【Nat. Nanotechnol.】中科大王育才|针对90%癌症死因，科学家发现小于200微米早期微转移瘤的纳米递药新机制#

a, Time-lapse IVM images showing NP extravasation from normal hepatic vessels and vessels adjacent to CT26 liver metastases. Scale bars, 100 μm; 50 μm (insets). b, Normalized NP fluorescence intensity in normal vessels and metastatic adjacent vessels (n = 4 biologically independent samples). P = 2.3 × 10−6 at 1 h, P = 3.3 × 10−8 at 2 h; P = 4.3 × 10−11 at 4 h. c, 3D reconstructions of normal vessels and metastatic adjacent vessels showing NP extravasation in metastatic adjacent vessels. Scale bars, 10 μm. d, TEM images of normal vessels and metastatic adjacent vessels. Arrows indicate endothelial junctions or intercellular gaps. Scale bars, 2 μm; 500 nm (insets). e, Gap size distribution in normal vessels and metastatic adjacent vessels (n = 30 gaps from six biologically independent samples). f, Focused-ion-beam SEM 3D reconstruction of a metastatic adjacent vessel from 280 serial electron microscopy sections, with a representative two-dimensional slice and magnified view highlighting the endothelial gaps. Scale bars, 1 μm. g, Gap sizes measured at different depths across reconstructed slices of metastatic adjacent vessels (n = 5). h, TEM images showing the AuNP distribution at 0.5 h post-injection in normal vessels and metastatic adjacent vessels. Scale bars, 2 μm; 200 nm (insets). i, TEM images showing AuNP distribution in tumour cells at 12 h post-injection. Scale bars, 2 μm; 500 nm (insets). j, Western blot quantification of ZO-1, VE-cadherin, JAM-A and Cx43 in normal and metastatic adjacent liver tissue (normalized to β-actin; n = 9 biologically independent samples for ZO-1; n = 5 biologically independent samples for VE-cadherin, JAM-A and Cx43). k, Immunofluorescence images of ZO-1 expression in normal vessels and metastatic adjacent vessels. Scale bars, 10 μm; 5 μm (insets). l, Immunofluorescence images of NP extravasation in metastatic adjacent vessels with different ZO-1 coverage. Scale bars, 5 μm. Correlation between NP extravasation and ZO-1 coverage was assessed using linear regression. The solid line indicates the regression fit, and the dotted lines represent 95% confidence intervals (n = 36 vessels from three biologically independent samples). P = 5.9 × 10−10. m, Schematic showing reduced ZO-1 integrity promotes endothelial gap formation, increases vessel permeability and enhances NP accumulation in micrometastases. Images are representative of at least three independent experiments. Data in b, j are presented as the mean ± s.d. Significant differences were assessed using a two-way ANOVA with Sidak’s multiple comparisons test (b), a two-tailed unpaired Student’s t-test (j) and a simple linear regression and Pearson’s r (l).#

为了看清药物渗漏的微观结构，团队利用透射电镜和聚焦离子束扫描电镜进行了高分辨率重建。结果显示，在靠近微转移瘤的正常毛细血管上，内皮细胞之间裂开了100到500纳米的细胞间隙，而正常肝脏血管的结构非常紧密。进一步检测表明，这种独特的间隙具有明显的尺寸选择性，能让100到310纳米大小的纳米颗粒高效通过。分子层面则证实，紧密连接蛋白ZO-1表达量的显著下调是导致这些间隙形成的核心原因。

这些正常血管为什么会突然裂开？研究团队将目光投向了物理力学环境。他们发现，随着微转移瘤内部细胞外基质的重构，肿瘤组织的硬度明显增加，从而对邻近的正常血管产生了持续的物理挤压应力。为了验证这一点，研究人员利用卡托普利抑制胶原蛋白沉积来减轻硬度。实验表明，当挤压应力被消除或减轻时，内皮细胞的ZO-1蛋白水平随之恢复，血管间隙消失，纳米颗粒的聚集量也大幅减少，这有力证明了机械挤压是启动该效应的源头。

a, Top: immunofluorescence images of α-SMA in normal livers and micrometastases. Bottom: 3D reconstruction of collagen deposition using second-harmonic generation imaging. Scale bars, 100 μm (α-SMA); 50 μm (collagen). b, Top left: fluorescence image of fresh normal liver and CT26 liver micrometastases overlaid with the bright-field image of the atomic force microscopy tip, probed for spatially resolved stiffness measurements. Scale bars, 50 μm. Bottom left: force maps corresponding to the yellow-boxed region. Top right: schematic illustrating the atomic-force-microscopy-based force mapping and stiffness measurement. Bottom right: stiffness values for layers of normal livers and CT26 micrometastases (n = 5 biologically independent samples). c, IVM images displaying NP distribution in ECM-mimetic Matrigel of varying stiffness at 12 h post-injection. Pseudocolour indicates normalized NP intensity. Scale bars, 100 μm. d, High-resolution ultrasound images showing liver deformation after vibratome slicing. The dashed line indicates the agarose reference line used for deformation quantification. e, 3D reconstructions of liver metastases and adjacent vessels for structural analysis. Scale bars, 15 μm. f, Immunofluorescence images of liver vessels after tissue transparency. The aspect ratio of LSECs in normal vessels and metastatic adjacent vessels (n = 30 LSECs from five biologically independent samples). Scale bars, 20 μm; 10 μm (insets). P = 6.7 × 10−12. g, Schematic showing an adjustable compression device that is compatible with vAIW. Compression is modulated by screw rotation, enabling the precise control of applied pressure. h, IVM images showing NP distribution in uncompressed and compressed livers of Tie2-GFP fluorescent-transgenic reporter mice. Scale bars, 10 μm. Normalized intensity of NPs leaked from vessels (n = 6 biologically independent samples). i, Top: schematic showing NP permeability through a monolayer of mLSECs (CP-M040) under external compressive stress for 12 h. Bottom: SEM images showing the junctions of mLSECs under compressive stress (15 and 50 nN). Scale bars, 2 μm. j, TEM images showing metastatic adjacent vessels of mice with and without captopril treatments (p.o., 10 mg kg−1). Scale bars, 1 μm; 500 nm (insets). Distribution of endothelial gap sizes (n = 30 gaps from three biologically independent samples). k, NP distribution in liver metastases from captopril-treated and untreated mice. Scale bars, 50 μm. Quantification of the normalized NP intensity in entire metastases and metastatic edges (n = 14 avascular micrometastases from three biologically independent samples). P = 2.0 × 10−7 for metastases; P = 1.9 × 10−11 for metastatic edge. Images are representative of at least three independent experiments. Data in b, f, h and k are presented as the mean ± s.d. Significant differences were assessed using a two-tailed unpaired Student’s t-test (b, f, h and k).#

团队通过转录组测序深入挖掘了背后的分子机理。数据表明，正常血管内皮细胞在受到肿瘤的挤压后，其表面的β1整合素被激活。这一机械传感器随后触发了下游的RhoA/ROCK信号通路，引发细胞骨架的重新排列与收缩。使用ROCK抑制剂法舒地尔或β1整合素抑制剂，都可以显著阻断这一过程，使血管恢复紧密状态。这就清晰地勾勒出了一条由“物理机械挤压到生物信号激活，再到血管屏障打开”的完整机制链条。

a, Bubble plot illustrating gene ontology (GO) enrichment pathways of significantly upregulated genes in metastatic adjacent livers versus normal livers. pos., positive. b, Heat map of genes encoding integrin subunits in normal and metastatic adjacent livers. Transcript levels in reads per kilobase of exon model per million mapped reads (RPKM). c, Immunofluorescence images of active β1 integrins and CD31 in normal and metastatic adjacent livers from mice treated with or without captopril. Scale bars, 10 μm; 5 µm (insets). d, Quantification of active β1 integrin coverage of normal and metastatic adjacent vessels (n = 20 vessels from four biologically independent samples). P = 3.6 × 10−7 between normal liver and metastasis; P = 1.9 × 10−5 between metastasis and metastasis + captopril. e, IVM images showing NP distribution in micrometastases of mice treated with or without a β1 integrin inhibitor (HY-100445A, 10 mg kg−1). Scale bars, 50 μm. f, Quantification of normalized NP intensity in entire micrometastases and at the metastatic edges (n = 12 avascular micrometastases from three biologically independent samples). P = 4.0 × 10−6 for metastases; P = 5.1 × 10−9 for metastatic edge. g, Immunofluorescence images of F-actin in mLSECs under compressive stress (50 nN) with or without ROCK inhibitor treatment. Scale bars, 10 μm. h, Immunofluorescence images of ZO-1 and CD31 in metastatic adjacent liver of mice treated with or without ROCK inhibitor fasudil. Scale bars, 10 μm; 5 µm (insets). Quantification of ZO-1 coverage on vessels (n = 20 vessels from four biologically independent samples). P = 9.3 × 10−13. i, NP distribution in micrometastases of mice treated with or without fasudil. Scale bars, 50 μm. j, Quantification of normalized NP intensity within entire metastases and metastatic edge (n = 12 avascular micrometastases from three biologically independent samples). P = 4.6 × 10−10 for metastases; P = 5.1 × 10−13 for metastatic edge. k, Schematic showing the proposed mechanism by which mechanical compression activates β1 integrin in LSECs, leading to RhoA/ROCK-mediated cytoskeletal contractility, junctional disruption and enhanced NP extravasation. Images are representative of at least three independent experiments. Data in d, f, h and j are presented as the mean ± s.d. Significant differences were assessed using a one-way ANOVA with Tukey’s multiple comparisons test (d) and two-tailed unpaired Student’s t-test (f, h and j).#

这一新机制直接转化为了强大的治疗优势。研究团队利用该效应，将抗癌药物喜树碱和伊立替康制成纳米制剂，或者利用脂质纳米颗粒包裹抑癌基因mRNA。实验结果显示，这些纳米药物凭借EPAV效应实现了在微转移灶的高效靶向富集。与传统化疗药相比，纳米药物极大地提高了微转移瘤的清除效果，不仅显著减少了肝脏微转移灶的数量，还大幅缩小了病灶面积，同时降低了对正常肝脏的毒副作用。

a, Immunofluorescence images of CD34 and nuclei in normal livers and metastatic adjacent livers from patients with colorectal cancer. The white arrowheads indicate the liver sinusoidal endothelial cells (LSECs). Scale bars, 10 μm. b, Quantification of the aspect ratio and maximum projection area of LSECs in normal vessels and metastatic adjacent vessels from patients with colorectal cancer (n = 12 LSECs from three biologically independent samples). P = 2.9 × 10−6. c, Bubble plot illustrating gene ontology enrichment analysis of significantly upregulated genes in metastatic adjacent livers versus normal livers. d, Heat map of genes encoding integrin subunits in normal livers and metastatic adjacent livers from patients with colorectal cancer. e, TEM images of normal vessels and metastatic adjacent vessels from patients with colorectal cancer. Scale bars, 1 μm; 500 nm (insets). f, Quantification of the number of gaps in normal vessels and metastatic adjacent vessels from patients with colorectal cancer based on TEM results (n = 5 biologically independent samples). P = 1.8 × 10−5. g, Distribution of gap sizes in normal vessels and metastatic adjacent vessels from patients with colorectal cancer (n = 5 biologically independent samples). Images are representative of at least three independent experiments. Data in b and f are shown as the mean ± s.d. Significant differences were assessed using a two-tailed unpaired Student’s t-test (b and f).#

最重要的是，该团队进一步分析了结直肠癌肝转移患者的手术切除标本。在人类临床样本中，他们同样观察到了微转移瘤边缘血管的挤压变形、内皮细胞间隙的形成以及ZO-1蛋白的明显缺失。同时，患者样本的基因表达也呈现出整合素通路激活的特征。这一发现极大地增强了该研究的临床转化价值，证明该效应在人类患者体内同样真实存在。

a, Schematic showing Cy5-conjugated mPEG-b-PLGA NPs (NPs-Cy5) used to encapsulate a fluorescent camptothecin derivative (CPT-BFL). b, IVM images showing the distribution of free CPT and NPs, and CPT@NPs within micrometastases. Solid lines outline metastatic lesions. Scale bars, 50 μm. c, Quantification of normalized intensity of CPT-BFL and CPT@NPs in normal livers and metastases (n = 8 metastases from three biologically independent samples). P = 5.2 × 10−7. d, Therapeutic schedule of CT26 liver metastasis-bearing mice. On days 0, 3, 6, 9 and 12, mice received solvent CPT or CPT@NPs (i.v., 3 mg kg−1) treatments. e, Magnetic resonance images and bright-field images of mouse livers bearing CT26 metastases subjected to different treatments. The white arrowheads indicate metastases. Scale bars, 1 cm. f, Number of liver metastatic foci following different treatments (n = 5 biologically independent samples). P = 9.0 × 10−7 between Ctrl and CPT@NPs. g, Ratio of metastatic area in the liver following different treatments (n = 5 biologically independent samples). P = 4.4 × 10−5. h, Therapeutic schedule of CT26 liver metastasis-bearing mice. On days 0, 3, 6, 9 and 12, mice received captopril (p.o., 10 mg kg−1), solvent IRT, Lipo IRT and captopril plus Lipo IRT (i.v., 5 mg kg−1) treatments. i, Gross appearance and H&E images of liver sections from mice subjected to different treatments. The black arrowheads indicate metastases, and the dashed outlines delineate the boundaries of metastatic tumours. Scale bars, 1 cm (appearance); 2 mm (H&E). j, Ratio of liver metastatic area following different treatments (Ctrl, captopril, n = 4 biologically independent samples; solvent IRT, Lipo IRT and captopril plus Lipo IRT, n = 5 biologically independent samples). k, Schematic showing the structure of mRNA-loaded lipid NPs (mRNA@LNPs). l, 3D images showing the spatial distribution of LNPs within LLC liver micrometastases at 12 h after injection. Scale bars, 50 μm. m, Expression of mCherry mRNA in LLC liver micrometastases at 18 h after injection of mCherry mRNA@LNPs. Scale bars, 10 μm. n, Therapeutic schedule of LLC liver metastasis-bearing mice. On days 0, 2, 4 and 6, mice received p53 mRNA@LNPs and PTEN mRNA@LNPs (at an equivalent mRNA dose of 0.5 mg kg−1) treatments. o, Gross appearance, fluorescence reflectance imaging and H&E images of liver metastases of mice after different treatments. Scale bars, 2 mm. p, Number of metastatic foci in livers after different treatments (n = 6 biologically independent samples). q, Ratio of metastatic area in livers after different treatments (n = 4 biologically independent samples). Images are representative of at least three independent experiments. Data in f, g, j, p and q are presented as the mean ± s.d. Significant differences were assessed using a two-tailed paired Student’s t-test (c) and one-way ANOVA with Tukey’s multiple comparisons test (f, g, j, p and q).#