撰文丨王聪 编辑丨王多鱼 排版丨水成文 代谢紊乱 与 认知能力下降 风险增加密切相关，西式 高脂肪饮食 （HFD） 已成为关键诱因。然而，其潜在的细胞和分子机制，目前仍不清楚。 2025 年 9 月 11 日，北卡罗来纳大学医学院 宋娟 团队在 Cell 子刊 Neuron 上发表了题为： Targeting glucose-inhibited hippocampal CCK interneurons prevents cognitive impairment in diet-induced obesity 的研究论文。 该研究证实， 短期高脂饮食 （stHFD） ，就会通过诱导大脑海马体齿状回 胆囊收缩素表达中间神经元 （CCK-IN） 的过度活跃， 从而破坏记忆处理。而靶向抑 制该神经元， 可预防饮食诱导肥胖所致认知障碍。 在这项新研究中，研究团队证明，仅仅 5 天的 短期高脂饮食 （stHFD） ，就会通过诱导大脑海马体齿状回 胆囊收缩素表达中间神经元 （CCK-IN） 的过度活 跃，从而破坏记忆处理。 研究团队发现， 齿状回 CCK-IN 神经元是葡萄糖抑制型神经元，在 短期高脂饮食 （stHFD） ...

Core Viewpoint - Cancer remains a leading cause of death globally, prompting the search for new targeted therapies, particularly antibody-drug conjugates (ADCs) which show promise in delivering chemotherapy directly to cancer cells while minimizing side effects [2][3]. Group 1: Current ADC Limitations - Current ADCs have a drug-to-antibody ratio (DAR) of only 2-8, limiting the range of chemotherapy drugs that can be used, as only highly potent drugs can be selected [2][6]. - The limited DAR means that ADCs cannot utilize a broader spectrum of less potent chemotherapy drugs, which constrains treatment options [6]. Group 2: Introduction of ABC Technology - The newly developed antibody-bottlebrush prodrug conjugates (ABC) offer modular synthesis and a significantly higher DAR, allowing for a wider range of effective payloads, including less potent chemotherapy drugs [3][9]. - ABC technology enables the delivery of hundreds of prodrug molecules via a single antibody, enhancing the customization and diversity of drug combinations [8][9]. Group 3: Experimental Results - In preclinical models, ABCs demonstrated superior efficacy in eliminating tumors compared to traditional ADCs and non-targeted prodrugs, even at very low doses [13][14]. - The study showed that ABCs outperformed FDA-approved ADCs like T-DXd and TDM-1, indicating a potential for enhanced treatment outcomes [14]. Group 4: Future Directions - The research team plans to explore combinations of different chemotherapy drugs with varying mechanisms to improve overall efficacy [14]. - There is potential for using various monoclonal antibodies, as over 100 have been approved, to create new targeted cancer therapies through the ABC platform [14].

Core Viewpoint - Early cancer detection is crucial for reducing mortality rates among cancer patients, and the study highlights the potential of phosphatidylserine-positive extracellular vesicles (PS+ EV) as a specific biomarker for multiple operable cancers [2][10]. Group 1: Importance of Early Cancer Detection - Most cancers are diagnosed at advanced stages, limiting treatment options and chances of cure [4]. - Current cancer diagnostics rely heavily on imaging techniques and histopathological analysis, which have inherent risks and insufficient sensitivity for early-stage cancers [4][5]. - There is a critical need for sensitive and accessible screening methods to detect various cancer types at earlier, more treatable stages [4]. Group 2: Liquid Biopsy and Extracellular Vesicles - Liquid biopsy is a minimally invasive and sensitive detection method that can identify cancer at earlier stages [5]. - Blood is the primary biological fluid for analyzing circulating tumor-derived components, including circulating tumor cells (CTC), extracellular vesicles (EV), and various acellular molecules [5][6]. - Tumor-derived EVs are released by metabolically active and proliferating cancer cells, providing a promising avenue for early diagnosis and clinical decision-making [6][7]. Group 3: Research Findings - The study identified phosphatidylserine (PS) as a tumor-specific EV biomarker, leading to the development of a blood biopsy method called "PSEV-MultiCancer" [7][11]. - In a clinical sample of 1869, including 1269 cancer patients, PSEV-MultiCancer achieved an area under the curve (AUC) of 0.932, with a positive detection rate of 84.7% [7]. - For early-stage (I-II) cancers, the sensitivity reached 74.7% and specificity was 89.8% [7]. Group 4: Validation and Implications - The method underwent blind validation in three independent clinical cohorts, yielding AUC values of 0.97, 0.99, and 0.89, with an average sensitivity of 84.1% and average specificity of 97.3% [8]. - These findings support PSEV-MultiCancer as a promising non-invasive early cancer detection tool, aiding timely therapeutic interventions for patients [10][11].

例如， 2025 年 2 月， Nature Medicine 期刊发表的一项研究证实了 微塑料/纳米塑料 （MNP） 在人类 肝脏 、 肾脏 和 大脑 中的存在，该研究还发现，大脑 中的 MNP 浓度高于肝脏和肾脏，且随着时间的推移，MNP 浓度随之增加。此外，患有痴呆症的人群的大脑中 MNP 浓度显著高于没有痴呆症的人。2025 年 4 月，南开大学 汪磊 / 孙红文 团队在国际顶尖学术期刊 Nature 上发表研究论文，证实了 空气 是微塑料/纳米塑料 （MNP） 进入植物的主要途径——空气中的 MNP 会进入植物叶片，从而最终进入我们的食物 【2】 。 关于 微塑料/纳米塑料 （MNP） 对人类健康潜在影响 的研究获得广泛关注，但同时也引发了有关这些研究的科学严谨性和可靠性的激烈争论。 2025 年 9 月 11 日，阿姆斯特丹自由大学的研究人员在国际顶尖医学期刊 Nature Medicine 上发表了题为 ： Health impacts of microplastic and nanoplastic exposure 的综述论文。 该综述全面总结了 微塑料/纳米塑料 （MNP） 对人类健康影响的 ...

撰文丨王聪 编辑丨王多鱼 排版丨水成文 营养不良 是 5 岁以下儿童死亡的主要原因。在全世界范围内，有近 1.5 亿的 5 岁以下儿童因缺乏营养而 发育迟缓 （身高低于同龄标准） 。生命最初 1000 天内的营养不良尤其有害，会导致不可逆的长期认知和 发育损害，以及将来的学业成绩不佳、经济劣势等。东南亚和撒哈拉以南的非洲的营养不良水平最高，其 中，马拉维共和国的发育迟缓问题最为突出，发生率高达 37%。 饮食不足是儿童发育迟缓的主要原因，但多年前的研究就已发现， 肠道微生物 的功能失调在其中也发挥着 重要作用 。然而， 传统的微生物组研究方法缺乏分辨率，难以确定哪些微生物发挥作用。 2025 年 9 月 9 日，索尔克研究所、加州大学圣地亚哥分校、贝勒医学院、圣路易斯华盛顿大学等机构的 研究人员合作，在国际顶尖学术期刊 Cell 上发表 了题为： Culture-independent meta-pangenomics enabled by long-read metagenomics reveals associations with pediatric undernutrition 的研究论文。 该研究利 ...

Core Viewpoint - The article highlights the significant advancement in personalized medicine through the successful application of a bespoke CRISPR gene editing therapy for a patient with a severe genetic disorder, marking a milestone in individualized treatment approaches [2]. Group 1: Personalized CRISPR Therapy - The first patient to receive a personalized CRISPR gene editing therapy was diagnosed with carbamoyl phosphate synthetase 1 (CPS1) deficiency shortly after birth, and researchers developed a tailored lipid nanoparticle (LNP) delivery system for the therapy within six months, leading to substantial clinical improvement [2]. - The success of this case serves as a paradigm for personalized therapies, showcasing the potential of CRISPR technology in treating various devastating diseases, including vascular disorders [2]. Group 2: Research on Vascular Disease - A study published by researchers from Harvard Medical School and Massachusetts General Hospital demonstrated the successful treatment of a severe vascular disease, multi-system smooth muscle dysfunction syndrome (MSMDS), using a customized CRISPR-Cas9 base editor in mouse models [3][4]. - MSMDS, characterized by mutations in the ACTA2 gene, currently lacks effective treatment options, with the most common mutation being a G to A single nucleotide change at the sixth exon of the ACTA2 gene [7]. Group 3: Development of Targeted Therapy - The research team identified that conventional adenine base editors (ABE) could cause "bystander editing," leading to ineffective treatment outcomes. Therefore, they developed a personalized therapy targeting the most common pathogenic mutation, ACTA2 R179H, by screening for a SpCas9 enzyme with enhanced targeting specificity [9]. - An engineered SpCas9-VRQR was successfully constructed, allowing for precise A to G editing while minimizing unintended edits, thus improving the efficacy of the base editor [9]. Group 4: Efficacy in Mouse Models - The team created a mouse model that exhibited phenotypes consistent with human patients, including vascular lesions and early mortality, to explore the in vivo therapeutic potential of their strategy [10]. - The use of an engineered smooth muscle-tropic adeno-associated virus (AAV-PR) vector to deliver the customized base editor resulted in significantly extended lifespans for MSMDS mice and rescued systemic phenotypes throughout their lifetimes [10]. Group 5: Regulatory Progress and Future Trials - The developed therapy has received orphan drug designation from the FDA for rare diseases, and the research team plans to conduct further toxicology studies, with the aim of initiating human trials by 2027 [13].

Core Viewpoint - The article discusses a novel "drop-printing" strategy for the transfer of ultra-thin films to complex biological surfaces, addressing the challenge of stress-induced damage in flexible electronics and brain-machine interfaces [4][6][7]. Group 1: Research Background - The rapid development of wearable electronics, brain-machine interfaces, and neural rehabilitation technologies has created a need for precise electronic devices that can conform to biological tissues [3]. - Traditional attachment methods often lead to significant internal stress in devices, particularly when applied to uneven surfaces like skin or neural tissues, risking damage to fragile components [3]. Group 2: Innovative Methodology - The "drop-printing" technique allows for the attachment of fragile, non-stretchable films to surfaces such as skin, polymers, and neural tissues without damage [4][6]. - This method utilizes droplets to create a lubricating layer between the film and the target surface, facilitating local sliding during the attachment process, which prevents excessive stretching and reduces stress concentration [6][7]. Group 3: Experimental Validation - In vivo experiments demonstrated the successful attachment of a 2-micron thick silicon-based electronic film to the surface of mouse neural and brain tissues using the drop-printing technique [4][6]. - The resulting neural electronic interface achieved high spatiotemporal resolution for infrared light modulation of internal nerves [6][7]. Group 4: Implications and Future Applications - The research presents a groundbreaking approach to flexible electronics, providing critical technological support for the development of brain-machine interfaces and other interdisciplinary fields [7].

Core Viewpoint - Neuroblastoma is the most common extracranial solid tumor in children, characterized by significant developmental plasticity and intratumoral heterogeneity, leading to a 5-year survival rate of less than 50% in high-risk cases [2][8]. Group 1: Research Background - The study published by Professor Wang Jian's team from Suzhou University and Professor Hu Yizhou's team from Karolinska Institute in the journal Developmental Cell explores the developmental plasticity of neuroblastoma through single-cell multi-omics and spatial transcriptomics [2]. - Neuroblastoma originates from neural crest progenitor cells, which can differentiate into various cell types, including sympathetic neurons and chromaffin cells, and is influenced by genetic abnormalities and the tumor microenvironment [8][9]. Group 2: Research Findings - The research identified an intermediate "bridge" cell state crucial for malignant transformation in high-risk neuroblastoma, characterized by extensive epigenetic pre-activation features and potential for multidirectional state transitions [9][10]. - A transcription factor-enhancer gene regulatory network (TF-eGRN) was defined, which characterizes tumor states and is influenced by the specific microenvironment of the bridge/progenitor cell state, promoting mesenchymal characteristics [10].

Core Viewpoint - Immunotherapy, particularly immune checkpoint blockade (ICB), has transformed cancer treatment but remains ineffective in "cold tumors" due to immune suppression in the tumor microenvironment (TME) [2][5][6] Group 1: Research Findings - A new biomimetic Ru/TiO₂ nanobiocatalyst system inspired by natural enzyme reaction systems (ERS) has been developed, capable of rapid, pH-dependent generation of reactive oxygen species (ROS) and oxygen (O₂), effectively converting cold tumors into hot tumors [3][6][7] - The Ru/TiO₂ system enhances anti-tumor immunity and suppresses tumor metastasis when used in conjunction with ICB therapy [3][7] - This research establishes a precedent for adaptive nanobiocatalysts in the TME and paves the way for the development of next-generation immunotherapies targeting drug-resistant cancers [3][6] Group 2: Mechanism of Action - The study demonstrates that Ru/TiO₂ can mediate immunogenic cell death (ICD) in melanoma cells through endoplasmic reticulum stress, while also inhibiting hypoxia-induced immune suppression [7] - The design of Ru/TiO₂ aims to reverse immune suppression and enhance immunogenicity, transforming "immune cold" tumors into "immune hot" tumors [7] Group 3: Clinical Implications - The findings suggest that the rational design of robust and efficient biocatalytic materials could extend beyond cancer treatment, opening new avenues for immune modulation in other diseases [3][6]

Core Insights - The article discusses the transformative impact of artificial intelligence (AI) on protein design, revolutionizing methods for drug discovery, biotechnology, and synthetic biology applications [2][3]. - A comprehensive and actionable roadmap for integrating advanced AI tools into protein design workflows is provided, highlighting AI's potential to innovate synthetic biology, accelerate drug development, and drive sustainable biotechnology [3][8]. Summary by Sections Overview of Protein Design - Protein design has long been a cornerstone of scientific innovation, driving breakthroughs in drug development, biotechnology, and synthetic biology. However, traditional methods are nearing their limits in addressing the vast complexity and diversity of protein sequences [5][6]. AI's Role in Protein Design - AI is emerging as a transformative force to tackle challenges previously deemed unsolvable, enhancing both directed evolution and rational design strategies. Directed evolution simulates natural selection through random mutations, while rational design relies on structural and functional data [6][7]. - The search space for protein design is immense, with a typical protein of 350 amino acids having approximately 10^455 possible sequences, making exhaustive exploration impractical [6][7]. Development of AI Tools - AI-driven advancements have led to the development of new tools that provide unprecedented speed, scale, and precision in both directed evolution and rational design. AI tools can accurately propose beneficial mutations and predict functions from sequences, significantly shortening experimental cycles [7][8]. - The integration of deep learning methods into protein design workflows is not only feasible but essential, transforming the process from trial-and-error to a predictive and efficient discipline [7][9]. AI-Driven Protein Design Roadmap - The article outlines a roadmap for integrating AI tools into protein design, categorizing them into seven toolkits that support various tasks throughout the workflow, from initial design to experimental validation [9][22]. - Each stage of the protein design process is matched with the most suitable AI toolkit, guiding designers in assembling end-to-end AI-driven workflows [9][24]. Case Studies - AI-driven directed evolution of adeno-associated virus (AAV) capsids involved introducing random mutations to generate a virtual library of 10^10 AAV2 sequences, resulting in 20,426 sequences being experimentally validated [27]. - AI-driven antibody directed evolution utilized the ESM protein language model to generate heavy and light chain variants, achieving binding affinity improvements of up to 160 times [27]. - Rational design of a novel luciferase involved using AI tools to optimize the structure and function, resulting in variants with excellent thermal stability and specificity [28]. Future Directions - The next generation of AI tools must be built on robust and diverse data foundations to address challenges in protein design, including the need for explainable AI methods to enhance trust and adoption [29][30]. - AI-driven protein design is poised to open a new era of precision therapeutics, enabling the targeting of previously "undruggable" proteins and accelerating the design-manufacture-test-analyze cycle [31][32].