用干细胞治疗自闭症、癫痫与精神分裂症
摘要
斯坦福大学精神病学与行为科学教授、斯坦福脑类器官发育项目主任 Sergiu Pașca 博士，探讨了自闭症谱系障碍的生物学与遗传学机制、其患病率上升的原因，以及他的实验室正在开发的突破性干细胞技术。他解释了类器官与组装体——从干细胞培育而成的三维人脑回路——如何使研究人员能够在细胞层面研究重度自闭症和精神分裂症等精神疾病，以及这些模型如何被用于开发首批靶向临床疗法。
核心要点
- 自闭症并非单一疾病。 它以行为表现为诊断依据，缺乏生物标志物，很可能代表着数百种不同的遗传病症被归入同一个总称之下——就像19世纪的”发烧”概念一样。
- 患病率接近3%，远高于曾被视为罕见病时的水平；全球范围内均有上升趋势，而非仅限于美国，这使得以美国特有环境因素来解释其成因的论点难以成立。
- 强遗传性是自闭症研究中最为稳健的发现。数百个基因——影响突触蛋白、离子通道及染色质包装——与特定形式的自闭症相关联。
- 目前仅约20%的重度自闭症患者能获得明确的遗传学诊断；其余患者仍属于”特发性”类别，确切病因不明。
- 山中因子（诱导多能干细胞 / iPS细胞）通过允许将任何患者的皮肤细胞重编程为神经元，彻底变革了神经科学研究——从而消除了对人类胚胎组织的依赖。
- 类器官能重现真实的发育时序，包括标志性的分子转换，如NMDA受体2B亚基向2A亚基的转变（发生于相当于出生前后的时期）——且无需任何外部激素信号的诱导。
- 在海外诊所提供的非正规干细胞注射对自闭症缺乏科学依据，并存在严重风险，包括感染、肿瘤生长及永久性神经损伤。
- 利用CRISPR的基因疗法前景可期，但对于脑部疾病仍面临重大挑战：如何将基因递送至正确的细胞、病毒载体的运载容量有限、操作不可逆，以及何时干预才能达到最佳效果的时机问题。
- 人类大脑的髓鞘化持续至生命的第三个十年，凸显出神经发育与其他器官相比是多么漫长且易受影响。
详细笔记
什么是自闭症？
- 自闭症是一种以行为表现为诊断依据的谱系障碍——通过评估特定发育年龄段特定行为的有无来进行诊断，没有生物学生物标志物。
- 诊断率约为每30至40人中有1人（接近3%），相较于数十年前被视为罕见病的时期已大幅上升。
- 男女比例约为4:1，但女性漏诊（“掩蔽效应”）以及神经系统韧性与成熟速度的性别差异可能是其中的影响因素。
- 重度自闭症——谱系中最严重的一端——常与智力残疾和癫痫共病，且通常与可识别的基因突变相关联。
- 该病症被理解为类似于历史上的”发烧”更为恰当：一种具有多种不同生物学病因的行为信号，需要针对不同病因施以不同的治疗。
患病率为何上升？
- 诊断标准的拓宽与诊断迁移（例如，此前被诊断为智力残疾的儿童如今符合自闭症诊断标准）解释了部分增长。
- 高度遗传性成分——在所有精神疾病中遗传性最高之列——但许多致病变异单独来看均较为罕见。
- 环境因素（例如，孕期服用沙利度胺会增加自闭症风险）有所贡献，但尚未确立单一的环境致病因素。
- 没有可靠证据支持疫苗是致病原因；支持该说法的原始论文已被撤稿，相关关联亦未能在流行病学研究中得到重复验证。
- 患病率的上升是全球性的，在韩国、斯堪的纳维亚等地区均有观察——这削弱了与美国特有环境暴露相关的论点。
自闭症的遗传学
- 数百个基因与自闭症相关，主要分为以下几大类别：
- 突触病变：突触蛋白基因突变
- 离子通道病变：离子通道基因突变（如钙离子通道）
- 染色质病变：DNA包装蛋白基因突变
- Timothy综合征是一种有据可查的单基因遗传形式：钙离子通道基因（CACNA1C）的单个核苷酸变异导致通道持续开放时间延长，进而引发自闭症、癫痫及心脏异常。
- 许多与自闭症相关的基因同样在外周组织（皮肤、感觉神经元）中表达，由此引发了一种可能性：发育过程中外周的干扰可能破坏中枢神经系统的连接形成——这一观点得到了哈佛大学 David Ginty 实验室精妙动物实验的支持。
- 突变可以是遗传性的，也可以是**新发（de novo）**的（父母双方均不携带）；人类自然会产生约80个新突变，其中约30个为截断蛋白质的突变。
基因疗法与CRISPR
- 基因疗法广义上涵盖：通过病毒载体递送缺失基因、直接递送功能性蛋白质，或使用CRISPR在DNA层面纠正突变。
- 镰状细胞贫血症已成功通过CRISPR得到治疗，因为靶细胞（血细胞）易于获取。
- 脑部递送的挑战远大于此：
- 鞘内注射（脊髓腔）或直接手术注射是可行方案，但效率较低。
- 腺病毒 / AAV是常用载体；经改造后无致病性，但无法携带过大的基因（例如，钙离子通道基因体积过大）。
- 免疫反应限制了重复给药——通常只能注射一次。
- 基因编辑一旦整合便不可逆转。
- 在DNA直接纠正之外，以RNA层面为靶点的下游策略目前更具可操作性，也是 Pașca 实验室正在采用的方向。
- 时机至关重要：对于神经发育障碍而言，干预能产生治疗效果的时间窗口目前尚不明确。
诱导多能干细胞（iPS细胞）与山中伸弥的发现
- 2006至2007年间，**山中伸弥（Shinya Yamanaka）**证明，通过导入四种转录因子（“山中因子”），成体皮肤细胞可以被重编程还原为多能干细胞——逆转了此前认为单向行进的发育进程。
- 这一发现消除了对胚胎干细胞的需求，以及由此引发的伦理争议。
- iPS细胞可以：
- 无限期保存（冻存后复苏不发生变化）
- 分化为几乎任何细胞类型，包括神经元
- 通过简单的皮肤活检从任何患者体内获取
类器官：培养皿中的大脑
- 类器官是从iPS细胞培育而成的三维自组织人类神经元簇——而非平面的二维培养物。
- 主要优势：
- 神经元在三维环境中发育，存活时间远长于平面培养皿（可存活数月至数年）
- 发育进程追踪真实的人类发育时序，包括分子转换，如NMDA受体2B→2A亚基转变（约在九个月时发生）——且不受任何分娩相关激素信号的影响
- 可从特定疾病患者体内获取，从而能够直接比较患者与健康对照者的神经元
- Pașca 实验室保持着迄今报道的持续时间最长的类器官培养记录——部分已培养超过800天。
- Timothy综合征类器官的重要发现：与对照组相比，患者来源神经元中钙离子内流时间延长，这是首次在人类神经元中直接证实了该通道缺陷。
干细胞注射：现有证据怎么说
- 脐带干细胞的分化潜能已受到限制——主要适用于血液疾病，尽管商业宣传如此，并非对未来所有干细胞疗法均具有普遍适用性。
- 非正规商业干细胞注射（在墨西哥、哥伦比亚及部分欧洲诊所提供）用于治疗自闭症等疾病：
- 缺乏科学依据（目前尚不知晓自闭症中存在某种”缺失”的特定细胞类型）
- 所用细胞来源及类型往往不明
- 存在感染、肿瘤形成及神经损伤的风险
- 任何报告中的改善很可能源于强烈的**（父母的）替代性安慰剂效应**、自然发育里程碑的达成，或类似”发烧效应”的非特异性炎症反应
English Original 英文原文
Curing Autism, Epilepsy & Schizophrenia with Stem Cells
Summary
Dr. Sergiu Pașca, professor of psychiatry and behavioral sciences at Stanford and director of the Stanford Brain Organogenesis Program, discusses the biology and genetics of autism spectrum disorder, why its prevalence is rising, and the groundbreaking stem cell technologies his lab is developing. He explains how organoids and assembloids—three-dimensional human brain circuits grown from stem cells—are enabling researchers to study psychiatric diseases like profound autism and schizophrenia at the cellular level, and how these models are being used to develop the first targeted clinical therapies.
Key Takeaways
- Autism is not one disease. It is behaviorally defined, lacks biomarkers, and likely represents hundreds of distinct genetic conditions grouped under a single umbrella term—much like “fever” once was in 19th-century medicine.
- Prevalence is near 3% of the general population, up significantly from when it was considered rare; increases are seen globally, not just in the United States, making US-specific environmental explanations insufficient.
- Strong genetic heritability is the most robustly supported finding in autism research. Hundreds of genes—affecting synaptic proteins, ion channels, and chromatin packaging—are associated with specific forms of autism.
- Only ~20% of profound autism patients receive a specific genetic diagnosis today; the rest remain in an “idiopathic” category where the precise cause is unknown.
- The Yamanaka factors (induced pluripotent stem cells / iPS cells) revolutionized neuroscience research by allowing any patient’s skin cells to be reprogrammed into neurons—eliminating the need for human embryonic tissue.
- Organoids recapitulate real developmental timing, including canonical molecular switches like the NMDA receptor 2B-to-2A subunit shift that occurs around the equivalent of birth—without any external hormonal cues.
- Unregulated stem cell injections offered in clinics abroad are scientifically unsupported for autism and carry serious risks, including infection, tumor growth, and permanent neurological damage.
- Gene therapy using CRISPR holds promise but faces major hurdles for brain disorders: delivery to the right cells, limited viral cargo size, irreversibility, and the timing problem of how early intervention must occur.
- Human brain myelination continues through the third decade of life, underscoring how prolonged and vulnerable neurodevelopment is compared to other organs.
Detailed Notes
What Is Autism?
- Autism is a behaviorally defined spectrum disorder—diagnosed by the presence and absence of specific behaviors at certain developmental ages, with no biological biomarker.
- Diagnosed in approximately 1 in 30–40 individuals (close to 3%), up from rare-disease status decades ago.
- Male-to-female ratio is roughly 4:1, though diagnostic under-detection in females (“masking”) and differences in nervous system resilience and maturation pace are likely contributing factors.
- Profound autism—the most severe end of the spectrum—is commonly co-occurring with intellectual disability and epilepsy, and is often linked to identifiable genetic mutations.
- The condition is better understood as analogous to historical “fever”: a behavioral signal with many distinct biological causes requiring different treatments.
Why Is Prevalence Rising?
- Broadened diagnostic criteria and diagnostic migration (e.g., children previously diagnosed with intellectual disability now meeting autism criteria) account for some increase.
- Highly heritable component—among the highest heritability of any psychiatric disorder—but many causative variants are individually rare.
- Environmental factors (e.g., thalidomide during pregnancy increases autism risk) contribute, but no single environmental cause has been established.
- No solid evidence supports vaccines as a cause; the foundational paper supporting that claim has been retracted and the association has not been replicated epidemiologically.
- Rising prevalence is global, observed in South Korea, Scandinavia, and elsewhere—undermining arguments tied to US-specific exposures.
Genetics of Autism
- Hundreds of genes are implicated, falling into major categories:
- Synaptopathies: mutations in synaptic proteins
- Channelopathies: mutations in ion channels (e.g., calcium channels)
- Chromatinopathies: mutations in DNA-packaging proteins
- Timothy Syndrome is a well-characterized monogenic form: a single nucleotide change in a calcium channel gene (CACNA1C) causes the channel to stay open longer, resulting in autism, epilepsy, and cardiac abnormalities.
- Many autism-associated genes are also expressed in peripheral tissues (skin, sensory neurons), raising the possibility that peripheral disruptions during development can perturb central nervous system wiring—supported by elegant animal experiments from David Ginty’s lab at Harvard.
- Mutations can be inherited or arise de novo (not present in either parent); humans naturally acquire ~80 new mutations, ~30 of which are protein-truncating.
Gene Therapy and CRISPR
- Gene therapy broadly encompasses: delivering a missing gene (via viral vector), delivering a functional protein directly, or correcting a mutation at the DNA level using CRISPR.
- Sickle cell anemia has been successfully treated using CRISPR because the target cells (blood cells) are easily accessible.
- Brain delivery is far more challenging:
- Intrathecal (spinal canal) or direct surgical injection are options but have low efficiency.
- Adenoviruses / AAVs are commonly used vectors; they are modified to be non-pathogenic but cannot carry very large genes (e.g., calcium channel genes are too large).
- Immune responses limit repeated dosing—typically one shot only.
- Gene edits are irreversible once integrated.
- A downstream RNA-level strategy (rather than direct DNA correction) is currently more practical and is the approach Pașca’s lab is pursuing.
- Timing is critical: the window in which correction is therapeutically meaningful for neurodevelopmental disorders is not yet well defined.
Induced Pluripotent Stem Cells (iPS Cells) and the Yamanaka Discovery
- In 2006–2007, Shinya Yamanaka demonstrated that adult skin cells could be reprogrammed back into pluripotent stem cells by introducing four transcription factors (the “Yamanaka factors”)—reversing what was thought to be a one-way developmental street.
- This discovery eliminated the need for embryonic stem cells and the associated ethical controversies.
- iPS cells can be:
- Maintained indefinitely (frozen and revived without change)
- Differentiated into virtually any cell type, including neurons
- Derived from any patient via a simple skin biopsy
Organoids: Brain-in-a-Dish
- Organoids are three-dimensional, self-organizing clusters of human neurons grown from iPS cells—not flat, two-dimensional cultures.
- Key advantages:
- Neurons develop in 3D, surviving far longer (months to years) than in flat dishes
- Development tracks real human developmental timing, including molecular switches like the NMDA receptor 2B→2A subunit transition at approximately nine months—occurring without any birth-related hormonal cues
- Can be derived from patients with specific diseases, allowing direct comparison of patient vs. healthy donor neurons
- Pașca’s lab maintains the longest-running organoid cultures ever reported—some kept for over 800 days.
- Key finding in Timothy Syndrome organoids: calcium influx in patient-derived neurons was prolonged compared to controls, directly demonstrating the channel defect in human neurons for the first time.
Stem Cell Injections: What the Evidence Says
- Umbilical cord stem cells are already restricted in potency—applicable mainly to blood disorders, not universally useful for future stem cell therapies despite commercial marketing.
- Unregulated commercial stem cell injections (offered in Mexico, Colombia, and some European clinics) for autism and other conditions:
- Lack scientific rationale (no specific cell type is known to be “missing” in autism)
- Often involve cells of unknown origin or type
- Carry risks of infection, tumor formation, and neurological damage
- Any reported improvements are likely due to strong placebo effects by proxy (in parents), natural developmental milestones, or non-specific inflammatory effects resembling the “fever effect”