Spleen contributes to restraint stress induced changes in blood leukocytes distribution

Psychological stress has great impacts on the immune system, particularly the leukocytes distribution. Although the impacts of acute stress on blood leukocytes distribution are well studied, however, it remains unclear how chronic stress affects leukocytes distribution in peripheral circulation. Furthermore, there is no report about the role of spleen in the blood leukocytes distribution induced by stress. Here we show that spleen contributes to the alteration of restraint stress induced blood leukocytes distribution. Our data confirmed that restraint stress induced anxiety-like behavior in mice. Furthermore, we found that restraint stress decreased the CD4/CD8 ratio and elevated the percentages of natural killer cells, monocytes and polymorphonuclear myeloid-derived suppressor cell. We demonstrated that activation of hypothalamic-pituitary-adrenal axis (HPA) and sympathetic nervous system (SNS) contributes to restraint stress induced alteration of blood leukocyte distribution. Interestingly, we found that splenectomy could reverse the change of CD4/CD8 ratio induced by restraint stress. Together, our findings suggest that activation of HPA axis and SNS was responsible for the blood leukocyte subsets changes induced by restraint stress. Spleen, at least in part, contributed to the alteration in peripheral circulation induced by restraint stress.

cycles of restraint stress significantly changed the percentages of leukocyte subsets while 1, 7 cycles of restraint stress did not. In addition, blockade of the HPA axis activation or SNS activation could partially reverse restraint stress induced blood leukocyte redistribution. Moreover, splenectomy 14 days before restraint stress prevented the changes of CD4/CD8 ratio induced by restraint stress. Taken together, these data showed activation of HPA axis and SNS was responsible for the blood leukocyte subsets changes induced by restraint stress and splenectomy partially prevented the changes of leukocyte subsets induced by restraint stress.

Restraint stress induced anxiety-like behavior.
Mice were subjected to 2-hour restraint stress each day for 21 consecutive days (Fig. 1A). To study the impact of the restraint stress, we measured anxiety-like behavior by using open field test at different time point. We found that mice undergoing 1 cycle or 7 cycles of restraint stress took similar time to first enter the center of the open filed, spent similar time in the center of the open field and showed similar frequency to enter the center of the open filed compare with control mice (Fig. 1B,C and D), which suggest that short-term restraint stress did not affect anxiety-like behavior. However, mice undergoing 21 cycles of restraint stress displayed increased anxiety-like behavior in the open field test. Our data showed that mice subjected to 21 cycles of restraint stress took longer to first enter the center of the open field than the controls (control mice 30.88 ± 9.74 s, stress mice 88.98 ± 16.17 s; p = 0.025; Fig. 1B). Meanwhile, the restraint stressed mice spent less time in the center of the open field (control mice 18.56 ± 3.3 s, stress mice 8.95 ± 1.26 s; p = 0.018 Fig. 1C). Furthermore, restraint stressed mice entered the center of the open field less often than the controls (control mice 17.69 ± 5.52, stress mice 10.00 ± 4.99 s; p = 0.003; Fig. 1D). There was no difference of the distance traveled in the open field between the restraint stressed mice and the control mice (control mice 2167.63 ± 89.28 cm, stress mice 1996.29 ± 68.83 cm; p = 0.086; Fig. 1E), indicating that differences were not due to changes in locomotion or activity. Body weight was measured before starting stress, the day after the mice in the open field between the restraint stressed mice and the control mice. (F) Body weight gain was reduced significantly after the mice undergoing 7 or 21 cycles of restraint stress. Data were shown as mean ± SEM (n = 10-15). *p < 0.05, **p < 0.01. subjected to 7 cycles and 21 cycles of stress. Comparing to the control animals, body weight gain was reduced significantly after the mice undergoing 7cycles (control mice 1.01 ± 0.17 g, stress mice 0.18 ± 0.27 g; p = 0.031), and 21cycles (control mice 3.18 ± 0.28 g, stress mice 2.13 ± 0.28 g; p = 0.019) of restraint stress (Fig. 1F), which suggests that restraint stress attenuates body weight gain. Our data imply that only long-term restraint stress induces anxiety-like behavior in mice.
Restraint stress increased circulating norepinephrine and corticosterone. Previous study indicates that chronic stress elevated level of plasma IL-6 and TNF-α the day after the last cycles of stress 20 . Therefore, we determined to assess the level of IL-1β, IL-6, TNF-α, IFN-γ, IL-4, and IL-10 in serum after undergoing restraint stress (Fig. 3A). Our results showed that only the level of serum IFN-γ increased 1 h following the 1st cycle of restraint stress (control 56.61 ± 15.21 pg/ml, 1 cycle of stress 239.54 ± 63.62 pg/ml; p = 0.028; Fig. 3B). The level of other cytokines was not significantly different as compared to the controls (Fig. 3C,D,E,F,G). Our data suggest that serum cytokines are not involved in the restraint stress induced anxiety-like behavior.
As norepinephrine and corticosterone were the physiologic parameters of stress 21 , norepinephrine and corticosterone were measured in restraint stressed mice and control animals. Immediately after the stressed mice subjected to 1, 7 or 21 cycles of restraint stress, approximately 1 ml of blood was drawn from the retro-orbital plexus of the stressed and control mice. The stressed mice after undergoing 1, 7 or 21 cycles of restraint stress displayed significantly elevated serum norepinephrine concentration as compared to the control animals (control 3.24 ± 0.45 ng/ml, 1 cycle of stress 6.95 ± 0.86 ng/ml, 7 cycles of stress 5.68 ± 0.70 ng/ml, 21 cycles of stress 5.85 ± 1.32 ng/ml; p < 0.05 vs. control; Fig. 3H). Similarly, restraint stress also elevated the level of serum corticosterone (control 404.94 ± 15.47 ng/ml, 1 cycle of stress 534.57 ± 51.39 ng/ml, 7 cycles of stress 613.23 ± 67.44 ng/ ml, 21 cycles of stress 640.78 ± 54.53 ng/ml; p < 0.05 vs. control; Fig. 3I).These results demonstrate that restraint stress may induce anxiety-like behavior through elevating circulating norepinephrine and corticosterone.
In addition to increased serum corticosterone, chronic restraint stress also increased the levels of serum norepinephrine. Recent studies indicate that stress-induced activation of sympathetic nervous system (SNS) mediates effects on immune through beta(β)-adrenergic receptor 18,22 . To determine if beta(β)-adrenergic receptor antagonism would attenuate the changes of percentages of leukocyte subsets induced by chronic restraint stress, mice were injected sc with vehicle or propranolol prior to restraint stress each day for 21 days (Fig. 5A). The changes of the CD4/CD8 ratio (stress × propranolol interaction, F(1,22) = 9.027; p = 0.007; Fig. 5D) and the percentages of NK cells (stress × propranolol interaction, F(1,22) = 7.602; p = 0.012; Fig. 5E) were also blocked by propranolol. But, propranolol could not reverse the changes of the percentages of CD11b + Ly-6 low monocytes (stress × propranolol interaction, F (1,22) = 1.230; p = 0.280; Fig. 5F) and CD11b + Ly-6C low Ly-6G + PMN-MDSCs (stress × propranolol interaction, F(1,22) = 2.061; p = 0.167; Fig. 5G). These data indicated that RU486 or propranolol treatment could partially reverse the alteration in peripheral circulation in stressed mice, which supports that activation of hypothalamic-pituitary-adrenal axis (HPA) and sympathetic nervous system (SNS) is involved in the altered leukocyte distributions in peripheral circulation.
Splenectomy reversed the changed CD4/CD8 ratio caused by restraint stress. Our previous study showed that 21 cycles of restraint stress changed the CD4/CD8 ratio, reduced the percentages of CD11b + Ly-6 low monocytes and CD11b + Ly-6C low Ly-6G + PMN-MDSCs in spleen (Yu Li et al., in the submission).
To determine if splenectomy before restraint stress could affect the changes of leukocyte subsets distribution, mice were splenectomized 14 days before restraint stress and then exposed to 21 cycles of restraint stress (Fig. 6A). Splenectomy before restraint stress attenuated the changes of CD4/CD8 ratio (stress × splenectomy interaction, F (1,19) = 6.635; p = 0.020; Fig. 6C). However, our results showed that splenectomy could not block the changes of  (Fig. 6D,E,F). Our findings suggest that spleen, at least in part, is involved in the restraint stress induced changes in blood leukocytes distribution.

Discussion
These results of our study demonstrate a novel role of spleen in the restraint stress-induced blood leukocyte subsets changes. First, we have shown that restraint stress induced HPA axis and SNS activation, a delay in weight gain, and an increase in anxiety-like behavior. Second, restraint stress did not alter the serum cytokine but one cycle restraint stress increased the serum IFN-γ. Third, corticosterone antagonism blocked the 21 cycles of restraint stress induced the increase of the percentages of CD11b + Ly-6 low monocytes and CD11b + Ly-6C low Ly-6G + PMN-MDSCs. Fourth, beta(β)-adrenergic receptor antagonism attenuates the changes of the CD4/CD8 ratio, and the percentages of NK cells induced by 21 cycles of restraint stress. Finally, splenectomy before restraint stress reversed the changed CD4/CD8 ratio following 21 cycles of restraint stress.
Our data shown the restraint stress induced a delay in weight gain, an increase in anxiety-like behavior, indicating our stress model was effective 3, 23, 24 .Our results showed chronic restraint stress did not alter serum cytokine but the level of serum IFN-γ increased 1 h following the 1st cycle of restraint stress. However, previous studies reported chronic stress increased the plasma level of IL-6and TNF-α 24, 25 , while another report showed that restraint stress caused a significant decrease in the concentrations of circulating IFN-γ, IL-12(p40), IL-12(p70), IL-4, and IL-5 26 . Recently, Smith et al. reported chronic restraint stress and chronic variable stress caused different effect on immune 27 . Therefore, we propose that the differences in the level of serum cytokines may be caused by different animal model and the timing of sampling.
Previous study shown acute stress increased an early increase of leukocyte followed by a decrease, and began to increase and return to pre-stress baseline after the stressor was terminated 28,29 , which is consistent with our results that 1 or 7 cycles of restraint stress did not change the percentage of leukocyte subsets in blood. Stress activates HPA axis and SNS to release corticosterone and catecholamines (adrenaline and norepinephrine). Previous  30 . The adrenergic receptors can be divided into two subgroups, the αand βadrenergic receptors, and the βadrenergic receptor plays the most important role in the immune system 31 . Numerous studies have shown that acute stress induces alters in absolute numbers or/and relative proportions of blood leukocytes by stress hormones 28,[32][33][34][35] . Our results indicated that RU486 could block the increase of the percentages of CD11b + Ly-6 low monocytes and CD11b + Ly-6C low Ly-6G + PMN-MDSCs while propranolol could block the change of the CD4/CD8 ratio, and the increase of NK cells induced by 21 cycles of restraint stress. In other words, chronic restraint stress induced alteration of blood leukocytes by activating HPA axis and SNS.

researches shown T and B cells, neutrophils, monocytes and macrophages express corticosterone receptor while T and B cells, NK cells, monocytes and macrophages express adrenergic receptors
The splenectomy study provides first evidence that spleen plays an important role in the change of CD4/ CD8 ratio induced by chronic restraint stress. Splenectomy could prevent the change of CD4/CD8 ratio induced by chronic restraint stress resembles the studies of liver fibrosis that splenectomy reduced Th2 lymphocytes 36 . Splenectomy before stress could not block the changes of CD11b + Ly-6 low monocytes and CD11b + Ly-6C low Ly-6G + PMN-MDSCs. These results are consistent with Mckin et al. reported that splenectomy did not prevent increased monocytes in circulation following repeated social defeat stress 37 . Bone marrow may be to the initial production of accumulation of CD11b + Ly-6 low monocytes and CD11b + Ly-6C low Ly-6G + PMN-MDSCs because stress increased production of myeloid cell in bone marrow 38 .
In summary, our findings provide suggestive evidence that activation of HPA axis and SNS is responsible for the blood leukocyte subsets changes induced by chronic restraint stress. Spleen, at least in part, contributes to the changes of blood leukocyte subsets induced by chronic restraint stress.

Materials and Methods
Mice. Male C57BL/6 mice (6-8 weeks old) were obtained from Experimental Animal Center of Xi'an Jiaotong University College of Medicine and allowed to acclimate the surroundings for 14 days before the experimental procedure started. Mice were housed 5 animals per cage under specific pathogen-free condition and had free access to tap water and standard mouse diet. The study was approved by Ethics Committee of Xi'an Jiaotong University College of Medicine according to the principles outlined in the Declaration of Helsinki.

Restraint stress.
Mice were subjected to restraint stress similar to the previously reported 39 . In brief, mice were placed in ventilated 50-ml conical tubes for 2 h each day between 1700 and 2100 h Beijing for 21 consecutive days. Both the stress and control mice were not available for food and water during the stress period. Weight was measured before starting stress, the day after the mice subjected 7 cycles, 21 cycles stress. Anxiety-like behavior. As previously described 40,41 , open field test was applied to measured anxiety-like behavior. Mice were placed individually into the corner of open field chamber (45 × 45 × 45 cm) under a bright light (500 lux) for 5 min immediately after the mice subjected the first cycle stress, the day after mice subjected 7 and 21 cycles stress. Computer imaging video tracking system (SMART, Panlab SL, Barcelona, Spain) was applied to record the track. The dependent variables latency to entering the center, time spent in center, times of entering center were applied to measure the anxiety-like behavior. The chamber was cleaned with 75% alcohol between subjects.
Serum norepinephrine and cortcosterone measurement. Immediately after the mice subjected 1, 7, and 21 cycles of restraint stress, approximately 1.00 ml of blood was drawn from the retro-orbital plexus. The blood samples were allowed to clot at room temperature for 1 h, afterward, centrifuged at 5000 rpm for 10 min and serum stored at −80 °C for subsequent. Serum norepinephrine and cortcosterone were measured with mouse norepinephrine ELISA kit (Westang, Shanghai, China) and corticosterone Enzyme Immunoassay kit (Arbor Assays, Michigan, USA) following the manufacturer's instructions, respectively. Splenectomy. Splenectomy was performed 14 days before the beginning of stress. Mice were anesthetized with chloral hydrate (100 ug/g body weight) and the skin in the left abdomen was shaved. A splenectomy was performed through a small incision in the abdomen. The associated vessels and nerves were ligatured. Afterward, the incision was sutured. Sham surgery was performed only made an incision in the abdomen without removal of the spleen.
Statistical analysis. Data of obtained from open field test, body weight, stress hormone and cytokine ELISA's, and flow cytometric procedures were analyzed by two-tailed non-parametric test (Mann-Whitney U-test); to determine the effects of stress and pharmacological treatments or splenectomy, data were analyzed by a two-way ANOVA (SPSS 13.0, Chicago, IL, USA). In all comparisons, differences were considered significant at p < 0.05. Data were shown as mean ± standard error of the mean (SEM).
Data availability statement. All data generated during and analysed during the current study are available from the corresponding author on reasonable request.