Multiple Sex- and Circuit-Specific Mechanisms Underlie Exercise-Induced Stress Resistance

Abstract

Prior physical activity reduces the risk of future stress-related mental health disorders including depression, anxiety, and post-traumatic stress disorder. Rodents allowed to engage in voluntary wheel running are similarly protected from behavioral consequences of stress. The present review summarizes current knowledge on mechanisms underlying exercise-induced stress resistance. A conceptual framework involving the development (during exercise) and expression (during stress) of stress resistance from exercise is proposed. During the development of stress resistance, adaptations involving multiple exercise signals and molecular mediators occur within neural circuits orchestrating various components of the stress response, which then respond differently to stress during the expression of stress resistance. Recent data indicate that the development and expression of stress resistance from exercise involve multiple independent mechanisms that depend on sex, stressor severity, and behavioral outcome. Recent insight into the role of the prefrontal cortex in exercise-induced stress resistance illustrates these multiple mechanisms. This knowledge has important implications for the design of future experiments aimed at identifying the mechanisms underlying exercise-induced stress resistance.

1. Introduction

Exposure to stress is a primary contributing factor to the development of mental health disorders such as depression, anxiety, and post-traumatic stress disorder. These stress-related mental health disorders are the most common mental health disorders in the world, accounting for 7% of the global burden of disease and 19% of all years lived with disability (Rehm and Shield 2019). In 2019, depression disorders were the second and anxiety disorders the eighth leading contributor to years lived with disability (Ferrari et al. 2022). Understanding factors that can facilitate the prevention (stress resistance) and recovery from (stress resilience) the effects of stress on mental health is thus of utmost importance.

Stress resistance and resilience factors include experiential factors over which we have some degree of control. One experiential factor that both enables protection against, and facilitates recovery from, stress-related disorders is physical activity. Indeed, sedentary behavior is recognized as a leading risk factor for development of future stress-related mental health disorders, whereas a history of physical activity reduces this risk (Chekroud et al. 2018; Laird et al. 2023; Harvey et al. 2018). Additionally, physical activity, alone or in combination with traditional behavioral or pharmacological treatments, can aid in the recovery from stress-related disorders (Heissel et al. 2023; Hegberg et al. 2019; Bjorkman and Ekblom 2022; Stonerock et al. 2015). A handful of pharmaceutical treatment options with varying degrees of efficacy are employed for the treatment of existing stress-related disorders (Cipriani et al. 2018). However, no drugs are currently available to reduce the risk of development of future stress-related mental health disorders. Therefore, the ability of physical activity to provide protection against the future onset of these disorders (here simply referred to as stress resistance) is of particular importance.

Understanding the features of, and mechanisms underlying, exercise-induced stress resistance could potentially inform strategies to maximize the stress-protective potential of physical activity and reveal circuit and/or molecular targets for novel prophylactic pharmaceuticals (exercise mimetics). With these objectives in mind, this review discusses current knowledge and issues pertaining to the mechanisms by which physical activity enables stress resistance. First, we discuss the understudied aspect of sex differences in exercise-induced stress resistance and the intriguing possibility that females are more responsive than males to stress resistance from exercise. Next, we present a conceptual framework for the study of exercise-induced stress resistance which includes the development (during exercise) and expression (during stress) of stress resistance. Applying this framework to our model system of acute inescapable stress clarifies the involvement of exercise signals, molecular mediators, and circuit adaptations to exercise-induced stress resistance. Finally, recent insights into sex- and circuit-determinants of exercise-induced stress resistance support the notion that exercise recruits multiple, independent, stress resistance mechanisms which act in parallel to enable protection from stress. These topics are presented with an eye toward limitations of current experimental approaches and ideas to guide future research.

2. Sex Differences in Exercise-Induced Stress Resistance

Women are more likely than men to suffer from stress-related mental health disorders (Pavlidi et al. 2023; Steel et al. 2014). Additionally, some stress resistance factors (Fallon et al. 2020) and pharmacological treatment strategies (Bigos et al. 2009) are reported to work better in males compared to females. Given the recent emphasis on prevention (Insel and Scolnick 2006; Simmons et al. 2021), it is particularly relevant that physical activity is a readily available and inexpensive stress prophylactic strategy for both sexes. Understanding potential sex differences in exercise-induced stress resistance could inform the implementation of physical activity programs designed to optimize benefits for a particular sex.

2.1. Limitations of Clinical Studies on Exercise-Induced Stress Resistance

Longitudinal studies including both female and male subjects indicate that exercise can reduce the risk for development of future stress-related disorders in both sexes. For example, Pearce and colleagues recently estimated that achieving the recommended minimum amount of weekly physical activity reduces the risk of developing future depression in both sexes by an impressive 25% (Pearce et al. 2022). However, the duration, frequency, or intensity of physical activity required to enable stress resistance, or the magnitude of the protective effect of exercise against a stressor of a given severity, could differ between the sexes in subtle ways that are difficult to detect in longitudinal studies. Intervention studies that assign groups of people to exercise and stress conditions could, in theory, be designed to detect potential sex differences. Clinical intervention studies, however, are limited to the use of relatively mild stressors, since it is unethical to expose people to the types of stressors that lead to stress-related mental health disorders. Characteristics of resistance to the effects of mild stress could differ from those of severe stress, so it cannot be assumed that results of intervention studies using mild stressors resemble what would be found in response to severe stress. Intervention studies can more readily be designed to detect potential sex differences in therapeutic effects of exercise, by assigning groups of humans with stress-related mental health disorders to exercise groups. However, mechanisms underlying therapeutic exercise are not necessarily the same as those of prophylactic exercise. These are some of the reasons why our understanding of sex differences in exercise-induced stress resistance remains limited.

2.2. Considerations for the Use of Pre-Clinical Models to Study Sex Differences in Exercise-Induced Stress Resistance

Given the limitations of clinical studies, preclinical models could be particularly useful in the study of sex differences in stress resistance. Preclinical models allow groups of rodents to be assigned to sedentary or exercise conditions, followed by exposure to no stress or a laboratory stressor resulting in behaviors resembling symptoms of stress-related mental health disorders. Acute, rather than chronic, stress models are advantageous here, as the exercise intervention can be precisely timed to occur prior to stress administration. It is difficult to differentiate effects of physical activity on stress resistance vs. recovery in studies that allow animals to continue exercising during chronic stress exposure.

There are various ways, both voluntary and involuntary, to increase the physical activity status of rodents. Since removing the control over aspects of one’s environment is aversive (Leotti et al. 2010), chronic stress produced by forced exercise could overshadow potential stress-protective effects. Voluntary physical activity paradigms are thus most appropriate. Voluntary physical activity paradigms are easily employed in the laboratory, as rodents demonstrate robust spontaneous wheel running behavior on wheels placed in their cages (voluntary wheel running; VWR). An additional benefit of preclinical studies is that they allow the measurement and manipulation of factors potentially involved in exercise-induced stress resistance, and so are useful for identification of underlying mechanisms.

An important, and often overlooked, consideration in the design of experiments aimed at determining sex differences in exercise-induced stress resistance is that the experiments need to be designed to statistically compare female and male subjects. Designs that simply include subjects of both sexes and determine effects of exercise in each sex separately are insufficient for determination of sex differences. Designs allowing interactions between sex and exercise to be compared statistically are necessary (Garcia-Sifuentes and Maney 2021). Additional factors such as housing conditions, breeding history, estrous phase, controlling for handling and potential stress effects of estrous phase monitoring, and hormonal status should also be considered (Becker et al. 2005).

2.3. Exercise-Induced Stress Resistance Is Sex Divergent

We have been using an acute, inescapable stress (IS) model, to study sex difference in exercise-induced stress resistance. The IS procedure involves restraining rats in tubes and delivering a series of 100 unpredictable, inescapable electric shocks to the tail. IS produces behaviors in rats that resemble symptoms of stress-related mental health disorders, including anxiety-like social avoidance, exaggerated fear-like potentiated shock-elicited freezing, and depression-like deficits in shuttle box escape (Christianson et al. 2008; Maier and Watkins 2005). Importantly, the behavioral consequences of IS are dependent on the uncontrollability of the stressor and do not occur if shocks of equal duration and intensity can be terminated by the subject (Christianson et al. 2008; Maier and Watkins 2005).

We and others have previously reported that long-term VWR (e.g., 4–6 weeks) prevents anxiety-, fear-, and depression-like behavioral consequences of IS in adult male (Dishman 1997; Dishman et al. 1997; Duman et al. 2008; Greenwood et al. 2003a, 2012a) and female (Tanner et al. 2019) rats. We have also observed a potential sex difference in the duration of VWR required to enable stress resistance; whereby 3 weeks of VWR prevents IS-induced social avoidance and exaggerated fear in females (Fallon et al. 2020), but fails to prevent IS-induced exaggerated fear or the shuttle box escape deficit in males (the effects of 3 weeks of VWR on IS-induced social avoidance in males was not assessed; (Greenwood et al. 2005a)). These time course studies, however, were performed in different laboratories, years apart, and did not allow statistical comparison of the effects of exercise between sexes.

To determine if sex differences in exercise-induced stress resistance exist, we recently allowed adult (age-matched), female and male Sprague-Dawley rats voluntary access to in-cage running wheels for either 3 or 6 weeks prior to exposure to a single session of IS. As expected, 6 weeks of VWR prevented IS-induced social avoidance and exaggerated fear in both sexes. Confirming prior work with females and males separately, we observed a sex difference in exercise-induced stress resistance, whereby 3 weeks of VWR prevented IS-induced exaggerated fear in females, but not in males. Surprisingly, however, 3 weeks of VWR did protect males from IS-induced social avoidance (Tanner et al. 2023).

This study reveals that long-term physical activity may be equally effective at enabling stress resistance in females and males, but the duration of physical activity required to enable stress resistance differs between the sexes in a stress outcome-dependent manner. A similar duration of physical activity seems to enable protection against anxiety-like consequences of stress between sexes, whereas exaggerated fear-like effects of stress seem to be more readily prevented by physical activity in females compared to males. Physical activity also seems to reduce depression-like behavior more readily in female mice than males. Both VWR (Elias et al. 2023) and treadmill training (Munive et al. 2016; Naghibi et al. 2021) have been reported to reduce immobility time in the tail suspension test in female, but not male, mice.

If these preclinical data apply to humans, they suggest that women at risk for mental health disorders should prioritize physical activity as a prophylactic strategy. Unfortunately, this sex difference will be difficult to confirm in humans, given the difficulty in performing clinical intervention studies on stress resistance. One longitudinal study following more than 100,000 adults over 3 years did find a sex difference like that observed in rodents. Kim et al. (2019) found that transitioning from a sedentary lifestyle to a physically active one reduced the incidence of new depression within a year in women, but not in men (Kim et al. 2019).

2.4. Sex Differences in Exercise-Induced Stress Resistance Resemble Sex Differences in Other Exercise Effects

The observation that females are more responsive than males to exercise-induced stress resistance does not seem to be an anomaly among sex differences in response to exercise. Physical activity can improve cognitive function more readily in female than in male rats and humans (Wang et al. 2015; Colcombe and Kramer 2003; Barha et al. 2017a; Barha and Liu-Ambrose 2018) and can reduce the risk for dementia to a greater extent in women than in men (Hogervorst et al. 2012; Laurin et al. 2001). A similar sex difference is found in studies examining the impact of physical activity on drug-seeking behaviors (Cosgrove et al. 2002; Lespine and Tirelli 2018; Smethells et al. 2020; Gallego et al. 2015), hypoalgesia (Rice et al. 2019), and mood (McDowell et al. 2016). Although conflicting reports exist (Peterson et al. 2014; Szuhany et al. 2015), growing evidence indicates that females are generally more responsive to the behavioral effects of physical activity than males. However, these effects may partly depend on exercise type (Barha et al. 2017a, b).

2.5. Potential Mechanisms Underlying Sex Differences in Exercise-Induced Stress Resistance

Relatively little work has explored whether mechanisms underlying exercise-induced stress resistance are sex divergent. Observed sex differences in the effects of physical activity on hippocampal microglial activation (Kohman et al. 2013), hippocampal neurotrophic factors (Naghibi et al. 2021; Uysal et al. 2015), and hippocampal volume (Varma et al. 2015) could all contribute to sex differences in exercise-induced stress resistance. Two factors that stand out as potentially driving sex differences in exercise effects, including stress resistance, are sex differences in exercise behavior and ovarian hormones.

Female rats run greater distances and faster speeds than males (Basso and Morrell 2017; Eikelboom and Mills 1988). Additionally, ovarian hormones during early life could alter the development of the female brain rendering it more susceptible to exercise effects (organizational effects). Alternatively, ovarian hormones during adulthood could interact with exercise, altering the way in which exercise impacts the brain (activational effects).

Sex differences in VWR behavior are dependent on activational effects of ovarian hormones (Gentry and Wade 1976; Slonaker 1924), particularly estrogen (Gorzek et al. 2007), thus the involvement of sex differences in VWR behavior and activational effects of ovarian hormones in mediating sex differences in exercise-induced stress resistance can be simultaneously tested by removing ovarian hormones. We removed ovaries from adult, female rats, and observed that female VWR behavior now resembled that of males (Tanner et al. 2023). Interestingly, despite the elimination of the female pattern of VWR and the absence of activational effects of ovarian hormones during exercise, 3 weeks of VWR still produced rapid protective effects against both IS-induced social avoidance and exaggerated fear (Tanner et al. 2023).

Whether ovarian hormones are similarly unnecessary for other sex differences in response to exercise is unknown. It does seem that ovarian hormones are not required for exercise to impact brain and behavior in females, in general. For example, VWR rapidly upregulates BDNF gene expression (Berchtold et al. 2001) and cell proliferation (Jin et al. 2008) in the hippocampus of ovariectomized female rats, and both VWR and treadmill training have been reported to protect ovariectomized rats from impairments in cognition produced by sleep deprivation (Saadati et al. 2015; Rajizadeh et al. 2020). Future work will be needed to determine mechanisms underlying sex differences in exercise-induced stress resistance. Since very little research on sex differences in underlying mechanisms has been completed, the following section on mechanisms focuses on work done in males.

3. A Conceptual Framework for Mechanisms Underlying Exercise-Induced Stress Resistance Includes the Development and Expression of Stress Resistance

A conceptual framework to aid in consideration of the variety of mechanisms underlying exercise-induced stress resistance is provided in Fig. 1. For physical activity to alter the brain and behavioral response to stress, physical activity must produce experience-dependent plasticity prior to stressor exposure which then allows the physically active organism to respond to the stressor differently than it would if it were sedentary. Mechanisms of exercise-induced stress resistance can, therefore, be thought of in terms of those mechanisms underlying the development of exercise-induced stress resistance (i.e., during exercise prior to stress) and the expression of exercise-induced stress resistance (i.e., during stress after a period of physical activity).

Fig. 1.

A conceptual framework for exercise-induced stress resistance involves the development and expression of stress resistance. During the development of stress resistance (left, blue box), neural circuits governing physical activity motivation, planning, execution, and reinforcement determine the extent to which the organism engages in physical activity. Repeated physical activity triggers exercise signals, either derived from peripheral organs and/or motor/motivation circuits themselves, which communicate the experience of physical activity to brain stress-responsive circuits responsible for stress-induced behavioral outcomes. Exercise signals converge on molecular mediators which produce experience-dependent plasticity within these circuits. During the expression of stress resistance (right, red box), stress circuits altered by prior physical activity now respond to stressful events in a manner that is protective against stress-induced behavioral outcomes

4. The Development of Stress Resistance from Exercise Involves Molecular Mediators, Neural Circuits, and Exercise Signals

As we saw with the protective effects of VWR against the behavioral consequences of IS, stress resistance from exercise develops over time with repeated exercise. The development of stress resistance from exercise refers to exercise-induced physiological adaptations occurring during exercise which mediate the later expression of stress resistance. The molecular mediators of these adaptations, the circuits in which these adaptations occur, and the signals through which exercise initiates these adaptations are all critical to the development of stress resistance (Fig. 1).

Physical activity status of the organism modulates many aspects of physiology, including central gene expression (Loughridge et al. 2013; Tong et al. 2001), alterations in neurotransmitter systems (Dishman 1997; Greenwood 2019; Greenwood and Fleshner 2011; Petzinger et al. 2015), synaptic plasticity (Cotman et al. 2007; de Sousa Fernandes et al. 2020), proliferation and survival of adult-born neurons (Vivar et al. 2013; van Praag 2008), and brain structure (Wilckens et al. 2021; Chen et al. 2020). Despite these many exercise adaptations, identifying which, if any, of these are causally related to stress resistance is challenging. The types of experiments required to determine causation are especially difficult to perform with long-term exercise studies. Loss-of-function experiments require candidate mechanistic factors to be altered in physically active organisms for the entire duration of the exercise period (during the development of stress resistance), but not during exposure to stress (during the expression of stress resistance). Similarly, gain-of-function experiments require factors to be increased for a long period of time prior to, but not during, stress. Compounding the issue is the possibility that physical activity utilizes several different mechanisms acting in parallel to enable stress resistance. For example, several molecular mediators could drive the experience-dependent neural plasticity required for stress resistance. Likely candidates include brain-derived neurotrophic factor (Duman et al. 2008; Cotman et al. 2007), mammalian target of rapamycin (Lloyd et al. 2017; Moya et al. 2020), and delta-FosB (Mul et al. 2018; Werme et al. 2002). Additionally, several exercise signals derived from peripheral organs (e.g., muscle, fat, liver, gut) or the brain itself (e.g., from circuits directly governing voluntary exercise) could be responsible for communicating the experience of exercise to these molecular mediators. Irisin (Islam et al. 2021), IL-6 (Severinsen and Pedersen 2020), and lactate (Xue et al. 2022), for example, have been shown to be important for periphery-to-brain communication during exercise.

Since physical activity most obviously involves the movement of skeletal muscle, it is not surprising that research efforts aimed at identifying factors responsible for communicating the experience of exercise to the brain circuits underlying behavioral effects of exercise (referred to here as exercise signals) have been focused on signals released from muscle or other peripheral organs. We must keep in mind, however, that these signals could also come from within the brain itself. Potential sources of central exercise signals include neural circuits involved in the motivation for physical activity, and/or the planning, performance, and/or reinforcing effects of physical activity. Our understanding of these circuits has been growing in recent years (Tanner et al. 2022; Dohnalova et al. 2022). Interestingly, there is evidence that circuits governing voluntary exercise vary by sex (Tanner et al. 2022). It is therefore plausible that sex differences in central signals could drive sex differences in the development of exercise-induced stress resistance.

5. Multiple Mechanisms Underly the Expression of Exercise-Induced Stress Resistance

The expression of stress resistance from exercise occurs during the discrete timepoint of stressor exposure. The changes in the brain established during the development of stress resistance enable stress-reactive neural circuitry orchestrating neurochemical, behavioral, and physiological stress responses to react differently during stress than they would if the organism were sedentary (Fig. 1). Results of recent experiments reveal a common take-home message: there are multiple mechanisms mediating the expression of stress resistance from exercise. These mechanisms can vary depending on sex, stressor severity, duration of prior exercise, and behavioral outcome. The multiple stress resistance mechanisms also seem to act in parallel to enable the expression of stress resistance from exercise. This is perhaps not surprising, given the wide variety of physiological and neural systems impacted by exercise.

5.1. Non-specific Effects of Exercise

In addition to changes in specific stress-responsive systems, non-specific effects of exercise, such as altered pain sensitivity or locomotor activity, could contribute to the expression of stress resistance from exercise. VWR can reduce pain sensitivity, raising the possibility that VWR could reduce sensitivity to stressors such as electric shock. VWR, however, does not alter sensitivity to shock in mice (Duman et al. 2008) or rats (Tanner et al. 2023), nor does it globally blunt stress responses to electric shock stressors, as would be expected if VWR reduced shock sensitivity. For example, VWR has no effect on IS-induced increases in corticosterone (Tanner et al. 2023; Fleshner 2000; Speaker et al. 2014), plasma cytokines (Speaker et al. 2014), or cFos in several stress-responsive brain regions (Greenwood et al. 2003b, 2005a). Regarding altered pain sensitivity contributing to protection against the behavioral effects of IS specifically, IS itself produces massive analgesia (Maier et al. 1982) that begins early during IS treatment (Maier and Keith 1987). Importantly, the behavioral sequelae of IS occur whether or not these antinociceptive effects of IS are present, indicating that IS outcomes are independent of pain sensitivity (MacLennan et al. 1982; Hemingway and Reigle 1987).

Animal models of stress-related mental health disorders often depend on subjects engaging in locomotor activity, whereby general increases in movement could be mistaken for anxiolytic or antidepressant effects. This raises the possibility that potential effects of VWR on locomotor activity could contribute to the observed behavioral effects. IS reduces locomotor activity during exposure to a novel environment for at least 48 h following IS, but VWR has no impact on this stress-induced reduction in locomotor activity in either sex (Tanner et al. 2023). This observation suggests that the stress-protective effect of VWR against IS outcomes is not attributable to alterations in locomotor activity. Future research utilizing different animal models should consider how these factors and others (such as increased fitness) contribute to the behavioral effects of exercise in that model.

5.2. The Role of the Hypothalamic-Pituitary-Adrenal Axis

Physical activity blunts the response of the hypothalamic-pituitary-adrenal (HPA) axis to stress (see (Nowacka-Chmielewska et al. 2022) for a recent review). Given the link between HPA activity and stress-related mental health disorders (Sapolsky 2000; McEwen et al. 2015), it is easy to conceptualize how constraint over stress-induced activity of the HPA axis could contribute to stress resistance from exercise. However, severely stressful events are more likely to precipitate the formation of stress-related disorders than mild stressors, and the impact of VWR on the HPA axis seems to be limited to mild-moderate stressors (Campeau et al. 2010). VWR, for example, fails to attenuate the HPA response to IS in either sex (Tanner et al. 2023; Fleshner 2000; Speaker et al. 2014). Attenuation of HPA responses could, therefore, contribute to exercise-induced resistance to the effects of relatively mild stressors, but protective effects of exercise against behavioral consequences of more severe stressors likely involve mechanisms other than blunting of the HPA axis. These data demonstrate that mechanisms underlying the expression of stress resistance from exercise can vary depending on stressor severity.

5.3. Constraint over Stress-Induced Activity of Dorsal Raphe Nucleus Serotonergic Neurons Is One Mechanism Underlying the Expression of Stress Resistance from Exercise

Since VWR prevents the anxiety- and depression-like consequences of acute exposure to severe, inescapable stressors, such as IS, it is of particular interest to consider the mechanisms involved in this effect. One of the benefits of using the IS model is that the neural circuits mediating the behavioral consequences of IS have been well-defined. Briefly, brain regions involved in processing specific aspects of IS, including the locus coeruleus (LC), lateral habenula, and bed nucleus of the stria terminalis (BNST), have converging excitatory inputs to the dorsal raphe nucleus (DRN). Activity of these inputs during severe stressors drives the activity of DRN serotonergic (5-HT) neurons, measured both by cFos (Grahn et al. 1999) and extracellular 5-HT released into the DRN from DRN axon collaterals (Maswood et al. 1998). The excessive extracellular 5-HT in the DRN during IS desensitizes inhibitory somatodendritic 5-HT_1A autoreceptors, leaving DRN 5-HT neurons in a sensitized state (Rozeske et al. 2011). Sensitized DRN 5-HT neurons now respond to future mild stressors with excessive 5-HT release in DRN projection sites. Among these DRN projection sites are the basolateral amygdala (Amat et al. 1998), where sensitized 5-HT release produced by prior IS leads to anxiety-like (Christianson et al. 2010) and exaggerated fear-like (potentiated shock-elicited freezing) behavior (Strong et al. 2011), and the dorsal striatum, where sensitized 5-HT release produces depression-like behavior, such as the shuttle box escape deficit (Strong et al. 2011). In addition to being necessary and sufficient for the behavioral consequences of IS (Maier and Watkins 2005), increased 5-HT activity in the DRN is also involved in the anxiety- and depression-like consequences of social defeat (Hammack et al. 2012; Amat et al. 2010). Moreover, manipulations that prevent the behavioral outcomes of IS, including behavioral control over the stressor (Amat et al. 2006), prophylactic ketamine (Amat et al. 2016), dominance rank (Coleman et al. 2023), and 6 weeks of VWR (Greenwood and Fleshner 2011), all prevent IS-induced DRN hyperactivity and sensitization.

In adult, male rats, 6 weeks of VWR attenuates IS-induced expression of the neural activation marker cFos within DRN 5-HT neurons (Greenwood et al. 2003a), indicating that 6 weeks of VWR prevents the excessive activation of DRN 5-HT neurons that occurs during IS. Six weeks of VWR also prevents IS-induced sensitization of extracellular 5-HT in the DS in response to a few foot shocks, assessed with microdialysis (Clark et al. 2015). These data suggest that the expression of stress resistance from 6 weeks of VWR involves the prevention of stress-induced hyperactivity and subsequent sensitization of DRN 5-HT neurons, at least in males. An increase in central 5-HT is thought to contribute to central fatigue (Davis and Bailey 1997). Constraint over DRN 5-HT activity could thus be a training adaptation that evolved to help delay fatigue during exercise. Whether DRN constraint is involved in the expression of stress resistance from exercise in females is a critical unanswered question.

Understanding how exercise comes to constrain the DRN 5-HT response to severe stressors could reveal novel insights into mechanisms underlying the development and expression of exercise-induced stress resistance. Two possibilities seem most plausible. First, exercise could alter inputs to the DRN that regulate 5-HT activity during stress. Second, exercise could produce adaptations within DRN 5-HT neurons themselves which alter their response to stress. There is evidence that exercise does both, as outlined below. Additionally, exercise modifies the response of DRN efferent circuits mediating specific behavioral consequences of IS to 5-HT and/or stress. Multiple independent mechanisms involving the 5-HT system thus seem to contribute to the expression of stress resistance from exercise (Fig. 2).

Fig. 2.

Circuit-specific determinants of exercise-induced stress resistance. Repeated physical activity initiates the development of the stress-resistant brain by acting on at least 3 unique neural substrates, each of which could be sufficient for the expression of stress resistance: (1) Repeated physical activity potentiates activity of inhibitory inputs to the dorsal raphe nucleus (DRN; dotted line) and potentially reduces activity of excitatory DRN inputs (solid lines) during stress. (2) Repeated physical activity alters factors intrinsic to DRN serotonergic (5-HT) neurons that influence their response to stress. (3) Repeated physical activity produces adaptations within DRN efferent circuits that are the proximal mediators of behavioral-specific stress outcomes. These stress-protective mechanisms develop at different rates during exercise, depending on sex and circuit-specific behavioral outcome

5.4. The Role of Inputs to the Dorsal Raphe Nucleus

One means by which exercise could attenuate the DRN 5-HT response to stress is by modifying the activity during stress of brain structures with modulatory control over DRN 5-HT neurons. VWR could, for example, attenuate excitatory drive or increase inhibitory drive to DRN 5-HT neurons from afferent structures during IS. Numerous DRN afferent structures are sensitive to physical activity and are thus prime candidates for exercise-induced modulation of DRN activity, as discussed previously (Greenwood 2019; Greenwood and Fleshner 2008, 2011; Nicastro and Greenwood 2016). Exercise produces a hyperdopaminergic state within midbrain-striatal pathways which we have argued simultaneously contributes to both a more positive outlook on stress and DRN constraint in physically active compared to sedentary organisms (Greenwood 2019). The medial prefrontal cortex (PFC) also stands out due to its established role in stress resistance and resilience (Maier et al. 2006). Based on our prior data, it was thought that the PFC was not involved in stress resistance from exercise (Christianson and Greenwood 2014; Greenwood et al. 2013). However, here we present new data which provides an updated view on the role of the PFC in exercise-induced stress resistance.

The PFC has an inhibitory influence over DRN 5-HT activity, mediated through a glutamatergic projection from the medial PFC to the DRN, where PFC glutamatergic terminals synapse on GABAergic neurons which inhibit 5-HT neural activity (Amat et al. 2005). Physical activity induces neuroplasticity in the PFC and enhances PFC function in both humans (Erickson and Kramer 2009; Soshi et al. 2021) and animals (Dong et al. 2018; Brockett et al. 2015). Given the established sensitivity of the PFC to exercise and the ability of the PFC to inhibit the DRN during IS, we sought to determine whether the PFC is necessary for the expression of exercise-induced stress resistance. After 3 weeks of wheel access to establish habitual VWR, male rats received sham surgery or bilateral lesions of the medial PFC, where DRN-projecting neurons are located. Rats ran an additional 3 weeks following surgery, for a total of 6 weeks of VWR, prior to exposure to no stress or IS. The idea was that if the PFC-to-DRN pathway inhibits the DRN during IS in VWR rats, then removing this inhibitory influence over DRN activity would prevent the expression of stress resistance and would restore the behavioral consequences of IS in previously physically active rats. To our surprise, PFC lesions had absolutely no effect on exercise-induced stress resistance. Six weeks of VWR still protected male rats from the behavioral consequences of IS despite PFC lesions (Greenwood et al. 2013).

These data indicate that the PFC is not necessary for the expression of exercise-induced stress resistance. But does this negative experimental result mean that physical activity does not utilize the PFC or the PFC-to-DRN pathway for stress resistance at all? A limitation of lesion and inactivation studies is that if negative results are found, it cannot be assumed that the inhibited region or circuit is not utilized for stress resistance in an intact animal. Indeed, if multiple independent mechanisms act in parallel to enable stress resistance from exercise, then inhibiting one should have no impact on another and stress resistance should remain intact. If multiple neural circuits contribute to DRN constraint during stress in previously physically active animals, the only way to know whether a certain circuit contributes to DRN constraint is to measure its activity during stress.

We designed an experiment to determine whether prior exercise recruits the PFC-to-DRN circuit during stress (Fig. 3a). Red fluorescent retro-beads were injected into the DRN of male rats (Fig. 3c). These beads reveal DRN-projecting PFC neurons by traveling retrograde from axon terminals in the DRN to cell bodies of origin in DRN afferent structures. After 6 weeks of sedentary or VWR conditions, rats were exposed to IS and cFos was quantified in DRN-projecting PFC neurons with immunohistochemistry (Fig. 3b; n = 6/group). We found that 6 weeks of VWR increased the percentage of DRN-projecting PFC neurons expressing cFos following IS (unpaired t-test p < 0.0001; Fig. 3d).

Fig. 3.

Prior physical activity recruits a prefrontal cortex (PFC)-to-dorsal raphe nucleus (DRN) pathway during inescapable stress. (a) Experimental timeline. Rats received red retro-beads in the DRN prior to 6 weeks of sedentary (Sed) or voluntary wheel running (VWR) conditions. All rats were exposed to inescapable tail shock stress and then euthanized 90 min later. PFC was processed for cFos immunohistochemistry (IHC) and the % of DRN-projecting PFC neurons containing cFos was quantified. (b) cFos (green) expression within DRN-projecting PFC neurons (red). Arrows point to DRN-projecting neurons expressing cFos. Distance noted is relative to Bregma. Scale bar = 200 μm. (c) Injection site of red retro-beads in the DRN. Distance noted is relative to Bregma. Scale bar = 200 μm. (d) Relative to Sed rats, prior VWR increased cFos expression within DRN-projecting PFC neurons during inescapable stress. ***p < 0.0001. (e) Cartoon illustrating that VWR rats with intact PFC utilize the PFC-to-DRN pathway during stress. (f) In PFC-lesioned rats, VWR rats no longer utilize the PFC-to-DRN pathway but are still protected from stress through an alternative mechanism. This illustrates that multiple mechanisms are involved in the expression of exercise-induced stress resistance. Figure created in BioRender. Brain images adapted from the rat brain atlas of Paxinos and Watson (Paxinos and Watson 1998)

These data provide an updated view on the role of the PFC in exercise-induced stress resistance. VWR does indeed recruit the PFC-to-DRN inhibitory circuit during stress, which likely contributes to DRN inhibition (Fig. 3e). However, since VWR still enables stress resistance even in the absence of the PFC, this activity is not necessary for stress resistance from exercise and exercise must utilize an additional mechanism that can enable stress resistance in the absence of the PFC (Fig. 3f).

5.5. The Role of Exercise Adaptations Within Dorsal Raphe Nucleus 5-HT Neurons

Another possible way in which prior physical activity could attenuate activity of DRN 5-HT neurons during stress is by impacting DRN 5-HT neurons directly (during the development of stress resistance), altering factors intrinsic to 5-HT neurons which modulate their response to stress (during the expression of stress resistance; Fig. 2). Reported adaptations within DRN 5-HT neurons in response to VWR, such as increased 5-HT content in the DRN and DRN projection sites (Dishman 1997; Dishman et al. 1997) and altered gene expression within DRN 5-HT neurons (Loughridge et al. 2013), are consistent with this possibility. For example, VWR increases mRNA coding for the 5-HT_1A inhibitory autoreceptor in the DRN in male rats (Greenwood et al. 2003a). An increase in 5-HT_1A autoinhibition could help constrain the DRN 5-HT response to stress. This mechanism may only contribute to stress resistance after long-term VWR in males, since 6 weeks, but not 3 weeks, of VWR increases 5-HT_1A mRNA (Greenwood et al. 2005b). Other DRN adaptations occur more rapidly (Greenwood et al. 2005b). Future work should determine if similar intrinsic changes in DRN 5-HT neurons occur in females and clarify whether these adaptations contribute to DRN constraint during stress in both sexes.

5.6. The Role of Exercise Adaptations Within Dorsal Raphe Nucleus Efferent Circuits

Constraint over stress-induced activity of DRN 5-HT neurons, either via modulation of the activity of DRN inputs or adaptations within DRN 5-HT neurons, is one potential mechanism underlying the expression of stress resistance from long-term exercise. However, it doesn’t appear to be the only one. In male rats, 3 weeks of VWR does not attenuate the activity of DRN 5-HT neurons during IS, as 6 weeks does (Greenwood et al. 2005a). How then, does 3 weeks of VWR prevent IS-induced social avoidance in males?

IS-induced social avoidance depends on IS-induced sensitization of 5-HT release in the amygdala (Christianson et al. 2010). Since 3 weeks of VWR fails to constrain the DRN response to IS, 3 weeks of VWR could instead be preventing IS-induced social avoidance by acting on DRN efferent circuits that mediate the anxiety-like consequences of sensitized 5-HT release. The BNST is implicated in anxiety (Walker et al. 2003), and stress-induced social avoidance involves a circuit from the amygdala to the BNST (Vantrease et al. 2022). Three weeks of VWR does not alter cFos in response to IS in the amygdala (Greenwood et al. 2005a), but it does alter gene expression (Fox et al. 2023) and the cFos response to IS (Greenwood et al. 2005a) in the BNST. Anxiogenic effects in a fear-potentiated startle paradigm produced by 2 weeks of VWR were recently shown to be mediated by a reduction in 5-HT_2C receptor expression in the BNST (Fox et al. 2023).

Together, these data are consistent with the possibility that rapid exercise-induced adaptations in the BNST could lead to protection from social avoidance and other anxiety-like behaviors, despite a failure of 3 weeks of VWR to prevent IS-induced sensitization of the DRN. Of course, longer-duration physical activity could also produce adaptations within DRN efferent circuits that underly fear- and depression-like effects of stress. Indeed, 6 weeks of VWR reduces the expression and sensitivity of 5-HT_2C receptors in the amygdala and dorsal striatum (Greenwood et al. 2012b). Since the behavioral consequences of IS are dependent on 5-HT_2C receptor signaling in these regions (Christianson et al. 2010; Strong et al. 2009, 2011), the expression of stress resistance induced by a long duration of prior physical activity could involve both DRN constraint and resistance to the post-synaptic effects of 5-HT in DRN projection sites mediating specific IS-induced behaviors. This prior work was done in males. Whether physical activity produces neural adaptations in DRN projection sites in females is currently unknown and should be addressed in future studies.

6. Conclusions

Physical activity enables robust protection against behavioral effects of stress, including stress-related mental health disorders in humans and anxiety-, fear-, and depression-like behaviors in animals. Mechanisms underlying stress resistance from exercise can be thought of in terms of those driving the development of stress resistance (during exercise) and those driving its expression (during stress). The development of stress resistance involves multiple exercise signals and molecular mediators and varies depending on sex and behavioral outcome. The expression of stress resistance from exercise similarly involves multiple independent mechanisms acting in parallel, as is demonstrated by the role of the PFC in exercise-induced stress resistance. This knowledge has important implications for the design and interpretation of causal experiments aimed at identifying underlying mechanisms. The inescapable stress model in rodents reveals a key role of DRN constraint in mediating exercise-induced stress resistance, but the specific involvement of exercise adaptations within DRN inputs, within the DRN itself, and within DRN projection sites require further clarification. Much of our knowledge of the mechanisms underlying exercise-induced stress resistance comes from work using male subjects. The recent realization that female rodents are more responsive than males to exercise-induced stress resistance highlights the need to expand our mechanistic knowledge in females. Whether this sex difference extends to humans is a critical unanswered question, but regardless of the answer, identifying the mechanisms driving this sex difference could reveal novel targets that could be harnessed by both sexes to bolster the stress-protective impact of exercise.