The Neuroendocrinology of Stress: A Never Ending Story
压力的神经内分泌学：一个永无止境的故事

Why should anyone be interested in stress? Besides the purely personal fascination of trying to understand how a phenomenon as pervasive as stress can acutely alter cognition, memory, cardiovascular activity and glucose, protein and fat metabolism, it is clearly an important aspect of global health to understand how chronic stressful stimuli can lead to increased morbidity and mortality from depression, cardiovascular disease and metabolic disorders.
为什么有人会对压力感兴趣？除了试图理解像压力这样普遍存在的现象如何迅速改变认知、记忆、心血管活动以及葡萄糖、蛋白质和脂肪代谢的纯粹个人魅力之外，了解慢性压力刺激如何导致抑郁症、心血管疾病和代谢紊乱的发病率和死亡率增加显然是全球健康的一个重要方面。

Since the neuroendocrine system is such a good candidate for mediator of many of the diseases linked to chronic stress, the first areas that need to be addressed are the central mechanisms underlying the neuroendocrine stress response and how they might change when acute stressors become repeated or chronic.
由于神经内分泌系统是许多与慢性应激相关的疾病的介质的良好候选者，因此需要解决的首要领域是神经内分泌应激反应的核心机制，以及当急性应激源变得反复或慢性时它们如何变化。

Acute and chronic stress
急性和慢性压力

The parvocellular cells of the paraventricular nucleus of the hypothalamus are the major information junction for the neuroendocrine response to stressors. Inputs from both limbic circuits and brain stem centres ensure these cells can be activated by both psychological and physical stressors (1), with a rapid increase in c-fos (2) followed by increased activation of many genes (3, 4), of which corticotrophin-releasing hormone (CRH) and arginine vasopressin (AVP) are the most important for neuroendocrine activation of the hypothalamic-pituitary-adrenal (HPA) axis. The peptide products, CRH and AVP, are secreted into the hypophyseal portal blood and act in a synergistic fashion on the corticotroph cells of the anterior pituitary to increase pro-opiomelanocortin (POMC) transcription, adrenocorticotrophin (ACTH) release into the circulation and consequent glucocorticoid hormone secretion from the adrenal cortex (Fig. 1).
下丘脑室旁核的细小细胞是神经内分泌对应激源反应的主要信息连接。来自边缘回路和脑干中心的输入确保这些细胞可以被心理和身体压力源激活 （ 1），c-fos （ 2） 迅速增加，随后许多基因 （ 3， 4） 的激活增加，其中促肾上腺皮质激素释放激素 （CRH） 和精氨酸加压素 （AVP） 对于下丘脑-垂体-肾上腺 （HPA） 轴的神经内分泌激活最为重要。肽产物 CRH 和 AVP 分泌到垂体门静脉血中，并以协同方式作用于垂体前叶的促肾上腺皮质激素细胞，以增加促黑皮质素 （POMC） 转录、促肾上腺皮质激素 （ACTH） 释放到循环中，以及随之而来的肾上腺皮质质激素分泌（图 1）。

In response to chronic stress, however, we find a markedly different hypothalamic response. In mycobacterial induced adjuvant arthritis (a very effective activator of HPA activity), although we find a chronic increase in POMC mRNA in the anterior pituitary, and both plasma corticosterone and adrenal weight (a well validated measure of chronic HPA activity), we actually find a paradoxical reduction both in paraventricular (PVN) CRH mRNA and in portal blood concentrations of CRH 41 (5). Interestingly, the decline in CRH mRNA and peptide were paralleled by a marked increase in AVP mRNA and portal blood AVP concentration (6), suggesting that, in this model of chronic stress, the central activation of HPA activity has been taken over by a predominant AVP rather than CRH drive.
然而，在应对慢性压力时，我们发现下丘脑反应明显不同。在分枝杆菌诱导的辅助性关节炎（一种非常有效的HPA活性激活剂）中，尽管我们发现垂体前叶中的POMC mRNA长期增加，血浆皮质酮和肾上腺重量（慢性HPA活性的良好验证的衡量标准）均存在矛盾的降低，但我们实际上发现室旁（PVN）CRH mRNA和CRH 41的门静脉血浓度均存在矛盾的降低（5）。有趣的是，CRH mRNA 和肽的下降与 AVP mRNA 和门静脉血 AVP 浓度的显着增加并行 （ 6），这表明在这种慢性应激模型中，HPA 活性的中枢激活已被占主导地位的 AVP 而不是 CRH 驱动接管。

We investigated this in more detail in a model of repeated daily acute stress (7). This confirmed that the ratio of AVP : CRH mRNA in the PVN increased with each episode of stress (8) and also that, eventually, there emerged an isolated AVP but not CRH response to subsequent restraints (7). This was in good keeping with the work from the laboratory of Aguilera and colleagues (9) on the HPA hormonal adaptation to repeated homotypic stressors (Fig. 1).
我们在重复的每日急性应激模型中对此进行了更详细的研究 （ 7）。这证实了 PVN 中 AVP ： CRH mRNA 的比率随着应激的每次发作而增加 （ 8），并且最终出现了孤立的 AVP，但对随后的约束没有 CRH 反应 （ 7）。这与 Aguilera 及其同事 （ 9） 的实验室关于 HPA 激素对重复同型应激源的适应的工作非常吻合 （ 图 1）。

Stressors and hormone secretion
压力源和激素分泌

How do the changes in hypothalamic CRH and AVP transcripts manifest themselves in terms of physiological function? To answer this, we first need to know how plasma glucocorticoids are regulated under basal conditions and then what happens under conditions of acute and chronic stress. Although this sounds reasonably straightforward, taking multiple samples from conscious freely-moving unrestricted rodents is quite challenging, particularly since any contact with the animal will result in activation of the HPA axis.
下丘脑CRH和AVP转录本的变化在生理功能方面如何表现出来？要回答这个问题，我们首先需要知道血浆糖皮质激素在基础条件下是如何调节的，然后在急性和慢性应激条件下会发生什么。虽然这听起来相当简单，但从有意识的自由移动的不受限制的啮齿动物身上采集多个样本是相当具有挑战性的，特别是因为与动物的任何接触都会导致HPA轴的激活。

Colin Ingram, Richard Windle and I therefore adapted the automated blood sampling system devised by Iain Robinson at Mill Hill (10) for use in our HPA studies. Using this equipment, we were able to demonstrate that there is not only a clear circadian rhythm of glucocorticoid secretion, but also that this circadian rhythm is actually made up from changes in the regulation of a much faster underlying ultradian rhythm consisting of pulses of corticosterone release that occur approximately one hourly (11).
因此，科林·英格拉姆（Colin Ingram），理查德·温德尔（Richard Windle）和我采用了伊恩·罗宾逊（Iain Robinson）在米尔希尔（Mill Hill）设计的自动血液采样系统（10），以用于我们的HPA研究。使用这种设备，我们能够证明不仅有明确的糖皮质激素分泌昼夜节律，而且这种昼夜节律实际上是由更快的潜在超节律的调节变化组成的，该节律由大约每小时发生的皮质酮释放脉冲组成（11）。

This hourly pulsatile secretion is characterised by alternating episodes of HPA activation and inhibition. Interestingly, after each pulse of secretion, there is actually a period of inhibition (in effect a refractory period) during which the HPA is no longer sensitive to activation by a mild stressor (11). This provides an explanation for the well-known variability in the HPA responsiveness to stressors, as the magnitude of the response will depend on the stage of the endogenous secretory cycle at which the animal is exposed to the stressor. Furthermore, when the frequency of pulsatility increases, as it does in chronic stress (vide infra), there is an increased proportion of time when the animals are in a stress nonresponsive state, giving rise to apparent stress hyporesponsiveness (12).
这种每小时一次的搏动分泌的特征是 HPA 激活和抑制的交替发作。有趣的是，在每次分泌脉冲之后，实际上有一个抑制期（实际上是一个不应期），在此期间，HPA不再对轻度应激源的激活敏感（11）。这为HPA对应激源反应的众所周知的变异性提供了解释，因为反应的大小将取决于动物暴露于应激源的内源性分泌周期的阶段。此外，当脉动频率增加时，就像在慢性应激中一样（见下文），当动物处于应激无反应状态时，时间比例增加，从而引起明显的应激反应性低（12）。

Of course, we now need to know what happens to the pattern of HPA activity in animals under conditions of chronic stress with a raised hypothalamic AVP and reduced CRH drive to the pituitary. These animals do indeed have a markedly altered regulation both of their circadian and ultradian rhythms: the circadian rhythm is flattened or lost and the frequency of corticosterone pulses almost doubles (12).
当然，我们现在需要知道在慢性应激条件下，动物的HPA活动模式会发生什么变化，下丘脑AVP升高，CRH对垂体的驱动减少。这些动物的昼夜节律和超节律确实发生了明显变化：昼夜节律变平或丢失，皮质酮脉冲的频率几乎翻了一番（12）。

Genetic and neonatal influences
遗传和新生儿的影响

There is great heterogeneity of stress responsiveness. The question we wanted to ask was whether this was related to the regulation of the underlying ultradian rhythm and, if so, how it might be related to genetic or epigenetic influences (Fig. 2).
应激反应性存在很大的异质性。我们想问的问题是，这是否与潜在的超节律的调节有关，如果是，它与遗传或表观遗传影响有何关系（图2）。

First we looked at the HPA axis in two strains of histocompatible rat that Esther Sternberg and colleagues had shown to have markedly different HPA responses to stress (13). The Lewis rat had a completely normal circadian and ultradian rhythm, and also showed a normal post-pulse refractory period. The stress hyper-responsive Fisher rat, on the other hand, had a markedly abnormal rhythm with high amplitude corticosterone pulses occurring throughout the 24 h and, quite remarkably, showed no post-pulse inhibition (14). We have therefore been able to demonstrate that the genetic difference between these two very similar strains of rat resulted in marked abnormalities in the regulation of their pulsatile corticosterone secretion and, consistent with this, in their stress responsiveness.

The second major influence that we wanted to investigate was the epigenetic effect of neonatal programming on adult HPA ultradian rhythmicity. As long ago as 1967 in a seminal paper in Science, Levine had demonstrated that the HPA axis could be programmed by early life events (15). Using a model developed by Nola Shanks in Michael Meaney’s laboratory (16), in which animals exposed to endotoxin during the first week of life become more stress responsive as adults, we investigated whether this change in HPA reactivity might be related to a change in the organisation of HPA pulsatility. We found that neonatal endotoxin resulted in both an increased frequency of pulses and increased corticosterone pulse amplitude (17), confirming that this neonatal stimulus had exerted long-term programming effects on the mechanisms generating basal pulsatile HPA activity in adult animals. We have now gone on to show that there are also organisational effects of the gonadal steroids that are normally secreted neonatally. Indeed, the presence or absence of circulating androgens perinatally can programme the activity of the HPA axis throughout the rest of the life of the animal (18, 19).

One of the things that fascinated us about the neonatal programming data was the fact that there were clear modifications of behaviour, as well as neuroendocrine hormone secretion. Thus, the animals that had been given neonatal lipopolysaccharide, showed much longer lasting activity responses to noise stress in adult life. We wanted to investigate this further and test the hypothesis that these altered responses might be related to a change in the regulation of serotonergic activity within the dorsal raphe nucleus. To test this, Chris Lowry and I collaborated with Paul Plotsky who used his maternal separation paradigm in which rats were exposed to either short (15 min) or long (180 min) periods of maternal deprivation during a critical period of development. Then, as adults, these neonatally-treated animals were exposed to the powerful psychosocial stressor of social defeat. Not only did we find that the different paradigms of maternal deprivation resulted in different behaviours, with the 180 min neonatal deprivation group showing more passive-submissive behaviour and less proactive coping behaviour (20), but also we were able to show that these behavioural effects were associated with marked differences in dorsal raphe tryptophan hydroxylase (TPH2) mRNA (21). Of particular interest was the fact that 15- and 180-min maternally-deprived rats that had undergone social defeat had much greater differences in TPH2 mRNA than unstressed maternally-deprived controls. This was a clear demonstration that the serotonergic differences following maternal separation were context specific and could be brought out by psychosocial stress in adult life. This, of course, has important implications for the effect of early life events on behavioural adaptation and on the susceptibility to affective and other stress-related disorders in adult life in man (Fig. 2).

Specificity of glucocorticoid signalling

The pulsatile nature of the glucocorticoid signal in the plasma provides scope for a digital, in addition to analogue, signal for tissue glucocorticoid receptors. To test whether tissues can detect different patterns of pulsatility, we need to demonstrate that rapid responses to individual glucocorticoid pulses can occur and define a mechanism through which different pulse frequencies could impart different information (Fig. 3).

We have approached this question by developing a model in which we can reproduce different pulse sizes and frequencies at will. This model, which was developed by a graduate student Crispin Wiles, consists of an adrenalectomised rat with an indwelling venous cannula connected to an automated infusion system. Using this system, Becky Conway-Campbell has been able to demonstrate an extremely rapid translocation of both the glucocorticoid receptor (GR) and the mineralocorticoid receptor (MR) to the nucleus with subsequent DNA binding. A remarkable novel finding was that GR, unlike MR, then rapidly dissociates from the DNA (22). This may well be related to the lower affinity of corticosterone for GR than for MR (23). Using subcellular fractionation and western blotting, we found that GR was lost from the nucleus within the time course of a single 1-h interpulse interval, whereas MR remains. Changes in pulse frequency will therefore have differential effects on MR and GR binding and also probably on MR and GR homodimer and heterodimer formation. Similarly, the prolonged increase in plasma glucocorticoids in response to an acute stressor will result in a different pattern of GR and MR binding to DNA. Furthermore, the presence of different transcription factors and kinases, etc., in cells of different tissues will provide scope for multiple cell-specific responses to different digital signals.

Another area that we now need to approach to understand tissue selective transcription responses is the regulation of chromatin accessibility as a rate-limiting step, which will either make genes available or unavailable for GR binding and the subsequent transcriptional response. Work from the laboratory of Gordon Hager certainly suggests that preclusion of GR binding by a closed chromatin confirmation may provide a target for drugs that could change chromatin structure and GR responsivity. Their demonstration that, in different cell types, there are different chromatin profiles in the vicinity of GR binding sites (24) provides an exciting way ahead for our understanding of tissue specific GR responses.