Changes in peripheral immune cell numbers and functions in octogenarian walkers – an acute exercise study

Twenty elderly male and female participants (mean age 81.3 ± 1.9 years) of the 2013 Nijmegen Four Days Marches volunteered to participate in our study. The Nijmegen Four Days Marches represents the largest mass participation walking event in the world with approximately 45,000 participants annually. Based on sex and age, individuals walk 30, 40 or 50 km per day for 4 consecutive days. Blood samples were drawn before and after (within 10 min after exercise cessation) the 4 day walking event.

Blood sampling logistics were performed as described previously [23]. In brief, all participants reported twice to our laboratory, which was located at the start/finish area of the event. Baseline measurements were performed 12–36 h preceding the start (Table 1). Thereafter, the participants walked 30 km a day, for four consecutive days at a self selected pace. Exercise was performed under temperate ambient conditions with daily maximum wet bulb globe temperatures ranging between 24–27 °C. The recovery phase (fluid intake, food intake, sleep) between walking stages was uncontrolled and not monitored. Walking duration was recorded every day, while speed was calculated accordingly. Immediately (within 10 min) after finishing on the fourth day, all baseline measurements were repeated. Heart rate, as part of exercise intensity, was measured during day 1 as described previously [24]. Mean heart rate during exercise was presented in absolute values (beats per minute, (bpm)) and as a percentage of the predicted maximal heart rate. Predicted maximal heart rate (HRmax) was calculated using the Tanaka’s formula: HRmax = 208 - 0.7*age [25]. During the experiment, dry bulb, wet bulb and globe temperatures were measured every 30 min using a portable climate monitoring device (Davis instruments Inc., Hayward, USA), which was positioned at the start/finish area. The wet bulb globe temperature index (WBGT) was calculated using the formula: WBGT = 0.1 (Tdry bulb) + 0.7 (Twet bulb) + 0.2 (Tglobe).

Body mass (Seca 888 scale, Hamburg, Germany) and body height were measured and body mass index (BMI) was calculated. A four-point skin fold thickness measurement (biceps, triceps, sub-scapular, supra-iliac) was obtained in order to calculate the lean body mass [26]. Resting heart rate and blood pressure were measured twice using an automated sphygmomanometer (M5-1 intellisense, Omron Healthcare, Hoofddorp, the Netherlands) after 5 min supine rest. Finally, all subjects completed a questionnaire about their physical activity and health status.

White blood cell differential counts, hemoglobin and hematocrit levels were directly analyzed on a Sysmex XE-5000 system (Sysmex Corporation, Kobe, Japan). Relative changes in plasma volume were calculated from blood hematocrit and hemoglobin concentrations using Dill and Costill’s equation [27]. C-reactive protein (CRP) and creatinine measurements were measured in batches of frozen sera. Sera aliquots were stored at -20 °C directly after collection and thawed directly before analysis. All analyses were coded and anonymized. Serum CRP and creatinine were both measured on the c16000 Architect (Abbott Diagnostics, Abbott Park, IL).

Serum levels of CMV-specific IgG was essentially done as previously described [7]. In brief, 96-well ELISA plates (Greiner) were coated with lysates of CMV-infected fibroblasts overnight. Lysates of non-infected fibroblasts were used as negative controls. Following coating, dilutions of serum samples were incubated for 1 h. Goat-anti-human IgG was added and incubated for 1 h. Samples were incubated with phosphatase for 15 min, and the reaction was stopped with NaOH. The plates were scanned on a Versamax reader (Molecular Devices). A pool of sera from 3 seropositive individuals with known titers of CMV-specific IgG was used to quantify CMV IgG titers in the test samples.

Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation with Lymphoprep (Axis-Shield) and stored in -180 °C until staining. PBMCs (10⁶ cells) were stained simultaneously employing 4 different staining panels to asses markers of T cell differentiation, Treg and proliferation, NK cell inhibitory receptors and NK cell activating receptors (Table 2). All cell subsets are expressed as cell counts per liter unless indicated otherwise. Cell counts of these subsets were based on data from the full leucocyte differential in combination with the flowcytometry data. As an example, we used the absolute lymphocyte count from the full leucocyte differential and the percentage of CD3 T cells in the lymphocyte gate as assessed by flowcytometry for calculation of the T cell count. To perform intracellular staining with monoclonal antibodies to Cytotoxic T-lymphocyte-associated protein-4 (CTLA-4), forkhead box P3 (FoxP3) and the proliferation marker ki-67, the cells were first fixed and permeabilized with a FoxP3 staining buffer set (eBioscience). Samples were measured on a LSR-II (BD) and data were analyzed with Kaluza software (Beckman Coulter).

PBMCs at 1 x 10⁶ cells/mL were stimulated with 1 μg/mL LPS (Sigma-Aldrich, St. Louis, MO, USA) or left unstimulated. Cells were cultured in polypropylene tubes (BD bioscience) in RPMI with 10% FCS for 24 h. After 24 h, supernatants were collected and stored at -20 °C. Culture supernatants were analyzed for production of the cytokines IL-1β, IL-6, IL-8 and TNF-α by enzyme-linked immunosorbent assay (ELISA, Duoset, R&D system, Minneapolis, MN, USA) according to the manufacturer’s instructions, and read with a Versamax reader (Molecular Devices, Sunnyvale, CA, USA). The assay sensitivity was 8 pg/mL for IL-1β, 31 pg/mL for IL-6 and TNF-α, and 312 pg/mL for IL-8. The net cytokine production was calculated as cytokine production of the stimulated sample minus the cytokine production of the non-stimulated sample.

All study participants (n = 20) successfully completed the Four Days Marches at a self selected pace (4.0 ± 0.7 km/h). On average the participants walked 7 h and 47 min daily and had an average heart rate during the first day of 106 ± 15 bpm, representing an average exercise intensity of 70 ± 10%. Thus, the walking exercise in these elderly is qualified as a daily bout of moderate intensity exercise for 4 consecutive days. An overview of all subject characteristics is presented in Table 1. An exercise-induced plasma volume expansion of 3.7 ± 3.3% was observed over the 4 days, which coincided with a small, but significant decrease of hematocrit from 0.40 L/L at baseline to 0.38 L/L directly after 4 days of exercise (p < 0.0001).

When examining the composition of the peripheral blood compartment, our data show that the walking exercise resulted in a clear leucocytosis with numerical increases of granulocytes, monocytes and lymphocytes (Table 3). When analyzing the lymphocyte compartment, clear numerical increases were noted for T cells and to a lesser extent B cells. In contrast, a decline in the number of NK cells was detected. The numerical increase in T cells was largely due to an increase in CD4+ T cells (Table 3). Although CD8+ T cell numbers showed a statistical significant increase after exercise, the absolute increase was very limited. The mean CD4/CD8 ratio, an age-appropriate value of 3 [9], was not significantly increased by the walking exercise (data not shown).

As carriage of CMV has pronounced effects on the immune system, we compared the effects of the walking exercise between CMV seropositive (n = 13) and CMV seronegative subjects (n = 7). Although an exercise induced leucocytosis was seen in both CMV seropositive and seronegative individuals, increases in granulocytes, monocytes and lymphocytes were statistically significant in CMV seropositive subjects but did not reach significance in CMV seronegative subjects (Table 4). Lymphocytosis in CMV seropositive subjects was largely caused by T cells, rather CD4 than CD8, and to a lesser extent B cells, whereas NK cell numbers remained unchanged. B cell numbers increased irrespective of CMV serostatus. Interestingly, a decrease of NK cells was seen only in CMV seronegative subjects. Thus, acute exercise induced responses were more clear in CMV seropositive octogenarians.

We next examined the proliferative properties of circulating immune cells as measured by Ki-67 expression. Post walking, both CD4+ and CD8+ T cells demonstrated reduced percentages of proliferating cells (Fig. 1a). Similar observations were seen in CMV seropositive and seronegative subjects (Additional file 1: Figure S1a). Thus, the data suggest an exercise-induced increase of T cells with low proliferative capacity [20].

To determine if certain T cell populations respond differently, we next investigated exercise-induced changes in CD4+ and CD8+ differentiation subsets. A flow-cytometric analysis employing CD45RO and CCR7 was applied to identify naïve T cells (T_Naïve), central memory T cells (T_CM), effector memory T cells (T_EM) and terminally differentiated T cells (T_TD) [28]. Numerical changes of differentiation subsets were most profound in the CD4+ compartment (Fig. 1b) and appeared to be linked to CMV carriage (Additional file 1: Figure S1b). Although significant increases were noted for all four differentiation subsets, the most profound increase was noted in the CD4+ T_Naïve subset (Fig 1b). Interestingly, we also noted a significant increase of CD4+ T_Naïve in CMV seronegative subjects (Additional file 1: Figure S1b). When analysed for the contribution of recent thymic emigrants (RTE), defined as CD31^pos CD4 + T_Naïve and the central naïve CD31^neg CD4 + T_Naïve subsets, we found both subsets significantly increased [29] (Fig. 1c). In CMV seronegative subjects the rise in CD4+ T_Naïve appeared due to the CD31^neg CD4 + T_Naïve subset (Additional file 1: Figure S1c)_. Moreover, the naïve CD4 + CD25^dim subset, recently described to develop in secondary lymphoid organs upon TCR priming, was found significantly increased ([30], data not shown). The combined data suggest the exercise-induced response of CD4 + T_Naïve and CD4+ memory subsets (T_EM >T_TD > T_CM).

In the CD8+ compartment, slight but significant numerical increases were noted for both the CD8 T_EM and the CD8 + T_Naïve subsets, whereas the T_CM and the T_TD subsets were not changed (Fig. 1b). These changes appeared associated with CMV carriage (Additional file 1: Figure S1b). Taken together, our findings show an increase of peripheral naive CD4+ T cells particularly in response to exercise. Carriage of CMV is largely associated with enhanced numbers of both CD4 and CD8 subsets.

As we noted numerical increases in particularly the CD4+ T_EM and the T_TD subsets and to a lesser extent the CD8 T_EM, but not the CD8 T_TD subset, we examined these subsets for expression of the exhaustion markers CTLA-4 and PD-1 [8]. Frequencies of CTLA-4 and PD-1 in the CD4+ T_EM and the T_TD subsets were largely comparable before and after the walking exercise; although a slight decrease of PD-1 expressing cells in the T_EM subset was detected (Additional file 2: Figure S2). In the CD8+ T_EM and the T_TD subsets, where only modest or no numerical increases were noted, respectively, only slight increases in the frequencies of CTLA4-, but not PD1-expressing cells were observed. Expression of these markers was generally lacking on T_Naïve and T_CM (data not shown). Thus, the walking exercise does not affect markers of immune exhaustion of either CD4 or CD8 T_EM and T_TD subsets.

Based on CD45RA and FoxP3 expression [31], CD45RA + FoxP3^low naïve (resting) Treg cells (nTreg) and CD45RA-FoxP3^high- memory (activated) Treg cells (memTreg) were identified in the peripheral blood of elderly walkers (Fig. 2a). Post exercise, a clear numerical increase of nTreg was observed whereas the numbers of memTreg remained stable (Fig. 2b). The increase in nTreg was seen irrespective of CMV serostatus, although the increase was more clear in CMV seropositive subjects (Additional File 3, Figure S3). Thus, as seen with the conventional CD4+ T_Naïve cells, the peripheral numbers of nTregs in elderly walkers also increased in response to acute exercise. In contrast, whereas exercise led to increases in conventional CD4+ memory T cells (Fig. 1b), exercise did not induce numerical increases of memTreg (Fig 2b).

Peripheral NK cell numbers were found reduced after exercise which was caused by a numerical decline of CD56^dim but not CD56^bright NK cells [15] (Table 3 and Fig. 3a). Notably, the decrease in CD56^dim NK cells was seen in CMV seronegative donors only (Fig 3a). As CD56^dim NK cells are most frequent in the blood (90% of total NK cells), we next investigated exercise-induced changes in the expression of inhibitory and activating NK receptors by this subset. In CMV seronegative subjects, we found comparable frequencies of cells positive for inhibitory receptors (KIR2DL1, KIR2DL2/3, KIR3DL1, NKG2A and NKG2C) and activating receptors (NKp30, NKp44, NKp46, 2B4 and NKG2D) within CD56^dim NK cells before and after walking (Fig. 3bc). However, in CMV seropositive subjects, exercise did modulate frequencies of NK cells with inhibitory and activating receptors. More specifically, KIR2DL1-, KIR2DL2/3- and KIR3DL1-positive cells decreased with exercise, but frequencies of NKG2C-positive cells were increased (Fig. 3b). Also, frequencies of activating receptor NKG2D+ cells were found increased (Fig. 3c). Thus, the walking exercise led to a numerical decline of CD56^dim NK cells in CMV seronegative subjects. In CMV seropositive subjects the walking exercise did not seem to affect the numbers of NK cells but down-modulated expression of most inhibitory receptors whereas the expression of activating receptors was largely unchanged, suggesting a less inhibited phenotype as a net result.

We next investigated the effects of exercise on PBMC function. Hereto, we analysed the LPS stimulated production of IL-1β, TNFα, IL-6 and IL-8 by PBMC before and after walking. The data show increases in production of IL-6 and IL-8, but not IL-1β and TNFα (Fig. 4). Similar data were obtained in CMV seropositive and seronegative subjects (data not shown).