced, and the profound immunotherapeutic effect was achieved. In addition, the immunotherapeutic effect of another clinically used checkpoint antibody, anti-PD-1, also benefited from the diterpenoid derivative. Of note, our mechanism study revealed that the targeted diterpenoid derivative increased the sensitivity of cancer cells to immune attack via THBS1 downregulation and the resultant destruction of THBS1-CD47 interaction. Collectively, co-delivering THBS1 inhibitor and checkpoint blockade is promising to boost cancer immunotherapy. We first time discovered that THBS1 suppression could strengthen checkpoint therapy.

Keywords: Checkpoint blockade; Co-delivery; Diterpenoid-based conjugate; Immunotherapy; Liposomes; Thrombospondin-1.

Graphical abstract

Figure 1

Preparation and characterization. (A) Schematic illustration of HNP preparation. Effect of mass ratio of HA-ORD and LTSLs on (B) zeta potential and (C) diameter of HNPs. (D) FRET between FITC-HA-ORD and Rho-LTSLs (FITC/Rho, w/w) in HNPs. (E) Membrane-fluidity assay (n = 3). The fluorescence anisotropy (r) of the hydrophobic membrane fluorescence probe encapsulated inside the lipid bilayer was used to demonstrate liposome membrane fluidity. The anisotropy index, r, is calculated by the formula, r = I_∥−I_⊥I_∥+2gI_⊥, I_∥ and I_⊥ individually indicate the emission intensities of parallel and perpendicular polarization, and g demonstrates the correct factor specified by a spectrometer. (F) TEM determination. (G) In vitro release profile of HNPs at 42 or 37 °C. Data are presented as mean ± SD, n = 3. ∗∗∗P < 0.001, compared with the same time point at 37 °C.

Figure 2

Tissue biodistribution and tumor penetration. (A and B) Biodistribution of nanoparticles in different tissues. These tissues were isolated at 24 h after dosing. The DiR dose was 0.5 mg/kg, n = 3, ∗∗∗P < 0.001, compared with DiR; ^###P < 0.001, compared with DiR-LTSLs. (C) Study of HNP integrity by confocal imaging. Dual-labeled HNPs: HNPs co-loaded with Rho-HA-ORD (red) and CF-LTSLs (green). After dosing HNPs, the tumor was not subjected to local hyperthermia (HT) stimulation. The tumors were collected for sectioning at 2 h post-injection via the tail vein. Distribution of (D) FITC-labeled and (E) CF-labeled nanoparticles (green) in the tumors collected at 2 h after dosing at the dye 0.5 mg/kg, according to the animals' body weight. The tumor suffered from the local HT stimulation for 1 h after administration. Red spots represent the microvessels inside the tumors marked with Cy7-labeled CD31 antibody. The nuclei were marked with DAPI (blue). (F) Semi-quantitative data of the fluorescence intensity ratio of nanoparticles or the conjugate to CD31. ImageJ was used to calculate the positive area and signal. The ratio was calculated by dividing the green signal by the red signal. Scale bars: 20 μm. Data are presented as mean ± SD, n = 3, ∗P < 0.05 vs. indicated.

Figure 3

Anti-tumor immunotherapy in the normal tumor model. (A) Alterations of tumor volume. n = 6. Quantified (B) apoptosis and (C) proliferation were determined by TUNEL assay and Ki67 immunochemistry, respectively. Lymphocyte expression in the (D to F) tumors and (G to I) spleens was collected at the treatment end. The immune cells were analyzed by collecting single-cell suspensions from the tissues, stained with antibodies and determined by flow cytometry. Data are presented as mean ± SD, n = 3, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, compared with saline; ^#P < 0.05, ^##P < 0.01, compared with mAb or HA-ORD; ^$P < 0.05, ^$$P < 0.01, compared with physical mixture; ^&P < 0.05, compared with low dose; ns, not significant. Various preparations (0.2 mL) were intravenously injected into B16F10 tumor-bearing mice every 3 days at an ORD dose of 5 mg/kg or a mAb dose of 1 mg/kg, followed by HT stimulation at 42 °C for 1 h to trigger drug release. For HNPs, low and high doses of 5/1 and 5/2 mg/kg for ORD/mAb were administrated.

Figure 4

Anti-tumor efficacy of HNPs co-loaded with anti-PD-1 and HA-ORD. (A) Tumor growth curve, (B) weight of isolated tumor collected at the end of treatment and (C) body-weight change during the treatment period. Different formulations (0.2 mL, free anti-PD-1, HA-ORD, physical mixture and HNPs co-loaded anti-PD-1 and HA-ORD) were intravenously administrated every 3 days for five doses at 2 mg/kg for anti-PD-1 and 5 mg/kg for ORD, according to the body weight. The physical mixture was the simple combination of HA-ORD with anti-PD-1 with a similar ratio of ORD/anti-PD-1 in HNPs. Data are presented as mean ± SD, n = 6, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, compared with saline; ^#P < 0.05, compared with anti-PD-1 or HA-ORD; ^&P < 0.05, compared with physical mixture; ns, not significant.

Figure 5

THBS1 role in anti-tumor immune response and HA-ORD-mediated immune depletion through THBS1 downregulation. (A) Schematic illustration of the active target (THBS1) identification. (B) Silver staining after separation by SDS-PAGE. The supernatant medium that was incubated with HA-ORD for 12 h was collected and isolated by SDS-PAGE. The ORD concentration was from 0 to 50 μg/mL. (C) THBS1 knockdown efficacy in vitro was verified with target RNA sequence by qPCR. The B16F10 was incubated with AD-THBS1-shRNA for 48 h at a multiplicity of infection (MOI) of 500 at 37 °C. Null control, negative control, ∗∗∗P < 0.001, compared with null control. A control vector constructed the null control (NC) group without gene intervention activity. (D–J) HA-ORD-mediated immune response in the normal or THBS1-knockdown tumor-bearing mice. (D) Tumor volume growth curve (n = 6). THBS1 was knockdown continuously by two-time intratumor injection of AD-THBS1-shRNA#2, first on Day 3 before treatment with HA-ORD and second on Day 10 during the treatment period. The virus particle for AD was 5 × 10⁹/mice. HA-ORD was intravenously injected every three days five times via the tail vein. Immune cells in the tumor (E to G) and spleen (H to J) were assayed at the end of the experiment. The ORD dose was 5 mg/kg. Data are presented as mean ± SD, n = 3, ^#P < 0.05, ^##P < 0.01, ^###P < 0.001, compared with saline or null control; ^&P < 0.05, ^&&P < 0.01, compared with THBS1-knockdown group of AD-THBS1-shRNA#2; ns, not significant.

Figure 6

The levels of THBS1 in tumor tissues isolated from normal tumor model. (A) Immunofluorescent staining of the isolated tumor slice and (B) the semi-quantitative data. The THBS1 was stained in red by a Cy3-labeled secondary antibody. The dosage of mAb and ORD was 1 and 5 mg/kg, respectively. For HNPs, low and high doses of 5/1 and 5/2 mg/kg for ORD/mAb were administrated, followed by local HT stimulation at 42 °C for 1 h to trigger drug release. The digital graphs were snapped by CLSM and semi-quantified using ImageJ. Data are presented as mean ± SD, n = 3, ∗∗P < 0.01, ∗∗∗P < 0.001, compared with saline or mAb; ^#P < 0.05, compared with HA-ORD; ns, not significant.

Figure 7

Anti-tumor immunotherapy in the normal or THBS1-knockdown tumor model. (A) Growth curve of tumor in the knockdown mice at a dose of 5/2 mg/kg for ORD/mAb. Immune cells in the tumor (B to D) and spleen (E to G) were collected from the THBS1-knockdown animals with HNP treatment. (H) Illustration of the two therapeutics in HNPs working collaboratively. THBS1 was knockdown continuously by two-times intratumor injection with AD-THBS1-shRNA#2. The virus particle for AD was 5 × 10⁹/mice. A control vector constructed the null control (NC) group without gene intervention activity. Data are presented as mean ± SD, n = 6, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, compared with saline; ^#P < 0.05, compared with HNPs or null control plus HNPs; ns, not significant.