【JACS】同济大学张兵波|检出限低至 6×10⁴ CFU！靶向激活型 MRI 探针实现肿瘤内细菌精准成像与抗转移治疗
文章标题：A Target-And-Activate Magnetic Resonance Imaging Strategy for Precision Imaging of Intratumoral Bacteria and Screening Antibiotics to Suppress Breast Cancer Metastasis
通讯作者：Bingbo Zhang
文章概要
引言
越来越多的研究证实，肿瘤内细菌尤其是金黄色葡萄球菌这类革兰氏阳性菌，会显著促进乳腺癌细胞的存活与扩散，是推动肿瘤转移的重要因素，因此靶向肿瘤内细菌也成为抑制癌症转移的全新研究方向。不过目前检测肿瘤内细菌主要依靠有创的组织活检，结合微生物培养或基因测序开展分析，这类方式不仅存在采样偏差，患者接受度较低，还无法动态追踪治疗过程中细菌载量的变化。无创的磁共振成像（MRI） 本是动态监测病灶的理想技术，但现有的细菌响应型 MRI 造影剂存在难以调和的短板，纳米探针会因肿瘤组织的高通透性和滞留效应发生非特异性富集，带来严重的背景信号干扰，小分子探针虽代谢快、背景噪音低，却又普遍存在检测灵敏度不足的问题。与此同时，当下多数可激活型成像探针仅能响应肿瘤微环境中的通用刺激物，无法精准识别肿瘤内的细菌。为解决这一系列难题，本研究提出了全新的靶向 - 激活型 MRI 成像策略，设计出多肽探针BS-FFGd，依靠双重特异性识别机制实现肿瘤内细菌的精准可视化，并进一步探索该探针在抗生素筛选、抑制乳腺癌转移方面的实际应用价值。
Scheme 1. A Target-And-Activate MRI (TAM) Strategy, BS-FFGd Probe Design and Application; (a) Schematic Diagram of Target-And-Activate MRI Strategy via a Sequential Process of Targeting, Activation, Retention and Amplification (TARA); (b) Chemical Structure of BS-FFGd Probe and; (c) Schematic Illustration of TAM Strategy for Precise Detection of Intratumoral Bacteria and in vivo Antibiotic Screening by MRIa
主要实验及结论
研究团队设计合成了分子量为 3303 Da 的多肽探针BS-FFGd，该探针集成了靶向序列、酶切位点、自组装核心与 MRI 信号基团，具备双重激活条件：一方面依靠特异性肽段结合革兰氏阳性菌表面的脂磷壁酸（LTA），另一方面可被肿瘤微环境中高表达的基质金属蛋白酶 - 2（MMP-2） 切割。完整状态下的 BS-FFGd 带有亲水基团，以单体形式存在并能被机体快速清除，当被 MMP-2 特异性切割后，亲水结构带来的空间位阻消失，探针会暴露自组装核心，进而在细菌表面原位组装成纳米纤维。研究借助高效液相色谱、质谱、圆二色谱、荧光检测、电镜等多种表征手段验证了这一过程，结果显示 MMP-2 对 BS-FFGd 的切割效率最高可达94.0%，且该切割作用具备高度特异性，肿瘤微环境里的氢离子、谷胱甘肽、其他蛋白酶等物质都无法触发切割反应。分子动力学模拟也从原子层面解释了背后机制，完整探针的亲水残基会阻碍分子间聚集，切割后的片段结合自由能显著降低，更容易形成稳定的聚集体结构。
Figure 1. MMP-2 specific cleavage–triggered self-assembly and structural characterization of BS-FFGd. (a) Schematic illustration of the self-assembly mechanism of BS-FFGd. (b) HPLC analysis of BS-FFGd after incubation with MMP-2 at different concentrations in TCNB buffer and the corresponding cleavage efficiency. (c) CD spectra of BS-FFGd before (gray line) and after (red line) incubation with MMP-2. (d) Dynamic ThT fluorescence monitoring of the assembly of BS-FFGd and NBS-FFGd after incubation with MMP-2. (e) TEM images of nanofibers formed from BS-FFGd after incubation with MMP-2. (f) Hydrodynamic size distribution of BS-FFGd before (gray line) and after (red line) incubation with MMP-2.
团队随后测试了探针的 MRI 成像性能与体外检测能力，未被激活的 BS-FFGd 纵向弛豫率为 9.03 mM⁻¹ s⁻¹，经 MMP-2 切割并自组装后，弛豫率大幅提升至16.51 mM⁻¹ s⁻¹，MRI 信号实现明显放大，而无法被酶切的对照探针 NBS-FFGd 始终没有出现信号变化，这是因为纳米纤维结构大幅延长了探针的旋转相关时间，符合顺磁弛豫理论。体外细胞与细菌实验充分印证了双重识别机制的必要性，仅分泌 MMP-2、不含脂磷壁酸的 4T1 乳腺癌细胞无法让探针稳定组装滞留，同时具备两种靶点的金黄色葡萄球菌则能诱导探针形成大量纳米纤维，针对革兰氏阴性大肠杆菌的测试也未观察到组装现象，证明探针对革兰氏阳性菌拥有优异特异性。体外检测实验还确定，BS-FFGd 对金黄色葡萄球菌的检出限可达 10⁵ CFU，成像信号强度会随着细菌浓度升高同步增强。
Figure 2. MD simulation of BS-FFGd and FFGd. Snapshots of BS-FFGd system (a) and FFGd system (b) at 0, 25, 50, and 100 ns. (c) Per-residue decomposition of binding free energy of BS-FFGd system (blue line) and FFGd system (red line). Time-dependent binding free energy curves of BS-FFGd system (blue line) and FFGd system (red line) (d) and the corresponding average binding free energy over the final 10 ns (e). Time-dependent Rg curves of BS-FFGd system (blue line) and FFGd system (red line) (f) and the corresponding average Rg over the final 10 ns (g). Data shown as means ± SD; *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant.
研究构建了荷瘤小鼠模型开展活体成像与定量检测实验，在双侧 4T1 肿瘤模型中，分别对比了临床常用造影剂 Gd-DTPA、对照探针 NBS-FFGd 与 BS-FFGd 的成像效果。数据表明，BS-FFGd 成像的对比度噪声比（ΔCNR）表现最优，相比 Gd-DTPA 提升约 4.7 倍，相比 NBS-FFGd 提升约 2 倍，仅在定植金黄色葡萄球菌的肿瘤区域出现高亮信号，无菌肿瘤、定植大肠杆菌的肿瘤都仅存在微弱背景信号。在不同细菌载量的单侧肿瘤模型中，团队进一步验证了探针的定量能力，证实其体内检出限低至 6×10⁴ CFU，肿瘤的 ΔCNR 数值与细菌菌落数、革兰氏阳性菌染色面积、脂磷壁酸阳性面积均呈现极强的线性相关性，相关系数 R 最高达到 0.96，可精准量化肿瘤内的细菌数量。活体分布与毒性检测结果显示，BS-FFGd 主要在细菌表面富集形成纤维网络，可通过肝肾、肝胆途径正常代谢，小鼠血常规、生化指标以及心、肝、脾、肺、肾等主要脏器的病理切片均未发现异常，证明该探针拥有良好的生物相容性。
Figure 3. MMP-2 triggered T1-weighted MRI enhancement of BS-FFGd. Longitudinal (a) and transverse (b) relaxivity of BS-FFGd and NBS-FFGd in the presence or absence of MMP-2 (2 μg/mL, 24 h). The cleavage efficiency of BS-FFGd was calculated to 86.6%. (c) T1-weighted MR images of BS-FFGd before and after incubation with MMP-2. (d) Schematic illustration of the MMP-2 cleavage–triggered MRI signal enhancement mechanism. Specific cleavage by MMP-2 leads to a transition from a rapidly rotating monomeric state (short τR, low MRI signal) to a slowly rotating aggregated state (long τR, high MRI signal). T1-weighted MR images (e) and corresponding longitudinal relaxation rates (f) of BS-FFGd after incubation with various analytes. MMP-2 concentration-dependent T1-weighted MR images (g) and corresponding longitudinal relaxation rates (h) of BS-FFGd. Data shown as means ± SD; *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant.
Figure 4. In vitro validation of specific and sensitive MRI of SA via a TAM Strategy. Schematic illustration of procedure for analysis of the targeting and in situ self-assembly of BS-FFGd on 4T1 cells (a) and SA (b). Nanofibers detected by ThT fluorescence in 4T1 cells (c) and SA (d) after incubation with NBS-FFGd or BS-FFGd. Scale bars, 20 μm (c); 5 μm (d). SEM images of 4T1 cells (e) and SA (f) treated with BS-FFGd (50 μM) at 37 °C for 2 h. Scale bars, 2 μm (e); 500 nm (f). SA concentration-dependent T1-weighted MR images (g) and T1-weighted MRI signal intensity (h) of NBS-FFGd (100 μM) after incubation with SA for 8 h. Time-dependent T1-weighted MR images (i) and T1-weighted MRI signal intensity (j) of NBS-FFGd after incubation with SA (108 CFU). SA concentration-dependent T1-weighted MR images (k) and T1-weighted MRI signal intensity (l) of BS-FFGd (100 μM) (data shown as means ± SD; *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant). (m) Time-dependent T1-weighted MR images of BS-FFGd. (n) Longitudinal relaxation rates of BS-FFGd and NBS-FFGd before and after incubation with SA. The panel a was created with BioRender.com.
依托 BS-FFGd 的活体成像与定量能力，团队将其搭建为抗生素筛选平台，选取对金黄色葡萄球菌有效的莫西沙星、效果不佳的头孢克肟开展对照治疗实验。连续给药 8 天后，MRI 成像结果直观区分出两种抗生素的药效，莫西沙星治疗组的肿瘤成像信号回落至正常水平，肿瘤内细菌载量下降约 5 个数量级，而头孢克肟治疗组的细菌数量仅减少约 1 个数量级，成像信号没有明显改善。研究持续观测小鼠肺部转移情况后发现，肿瘤内定植的金黄色葡萄球菌会显著加重乳腺癌肺转移，未接受有效抗生素治疗的小鼠肺部转移结节数量大幅增加，使用莫西沙星清除肿瘤内细菌后，肺转移结节数量下降约 3 倍，恢复至无菌肿瘤小鼠的水平，同时抗生素干预并不会影响原发肿瘤的生长，直接证实了肿瘤内金黄色葡萄球菌是促进乳腺癌转移的关键推手。
Figure 5. TAM Strategy via BS-FFGd for precision MRI of intratumoral SA in bilateral tumor models. (a) Schematic of the experimental timeline in the bilateral 4T1 tumor model. The right tumor was injected with SA, while the left tumor received saline or EC as a control. (b) Representative T1-weighted MR images at different time points after intravenous administration of Gd-DTPA, NBS-FFGd, or BS-FFGd, red dashed circles indicate noncolonized tumors, and yellow dashed circles indicate tumor colonized with S. aureus. Quantitative analysis of tumor to muscle ΔCNR for (c) Gd-DTPA, (d) NBS-FFGd, and (e) BS-FFGd over time, comparing SA-colonized (4T1 + SA) versus noncolonized (4T1) tumors (data shown as means ± SD; *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant). Representative bio-SEM images of SA-colonized tumors before (f) and after (g) BS-FFGd treatment. Scale bars, 500 nm. The panel a was created with BioRender.com.
Figure 6. BS-FFGd for precise detection of intratumoral bacterial loads and screening antibiotics. (a) Schematic of the experimental design for assessing the in vivo sensitivity of BS-FFGd. (b) Representative T1-weighted MR images of tumors colonized with different bacterial loads (Groups I–VIII) before and 90 min after BS-FFGd administration, yellow dashed circles indicate tumor colonized with_S. aureus_. (c) Quantitative analysis of tumor to muscle ΔCNR for the different bacterial load groups. (d) Analysis of the tumors (Groups I and IV–VIII), including bacterial colony plates, Gram stain, and IHC for MMP-2 (green), LTA (red) and DAPI (blue). Scale bar, 20 μm. Linear correlation analysis between tumor ΔCNR at 90 min and final bacterial colony counts (Log10 CFU) (e), Gram-positive bacteria positive stained area (%) (f), and LTA positive area (%) (g). (h) Schematic of the experimental design for screening antibiotics. Mice bearing SA-colonized tumors (107 CFU) were treated by oral gavage with cefixime or moxifloxacin (30 mg/kg). (i) Representative T1-weighted MR images of tumors from the four treatment groups (G1-G4) at Day 0 and Day 8. (j) Quantification of tumor to muscle ΔCNR at Day 0 and Day 8 for all groups. (k) Representative photos of bacterial colony plates from excised tumors. (l) Quantification of intratumoral bacterial loads (Log10 CFU). (m) Representative photos and H&E staining of lungs from all groups on Day 21 to assess metastasis. (n) Quantification of lung metastatic nodules. Data shown as means ± SD; *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant.
总结及展望
本研究提出的靶向 - 激活型 MRI 新策略，融合了分子靶向、酶响应切割与原位超分子自组装技术，成功突破了传统 MRI 探针在肿瘤内细菌检测中特异性与灵敏度无法兼顾的瓶颈。研发的 BS-FFGd 多肽探针依靠双重识别机制，实现了肿瘤内革兰氏阳性菌的高特异、高灵敏成像与精准定量，同时完成了活体层面的抗生素药效筛选，通过有效清除肿瘤内细菌，显著抑制了乳腺癌的转移进程，搭建起 “精准诊断 - 疗效评估 - 靶向干预” 一体化的诊疗体系，为肿瘤 - 微生物互作研究、抗菌抗肿瘤治疗提供了全新工具。该研究也存在一定局限性，这套成像平台的组织穿透能力与氟代脱氧山梨醇 PET 成像相当，但在检测灵敏度和定量精度上仍有提升空间，后续团队会进一步优化探针结构，提升弛豫性能。为推进临床转化并规避钆元素长期滞留的潜在风险，研究人员还计划采用铁基螯合物等生物相容性更佳的顺磁性材料替代现有钆基信号基团。另外，BS-FFGd 采用模块化多肽设计，具备很强的拓展性，未来可以通过改造靶向肽段、自组装核心，将该策略应用于更多细菌种类与不同疾病模型，也可整合更多肿瘤微环境响应元件，让这套靶向激活成像技术拥有更广阔的应用场景。