Epidemiological studies have correlated a variety of stressors with an increased risk of fatal secondary bacterial respiratory infections in humans  and respiratory disease in animals [11, 12]. Contradictory evidence has been reported regarding the contribution of transport stress to undifferentiated BRD in feedlot calves [14, 15] but weaning was significantly correlated with an increased incidence of undifferentiated BRD . The impact of weaning on mortality was minimal in this study, however, due to concurrent vaccination and treatment with antibiotics. The present investigation determined that stress from weaning and maternal separation (WMS) altered fatal viral-bacterial synergy in naïve calves when weaning was initiated at the time of a primary BHV-1 infection. This coincidence of WMS stress and a primary BHV-1 infection in naïve calves significantly increased mortality following a secondary M. haemolytica respiratory infection in two independent trials with only 10-20% of calves surviving the secondary bacterial challenge (Figure 1b and 1c). Furthermore, the effect of WMS on the fatal viral-bacterial synergy was of limited duration with only 10% mortality when the secondary bacterial infection was initiated 12 days after virus infection and weaning (Figure 1c). A previous investigation reported that calves weaned immediately before transport to a feedlot developed significantly more undifferentiated BRD when compared to calves adapted to weaning for 45 days . The current observations (Figure 1c) indicate, however, that the effect of WMS stress on viral-bacterial synergy is of short duration. Therefore, management interventions which either mitigate stress responses, minimize exposure to respiratory pathogens, or enhance specific immunity to respiratory pathogens may be of greatest value within the first four days after weaning.
The rapid onset of death with reduced lung pathology (Figure 1) suggested that both systemic and local responses in the lung may have contributed to increased mortality in WMS calves. Previous analysis of stress responses in mice demonstrated that individual stressors may either enhance [5, 6] or inhibit [6, 7] local responses in the lung following a viral respiratory infection but the effect of stress on systemic responses was not analyzed. We used molecular and cellular analyses to determine if WMS had significant effects on innate immune responses to viral infection and observed a significant increase in both CD14 expression (Table 2) and LPS responsiveness during viral infection and prior to secondary bacterial challenge (Figure 4c). Therefore, WMS stress significantly increased innate immune responses during BHV-1 infection (Figure 3a). These increased innate immune responses are directly linked to the recognition of M. haemolytica infection through lipopolysaccharide (LPS) [42, 43]. There is increasing evidence that respiratory viral infections can modulate expression of receptors involved in the recognition of LPS. TLR4 expression was increased on blood monocytes following human respiratory syncytial virus (RSV) infection of young children  and Porcine Reproductive-Respiratory Syndrome (PRRS) virus induced increased expression of both CD14 and LPS binding protein in the lung . In agreement with these observations, our transcriptional analyses revealed that BHV-1 infection of WMS calves increased PBMC expression of both CD14 and TLR4 and increased expression of these receptors correlated significantly (p < 0.01) with fatal secondary bacterial infections (Table 3). Thus, modulating expression of TLRs and associated adaptor molecules by primary viral infections may be a general mechanism of viral-bacterial synergy. In support of this conclusion, we also observed increased expression of TLR2, the receptor for peptidoglycans, following BHV-1 infection (data not shown). Thus, primary viral respiratory infections could potentially enhance pro-inflammatory responses to both Gram-negative and Gram-positive bacterial infections. It is also interesting to note that increased expression of TLR4 by itself in PA calves was not sufficient to significantly increase LPS-induced TNF secretion (Figure 4c). Therefore, increased responsiveness to LPS may depend on an up-regulation of all components of the TLR4 signaling complex.
Identifying the mechanism(s) by which specific stressors, such as WMS, enhance TLR expression and function during viral infections may increase our understanding of how individual stressors increases mortality following a secondary bacterial respiratory infection. BHV-1 is a potent inducer of IFN-α and IFN-γ secretion in the upper respiratory tract [33, 46] and both cytokines have been identified as mediators of LPS-sensitization following viral infection [41, 47]. Although earlier studies did not identify the mechanism by which IFN induce LPS-sensitization, a recent investigation revealed that IFN-γ increases expression of both TLR4 and CD14 and enhances LPS-induced responses of human macrophages . Thus, significant correlations between mortality and IFN-γ secretion, mortality and TLR4 expression, and mortality and CD14 expression is consistent with a causal relationship between weaning enhanced anti-viral responses and increased viral-bacterial synergy. A previous analysis of blood leukocyte populations during BHV-1 infection revealed that the number of blood monocytes remain relatively constant during infection . Thus, increased expression of TLR4 and CD14 in PBMCs cannot be explained by a simple increase in monocyte frequency following viral infection. A possible link between IFN secretion at the site of viral infection and altered gene expression in PBMC is also supported by increased 2'5' OAS gene expression in PBMC following viral infection (Table 1). However, co-production of both IFN-α and IFN-γ following BHV-1 infected calves makes it difficult to state whether one or both of these IFNs contributed to altered TLR4 and CD14 expression and function.
If IFN contributes to the enhanced mortality observed with stress then the question arises as to how stress enhances IFN-γ production in WMS calves (Figure 2d). Glucocorticoid responses to stress can inhibit cytokine production  and may have an effect on BHV-1 replication . There were, however, no significant differences in either cortisol level (Figure 1a) or virus shedding (Figure 2a) when comparing WMS and PA calves. Therefore, there was no apparent connection between cortisol production and altered viral-bacterial synergy in WMS calves. Furthermore, gene expression analysis revealed IFN-γ expression did not change in PBMC of either experimental groups following viral infection (data not shown). This observation may be consistent with a previous report that BHV-1 infection did not activate NK cell activity in blood, but resulted in the rapid recruitment of active NK cells to the site of infection . Enhanced recruitment and activation of NK cells in the respiratory tract of WMS calves may be one mechanism by which stress could increase IFN-γ secretion. Specific psychological stressors, such as restraint, can inhibit leukocyte migration to the murine lung , but social re-organization was shown to increase leukocyte migration into the lung and immunopathology . Thus, it is conceivable that the combination of stressors represented by weaning and maternal separation enhanced leukocyte recruitment to and IFN production at the site of viral infection. Increased IFN at the site of BHV-1 replication would also be consistent with the similar levels of BHV-1 replication in both experimental groups (Figure 2a) since BHV-1 is highly resistant to the antiviral effects of IFN .
It was hypothesized previously that pro-inflammatory responses induced by a primary BHV-1 infection contribute to the viral-bacterial synergy of a fatal M. haemolytica infection . Elevated body temperature and serum haptoglobin levels in the WMS group (Figure 2) may be explained by increased IFN production since expression of pro-inflammatory cytokines, such as IL-1 and TNF-α, were not significantly different in WMS versus PA calves (Table 2) or calves that died versus survived following secondary bacterial infection (Table 3). Viral infection did, however, induce significantly increased IL-10 expression levels in the PBMC of animals with fatal viral-bacterial synergy (Table 2). IL-10 can induce IFN-γ secretion by NK cells  and increased IL-10 may provide positive feedback to increase IFN-γ secretion during viral infection (Figure 3c). Thus, increased mortality in WMS calves may involve dys-regulation of the pro-inflammatory response. The more prolonged elevation of serum cortisol in the PA group (Figure 1a) may be another factor contributing to the lower pro-inflammatory responses in this group since corticosteroids have potent anti-inflammatory activity . It should also be noted, however, that both treatment groups in the present experiment and previous experiment  experienced multiple stressors, including transportation, social mixing, altered environment, and restraint during sample collection or treatment. Therefore, it remains to be determined if weaning and maternal separation, in the absence of any other stressors, is sufficient to significantly increase viral-bacterial synergy. Specific stressors or combinations of stressors may influence the activity of either the hypothalamus-pituitary adrenal axis or the sympathetic adrenal medullary axis in distinct ways and have very different effects on the immune response to respiratory viral infections [5–7]. It may be that one or more concurrent stressors were required to induce stress responses of sufficient magnitude or duration to significantly (p < 0.001) increased BRD mortality (Figure 1). The BHV-1 and M. haemolytica infection model provides a system to begin investigating the temporal relationship between individual or combined stressors on viral-bacterial synergy and to identify mechanisms by which specific stressors alter susceptibility to BRD. This would be the first step in identifying behavior modification protocols or therapeutic agents that effectively mitigate the effects of stress on disease susceptibility.