Regulation of natural killer cell redistribution by prefrontal cortex during stochastic learning

Abstract

Acute stress elicits redistribution of lymphocyte subsets, especially natural killer (NK) cells, probably for preparatory defense against potential invasion of antigens in fight–flight situations. We previously reported that regulation of lymphocyte redistribution is based on the evaluation of the controllability of a stressor (Kimura, K., Ohira, H., Isowa, T., Matsunaga, M., Murashima, S. 2007. Regulation of lymphocytes redistribution via autonomic nervous activity during stochastic learning. Brain Behav. Immun. 21, 921–934; Ohira, H., Isowa, T., Nomura, M., Ichikawa, N., Kimura, K., Miyakoshi, M., Iidaka, T., Fukuyama, S., Nakajima, T., Yamada, J. 2008. Imaging brain and immune association accompanying cognitive appraisal of an acute stressor. Neuroimage 39, 500–514). Specifially, lymphocyte redistribution is somewhat attenuated when a stressor is uncontrollable, probably to save biological energy in a situation where appropriate coping is unclear. We infer that this phenomenon might reflect top-down regulation over peripheral immune function by higher-ordered brain regions. To investigate the neural basis of such a phenomenon, we simultaneously recorded regional cerebral blood flow using ¹⁵O-water positron emission tomography and cardiovascular (blood pressure and heart rate), neuroendocrine (epinephrine, norepinephrine, and adrenocorticotropic hormone), and immune (proportions of NK cells and helper T cells in blood) indices in 16 male subjects who performed a stochastic learning task with manipulation of controllability (controllable vs. uncontrollable). Consistent with previous studies, the proportion of peripheral NK cells was attenuated in an uncontrollable stress condition. The dorsolateral prefrontal and orbitofrontal cortices were activated in the uncontrollable situation but not in the controllable condition, and additionally, these prefrontal brain regions significantly correlated with the degree of redistribution of NK cells in the uncontrollable condition. The results of the study suggest these brain regions are involved in both evaluation of the controllability of a stressor and regulation of immune function.

Introduction

To survive, all organisms must adapt to the challenges of dynamically changing environments by forming appropriate behavioral and physiological responses. One of the critical factors for survival in such challenging situations is immunological adaptation, especially rapid and efficient trafficking of specific leukocyte subpopulations into sites of wounding and antigen entry (Moser and Loetscher, 2001, Sprent and Tough, 1994). Psychoneuroimmunology studies have revealed an increase in circulating numbers of the lymphocytes representing innate immunity such as natural killer (NK) cells, whereas the numbers of lymphocytes representing acquired immunity such as helper T cells, do not change or slightly decrease during acute phases of psychological stress (Dhabhar et al., 1995, Bosch et al., 2003, Isowa et al., 2004, Isowa et al., 2006, Kimura et al., 2005, Landmann et al., 1984, Meehan et al., 1993, Minton and Blecha, 1990, Schedlowski et al., 1993, Schedlowski et al., 1996, Stefanski, 2000). Increasing numbers of peripheral NK cells that can non-specifically react with any antigen may be interpreted as a preparatory step in response to a potential invasion by bacteria from injuries accompanying fight/flight behaviors (Engler et al., 2004). This redistribution of NK cells during acute stress is mediated by activation of both the rapidly working sympathetic nervous system, and the relatively slowly working hypothalamic–pituitary–adrenocortical (HPA) axis (Bauer et al., 2001, Bauer et al., 2002, Bosch et al., 2005, Mills et al., 1995, Pike et al., 1997, Stevenson et al., 2001).

Any stable pattern of psychological and physiological responses to acute stress would not be effective. Rather, continuous assessment of environmental demands and dynamic modulation of responses to deal with those demands are critical for adaptation. Psychological models of stress adaptation (Blascovich et al., 1999, Lazarus and Folkman, 1984) have focused on the roles of cognitive appraisal. In particular, in response to a stressful event, the controllability of the event and the individual's coping resources during the event are evaluated. As a result of such appraisal of stressor controllability, subjective emotions and behaviors can be affected. Furthermore, autonomic, endocrine, and immune systems can react differently to a particular stressor according to the appraisal of stressor controllability (Gaab et al., 2003, Maier and Watkins, 2005, Peters et al., 1999, Peters et al., 2003). Specifically, we previously reported that redistribution of NK cells as well as mediating autonomic and endocrine responses to a stressor was attenuated or downregulated when the stressor was uncontrollable (Isowa et al., 2006, Kimura et al., 2007). Attenuation of autonomic and endocrine responses probably reflects a kind of “energy-saving strategy” to prevent the wasting of limited biological resources in a situation where the most adaptive form of coping is unclear. Although redistribution of immune cells including NK cells itself should contribute to the adaptation described above, we recently found that such immune cells, transiently increased by acute stress, secrete inflammatory cytokines within a short time (20 min.) (Yamakawa et al., in press). Because increases in the levels of inflammatory cytokines should lead to higher load for organisms, attenuation of the redistribution of NK cells in an uncertain situation should also be adaptive.

These findings suggest that hypothalamic or midbrain activities determining peripheral physiological systems may be modulated by higher brain cortices to cope with demands from environments. Motivated by such an inference, we examined the neural basis of the modulation of lymphocyte redistribution accompanying appraisal of the controllability of an acute stressor (Ohira et al., 2008). By simultaneous measurement of regional cerebral blood flow (rCBF) using ¹⁵O-water positron emission tomography (PET) and physiological parameters of cardiovascular, neuroendocrine, and immune activities during an acute stressor, we found that regions in the prefrontal cortex (PFC), specifically the orbitofrontal cortex (OFC), medial prefrontal cortex (MPFC) and anterior cingulate cortex (ACC), were involved in the appraisal of stressor controllability. The OFC was especially involved in the downregulation of NK cell redistribution in an uncontrollable stress situation. Several previous findings supported our results. The OFC evaluates contingencies between actions and outcomes in both humans (O'Doherty et al., 2001, O'Doherty et al., 2003) and animals (Roberts, 2006). Also the MPFC and ACC monitor one's own actions and action regulation in humans (Bush et al., 2000, Ridderinkhof et al., 2004). Thus the network including the OFC, MPFC, and ACC can serve as a neural basis for the appraisal of stressor controllability. Furthermore, we speculated that the OFC may affect peripheral immune functions via modulation of autonomic activities, because those brain regions have neural projections to limbic and midbrain areas (Kringelbach, 2005, Kringelbach and Rolls, 2004).

To our knowledge, our study (Ohira et al., 2008) is the first to report an association of the brain and immune activities accompanying the appraisal of stressor controllability. Thus the first goal of the present study was a conceptual replication of our previous findings. In our previous study, we used a mental arithmetic task as an acute stressor and varied controllability of the task using real and bogus feedback about subjects' performance. However, this method was dependent on the subjects' conscious perception, so rigorous manipulation of the controllability in a behavioral sense was difficult. Thus, in the present study, we used a stochastic learning task as an acute stressor. This task requires subjects to learn contingency between their actions and outcomes of success or failure, and we have confirmed that this task can work as an acute stressor by introducing time pressure and monetary reward and punishment (Kimura et al., 2007). The merit of this task is that we can strictly manipulate controllability by changing the contingency between actions and reward or punishment.

The second goal of the present study was to examine potential mediators of the on-line modulation by the PFC of peripheral physiological responses, including NK cell redistribution, according to stressor controllability. Some theorists have argued that the vagus nerve or parasympathetic nervous system may play a critical role in such flexible regulation (Thayer and Brosschot, 2005). Because of differences in the temporal kinetics of neuroeffectors, vagal effects occur faster than sympathetic effects (Saul et al., 1990); thus, the former should be more suitable for fast and delicate regulation. To examine this hypothesis, we measured heart rate variability (HRV) as an index of vagal activity during the stochastic learning task. Furthermore, we also measured blood epinephrine and norepinephrine as indices of sympathetic activity and adrenocorticotropic hormone (ACTH) as an index of the HPA activity during the task. By correlating these vagal, sympathetic and HPA indices with immune indices and brain activity, we examined which mechanisms accounted for more variance in the modulation of lymphocyte redistribution according to stressor controllability.