β₂-Adrenergic receptor signaling mediates the preferential mobilization of differentiated subsets of CD8+ T-cells, NK-cells and non-classical monocytes in response to acute exercise in humans

Abstract

Acute exercise preferentially mobilizes cytotoxic T-cells, NK-cells and non-classical monocytes to the bloodstream under the influence of hemodynamic forces and/or (β₂-adrenergic receptor (β₂-AR) signaling. However, the relative contribution of these mechanisms to the redeployment of the most exercise-responsive cell types is largely unknown. We determined the lymphocyte and monocyte subtypes mobilized to blood during exercise via β2-AR signaling whilst controlling for (β₁-AR mediated reductions in hemodynamic forces. In a randomized, double blind, complete cross-over design, 14 healthy cyclists exercised for 30-minutes at +10% of blood lactate threshold after ingesting: (1) a placebo, (2) a β₁-preferential antagonist (10mg bisoprolol), or (2) a non-preferential (β₁+β₂-antagonist (80mg nadolol) across three trials separated by >7-days. Bisoprolol was administered to reduce hemodynamic forces (heart rate and blood pressure) during exercise to levels comparable with nadolol but without blocking (β₂-ARs. The mobilization of total NK-cells, terminally differentiated (CD57+) NK-cells, central memory, effector memory and CD45RA+ effector memory CD8+ T-cells; non-classical monocytes; and γδ T-cells were significantly blunted or abrogated under nadolol compared to both bisoprolol and placebo, indicating that the exercise-induced mobilization of these cell types to the blood is largely influenced by (β₂-AR signaling. Nadolol failed to inhibit the mobilization of classical monocytes, CD4+ T-cells (and their subsets) or naïve CD8+ T-cells, indicating that these cell types are mobilized with exercise independently of the (β₂-AR. We conclude that the preferential mobilization of NK-cells, non-classical monocytes and differentiated subsets of CD8+ T-cells with exercise is largely dependent on catecholamine signaling through the (β₂-AR. These findings provide mechanistic insights by which distinct lymphocyte and monocyte subtypes are preferentially mobilized to protect the host from anticipated injury or infection in response to an acute stress response.

1. Introduction

Short-term activation of the biological stress response through physical exercise evokes an almost instantaneous mobilization of lymphocytes and monocytes to the peripheral circulation (Dhabhar, 2018; Simpson et al., 2017). This is considered an archetypical feature of the ‘flight or fight’ response, whereby the immune system prepares the body for combat or potential injury by mobilizing immune cells to the blood where they can be quickly transported to sites of injury (Dhabhar, 2018). Exercise preferentially mobilizes lymphocyte and monocyte subtypes with high levels of surface adhesion molecules and phenotypes associated with increased cytotoxicity, antigen experience, inflammation and antigen presentation (Campbell and Turner, 2018; Simpson et al., 2016). For instance, NK cells, γδ T-cells, CD8+ T-cells and ‘non-classical’ monocytes are mobilized in relatively greater numbers than CD4+ T-cells, B-cells or ‘classical’ monocytes (Anane et al., 2009; Campbell et al., 2009; Simpson et al., 2007; Simpson et al., 2009; Steppich et al., 2000), while the more mature subtypes of CD8+ T-cells (i.e. central memory (CM), effector memory (EM) and the ‘terminally’ differentiated (EMRA) cells) and NK-cells (i.e. CD56dim/CD57+) are preferentially mobilized over their less mature counterparts (Anane et al., 2010; Bigley et al., 2014; Campbell et al., 2009; Spielmann et al., 2014; Spielmann et al., 2016).

Increased hemodynamic forces and/or the actions of plasma catecholamines and β-adrenergic receptors (β-AR) are accepted as major mechanisms underpinning lymphocyte and monocyte mobilization to blood during exercise (Simpson et al., 2015; Walsh et al., 2011). Elevations in cardiac output and blood pressure increase hemodynamic shear stress, causing the detachment of adhered leukocytes from the vascular endothelium and mobilizing leukocytes from slow moving marginal pools into the main axial flow of the bloodstream. β-ARs are also important because, not only are they expressed on the surface of lymphocytes and monocytes mobilized by stress and exercise (Campbell and Turner, 2018; Dimitrov et al., 2010; Dimitrov et al., 2013; Hong et al., 2013; Kruger et al., 2008; Landmann, 1992; Simpson et al., 2016; Steppich et al., 2000; Van Tits et al., 1990), but lymphoid organs such as the spleen are a major source of the leukocytes mobilized to blood and are under adrenergic control (Nielsen et al., 1997; Van Tits et al., 1990). Administering a non-preferential β-AR antagonist (e.g. propranolol) prior to exercise, and indeed other tasks that activate the biological stress response such as public speaking and catecholamine infusion, has been shown to inhibit acute stress induced lymphocyte and monocyte mobilization (Benschop et al., 1994; Foster et al., 1986; Mills et al., 2000; Mills et al., 1999; Murray et al., 1992; Schedlowski et al., 1996; Steppich et al., 2000), indicating that the response is highly regulated by catecholamines and β-ARs. However, a glaring limitation of many of these studies was the failure to accurately determine the subtype of β-AR (i.e. β₁-versus β₂) signaling on the stress-induced mobilization of lymphocytes and monocytes whilst accounting for concomitant reductions in heart rate and blood pressure that occurs predominantly due to β₁-AR blockade (Gullestad et al., 1991). Moreover, the effects of β-AR blockade on the mobilization of the more discrete subtypes of monocytes (i.e. non-classical, intermediate and ‘classical subsets), CD8+ T-cells (naïve, CM, EM and EMRA subsets) and NK-cells (‘terminally differentiated’ CD57+) with exercise is largely unknown. It is important to identify the subtype of β-AR signaling responsible for this response as it could reveal useful biological targets for mobilizing highly specialized lymphocyte and monocyte subtypes in vivo for immune cell therapeutics (Simpson et al., 2017).

The aim of this study, therefore, was to determine the relative contribution of β₁-versus β₂-AR signaling to the mobilization of a wide range of highly specialized lymphocyte and monocyte subtypes known to be preferentially mobilized to the blood compartment with exercise. We used a non-preferential β₁+β₂-AR antagonist (nadolol) to block the β₂-AR and a preferential β₁-AR antagonist (bisoprolol) as a control to account for β₁-AR mediated reductions in hemodynamic forces (i.e. heart rate and blood pressure) that occur with non-preferential β₁+β₂-AR blockade during exercise. We report here that β₂-AR signaling is the primary mechanism by which the most stress-responsive lymphocyte and monocyte subtypes are mobilized to the bloodstream during exercise.

2. Methods

2.1. Participants

Fourteen (13 males, 1 female) healthy, non-smoking, physically active adults aged 24–43 years were recruited to participate in this study. Participants had no contraindications to performing vigorous exercise as determined by reporting zero positive responses on the ACSM-AHA pre-exercise screening questionnaire (Balady et al., 1998). They were also expected to provide a score of 5–7 (corresponding to regular ‘vigorous’ activity) on a 0–7 point scale when asked to describe their physical activity habits within the preceding month of the study (Jackson et al., 1990). Participants were excluded if they were using prescription medications, pregnant or breastfeeding, had allergies to β-blocker medication, or presented with symptoms of asthma. All participants refrained from performing vigorous exercise within 48h of each laboratory visit and were asked to arrive at the laboratory following an overnight fast having consumed nothing but water. All participants provided a signed written statement of informed consent prior to participation in the study. Participant characteristics are presented in Table 1. All study procedures were approved by the Committee for the Protection of Human Subjects (CPHS) at the University of Houston.

Table 1.

Exercise and demographic information of the study participants (n = 14; 1 female).

	Mean±SD	Range
Age (yrs)	31±6	24–43
Height (cm)	175.3±6.7	160.0–185.0
Weight (kg)	81.8±11.4	68.1–111.2
Cycling power at breakpoint BLT (W)1	133±30	75–176
Cycling power at 4mmol BLT (W)1	185±33	161–242
Resting HR (bpm)	68±11	48–91
Resting BP (mmHg)
Systolic	119±8	104–133
Diastolic	75±5	68–85
Workload at +10% breakpoint BLT (W)	148±30	83–195
Average Exercise HR (bpm)
Placebo	154±13	123–171
Nadolol	116±10#	98–133
Bisoprolol	119±9#	98–131
Average Exercise BP (mmHg)
Placebo
Systolic	163±15	138–181
Diastolic	78±4	70–88
Nadolol
Systolic	138±13#	113–158
Diastolic	76±5	66–85
Bisoprolol
Systolic	146±12#	118–163
Diastolic	77±11	60–107
End of exercise RPE (6–20) 2
Placebo	15±2	12–18
Nadolol	16±2	13–18
Bisoprolol	16±2	13–20
End of exercise Blood Lactate (mmol/L)
Placebo	2.9±1.0	1.2–4.8
Nadolol	2.8±0.8	1.6–4.1
Bisoprolol	3.7±1.2	1.8–5.8

^#Significant difference from placebo (p < 0.05).

¹The blood lactate threshold (BLT) was determined as the breakpoint or 4mmol (Weltman, 1995).

²Rating of perceived exertion (RPE) was determined using Borg’s 6–20 scale (Borg, 1998).

2.2. Experimental design

Participants reported to the laboratory on 5 separate occasions following an overnight fast. During visit 1, participants underwent a comprehensive screening procedure to ensure that they met the study inclusion / exclusion criteria. Eligible participants returned for a second visit to complete a blood lactate threshold test on a stationary leg cycling ergometer (Velotron, Racermate Inc., Seattle WA, USA; or CompuTrainer, Racermate Inc., Seattle, WA) following a protocol we have described previously (Bigley et al., 2014). Visits 3–5 served as the experimental trials of the study, with each visit consisting of a 30-minute fixed intensity leg cycling trial that the participant performed at a power output (watts) corresponding to +10% of their pre-determined individual blood lactate threshold in accordance with the breakpoint definition (Weltman, 1995). The experimental cycling trials were conducted under three separate conditions following oral administration of a placebo, a non-preferential β₁+β₂-AR antagonist (80mg nadolol), or a preferential β-₁-AR. antagonist (10mg bisoprolol). Drugs were administered 3h prior to exercise because peak serum concentrations of both bisoprolol and nadolol occur within 2–4h following oral consumption, with half-lives of 10–12h and 14–24h, respectively (Borchard, 1998). The experimental exercise trials were conducted using a randomized complete cross-over and double-blind experimental design. All exercise bouts were completed between 08:00 and 11:00, with each participant completing all 4 of their individual exercise bouts at the exact same time of day. There was a minimum of 7-days (maximum of 14-days) between each of the experimental trials to allow for drug washout.

Intravenous blood samples were collected into lithium heparin, EDTA, and serum separator gel vacutainers from an antecubital vein at baseline (prior to administration of the drug/placebo), immediately before exercise (Pre-Ex), immediately after exercise (Post-Ex), and 1h following exercise cessation (1h post-ex). Oxygen uptake, heart rate (Quark CPET, COSMED, Pavona di Albano Laziale, Italy), and exercise ECG (Quark C12x, COSMED, Pavona di Albano Laziale, Italy) were monitored continually throughout each trial, and blood pressure (determined manually with a stethoscope and inflatable cuff), blood lactate (P-GM7 Micro-Stat analyzer, Analox instruments Ltd., London, UK).], and rating of perceived exertion (RPE) was recorded every 10min. Serum samples were aliquoted and stored at −80 °C until analysis in duplicate for catecholamines and cortisol using commercially available ELISA kits (2-CAT ELISA, LDN, Nordhorn Germany; Cortisol ELISA, LDN, Nordhorn, Germany).

2.3. Rationale for drug selection and dosage

An ideal situation would have been to use the β₁-AR antagonist bisoprolol with and without a ‘selective’ β₂-AR antagonist to differentiate β₁ from β₂-AR involvement in the mobilization of lymphocytes and monocytes with exercise. This would have allowed us to directly compare the effects of β₁-AR versus β₂-AR antagonism under similar conditions of hemodynamic shear stress. However, there are no ‘selective’ β₂-AR antagonists currently on the market and experimental β₂-AR antagonists like ICI 188 551 (Gullestad et al., 1991; Van Tits et al., 1990) are no longer used in humans. We deemed it imperative to use a ‘selective’ β₁-AR antagonist that will alter hemodynamic responses to exercise by a similar magnitude to the ‘non-selective’ β₁+β₂-AR antagonist that is being used to block the β₂-AR. The selective β₁-AR antagonist metoprolol and the non-selective β₁+β₂-AR antagonist propranolol have previously been shown to elicit divergent effects on exercising heart rate and blood pressure (Mills et al., 1999) and were, therefore, deemed suboptimal. We opted to use bisoprolol because it has a ~14-fold preference for the human β₁-AR over the β₂-AR, compared to metoprolol which only has a ~2-fold preference (Baker, 2005). We also used nadolol as it has a ~23-fold preference for the β₂-AR over the β₂-AR compared to propranolol, which only has a β₂-AR preference in the order of ~8-fold (Baker, 2005). Importantly, we have shown previously that nadolol and bisoprolol at standardized doses of 80mg and 10mg, respectively, lowered hemodynamic responses to exercise by a similar magnitude in individuals weighing ~70 to ~110kg (Agha et al., 2018). This allowed us to compare the effects of ‘selective’ β₁-AR blockade to ‘non-selective’ β₁+β₂-AR blockade under similar conditions of hemodynamic shear stress.

2.4. Direct immunofluorescence assays and flow cytometry

Blood granulocyte, lymphocyte, and monocyte (mid-size cells) counts were determined using an automated hematology analyzer (Mindray BC-3200, Nanshan, Shenzhen, PR China). Multi-color flow cytometry was used to enumerate monocyte and monocyte subtypes in whole blood and lymphocyte and lymphocyte subtypes in isolated peripheral blood mononuclear cells (PBMCs) following isolation by density gradient centrifugation (Histopaque-1077, Sigma-Aldrich, St. Louis, MO, USA). All monoclonal antibodies (mAbs) used for the direct immunofluorescence assays were purchased from eBioscience (San Diego, CA, USA) unless otherwise stated. For monocyte phenotyping, 50 μl of whole blood was labeled with FITC-conjugated anti-CD14 mAb and a PE-conjugated anti-CD16 mAb. Aliquots of 1×10⁵ PBMCs were labeled with an FITC/Alexa Fluor 488 mAb against CD28, CD45RA, KLRG1, CD8 or NKG2C (Alexa Fluor 488; R&D systems, Minneapolis, MN, USA); PE-conjugated mAb against CD27, CD62L, or CD57; PerCP-Cyanine 5.5-conjugated mAb against CD4 or CD8 or PE-Cyanine 5.5-conjugated mAB against CD56; and APC-conjugated mAB against CD3.

Following mAb staining and a 30-minute incubation, PBMC samples were analyzed by flow cytometry. Whole blood samples were incubated with red blood cell lysing buffer for 30min, washed, and then run on the flow cytometer. Blood lymphocyte and monocyte phenotyping was performed using a BD Accuri C6 flow cytometer (BD Accuri, Ann Arbor, MI) or a MACSQuant ® analyzer 10 (Miltenyi Biotech, Bergisch Gladbach, Germany) flow cytometer. A fixed acquisition volume was obtained from the whole blood samples used to assess monocytes, and 20,000 lymphocytes were collected from each PBMC sample. Forward and side scatter plots were used to identify lymphocyte and monocyte cell populations, which were each gated electronically. T-cells were identified as those within the lymphocyte gate that expressed the CD3 surface antigen. The CD4+ T-cells (CD3+/CD4+) and CD8+ T-cells (CD3+/CD8+) were gated within the CD3+ lymphocytes, and differentiation status was identified based on their expression of CD28/CD27, CD45RA/CD62L, and KLRG1/CD57. NK cells were identified as those within the lymphocyte gate that expressed the CD3-/CD56+ phenotype. Finally, γδ T-cells were identified as those within the lymphocyte gate that expressed the CD3 +/CD4−/CD8− phenotype. The phenotypic nomenclature used to define the lymphocyte and monocyte subtypes are shown in Table 2. To enumerate the total number of lymphocyte and monocyte subtypes, the percentage of all lymphocytes or monocytes expressing the surface markers of interest was multiplied by the total lymphocyte or monocyte count.

Table 2.

Terms and phenotypes used to identify lymphocyte and monocyte subsets in blood in response to exercise.

	Phenotype	Reference
CD4+ and CD8+ T-Cell Subsets
Naïve	CD45RA+ CD62L+	(Sallusto et al., 1999)
Early-Differentiated	CD28+ CD27+;	(Appay et al., 2002)
	KLRG1− CD57−	(Kared et al., 2016)
Central Memory (CM)	CD45RA− CD62L+	(Sallusto et al., 1999)
Effector Memory (EM)	CD45RA− CD62L−	(Sallusto et al., 1999)
Intermediate-Differentiated	CD28− CD27+;	(Appay et al., 2002)
	KLRG1+ CD57−	(Kared et al., 2016)
Effector Memory RA (EMRA)	CD45RA + CD62L−	(Sallusto et al., 1999)
Late-Differentiated	CD28− CD27−;	(Appay et al., 2002)
	KLRG1+ CD57+	(Kared et al., 2016)
γδ T-cells	CD3+ CD4− CDS−	(Cron et al., 1989)
NK Cell Subsets
Less Mature / Mature‘terminally differentiated’NK-cells	CD57−/CD57+	(Bjorkstrom et al., 2010;Bozzano et al., 2017; Karedet al., 2016)
Activation Receptor expressedby adaptive ‘memory-like’ NK-cells	NKG2C+	(Bozzano et al., 2017; Della Chiesa et al., 2015)
Monocyte Subsets
Classical Monocytes	CD14++ CD16−	(Ziegler-Heitbrock, 2015)
Intermediate Monocytes	CD14++ CD16+	(Ziegler-Heitbrock, 2015)
Non-Classical Monocytes	CD14+ CD16+	(Ziegler-Heitbrock, 2015)

2.5. Statistical analyses

Maximum likelihood linear mixed models were built to assess the differences in absolute number of cells, which included main effects for time (baseline, pre-exercise, post-exercise, and 1h post-exercise) and trial (placebo, nadolol, and bisoprolol) and an exercise x trial interaction effect. Planned contrasts were also included in the model a priori to test for overall time effects within each trial and between specific time points, as well as trial effects within a specific time point. Statistical significance was taken at p≤0.05, and all statistical analyses were performed using SPSS Version 22 (IBM; Chicago, IL).

3. Results

3.1. Nadolol and bisoprolol lower exercising heart rate and blood pressure to comparable levels

All participants successfully completed the exercise trials. Participant physical characteristics and exercise data are presented in Table 1. Main trial effects were found for averaged exercising heart rate (F = 62.2), systolic blood pressure (F = 13.5) and blood lactate (F = 3.5), but not RPE (F = 1.4) or diastolic blood pressure (F = 0.2). Planned contrasts revealed that heart rate and systolic blood pressure were significantly higher in the placebo trial compared to both the bisoprolol and nadolol trials. No statistical differences were found between the bisoprolol and nadolol trials for exercising heart rate (p = 1.0) or systolic blood pressure (p = 0.367). Planned contrasts did not reveal statistical differences in blood lactate between individual trials, but there was a trend for lactate levels to be higher in the bisoprolol trial compared to both the nadolol (p = 0.064) and placebo (p = 0.111) trials.

3.2. Nadolol augments the exercise-induced increase in circulating epinephrine levels

Changes in serum catecholamines and cortisol in response to exercise are shown in Fig. 1. Main effects of time (F = 18.4), trial (F = 5.6) and time × trial (F = 2.8) were found for serum epinephrine. Planned contrasts revealed that the trial effect was significant at Post-Ex only (F = 12.2), with nadolol being significantly greater than both placebo and bisoprolol. A main time effect was found for serum norepinephrine (F = 42.0). Planned contrasts revealed no trial effects within each time point, but norepinephrine values Post-EX were significantly higher than at all other time points within each trial. A main trial effect was found for serum cortisol (F = 3.2) without a significant main effect of time (F = 2.2; p = 0.086). Planned contrasts did, however, reveal significant time effects within the nadolol (F = 2.8) but not the placebo (F = 0.5) or bisoprolol (F = 0.3) trials. Despite this, no statistical differences in serum cortisol were found across trials within each time point, or between specific time points within a trial.

Fig. 1.

The effects of an acute bout of exercise on serum levels of epinephrine, norepinephrine and cortisol In healthy participants after Ingesting a placebo, a β₁-AR antagonist (bisoprolol, 10mg), or a β₁+β₂-AR antagonist (nadolol, 80mg) 3h prior to exercise. Significant difference from placebo indicated by ## (p < 0.01). Significant difference from bisoprolol indicated by + + (p < 0.01). Significant difference from baseline indicated by ^ (p < 0.05) and ^^ (p < 0.01). Significant difference from Pre-Ex indicated by * (p < 0.05) and ** (p < 0.01). Significant difference from Post-Ex indicated by ^vv (p < 0.01).

3.3. Nadolol but not bisoprolol blunts NK cell and γδ T-cell mobilization in response to acute exercise, while both bisoprolol and nadolol augments granulocyte mobilization during exercise recovery

Changes in circulating leukocytes, granulocytes, total monocytes, total lymphocytes, total T-cells, total CD4+ T-cells, total CD8+ T-cells, γδ T-cells, total NK-cells, and NK-cells expressing CD57 or NKG2C are presented in Table 3. Significant within trial time effects were found for all cell types in all three trials with the exception of CD4+ T-cells and NKG2C+ NK-cells. While CD4+ T-cells did not change significantly over time in any trial, the time effect for NKG2C+ NK-cells was present within the placebo and bisoprolol trials but not the nadolol trial. This effect was driven by a trend for fewer NKG2C+ NK-cells in blood at Post-Ex in the nadolol trial (p = 0.055). Significant trial effects within the Post-Ex time point were found for total lymphocytes, total NK-cells and CD57+ NK-cells, with a trend observed for total CD8+ T-cells (p = 0.099). These effects were driven by a lower number of cells in the nadolol trial compared to the placebo and/or bisoprolol trial. Significant time effects within each trial were found for γδ T-cells, and although effect sizes were lowest in the nadolol trial, there were no significant trial effects within a particular time point. However, further analysis revealed that γδ T-cells at Post-EX were significantly higher than baseline and Pre-Ex within the placebo and bisoprolol trials but not the nadolol trial. Within the 1h post time point, significant trial effects were found for total leukocytes and granulocytes. These effects were driven by a greater number of cells in the nadolol compared to the placebo trial. Planned contrasts showed that leukocytes and granulocytes remained elevated at 1h post compared to baseline and/or Pre-Ex in the bisoprolol and nadolol trials, whereas cell numbers had returned to near baseline levels in the placebo trial.

Table 3.

The effects of an acute bout of exercise on the total numbers of circulating leukocytes and their subpopulations in healthy participants after ingesting a placebo, a β₁-AR antagonist (bisoprolol, 10mg), or a β₁+β₂-AR antagonist (nadolol, 80mg) 3h prior to exercise (mean ± SD). Blood samples were taken at baseline (before drug administration), pre-exercise (3h following drug administration and immediately prior to exercise), immediately post-exercise, and 1h post-exercise during each of the three trials.

						Main Effects
		Placebo	Bisoprolol	Nadolol	Trial effect F (p value)	Time	Trial	Time × Trial
Leukocytes (×103/μl)
	baseline	6.2±1.5	5.9±1.0	6.4±1.8	F=0.2 (0.806)
	pre-ex	6.4±1.6	6.6±1.4	6.6±1.9	F=0.03 (0.968)	F=30.0	F=0.6	F=1.2
	post-ex	10.0±2.5*,^	10.0±2.2*,^	9.4 ± 2.9*,^	F=0.5 (0.616)	p<0.001	P=0.536	P=0.317
	1h post	7.5±1.4V	8.0±2.6^, V	9.4±2.9#, *,^	F=3.4 (0.035
Time effect	F-statistic	F=10.7	F=11.7	F=10.0
	p value	<0.0001	<0.0001	<0.0001
Granulocytes (×103/μl)
	baseline	3.7±1.3	3.4±0.8	4.0±1.6	F=0.3 (0.719)
	pre-ex	4.1±1.5	4.2±1.4	4.3±1.6	F=0.04 (0.959)	F=19.5	F=1.4	F=0.8
	post-ex	5.9±2.2*,^	5.7±2.2^	5.8±2.3^	F=0.05 (0.950)	p<0.001	P=0.255	P=0.580
	1h post	5.5±1.3	6.0±2.4^	7.2±2.6#, *,^	F=3.3 (0.040)
Time effect	F-statistic	F=5.0	F=6.6	F=9.6
	p value	0.003	<0.0001	<0.0001
Monocytes (×103/μl)
	baseline	0.4±0.2	0.5±0.2	0.5±0.2	F=0.3 (0.738)
	pre-ex	0.4±0.2	0.4±0.2	0.4±0.2	F=0.2 (0.849)	F=35.5	F=0.1	F=0.9
	post-ex	0.8±0.4*,^	0.9±0.3*,^	0.7±0.3*,^	F=1.7 (0.194)	p< 0.001	P=0.912	P=0.529
	1h post	0.4±0.1V	0.4±0.1V	0.4±0.1V	F=0.5 (0.589)
Time effect	F-statistic	F=10.7	F=19.0	F=7.7
	p value	<0.0001	<0.0001	<0.0001
Lymphocytes (×103/μl)
	baseline	2.0±0.4	2.0±0.4	1.9±0.4	F=0.3 (0.730)
	pre-ex	1.9±0.3	2.0±0.5	1.9±0.4	F=0.5 (0.615)	F=84.1	F=2.3	F=1.4
	post-ex	3.2±0.6*,^	3.4±0.6*,^	2.8±0.7+, *,^	F=5.4 (0.005)	p<0.001	P=0.102	P=0.200
	1h post	1.6±0.3V	1.7±0.4V	1.8±0.5V	F=0.4 (0.654)
Time effect	F-statistic	32.8	37.9	16.2
	p value	<0.0001	<0.0001	<0.0001
CD3+ (×103/μl)
	baseline	1.5±0.4	1.5±0.3	1.4±0.4	F=0.2 (0.805)
	pre-ex	1.3±0.2	1.4±0.4	1.4±0.3	F=0.2 (0.794)	F=28.8	F=0.1	F=0.3
	post-ex	2.0±0.5*,^	2.0±0.4*,^	1.9±0.6*,^	F=0.1 (0.918)	p<0.001	P=0.915	P=0.922
	1h post	1.2±0.3V	1.3±0.3V	1.4±0.4V	F=0.5 (0.589)
Time effect	F-statistic	F=11.3	F=10.4	F=7.7
	p value	<0.0001	<0.0001	<0.0001
CD4+ (cells/μl)
CD3+ CD4+	baseline	746±255	663±195	743± 307	F=0.6 (0.579)
	pre-ex	656±151	678± 229	742± 233	F=0.5 (0.596)	F=3.2	F=3.2	F=0.3
	post-ex	798±252	741±222	925±399	F=2.3 (0.106)	p=0.025	P=0.042	P=0.919
	1h post	662±129	632±141	745± 255	F=0.9 (0.420)
Time effect	F-statistic	F=1.2	F=0.5	F=2.1
	p value	0.313	0.657	0.101
CD8+ (cells/μl)
CD3+ CDS +	baseline	484±166	480±153	397±119	F=1.3 (0.281)
	pre-ex	423±130	477±164	399±105	F=0.9 (0.430)	F=22.7	F=2.5	F=0.7
	post-ex	701±258*,^	688±231*,^	581±194*,^	F=2.4 (0.099)	p<0.001	P =0.082	P =0.676
	1h post	377±118V	394±137V	400±149V	F=0.1 (0.925)
Time effect	F-statistic	F=11.0	F=8.5	F=4.5
	p value	<0.0001	<0.0001	0.005
γδ T-cells (cells/μl)
CD3+ CD4−CD8−	baseline	88±41	101±50	86±45	F=0.2 (0.785)
	pre-ex	84±34	103±55	78±35	F=0.7 (0.493)	F=14.3	F=3.6	F=0.4
	post-ex	162±83*^	176±77*^	134±58	F=1.9 (0.156)	p<0.001	P=0.030	P=0.908
	1h post	70±35V	111±119V	79±32	F=1.8 (0.166)
Time effect	F-statistic	F=6.8	F=5.2	F=2.9
	p value	<0.0001	0.002	0.036
NK-cells (cells/μl)
CD3−CD56 +	baseline	205±87	229±103	194±88	F=0.2 (0.833)
	pre-ex	231±81	275±111	164± 67	F=1.8 (0.164)	F=115.60	F=14.0	F=5.3
	post-ex	708±219 +*^	901±403 #*^	466±199 #+*^	F=27.7 (<0.0001)	p<0.001	p<0.001	p<0.001
	1h post	117±41V	143±66V	102±34V	F=0.3 (0.777)
Time effect	F-statistic	F=41.0	F=69.9	F=15.1
	p value	<0.0001	<0.0001	<0.0001
CD57+ NK (cells/μl)
CD57+ CD3−CD56+	baseline	114±71	130±87	111±64	F=0.1 (0.915)
	pre-ex	123±70	157±82	96±52	F=0.8 (0.450)	F= 66.6	F=8.8	F=3.6
	post-ex	413±197+*^	563±309#*^	269±137#+*^	F=18.5 (<0.0001)	p<0.001	p<0.001	p=0.002
	1h post	56±26V	78±57V	45±24V	F=0.2 (0.788)
Time effect	F-statistic	F=23.5	F=45.8	F=7.4
	p value	<0.0001	<0.0001	<0.0001
NKG2C+ NK (cells/μl)
NKG2C+ CD3−CD56+	baseline	48±57	54±70	36±46	F=0.2 (0.865)
	pre-ex	45±52	58±86	27±29	F=0.4 (0.643)	F=8.2	F=2.5	F=0.4
	post-ex	127±135	145±190^	68±76	F=3.0 (0.055)	p<0.001	p=0.085	p=0.869
	1h post	25±26	41±61V	20±21	F=0.2 (0.796)
Time effect	F-statistic	F=4.0	F=4.6	F=0.8
	p value	0.008	0.004	0.506

Statistical differences indicated by the following:

^#Significant difference from placebo (p < 0.05).

⁺Significant difference from bisoprolol (p < 0.05).

^*Significant difference from pre-exercise (p < 0.05).

^{^}Significant difference from baseline (p < 0.05).

^vSignificant difference from post-exercise (p < 0.05).

3.4. Nadolol but not bisoprolol blunts the mobilization of highly dtffereniiated CD8+ T-cells caused by exercise

Changes in the total number of CD8+ and CD4+ T-cells exhibiting a naïve, CM, EM, and EMRA phenotype are presented in Fig. 2. Main time effects were found for CD8+ naive (F = 5.1), CM (F = 14.3), EM (F = 17.8), and EMRA (F = 11.8) subtypes. Main trial effects were found for CD8+ EMRA cells (F = 3.5) and planned contrasts revealed that the effect occurred at the Post-Ex time point only (F = 5.2), with the nadolol trial being significantly lower than both the placebo and bisoprolol trials. For CD8+ CM and EM subtypes, the effect sizes for time were lower within the nadolol trial (F = 3.6; p = 0.015 and F = 3.0; p = 0.032, respectively) compared to both the placebo (F = 5.5; p = 0.001 and F = 7.4; p < 0.0001, respectively) and bisoprolol (F = 6.3; p < 0.0001 and F = 8.9; p < 0.0001, respectively) trials. Planned contrasts revealed that the number of CD8+ CM and EM cells in blood at Post-Ex was significantly higher than at all other time points for placebo and bisoprolol, but within the nadolol trial, CD8+ CM and EM cells were higher at Post-Ex compared to baseline only. Similar results were observed using the CD28/CD27 and KLRG1/CD57 phenotypic nomenclatures in that the mobilization of those CD8+ subsets with the most differentiated phenotypes (CD27−/CD28−; KLRG1+/CD57+) in response to exercise was significantly lower in the nadolol trial compared to both the bisoprolol and placebo trials (data not shown). A significant main effect for time was found for CD4+ CM (F = 5.4), EM (F = 2.7) and EMRA (F = 3.6) cells but not CD4+ naïve (F = 1.6) cells in accordance with CD45RA and CD62L phenotype. However, planned contrasts revealed no significant time effects within each trial, or trial effects within a specific time point, for CD4+ CM, EM or EMRA cells (data not shown).

Fig. 2.

The effects of an acute bout of exercise on the total number (cells/μl of whole blood) of CD8+ T-cell subsets in healthy participants after ingesting a placebo, a β₁-AR antagonist (bisoprolol, 10mg), or a β₁+β₂-AR antagonist (nadolol, 80mg) 3h prior to exercise. Naive, central memory (CM), effector memory (EM) and CD45RA+ effector memory (EMRA) T-cells were identified using the surface markers CD62L and CD45RA. Significant difference from placebo indicated by # (p < 0.05). Significant difference from bisoprolol indicated by + (p < 0.05). Significant difference from baseline indicated by ^ (p < 0.05) and ^^ (p < 0.01). Significant difference from Pre-Ex indicated by * (p < 0.05) and ** (p < 0.01). Significant difference from Post-Ex indicated by ^v (p< 0.05) and ^vv (p < 0.01).

3.5. Nadolol but not bisoprolol blunts the mobilization of non-classical monocytes caused by exercise, while both bisoprolol and nadolol evokes the mobilization of intermediate monocytes

Changes in the total number of classical, intermediate and non-classical monocytes are presented in Fig. 3. Main effects of time were found for three monocyte subtypes (F = 36.1, F = 11.3, and F = 24.1, respectively) but no main trial effects were found. Planned contrasts revealed that time effects for classical monocytes were statistically significant within all trials, but effect sizes were lower for the nadolol (F = 6.6) trial compared to the placebo (F = 13.6) and bisoprolol (F = 17.7) trials. This resulted in a trend for a trial effect within the Post-Ex time point (F = 2.3; p = 0.106). Time effects for intermediate monocytes were statistically significant within both the bisoprolol (F = 5.1) and nadolol (F = 6.0) trials but not the placebo (F = 1.5) trial. Further analysis showed that the numbers of intermediate monocytes Post-Ex were significantly higher compared to Pre-EX and/or baseline within the nadolol and bisoprolol trials but not the placebo trial. Time effects for non-classical monocytes were statistically significant within all trials, but effect sizes were lower for the nadolol (F = 3.7) trial compared to the placebo (F = 9.3) and bisoprolol (F = 13.0) trials. This resulted in a trend for a trial effect within the Post-Ex time point (F = 2.7; p = 0.071), during which time the total number of non-classical monocytes were significantly elevated above all other time points within both the placebo and bisoprolol trials. There was a trend for non-classical monocytes to be elevated at Post-Ex compared to Pre-Ex within the nadolol trial (p = 0.053).

Fig. 3.

The effects of an acute bout of exercise on the total number (cells/μl of whole blood) of classical, intermediate and non-classical monocytes in healthy participants after ingesting a placebo, a β₁-AR antagonist (bisoprolol, 10mg), or a β₁+β₂-AR antagonist (nadolol, 80mg) 3h prior to exercise. Significant difference from baseline indicated by ^ (p < 0.05) and ^^ (p < 0.01). Significant difference from Pre-Ex indicated by * (p < 0.05) and ** (p < 0.01). Significant difference from Post-Ex indicated by ^v (p < 0.05) and ^vv (p < 0.01).

3.6. Nadolol but not bisoprolol inhibits the mobilization of NK-cells caused by exercise regardless of NK-cell maturation status

Changes in the total number of NK-cells contrasted by NKG2C and CD57 expression are presented in Fig. 4. Main effects of time (F = 72.3; F = 42.5), trial (F = 7.0; F = 4.1) and time × trial (F = 2.5; F = 2.4) were found for the NKG2C−/CD57− and NKG2C−/CD57+ subsets, respectively. Significant main effects for time (F = 26.3) and trial (F = 3.9) were also found for the NKG2C+/CD57− subset, whereas only a main effect of time (F = 5.5) was found for the NKG2C+/CD57+ subset. Planned contrasts revealed significant trial effects within the Post-Ex time point for the NKG2C−/CD57− (F = 12.9), NKG2C−/CD57+ (F = 11.0) and NKG2C+/CD57− (F = 5.4) subsets. Further analysis showed that this was due to lowered cell numbers in the nadolol compared to the placebo and/or the bisoprolol trial at this exercise time point. Time effects for NKG2C+/CD57+ NK-cells were statistically significant within both the placebo (F = 2.8) and bisoprolol (F = 3.1) trials but not the nadolol (F = 0.5) trial. Further analysis showed that the numbers of NKG2C+/CD57+ NK-cells Post-Ex were significantly higher compared to 1h-Post in the placebo and bisoprolol trials but not the nadolol trial.

Fig. 4.

The effects of an acute bout of exercise on the total number (cells/μl of whole blood) of NK-cell subtypes in accordance with NKG2C and CD57 expression in healthy participants after ingesting a placebo, a β₁-AR antagonist (bisoprolol, 10mg), or a β₁+β₂-AR antagonist (nadolol, 80mg) 3h prior to exercise. Significant difference from placebo indicated by # (p < 0.05) and ## (p < 0.01). Significant difference from bisoprolol indicated by ++ (p < 0.01). Significant difference from baseline indicated by ^^ (p < 0.01). Significant difference from Pre-Ex indicated by * (p < 0.05) and ** (p < 0.01). Significant difference from Post-Ex indicated by ^v (p < 0.05) and ^vv (p < 0.01).

4. Discussion

This is the first study to determine the effects of preferential systemic β₂-AR activation on the redeployment of discrete lymphocyte and monocyte subtypes known to be preferentially mobilized by exercise. We did this by comparing responses in the presence of the ‘selective’ β₁-AR antagonist, bisoprolol, versus responses in the presence of the ‘non-selective’ β₁+β₂-AR antagonist, nadolol using a randomized, double-blind and complete cross-over experimental design. We showed that nadolol, but not bisoprolol, completely abrogated or inhibited the redeployment of differentiated subsets of CD8+ T-cells, NK-cells and non-classical monocytes, thereby suggesting that this preferential mobilization is mediated by the β₂-AR activation. This occurred despite plasma epinephrine levels being substantially higher after exercise in the nadolol trial compared to the placebo and bisoprolol trials. However, because the mobilization of some cell populations (i.e. total lymphocytes, NK-cells; non-classical monocytes) was only blunted but not abrogated by nadolol, it is likely that hemodynamic shear stress still partially contributes to the mobilization of some effector lymphocytes and monocytes with exercise.

It has been suggested that distinct subpopulations of lymphocytes and monocytes are preferentially mobilized to protect the host in anticipation of injury or infection following an acute stress response (Dhabhar, 2018). This contention is supported by a number of consistent findings in the literature showing that a single exercise bout preferentially mobilizes subpopulations of NK-cells, CD8+ T-cells, γδ T-cells and monocytes that display phenotype characteristics associated with high differentiation, effector function, and tissue migration (Campbell and Turner, 2018; Kruger et al., 2008; Kruger and Mooren, 2007; Simpson et al., 2016; Simpson et al., 2015); cell types that are arguably useful in fight or flight situations (Dhabhar, 2018). Indeed, stress and exercise-induced NK-cell mobilization has been shown to inhibit tumor growth in experimental rodent models (Pedersen et al., 2016), whilst we and others have advocated harnessing the biological stress response to mobilize progenitor cells and specialized subsets of T-cells and NK-cells to augment immune cell therapeutics (Agha et al., 2018; Emmons et al., 2016; Niemiro et al., 2017; Simpson et al., 2017; Spielmann et al., 2016). The present findings show that β₂-AR signaling is largely responsible for the preferential redeployment of NK-cells, non-classical monocytes, γδ T-cells and differentiated CD8+ T-cells (i.e. the CM, EM and EMRA subtypes) with exercise. Specifically, by including a ‘selective’ β₁-AR antagonist to account for reductions in heart rate and blood pressure that accompany ‘non-selective’ β₁+β₂-AR blockade during exercise, we determined that ~60% of highly differentiated CD8+ T-cells, ~40% of NK-cells and ~40% of non-classical monocytes are mobilized to the blood solely due to β₂-AR signaling. This could have important implications for allogeneic peripheral blood stem cell transplantation, as increased proportions of NK-cells and γδ T-cells combined with lowered proportions of naïve CD4+ T-cells and B-cells in donor allografts has been shown to elicit better clinical outcomes in patients (e.g. lower relapse incidence, better engraftment, and lower incidence/severity of graft-versus-host disease) (Gu et al., 2018).

The current findings indicate that reductions in lymphocyte β₂-AR sensitivity seen with aging (Feldman et al., 1984; Kawamoto et al., 1989), chronic stress (Mausbach et al., 2008), and/or obesity (Hong et al., 2014) might diminish the beneficial effects of exercise on immune surveillance that are conferred by the frequent redistribution of immune cells between the blood and tissues with each bout of exercise (Campbell and Turner, 2018; Kruger et al., 2008; Kruger and Mooren, 2007; Simpson et al., 2016). Indeed, older adults are known to mobilize fewer numbers of CD4+ and CD8+ T-cells than the young in response to a single exercise bout (Ceddia et al., 1999; Mazzeo et al., 1998; Spielmann et al., 2014), while catecholamines secreted during exercise can provide anti-inflammatory signals to monocytes via the β₂-AR (Dimitrov et al., 2013). Cross-sectional data has shown that obesity and aerobic fitness have opposing effects on monocyte β₂-AR sensitivity and TNFα secretion in response to LPS stimulation in vitro (Hong et al., 2014), indicating that exercise training interventions could increase β₂-AR sensitivity and, consequently, heighten immune surveillance and lower inflammation. Exercise training has been shown to increase β₂-AR sensitivity in cardiac tissue (Leosco et al., 2007) and adipose tissue (Polak et al., 2005), and, conversely, overtraining has been associated with reductions in skeletal muscle β₂-AR sensitivity and muscle strength (Fry et al., 2006). While the effects of fitness status on immune cell β₂-AR density are inconsistent (Butler et al., 1982; Fujii et al., 1998; Hong et al., 2013; Maki et al., 1987), it remains to be seen if exercise training can alter β-AR sensitivity in lymphocytes and monocytes. Not only is this a viable mechanism by which exercise training could improve immune competence and reduce inflammation, but might also have important implications for individuals undertaking exercise regimens to increase/maintain health whilst taking non-selective β-blockers. Indeed, it would appear that drugs such as nadolol, carvedilol or propranolol might inhibit the beneficial β₂-AR mediated effects of exercise on immune surveillance and inflammation and that it would be advantageous to prescribe β_rAR ‘selective’ over ‘non-selective’ β₁+β₂-AR antagonists in certain medical situations if the same therapeutic benefit can be attained.

We found that the β₂-AR-mediated mobilization of NK-cells with exercise, unlike CD8+ T-cells, was not governed by maturation status when defined by expression of the ‘terminal differentiation’ marker CD57. However, ‘terminally differentiated’ NK-cells expressing NKG2C were less responsive to exercise and non-selective β₁+β₂-AR blockade than their NKG2C negative counterparts, although this was not a general feature of NKG2C expression because the CD57– NK-cells expressing NKG2C remained highly responsive to exercise and non-selective β₁+β₂-AR blockade. We reported previously that NKG2C+ NK-cells have low β₂-AR sensitivity and are consequently mobilized to blood in relatively low numbers during exercise (Bigley et al., 2015). In the current study, we found exercise to mobilize total NKG2C+ NK-cells following bisoprolol but not nadolol or placebo administration, indicating that selective β₁-AR blockade during exercise augmented catecholamine signaling through the β₂-AR pathway to evoke their mobilization.

The only cell type to show an augmented mobilization with exercise under conditions of β₁-AR blockade independently of β₂-AR signaling was intermediate monocytes. Both bisoprolol and nadolol, but not placebo, caused intermediate monocytes to mobilize with exercise, suggesting a possible involvement of changes in hemodynamic forces or α-AR signaling. It has been shown previously that the mobilization of certain monocyte subtypes with exercise is inversely related to hypertension severity (Dimitrov et al., 2013). Thus, the lowering effects of bisoprolol and nadolol on systolic blood pressure, even in people with normal blood pressure, may have facilitated the mobilization of intermediate monocytes with exercise reported here. It is also possible that β-AR blockade increased epinephrine availability to activate α-ARs, which are known to be expressed on human blood monocytes (Grisanti et al., 2011). This is supported somewhat by the trend for increased intermediate monocyte mobilization with exercise under nadolol compared to bisoprolol, which also evoked the largest epinephrine response. Our catecholamine findings corroborate a previous study showing larger epinephrine but not norepinephrine responses to maximal exercise following nadolol but not atenolol (selective β₁-AR antagonist) administration (Wolfel et al., 1990). This might be an adaptive response to override the effects of systemic non-selective β-AR blockade (i.e. increased secretion) or to fewer available binding sites for circulating epinephrine (i.e. decreased removal) following nadolol administration. Granulocyte mobilization was blunted by both bisoprolol and nadolol, indicating that increases in hemodynamic shear stress largely govern their mobilization to the blood during exercise. However, we did find that granulocytes continued to mobilize into the recovery phase of exercise following nadolol, and to a lesser extent bisoprolol, but not or placebo administration. This could be a compensatory and cortisol-driven mechanism for the lack of lymphocyte and monocyte trafficking provided by β₂-AR signaling during an acute stress response. Indeed, we found a larger time effect for cortisol within the nadolol trial, which could be driving the mobilization of granulocytes during exercise recovery (Bishop et al., 2001).

Several studies have shown that ‘non-selective’ β-blockers like propranolol can inhibit lymphocyte and monocyte mobilization in response to exercise (Foster et al., 1986; Mills et al., 1999; Steppich et al., 2000). However, only two studies, to our knowledge, have compared the effects of ‘non-selective’ β-AR blockade (propranolol) with ‘selective’ β₁-AR blockade (metoprolol), reporting that the mobilization of total NK-cells, total CD8+ cells and CD62L−/CD8+ cells was inhibited with propranolol but not metoprolol (Mills et al., 1999; Murray et al., 1992). While this suggests β₂-AR involvement, both studies used a between subjects design and the drugs were given in a non-blinded manner without a true placebo condition (Mills et al., 1999; Murray et al., 1992). We considered a between-subjects design to be suboptimal due to more recent findings that prior exposure to cytomegalovirus – a latent herpesvirus that infects 49–58% of all adults in the United States aged 20–49 years (Bate et al., 2010) – can have both profound and divergent effects on the mobilization of CD8+ T-cells and NK-cells with exercise, thus making infection history a major confounding factor (Simpson et al., 2016). Moreover, in the study by Mills et al. (1999), large differences in exercise duration, blood pressure and heart rate were observed between the two drug trials making it difficult to determine the relative contribution of β₂-AR signaling independently from concomitant reductions in hemodynamic forces (Gullestad et al., 1991). We do acknowledge that the imbalanced distribution of male and female participants and the relatively small sample size are limitations of our current study.

In conclusion, this is the first study to show that those lymphocyte and monocyte subtypes considered to be highly ‘exercise responsive’ (i.e. CM, EM and EMRA CD8+ T-cells, NK-cells, and non-classical monocytes) have the greatest dependency on β₂-AR signaling for their mobilization to the bloodstream with exercise. These findings provide mechanistic insights by which distinct lymphocyte and monocyte subtypes are preferentially mobilized to protect the host from anticipated injury or infection in response to an acute stress response. Future studies should consider the β₂-AR as a therapeutic target for improving immune and anti-inflammatory responses to exercise, especially in conditions associated with β₂-AR desensitization such as aging, obesity and chronic stress.