participant exposed to saline became ill, and about a third of the participants exposed to the cold viruses became ill. Results revealed that three measures of stress were related to disease onset: (1) a stressful life event scale to measure the cumulative event load, (2) a perceived-stress scale to assess perceptions of overload-induced stress, (3) and a measure of negative affect. For all three measures, participants who reported high stress were more likely to develop an infection than those who reported low stress (Cohen et al., 1991). Subsequent research has determined that risk of infection was increased most by stressors lasting over a month, especially those involving disruptions in personal ties (i.e., social conflicts, unemployment, and underemployment; Cohen et al., 1998; Glaser et al., 1999).

Social influences on infectious disease processes have also been linked with gene expression. In an experimental model of pyschosocial stress, the stress of social disruption caused by reorganizing established murine hierarchies during a respiratory viral infection significantly increased mortality compared to the stress of physical restraint in home cage control animals (Padgett and Sheridan, 1999). The increased severity of the infection leading to mortality was associated with hyper-inflammatory responses due to overexpression of key cytokine genes. Enhanced gene expression led to increased cell trafficking and accumulation in the lungs of infected animals, which in turn led to tissue consolidation and compromised lung function. These results contrasted with animals undergoing the stress of physical restraint, which is devoid of social interactions, but nonetheless activates the HPA axis in a similar fashion to the psychosocial stressor. However, hypo-inflammatory responses during respiratory viral infection were observed in these animals, due to suppression of cytokine responses by glucocorticoid hormones. Under these conditions, no increase in mortality was observed. Such findings underscore the unique significance of social interactional stressors in activating overexpression of cytokine genes implicated in increased mortality risks.

The specific nature of the personal ties, particularly their hierarchical features, also plays a major role in the individual response to stress and susceptibility to infectious disease. For example, in the social disruption paradigm described above, dominant male mice, when latently infected in the trigeminal ganglia with herpes simplex virus (HSV, a model for recurrent herpes labialis in humans), were twice as likely as subordinate animals to reactivate and shed infectious virus when their social environment was disrupted by reorganization (Padgett et al., 1998). The psychosocial nature of the stressor was also important in this model, simply stressing latently infected mice by restraint did not cause reactivation. Again, it is social interactional stress that activates physiological signals modulating the expression of individual host/pathogen genes. In the model of latent HSV infection, the inactive viral genome represents an environmental (or for-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement