综述：FKBP5/FKBP51介导的信号通路在神经精神疾病中的作用：生物标志物开发与靶向治疗的见解

本综述深入探讨了FKBP5/FKBP51在神经精神疾病中的核心作用，系统阐述了其介导的信号通路（如HPA轴调控、糖皮质激素受体GR功能调节）如何影响应激反应与疾病病理。文章聚焦于该靶点在重度抑郁症（MDD）、创伤后应激障碍（PTSD）及焦虑症等疾病中的机制，并前瞻性地讨论了以其为靶点开发新型生物标志物（Biomarker）和靶向治疗策略的巨大潜力，为相关领域的研究与转化提供了重要见解。

　　

Abstract

自然暴露对心理健康的益处已得到广泛证实。然而，单纯观看自然图像是否能产生类似的恢复效应及其涉及的大脑机制仍不清楚。为此研究，我们招募了131名健康大学生，并将其随机分配到自然图像观看组（NG）或城市图像观看组（CG），并进一步选择了49名参与者（NG=26，CG=23）在执行行为任务时接受功能性磁共振成像（fMRI）扫描。首先，我们比较了在应激诱导和图像观看后，NG和CG在情感和压力相关的主观评分和唾液皮质醇水平的变化。接下来，我们检查了在图像观看期间，两组之间默认模式网络（DMN）的功能连接（FC）模式的差异。最后，我们探讨了观看自然图像后观察到的恢复效应与FC改变之间的相关性。在压力下，NG参与者报告了更大的积极情感主观评分变化（t=2.610, p=0.010）、更低的消极情感（t=?3.008, p=0.003）和更少的状态反刍思维（t=?2.103, p=0.037）。神经数据还表明，DMN子系统与注意力和执行区域的连接在自然体验过程中调节压力相关反应方面起着至关重要的作用。内侧DMN子系统与其他网络之间增加的FC与情感和状态反刍思维的行为恢复分数显著相关。这些发现表明，观看自然场景图像有助于压力恢复，突显了室内自然观看有助于缓解城市环境中心理挑战的潜力。

1. Introduction

随着城市化进程加快和城市生活挑战的增加，人们发现自己处于压力日益增大的环境中。 prolonged and excessive stress can lead to illness and negatively affect mental health. Research indicates that mood and anxiety disorders are more prevalent in urban centers, and their incidence has been increasing. A study investigating the relationships between urban environments and psychiatric symptoms found that exposure to urban environments dominated by high degrees of poverty and air pollution is associated with increased negative emotional symptoms. In recent years, the restoration benefits of natural environments have been emphasized as a means to reduce the effects of stress on human cognition and mood. Human beings' physical and mental health has been closely linked to the natural environment. A growing body of research has demonstrated that spending time in natural environments can improve mental health. Meta-analyses revealed moderate to large positive effects of natural environments on anxiety, fatigue, positive affect, and vigor. Walking in green spaces and built green environments such as parks has been linked to improvements in depressive mood. Moreover, immersion in nature, such as in forested areas, has been found to significantly reduce symptoms of depression and anxiety and to improve mood. Relevant neuroimaging studies have shown that rumination decreases after a walk in a natural environment, and neural activation in both the subgenual prefrontal cortex and the amygdala is also reduced. Thus, having a connection to nature plays an important protective role for the mental health of urban dwellers under stress.

There is no doubt that one of the consequences of urbanization is the loss of the natural environment and the depletion of natural resources. If more accessible ways to connect with nature existed beyond hiking and walking, opportunities for mental health improvement would be greatly expanded. Recently, many studies have focused on the effects of indirect contact with nature by viewing images or videos and using virtual reality technology. Several studies have shown that exposure to the natural environments by viewing images has a potential role in human health and well-being, facilitating the restoration of attention, improving executive functioning, and alleviating negative affect and stress. However, it has also been suggested that virtual nature simulations, even those incorporating videos, may not fully replicate the experience of being in nature outdoors or effectively restore attentional resources. This suggests that the impact of viewing nature images on mental health still requires further exploration. In particular, static nature images are more convenient and accessible than videos or virtual reality, making them a promising tool for stress relief. If the effectiveness of viewing nature images in promoting stress recovery can be demonstrated, such could offer a simple, practical way for urban dwellers to alleviate stress—such as by setting nature images as their computer screensaver or cellphone wallpaper to quickly and easily relax during daily life.

The biophilia hypothesis proposes that humans have an innate tendency to interact positively with nature. Previous studies have found that natural views held viewers’ attention and interest more effectively than urban scenes. The Attention Restoration Theory (ART) posits that exposure to nature can provide a specific stimulus for attention to recover from the mental fatigue associated with the urban environments. Recent neuroimaging studies have attempted to explain biophilia in terms of the mechanisms of how the brain responds to exposure to nature. A study using functional magnetic resonance imaging (fMRI) found that viewing green urban landscapes elicits changes in the activity of the ventral posterior cingulate cortex (PCC), which modulates attentional and executive regions in a predominantly feedforward manner to regulated behavioral stress-related responses. Another study also observed that functional connectivity (FC) was significantly greater in the default mode network (DMN) and other networks such as the dorsal attention network (DAN) when participants viewed photographs of natural settings rather than built environments. These findings suggest that the involvement of the DMN and attention-related regions in response to viewing nature may help explain how natural environments allow individuals to disengage from effortful directed attention, supporting the core principles of ART.

Neuroimaging literature has uncovered three key networks particularly linked to stress responsiveness and regulation: the salience network (SN), the executive control network (ECN), and the default mode network (DMN). Studies examining stress-induced changes in FC have reported that stress-induced cortisol increases are associated with increased connectivity within the SN but with decreased coupling of the DMN at both local (within the network) and global (synchronization with brain regions also outside the network) levels. More recently, exposure to stress was shown to induce a dynamic functional interplay between heart rate variability and neural network connectivity, and thus heart rate variability was associated with DMN–ECN functional coupling before, but not after, stress exposure. Collectively, these findings highlight how stress alters brain network dynamics, particularly implicating altered FC both within the DMN and involving its interactions as a key neural signature of the stress response.

The DMN is a collection of distributed and interconnected brain regions that are typically inhibited when an individual is focused on external stimuli; however, in the absence of attention to external stimuli, the DMN switches or ''defaults'' to internally focused thought processes, such as self-referential processing and introspection. As research continues, the DMN has evolved from being viewed as a unitary network to exhibiting distinct regional functions, leading to its conceptualization of it as containing at least three subsystems, the core DMN (cDMN), dorsal DMN (dDMN), and medial DMN (mDMN). The cDMN consists of the anterior medial prefrontal cortex (aMPFC) and PCC, which are hypothesized to integrate the function of the other subsystems and is involved in the introspection about one's own mental states. The dDMN encompasses the dorsal medial prefrontal cortex (dMPFC), the temporoparietal junction (TPJ), the lateral temporal cortex (LTC), and the temporal pole (TempP), whose function are associated with mentalizing and social cognition, as well as story comprehension and semantic/conceptual processing. The mDMN includes the ventral medial prefrontal cortex (vMPFC), posterior inferior parietal lobule (pIPL), retrosplenial cortex (Rsp), parahippocampal cortex (PHC), and the hippocampal formation (HF), which are involved in episodic/contexture retrieval, mental simulations of the future, and autobiographical recall. Alterations in DMN connectivity have consistently been implicated in various psychiatric disorders and particularly in stress-related disorders; for example, DMN dysfunction has been implicated in both rumination and major depression. Previously, a meta-analysis including 14 fMRI studies linking DMN subsystems to rumination highlighted prominent roles of the cDMN and dDMN subsystems. Similarly, a study inducing rumination in healthy participants found that, compared to when participants were in a distracted state, rumination generally decreased within-DMN FC while increasing FC between the cDMN and mDMN subsystems and decreasing FC between the cDMN and dDMN subsystems. Thus, an improved understanding of the interactions between different DMN regions and subsystems may be important to explore stress-related neural underpinnings.

As mentioned previously, the restorative effects of viewing natural images have yet to be further investigated, and exploring the interactions between different DMN regions and DMN subsystems may help to disentangle key neural mechanisms of stress response and regulation. If it could be experimentally demonstrated that viewing nature images achieves a restorative effect similar to that of physical contact with the natural environment, we could then expect to observe the activation of regions within the DMN and the interactions between DMN subsystems, and the connectivity of the DMN subsystems and other networks. Accordingly, our study aimed to examine the recovery effects of viewing nature images after stress induction compared to viewing city images. Specifically, we conducted a behavioral experiment to determine whether exposure to natural stimuli would accelerate recovery from stress and improve mood (affect) compared to viewing urban stimuli following a stress-induction procedure. Additionally, we explored the FC patterns of the DMN using fMRI to assess the recovery effects of exposure to natural stimuli following stress induction. We hypothesized that participants in the stress condition who viewed nature images (nature group [NG]) would report more significant restoration of mood and reduced stress levels, as well as changes in salivary cortisol levels, compared to their counterparts who viewed city images (city group [CG]). Notably, rumination—characterized by sustained negative emotional immersion and self-referential focus—has been shown to alter DMN connectivity in specific ways. In contrast, natural viewing is posited to act as a restorative process that alleviates stress by redirecting attention toward the external environment and facilitating the recovery of directed attention, a mechanism involving functional coupling between the DMN and attention-related regions. Given that rumination (a stress-maintaining state) is associated with decreased within-DMN FC and specific interactions between DMN subsystems, we reasoned that natural viewing (a stress-reducing state) might exhibit opposing or distinct patterns. While our investigation of the neural mechanisms underlying how natural viewing accelerates stress recovery was exploratory, we expected FC differences within the DMN and between DMN subsystems to emerge between NG and CG during natural viewing. These differences might involve alterations in the coupling between DMN subsystems and whole-brain networks, such as enhanced connectivity between the DMN subsystems and attention-related regions.

2. Materials and methods

2.1. Participants

通过线上和线下发布招募参与者，并通过包含患者健康问卷（PHQ-9）和人口统计信息的链接进行注册。共招募了131名健康大学生（33名男性和98名女性；年龄18-28岁），其中49名（18名男性和28名女性）被进一步随机选择并受邀在一个月后执行行为任务时接受fMRI扫描。参与者的分配过程见图1a。

入选者需年满18岁且PHQ-9得分<10分。反之，我们排除了处于月经期的女性参与者（以避免其对皮质醇的 reported effect）以及主修心理学或精神病学的个人。同时，接受fMRI成像的参与者还需满足以下标准：（i）体内无金属植入物，如起搏器、人工耳蜗、纹身和假体；（ii）无幽闭恐惧症。所有受试者被告知在实验前一天晚上6点后不要饮用含咖啡因的饮料，如咖啡或茶，并在实验当天检查前避免进行非常剧烈的运动。南方医科大学珠江医院的机构审查委员会批准了整体研究方案，所有参与者均提供了签署的书面知情同意书，并在实验结束时获得报酬。

2.2. Procedure

所有材料均以脚本形式通过E-Prime 3.0软件运行。通过该软件，呈现了一系列基线量表、自然或城市图片以及主观评分。

参与者被随机分配到NG或CG。到达实验室后，参与者首先完成知情同意书和以下问卷：基本人口统计信息、贝克抑郁量表、积极和消极情感量表、感知压力量表、自然连接量表、心境状态量表和状态-特质焦虑量表。此外，还收集了积极情感、消极情感、放松、紧张和状态反刍思维项目的基线测量值。主观评分基于从“完全没有”（0分）到“非常”（100分）的视觉模拟量表。状态反刍思维通过简要状态反刍量表的第6和第8项得分之和来测量。在主观评分间隔15分钟后，参与者接受唾液皮质醇测量。然后，他们完成特里尔社会应激测试（TSST）以诱导心理压力。

在TSST之后，参与者执行了一系列图像观看任务。每项任务后，参与者被要求完成主观评分和唾液皮质醇测量，然后再进行下一项任务。整个过程见图1b。

2.3. Stress-induced paradigm

特里尔社会应激测试（TSST）。采用经典的TSST来诱导参与者的应激反应。整个过程包括5分钟的准备演讲任务、5分钟的演讲任务和5分钟的连续减法任务。演讲任务要求持续约5分钟，期间参与者被指示如下：“你现在将进行一项演讲任务，描述你未来的工作以及你认为自己适合该工作的原因。你的整个演讲将被录制，两名评估员将评估你的言语和非言语行为。你有5分钟时间准备面试。在此准备期间，你可以使用笔和纸做笔记，但这些笔记不允许在实际面试中使用。”每当参与者在不到5分钟内完成演讲时，面试官会以标准化方式回应。首先，他们告诉参与者“你还有一些时间。请继续！”如果受试者在5分钟结束前第二次完成，面试官会沉默20秒，然后提出准备好的问题。完成演讲任务后，参与者继续进行连续减法任务，期间面试官指示：“你现在将完成一个5分钟的算术任务。请从1022开始连续减去13，并尽可能快速准确地说出每个答案，用英语。如果你出错，你需要重新开始。”我们要求用英语口头报告算术结果是为了诱发更大的压力。

2.4. Stress recovery paradigm

实验图片材料。从互联网上为实验条件选择了两组图像材料。所有图片的分辨率均为1200×800像素。选择的自然环境图片以树木、草地或植物等自然元素为特色，没有任何人造结构，如建筑物、车辆或道路。相反，描绘城市环境的图像包括建筑物、道路和砖墙等人造特征。为了控制潜在的混淆变量，需要注意的是，建筑环境的一个关键特征是缺乏自然特征和城市设施的突出性。实验图片材料的自然度和城市化评级结果见补充图1。

在压力恢复阶段，参与者进行约10分钟的图片观看。每张图片显示20秒，每组共呈现30张图像。参与者被要求在整个呈现过程中观看每张图片，并想象自己处于图片所描绘的环境中（图1c）。

2.5. Imaging acquisition and preprocessing

使用3.0 Prisma-fit磁共振成像设备和64通道头线圈（Siemens, Erlangen, Netherlands）对参与者进行扫描。使用梯度回波平面成像序列执行功能成像，以收集任务状态下大脑的血氧水平依赖（BOLD）信号。扫描参数如下：重复时间（TR）=1500 ms，回波时间（TE）=31 ms，层厚=2.4 mm，层数=60，时间点=530，视野=88×88×60 mm，翻转角（FA）=70°，体素大小=2.4×2.4×2.4 mm³。使用矢状位高分辨率T1加权三维MPRAGE序列获取三维T1加权图像，扫描参数如下：TR=1800 ms，TE=2.07 ms，层厚=0.8 mm，层数=208，视野=320×320×208 mm，FA=9°，体素大小=0.8×0.8×0.8 mm³。

基于统计参数映射（SPM12）和MATLAB R2018b版（MathWorks, Inc., Natick, MA, USA），使用GRETNA工具箱对基于任务的fMRI数据进行预处理。对于每位参与者，前7个volume被丢弃，总共剩下532个volume。然后，将数据与第一个volume进行重新对齐以进行头动校正，并排除了头动>3.0 mm或x、y、z方向旋转>3.0°的三名参与者。随后，将数据分割为灰质、白质和脑脊液，并空间标准化到蒙特利尔神经学研究所（MNI）模板，重采样体素大小=3×3×3 mm；然后对所得图像进行线性去趋势和时间带通滤波（0.01?0.1 Hz）。

2.6. FC analysis

使用DPABI软件进行FC分析，并使用基于MATLAB的BrainNet Viewer软件将结果可视化。为了分析整个DMN和每个单独DMN子系统的平均FC，我们使用了基于种子的分析，根据已报道的DMN感兴趣区域（ROIs），我们使用一个球体（半径5 mm）作为DMN三个子系统的种子。具体来说，cDMN包括aMPFC（MNI坐标，?6, 52, ?2）和PCC（MNI坐标，?8, ?56, 26）；dDMN包括dMPFC（MNI坐标，0, 52, 26）、TPJ（MNI坐标，?54, ?54, 28）、LTC（MNI坐标，?60, ?24, ?18）和TempP（MNI坐标，?50, 14, ?40）；mDMN包括vMPFC（MNI坐标，0, 26, ?18）、pIPL（MNI坐标，?44, ?74, 32）、Rsp（MNI坐标，?14, ?52, 8）、PHC（MNI坐标，?28, ?40, ?12）和HF（MNI坐标，?22, ?20, ?26）（图2）。由于镜像（右/左）种子区域之间存在强相关性，并且顶叶区域存在强偏侧性，我们仅使用了11个左侧化的ROI。我们首先从这11个ROI中提取时间序列，并为每位参与者创建一个11×11的相关矩阵。在对相关值进行Fisher's r-to-z转换后，我们平均了11个ROI的FC以获得整个DMN的FC，然后平均了每个子系统内/间的每对FC的总和。最终的平均FC被定义为DMN子系统内/间的FC。

进一步，为了评估DMN子系统与其他脑网络之间的FC，我们为每位受试者计算了种子-全脑r-FC图，然后使用Fisher's r-to-z转换将其转换为z-FC图。然后，我们将所有脑区域分为七个网络，包括视觉网络（VN）、躯体运动网络（SMN）、背侧注意网络（DAN）、腹侧注意网络（VAN）、边缘网络（LN）、额顶网络（FPN）和DMN。

2.7. Data analysis

使用IBM SPSS 26进行统计分析。基于三个阶段量化行为表现的主观评分变化和生理上的唾液皮质醇变化：t₀在基线阶段测量，t_s在TSST应激诱导后测量，t_x在图像观看压力恢复后测量。我们将t₀, t_s和t_x标准化为R₁ = [(t_s?t₀)/(t_s+t₀)]以测试诱导应激的成功，而R₂ = [(t_x?t_s)/(t_x+t_s)]用于测量压力恢复效果。对于R₁，消极情感、紧张、状态反刍思维和唾液皮质醇的值越高，以及积极情感和放松的值越低，表明TSST范式有效诱导了压力。相反，对于R₂，消极情感、紧张、状态反刍思维和唾液皮质醇的值越低，以及积极情感和放松的值越高，表明图像观看范式成功促进了压力恢复。使用描述性统计呈现参与者的特征，包括两组招募的参与者数量，并注明连续变量的均值和标准化偏差。使用单样本t检验（p < 0.05）研究应激诱导和图像观看后标准化主观评分得分和唾液皮质醇水平是否发生变化，而使用双样本t检验（p < 0.05）检查两组之间标准化主观评分得分和唾液皮质醇水平的变化是否不同。

使用双样本t检验比较NG和CG之间DMN的FC、DMN子系统内和之间的FC以及DMN子系统与其他网络之间的FC。在组间比较过程中，将参与者的年龄、性别和头动视为 nuisance variable。对于种子-全脑FC分析，显著的双尾体素水平阈值设置为p < 0.005，簇阈值p < 0.05。根据需要，使用高斯随机场校正来校正多种子和子系统之间的多重比较。最后，我们进行了Spearman相关分析，以探索NG中所有已识别FC与压力恢复效果评分之间的关系，分别针对积极情感、消极情感、放松、紧张和状态反刍思维。

3. Results

3.1. Stress-induced effects

在年龄、性别、基线量表以及主观评分得分方面，NG和CG之间未观察到显著差异（p > 0.05）。我们将基线阶段和TSST应激诱导后的主观评分和唾液皮质醇水平标准化，以计算应激诱导效果。在诱导心理压力后，NG和CG的主观评分得分和唾液皮质醇水平（R₁）均发生显著变化（p < 0.001）。双样本t检验显示，两组在主观得分和唾液皮质醇水平的变化上没有显著差异（p > 0.05）。总之，这些发现表明压力被成功诱导。

3.2. Recovery effects after image viewing

为了量化图像观看后的恢复效果，我们将TSST应激诱导和图像观看后的主观评分和唾液皮质醇水平标准化。参与者观看图像后，NG的所有主观评分得分（R₂）均发生显著变化（p < 0.001），但唾液皮质醇水平除外；而CG的主观评分得分和唾液皮质醇水平（R₂）也发生显著变化（p ≤ 0.001），但消极情感除外。双样本t检验显示，NG和CG在图像观看后的主观评分在积极情感（均值±标准差：NG, 0.22±0.23; CG, 0.11±0.26, t=2.610, p=0.010）、消极情感（均值±标准差：NG, ?0.29±0.42; CG, ?0.07±0.40, t=?3.008, p=0.003）和状态反刍思维（均值±标准差：NG, ?0.32±0.35; CG, ?0.19±0.33, t=?2.103, p=0.037）方面存在显著差异。然而，我们没有发现放松（t=?0.254, p=0.800）和紧张（t=?1.151, p=0.252）的主观评分存在显著差异。此外，NG和CG之间的唾液皮质醇水平没有显著差异（t=0.532, p=0.596）。这些结果表明，自然观看对压力后的情感和状态反刍思维具有恢复作用。

3.3. FCs of the overall DMN and within and between DMN subsystems

49名参与者被随机选择在执行行为实验任务时完成fMRI任务。由于其中3人因头动过大被排除，最终46名参与者（NG=25, CG=21）被纳入分析，行为结果见补充表4。

我们使用基于种子的分析来检查该亚群中整个DMN的平均FC以及每个子系统内/间的FC。NG和CG之间未观察到整体DMN的FC存在显著差异（p > 0.05）。同样，NG和CG之间在DMN子系统内和之间的平均FC也未观察到显著差异（p > 0.05）。

3.4. FCs between DMN subsystems and the whole-brain network

为了进一步评估DMN子系统与其他脑网络之间的FC，我们为每位受试者计算了种子-全脑r-FC图，并将其转换为z-FC图。我们使用双样本t检验比较NG和CG之间DMN子系统与其他网络之间的FC，控制年龄、性别和头动作为 nuisance variables。对于种子-全脑FC分析，显著的双尾体素水平阈值设置为p < 0.005，簇阈值p < 0.05，使用高斯随机场校正来根据需要校正多种子和子系统之间的多重比较。NG和CG的FC差异结果如图4a所示。然后，我们通过将所有脑区域分为七个网络来说明结果。

3.4.1. The cDMN

与CG相比，NG显示出aMPFC与VN（右侧枕下回[IOG]和左侧枕中回[MOG]）的FC显著增强。后续的Spearman相关分析发现，在NG中，cDMN和VN（aMPFC–MOG.L）的连接强度与R₂消极情感恢复得分呈正相关（r=0.542, p=0.005）。

3.4.2. The dDMN

与CG相比，NG显示出TPJ与DMN（右侧海马旁回[PHG]）以及TempP与SMN（左侧辅助运动区[SMA]）的FC显著增强。

3.4.3. The mDMN

相对于CG，NG显示出vMPFC与FPN（右侧缘上回[SMG]和右侧额中回[MFG]）和VAN（左侧SMG）的FC分别增加，以及Rsp与FPN（右侧顶下小叶[IPL]、左侧MFG和右侧背外侧额上回[SFGdor]）的FC增加。在NG中，mDMN–FPN（Rsp–IPL.R）连接强度与R₂消极情感恢复得分呈正相关（r=0.411, p=0.041）。

与CG相比，NG还显示出PHC与DAN（右侧SMG）、VAN（左侧SMG、左侧MFG和右侧MFG）和FPN（右侧IPL）的FC分别增加，但PHC与VN（左侧舌回[LING]、左侧MOG和右侧LING）的FC减少。在NG中，mDMN–DAN（PHC–SMG.R）连接强度与R₂积极情感恢复得分呈正相关（r=0.444, p=0.030）。分别地，mDMN–VAN（PHC–SMG.L; PHC–MFG.L）连接强度与R₂积极情感恢复得分呈正相关（r=0.452, p=0.027; r=0.629, p=0.001），而mDMN–VAN（PHC–MFG.R）连接强度与R₂状态反刍思维恢复得分呈正相关（r=0.417, p=0.038）。此外，在NG中，mDMN–VN（PHC–LING.L）连接强度与R₂积极情感恢复得分呈正相关（r=0.489, p=0.015），mDMN–VN（PHC–LING.R）连接强度与R₂状态反刍思维恢复得分呈正相关（r=0.526, p=0.007）。

4. Discussion

这项研究考察了暴露于自然对压力恢复的影响，并揭示了在整体和子系统水平上DMN的FC模式