Annette Sommershof, Michael Basler, Carsten Riether, Harald Engler and Marcus Groettrup
Submitted to Journal of Immunology
29
Abstract
Chronic stress is suspected to increase the susceptibility to infections but experimental evidence is scarce. We examined the effects of chronic social stress on virus-specific cytotoxic T lymphocyte (CTL) responses in mice after infection with lymphocytic choriomeningitis virus (LCMV). Mice subjected to social stress on six consecutive days prior to infection showed a significant reduction of IFN-γ producing TCD8+ splenocytes and markedly lowered plasma concentrations of IFN-γ during the late stage of the infection. In contrast, the generation of LCMV-specific CTL responses was not altered in mice undergoing the same stress procedure concurrently with infection. Furthermore, stress exposure six days before and additional three days after LCMV infection profoundly reduced the expansion of TCD8+ cells in the spleen, due to a diminished in vivo proliferation capacity as shown by BrdU incorporation. Pharmacological blockade of glucocorticoid receptors with RU-486 completely abrogated the stress-associated decline of TCD8+ expansion. Stressed mice showed a significantly reduced expression of the early T cell activation marker CD69 as well as impaired in vitro cytokine secretion of IFN-γ and IL-2. Additionally, social stress led to an altered migration capacity of TCD8+ cells as demonstrated by adoptive T cell transfer experiments. Taken together, this study shows that chronic social stress fundamentally suppresses the functional capacities of T cells during LCMV infection providing a mechanism by which stress can increase the susceptibility to viral infections.
Introduction
Chronic stress is suspected to increase the susceptibility to infectious diseases by affecting the function of cells of the innate and adaptive immune system. For example, studies in humans have shown that stressful life conditions were associated with an increased risk for influenza virus infection (247) and rhinovirus infection (248) as well as herpes simplex virus recurrences (249). However, experimental evidence is scarce that chronic stress indeed enhances the susceptibility to viral infections, e.g., by suppressing anti-viral immune responses.
Only few studies have investigated the effects of chronic stress on T cell-mediated immunity during viral infection. For example, mice exposed to chronic restraint displayed a decreased generation of virus-specific T-lymphocytes (CTL) in response to primary HSV-1 infection (250), and altered memory cytotoxic T lymphocyte activation (251-253). However, most of the studies investigating the impact of chronic stress on antiviral responses were performed using stressors without relevant behavioral context. Although these studies provided important insights into the immunmodulatory capacities of chronic stress, animal models of social stress are considered to be biologically more relevant, also with respect to stressful situations in humans. In this context, some recently performed studies examined the impact of social stress on the pathophysiological outcome of viral infections such as Theiler’s virus (243, 246) and influenza virus (240) or mycobacterial BCG (bacillus
Calmette-Guérin)-Chapter 2
30 infection (244) but very little is known about the impact of chronic social stress on T cell responses to viral infections.
A series of studies has investigated the consequences of reapeated social stress exposure on T cell number and function. For instance, it has been demonstrated that social stress is accompanied by reduced numbers of T cells in the blood, spleen, and bone marrow of mice (234). In this context, it has been shown that social defeat in rats lead to decreased homing of adoptively transferred peripheral blood T cells into lymphoid organs, suggesting altered migration properties (254). Finally, it has been reported that social stress influences proliferation, cytokine production and T cell mediated cytotoxicity. For example, socially stressed mice exhibit suppressed T cell proliferation responses to the mitogen ConA (235) and increased production of pro-inflammatory cytokines IL-6 and IFN-γ as well as decreased production of anti-inflammatory cytokine IL-10 (235).
The aforementioned results indicated that social stress substantially alters the trafficking and functional capacities of T lymphocytes. However, it is important to note that these results were observed in the absence of viral infection, thus information on whether and how social stress affects the outcome of TCD8+ cell-mediated responses during a virus challenge are largely unknown. To analyse the effects of chronic social stress on an anti-viral T cell response, we performed a systemic infection using lymphocytic choriomeningitis virus (LCMV). LCMV is a natural mouse pathogen inducing a strong cytotoxic T cell (CTL) response that is responsible for virus elimination (255-256). The CTL response of C57BL/6 mice is focused on the three dominant GP33-41/Db, GP34-41/Kb and NP396-404/Db as well as several sub-dominant (GP276-286/Db, GP92-101/Db, GP118-125/Kb and NP205-212/Kb) T cell epitopes (2, 257). In addition, antigen processing and T cell epitope production are well characterized in the LCMV infection model and therefore it represents an optimal model to study potential T cell alterations in response to social stress.
In the present study we provide evidence that chronic social stress compromises the activation and expansion of LCMV-specifc TCD8+ cells in the spleen. We further show that glucocorticoids play a fundamental role in these alterations by intrinsically inhibiting TCD8+
cell cytokine production and proliferation. Additionally, we extend previous findings that social stress leads to a profoundly impaired migration pattern of TCD8+ cells.
Results
Experimental design
To analyze the impact of social disruption stress on the LCMV-specific cytotoxic T cell response, we compared different stress protocols (Figure 1). In a first trial, the stress procedure was either applied concurrently with LCMV infection (experiment 1a) or the experimental mice were first subjected to the stressor and received the infection afterwards (experiment 1b). In a second set of experiments, we examined the effect of a prolonged stress
31 procedure on the outcome of LCMV infection, including T cell expansion and migration as well as antigen presentation. For this purpose, mice were exposed to six consecutive days of SDR followed by LCMV infection and three additional days of SDR (experiment 2).
Figure 1. Scheme of the experimental design
Thymus atrophy and adrenal hypertrophy
Organ masses of thymus and adrenal glands were determined after six days of SDR (Figure 2 A, B). Repeated exposure to the social stressor resulted in atrophy of the thymus and hypertrophy of the adrenals as shown by a significant reduction of thymus mass (SDR:
42.5±1.4 mg vs. control: 69.6±2.2 mg; p≤0.001) and a marked enlargement of the adrenal glands (SDR: 4.9±0.1 mg vs. control: 4.0±0.1 mg; p≤0.001). Both thymic atrophy and adrenal hypertrophy are consequences of frequent or persistent adrenocortical activation and are classical indicators of chronic stress (225).
Figure 2. Social disruption stress (SDR) caused thymic atrophy and adrenal hypertrophy
Organ masses of thymus (A) and adrenal glands (B) after six consecutive days of SDR. Data represent mean
± SEM of one out of two experiments with at least five mice per experiment.
Effects of social stress on the LCMV-specific T cell response
In a first set of experiments (Figure 3 A), the effect of SDR on the LCMV-specific CTL response was examined in two different timing schedules, comparing the impact of a six-day stress procedure applied prior to virus infection (experiment 1b) with an experimental set-up in which SDR was administered concurrently with the infection (experiment 1a).At the peak of the primary response (day 8), intracellular cytokine staining (ICS) was performed to
Chapter 2
32 determine the frequency of IFN-γ producing TCD8+ cells specific for six defined LCMV epitopes: two dominant (GP33-41/Db/Kb and NP396-404/Db) and four subdominant epitopes (GP276-286/Db, GP92-101/Db, GP118-125/Kb and NP205-212/Kb).
Mice that were subjected to social stress prior to virus infection (experiment 1b) exhibited a significant reduction of IFN-γ producing TCD8+ cells specific for the dominant LCMV epitopes GP33-41/Db/Kb (SDR: 2.5±0.1% vs. control: 4.6±0.4% of total TCD8+ cells; p≤0.001) and NP396-404/Db (SDR: 1.9±0.1% vs. control: 3.1±0.2%; p≤0.001) compared to control mice.
The CTL-responses to the subdominant epitopes GP276-286/Db, NP205-212/Kb, GP92-101/Db and GP118-125/Kb were not significantly altered. We further compared these results with the quantification of GP33-specific TCD8+ cells by MHC-tetramer staining. As shown in Figure 3 D, SDR mice exhibited similar percentages of TCD8+/GP33-tetramer+ cells compared to control mice (SDR: 16.8 ± 0.5 % vs. control: 15.7 ± 0.8% of total TCD8+ cells), while the frequency of IFN-γ producing TCD8+ cells was reduced (SDR: 6.3 ± 0.3% vs. control: 10.0 ± 0.8%; p≤0.001). These results demonstrate that SDR profoundly impacts the function of splenic TCD8+ cells by inhibiting antigen-specific IFN-γ secretion.
In contrast, the LCMV-specific CTL response was not significantly altered in mice receiving the same stress procedure concurrently with the infection (experiment 1a). The frequency of IFN-γ producing TCD8+ cells specific for GP33-41/Db/Kb (SDR: 8.5±0.2% vs. control:
7.0±0.8%), NP396-404/Db (SDR: 6.2±0.5% vs. control: 4.8±0.7%) and GP276-286/Db (SDR:
3.8±0.2% vs. control: 3.2±0.2%) were slightly but not significantly increased.
Effects of social stress on plasma IFN-γ levels in LCMV-infected mice
We further determined plasma levels of IFN-γ in SDR and control mice on days 2, 4, and 6 post LCMV infection. As shown in Figure 3 E, plasma concentrations of IFN-γ increased gradually in both groups during the course of LCMV infection. However, while the concentration of IFN-γ was not altered in SDR mice οn day 4 post infection, we observed an almost 50% reduction in circulating IFN-γ levels on day 6 of infection (SDR: 306±64 pg/ml vs. control: 514±65 pg/ml; p=0.06) compared to controls.
33 Figure 3: Epitope-specific CTL responses in LCMV-infected control and SDR mice
Mice were subjected to six days of SDR and were subsequently infected with 200 pfu LCMV i.v. Epitope-specific CD8+ T cell responses were quantified on day 8 after infection by intracellular cytokine assay and tetramer staining. (A) Scheme of the experimental design. (B, C) Percentage of IFN-γ secreting CD8+ cells from LCMV-infected SDR and control mice after 5h of in vitro re-stimulation in the presence of indicated peptides. ∅ represents background values of splenocytes without peptide stimulation. (C) All values represent the mean percentage of IFN-γ positive cells of CD8+ cells ± SEM of 5 mice. (D) Production of IFN-γ by GP33-specific CD8+ cells after peptide-stimulation is compared with the percentage of CD8+/GP33-tetramer+ cells. (E) Plasma levels of IFN-γ in LCMV-infected SDR and control mice. On day 2, 4, and 6 post-infection peripheral blood was taken from SDR and control groups and cytokine levels for IFN-γ were determined using a multiplexed bead-based assay. All data are representative of one out of two experiments with at least five mice per experiment.
Prolonged stressor exposure leads to reduced expansion of TCD8+ splenocytes During the course of an LCMV infection, TCD8+ lymphocytes undergo multiple cycles of cell division after exposure to the antigen, as indicated by a progressive increase of the TCD8+
population within 8 days. Prolongation of the stress procedure for three additional days after LCMV infection caused a pronounced reduction of TCD8+ cells in the spleen of SDR mice on day 6 (SDR: 15.8±0.2% vs. control: 20.2±0.7% of splenocytes; p≤0.001) and day 8 (SDR:
31.3±2.5% vs. control: 52.4±0.6%; p≤0.001) of infection (Figure 4 B). In contrast, there were no significant differences in the TCD8+ population on day 2 (SDR: 9.3±0.2% vs. control:
9.6±0.2%) and day 4 (SDR: 7.3±0.2% vs. control: 7.8±0.1%) after infection, indicating that
Chapter 2
34 the alterations occur during the antigen-driven T cell expansion phase and are not due to initially reduced TCD8+ cell numbers. In contrast to the profound reduction of splenic TCD8+
cells, alterations in the LN were much less pronounced and not significant (SDR: 30.1±0.9%
vs. control: 35.4±1% of splenocytes; p≤0.001), whereas no differences were observed in the blood (SDR: 55.8±1.9% vs. control: 56.7±2%), liver (SDR: 47.2±0.5% vs. control:
48.9±2.2%) or lung (SDR: 54.8±2.6% vs. control: 55.2±3.8%).
Social stress affects the proliferation capacity of TCD8+ lymphocytes
In order to determine whether the decreased TCD8+ cell expansion found in SDR mice is due to a higher propensity of TCD8+ cells to undergo apoptosis, we performed stainings with Annexin V and propidium iodide (PI). The frequency of early (Annexin V+/PI-) and late (Annexin V+/PI+) apoptotic cells was determined by flow cytometry on days 2 and 8 post infection.
Analyses revealed no differences in the frequency of TCD8+ cells and total splenocytes undergoing apoptosis between SDR mice and non-stressed control mice (see Table I).
Table I: Percentages of apoptotic cells in SDR and control mice Percentages of total and TCD8+
apoptotic splenocytes on day 2 and 8 post infection in SDR and control mice. All values represent the mean±SEM of five individual mice and are representative of one out of two experiments.
To further examine the deficit in TCD8+ cell expansion, we analysed in vivo T cell proliferation by performing a BrdU pulse-labeling assay. Mice were treated with 1 mg BrdU i.p. on day 5 of LCMV infection and BrdU incorporation in splenocytes was assessed by intracellular staining 24 hours later. Analysis of BrdU staining in control mice demonstrated that approximately 50% of TCD8+ cells divide between day 5 and day 6 of LCMV infection (Figure 4 C). In contrast, the frequency of BrdU positive TCD8+ splenocytes in SDR mice was significantly reduced by 28% (SDR: 34.2±1.0% vs. control: 47.2±1.1% of TCD8+ cells;
p≤0.001). Our findings strongly suggest that the massive decrease in TCD8+ splenocyte expansion observed in SDR mice is predominantly a result of diminished proliferation capacity rather than an altered death rate of TCD8+ cells.
Stress reduces expression of the T cell activation marker CD69 on TCD8+ cells Next, we investigated whether the reduced proliferation of TCD8+ cells is accompanied by inefficient T cell activation. Therefore, we quantified the expression of different T cell activation markers throughout the early stage of LCMV infection by flow cytometry. Our data revealed that SDR and control mice exhibited comparable surface expression of CD44 and CD62L (day 2 and day 4 post infection, data not shown). However, we found significantly
35 lowered expression of the early T cell activation marker CD69 on day 2 post LCMV-infection in SDR compared to control mice (Figure 4 D; SDR: 35.7±1.1% vs. control: 51.3±2.1% of TCD8+ cells; p≤0.001).
Figure 4: Effects of social stress on TCD8+ cell expansion
Mice remained undisturbed or were subjected to six days of SDR, then infected with 200 pfu of LCMV i.v., and SDR mice were stressed for additional three days post infection. (A) Scheme of the experimental design (B) The percentage of TCD8+ lymphocytes was determined on days 2, 4, 6, and 8 post LCMV infection by flow cytometry. (C) Reduced TCD8+ proliferation in LCMV-infected SDR mice compared to control mice.
On day 5 after LCMV infection, both groups received an in vivo pulse labelling with BrdU. Twenty-four hours later, spleens were removed and BrdU incorporation was assessed by flow cytometry after CD8 surface staining and intracellular staining for BrdU. Representative BrdU pulse profiles are illustrated for SDR and control mice obtained at day 6 of LCMV-infection. ∅ represents background values of splenocytes from mice that did not receive BrdU injection. (D) CD69 surface expression on TCD8+
splenocytes in SDR compared to control mice. On day 2 post infection, TCD8+ splenocytes were analysed for CD69 surface expression by flow cytometry. The histogram represents CD69 expression of TCD8+
splenocytes from SDR (black line) and control (grey line) mice compared to non-infected naïve mice (dotted line). The negative control represents background values of the TCD8+ splenocyte population that was not stained for CD69 (dashed line). Data are presented as the mean percentage of CD69-positive cells ± SEM of TCD8+ cell population. (E, F) TCD8+ cell expansion on day 8 post-infection in SDR and control mice treated with slow-release pellets of RU486 (E), nadolol (F) or placebo. Data are presented as the mean percentage of CD8-positive cells ± SEM of the lymphocytepopulation. Two-factor ANOVA revealed a significant stress × drug interaction for RU486 (p < 0.001). All data are representative of one out of two experiments with five mice per experiment.
Chapter 2
36 Glucocorticoids mediate the stress-induced decrease in TCD8+ cell expansion
To identify the endocrine factors mediating the stress-induced decrease in TCD8+ cell expansion in the spleen, SDR and control mice were implanted with continuous release pellets containing either the glucocorticoid (GC) receptor antagonist RU486 or the β-adrenergic receptor antagonist nadolol. Unstressed control mice treated with RU486 or nadolol exhibited similar TCD8+ cell expansion in the spleen on day 8 of LCMV-infection compared to mice receiving placebo pellets (Figure 4 E, F), demonstrating that neither of the two drugs influenced the LCMV-infection by itself. More importantly, splenic TCD8+ cell numbers in RU486-treated SDR mice did not differ from placebo- or RU486-treated control mice (Figure 4 E, SDR RU486: 51.6±1.4%, control RU486: 53.5±1.3%, control placebo: 53.2±1.8%), indicating that treatment with the GC receptor antagonist completely abolished the stress-induced effects on TCD8+ cell expansion. In contrast, the frequency of TCD8+ splenocytes in nadolol-treated SDR mice was not significantly altered compared to placebo-treated SDR mice (Figure 4 F, SDR nadolol: 32.7±5.6% vs. SDR placebo: 29.6±3.4%), demonstrating that β-adrenergic receptor blockade was not effective in preventing the stress-associated changes of TCD8+ cell expansion in the spleen.
Viral clearance
In order to clarify whether the impaired generation of TCD8+ cells in the spleen resulted in an altered viral clearance, mice were infected with 200 pfu LCMV and virus titers were determined in the spleen of stressed and non-stressed mice on day 2, day 4, day 6 and day 8 post-infection.
Figure 5: Effects of SDR on viral titers in the spleen of LCMV-infected mice
After 6 days of SDR exposure mice were infected with 200 pfu LCMV and subsequently stressed for three further days. On days 2, 4, 6 and 8 post infection, virus titers were determined in the spleen. Viral titers are shown as log10 plaque forming units (pfu) per spleen. Symbols represent individual mice.
Our results demonstrated that splenic viral titers were not altered at early time points of infection (day 2 and day 4 p.i.) in SDR mice (Figure 5). We noted for both groups an extensive decrease of the viral load during the later time course of the infection period (day 6 and day 8 p.i.). However, viral titers in the spleen of SDR mice were not consistently altered
37 compared to control mice on day 8 p.i., although we noticed a tendency for enhanced virus titers among stressed mice (SDR: 18.5±7.4*103 pfu vs. control: 5.6±1.2*103 pfu).
Social stress did not alter phenotype or antigen presentation of DCs in the spleen To explore whether the reduced activation of TCD8+ cells in the spleen is the result of an altered antigen presentation, we compared the frequencies of splenic CD11c+ DCs and F4/80+ macrophages on day 2 and day 4 post-infection. As shown in Figure 6 A, SDR mice exhibited similar percentages of CD11c (SDR: 5.4±0.1% vs. control: 5.7±0.2%) and F4/80 (SDR:
8.9±0.4% vs. control: 9.7±0.6%) positive cells compared to unstressed control mice. Further analysis of the activation state of CD11c+ cells, analysed by MHC class I (H-2Db and H-2Kb) expression and CD80, CD86 surface expression revealed no differences between SDR and control mice (Figure 6 B), indicating that SDR neither affected the percentage of antigen presenting cells in the spleen nor the activation of dendritic cells.
Figure 6: Effects of social stress on DCs and antigen presentation
(A) Percentages of splenic CD11c+ DCs and F4/80+ macrophages in control and SDR mice on day 4 post infection as determined by flow cytometry. (B) Activation state of CD11c+ cells in SDR compared to control mice. On day 4 post infection, CD11c+ splenocytes were assessed for surface expression of H-2Db, H-2Kb, CD80 and CD86 by flow cytometry. Representative histogram of each staining is shown for SDR (black line) and control (grey filled) mice compared to non-infected naïve mice (dotted line). Dashed lines represent staining with isotype control. Data in the bar plot are presented as the mean percentage of positive cells ± SEM of CD11c+ cells. (C, D) Antigen presentation in the spleen of SDR mice compared to control mice. On day 4 of LCMV-infection, splenocytes from stressed (squares) and unstressed (diamonds) mice were used either (C) directly as antigen presenting cells for a GP33-specific CTL line or (D) were pulsed with GP33 peptide and then used as stimulator cells. Activation of the Thy1.1+ CTL line was analyzed at different E:S ratios (CTLs:splenocytes) by staining for Thy1.1+ effector cells and intracellular IFN-γ. Data are presented as the mean percentage of IFN-γ positive cells ± SEM of Thy1.1+ cells. All data are representative of one out of two experiments with five mice per experiment.
Chapter 2
38 Recent findings have suggested that glucocorticoids impair the competence of DCs to process and present virally expressed antigens, thereby altering their ability to activate T cells (211).
To further address the possibility that splenocytes from SDR-mice present LCMV-derived epitopes less efficiently, we performed an ex vivo antigen presentation assay. GP33-41/Db/Kb presentation of splenocytes (day 4 post-infection) was assessed by their capacity to activate a GP33-specific CTL line. Splenocytes from non-infected naïve mice were used as a negative
To further address the possibility that splenocytes from SDR-mice present LCMV-derived epitopes less efficiently, we performed an ex vivo antigen presentation assay. GP33-41/Db/Kb presentation of splenocytes (day 4 post-infection) was assessed by their capacity to activate a GP33-specific CTL line. Splenocytes from non-infected naïve mice were used as a negative