Changes in Immune Function during Initial Military Training

ABSTRACT

Purpose

Initial military training (IMT) is a transitionary period wherein immune function may be suppressed and infection risk heightened due to physical and psychological stress, communal living, and sleep deprivation. This study characterized changes in biomarkers of innate and adaptive immune function, and potential modulators of those changes, in military recruits during IMT.

Methods

Peripheral leukocyte distribution and mitogen-stimulated cytokine profiles were measured in fasted blood samples, Epstein–Barr (EBV), varicella zoster (VZV), and herpes simplex 1 (HSV1) DNA was measured in saliva by quantitative polymerase chain reaction as an indicator of latent herpesvirus reactivation, and diet quality was determined using the healthy eating index measured by food frequency questionnaire in 61 US Army recruits (97% male) at the beginning (PRE) and end (POST) of 22-wk IMT.

Results

Lymphocytes and terminally differentiated cluster of differentiation (CD)4+ and CD8+ T cells increased PRE to POST, whereas granulocytes, monocytes, effector memory CD4+ and CD8+ T cells, and central memory CD8+ T cells decreased (P ≤ 0.02). Cytokine responses to anti-CD3/CD28 stimulation were higher POST compared with PRE, whereas cytokine responses to lipopolysaccharide stimulation were generally blunted (P < 0.05). Prevalence of EBV reactivation was higher at POST (P = 0.04), but neither VZV nor HSV1 reactivation was observed. Diet quality improvements were correlated with CD8+ cell maturation and blunted proinflammatory cytokine responses to anti-CD3/CD28 stimulation.

Conclusions

Lymphocytosis, maturation of T-cell subsets, and increased T-cell reactivity were evident POST compared with PRE IMT. Although EBV reactivation was more prevalent at POST, no evidence of VZV or HSV1 reactivation, which are more common during severe stress, was observed. Findings suggest increases in the incidence of EBV reactivation were likely appropriately controlled by recruits and immune-competence was not compromised at the end of IMT.

The synergistic actions of the innate and adaptive immune systems are critical to optimal immune function, and function in the prevention and recovery from infection. Underlying that synergy are numerous immune cells, cytokines, chemokines, proteins, and bioactive molecules, among other components that recognize and neutralize pathogens (1). Ultimately, these compounds and pathways facilitate interactions between the nonspecific innate immune system, which recognizes and immediately initiates a response against invading pathogens, and the adaptive immune system which initiates a slower but pathogen-specific response within several days (2).

Poor nutrition, as well as physical, psychosocial, and environmental stressors, compromise functions of both the innate and adaptive immune systems thereby increasing risk of illness and infection (3,4). These factors likely contribute to the high incidence of illness and infection rates in some otherwise healthy populations, such as athletes and military personnel engaged in physically and psychologically demanding training, and those living in congregate settings (5–7). For example, during initial military training (IMT), civilian recruits entering into military service experience close living conditions, limited sleep, periods of high physical activity, and exposure to environmental stressors (8). Infectious respiratory diseases, such as those caused by influenza and adenoviruses and, more recently, SARS-CoV-2, are particularly prevalent in IMT settings, and are the most common source of missed work days among young adult Army service members (5). Similar issues occur in other environments characterized by crowded living arrangements, psychosocial stressors, and strenuous activity (8–10). Respiratory infections are associated with missed training and competition days in high-performance athletes (11) and school days in students (12). However, implementing immune-enhancing countermeasures in these populations and environments has had mixed success (13,14). An improved understanding of changes in the innate and adaptive immune systems within these populations and environments, and the factors influencing those responses, may provide useful insight that informs the development of improved immune-enhancing countermeasures.

The objective of the current study was to identify changes in biomarkers of the immune system in US military recruits during 22 wk of IMT. Distribution of circulating leukocytes, T-cell reactivity to mitogen stimulation, reactivation of herpes viruses, and salivary secretory immunoglobulin A (SIgA) were measured to collectively provide insight into changes in components of both the innate and adaptive immune systems. These outcomes were selected because they can be assessed within austere training environments (15–17), are sensitive to stressors within those environments (18,19), and respond to immunomodulatory interventions, such as diet modification and/or supplementation (2,13). We hypothesized that after the challenging training period, immune markers would indicate compromised immune function and that increases in diet quality and physical activity would be correlated with a less severe decline.

METHODS

Participants and study design

This observational, longitudinal study was conducted in US Army infantry recruits participating in IMT. Initial military training is the period during which civilian recruits enter into military service. The training period integrates physical fitness with military skills instruction in a tightly controlled, psychologically and physically demanding environment. To better prepare Soldiers for the rigors of modern combat, The U.S. Army recently extended the timeline of IMT for combat arms Soldiers from 14 to 22 wk, combining Basic Combat Training (BCT) with Advanced Individual Training (AIT) in one location referred to as One Station Unit Training (OSUT). The BCT portion (10 wk) of OSUT consists of cardiorespiratory aerobic (i.e., loaded road marches, sprinting, distance running) and strength training (i.e., calisthenics, pushups, sit-ups) in the form of physical training sessions 4 to 6 d·wk⁻¹ and military-related activities (i.e., obstacle courses, weapons training) along with didactic classroom instruction (20). Advanced Individual Training is occupation-specific training, providing specialized instruction in the Soldier’s career field (21). Physical training 4–6 d·wk⁻¹, military-related activities, and didactic classroom instruction continue during AIT. Unpublished pilot data from our group indicated that total daily energy expenditure measured by doubly-labeled water ranged from an average (± SD) of 3702 ± 557 kcal·d⁻¹ to 4546 ± 486 kcal·d⁻¹ (mean ± SD; n = 14) in male recruits during the various phases of OSUT. Others have reported doubly-labeled water or actigraphy measured total daily energy expenditures of ~3500 to 4300 kcal·d⁻¹ in male recruits and ~3000 to 3400 kcal·d⁻¹ in female recruits during IMT (22,23). Throughout the 22-wk OSUT period, recruits reside in barracks, are generally given the opportunity to sleep 7 h per night, have a communal style of living, and receive three meals per day either from the dining facility or in the form of operational rations.

Data for this study were collected at Ft. Benning GA at two time points: February 2020 (PRE) and July 2020 (POST) to assess changes in physiological outcomes and markers of immune function over the course of the 22-wk OSUT period. The PRE data collection took place at the end of the period during which recruits arrive at OSUT and are in-processed, and before the start of any training activities. The POST data collection took place during the final week of training. Data collection at both PRE and POST occurred in the morning before any physical activity with participants in a fasted state. A mid-point data collection was also planned for the end of the BCT phase and total daily energy expenditure measurements were planned for various phases of OSUT, but could not be completed due to Department of Defense-wide COVID-19 related travel restrictions.

Army recruits age 17 yr and older participating in OSUT training at Ft Benning were eligible to participate in this study. Those who donated blood within the 4 months before starting the study were excluded. No additional exclusion criteria were applied. The study was reviewed and approved by the U.S. Army Medical Research and Development Command (Ft. Detrick, Fredericksburg, MD) Institutional Review Board. Investigators adhered to the policies regarding the protection of human subjects as prescribed in Army Regulation 70-25. All volunteers provided written informed consent after receiving a verbal briefing.

Demographic characteristics

Demographic information (sex, age, race, ethnicity, marital status) was collected with a standardized self-administered questionnaire at PRE (Table 1). The average amount of time engaged in vigorous and moderate physical activities and sleeping over the previous week was assessed at PRE and POST using a modified version of the International Physical Activity Questionnaire (24) included within the Military Eating Behavior Survey (25). The same questionnaire included a modified version of the Perceived Stress Scale (26), which was used to measure subjective stress levels over the previous month.

TABLE 1.

Baseline demographic characteristics of study population.

	All Volunteers (n = 98)	OSUT Completers (n = 61)
Sex
Male	77 (79)	59 (97)
Female	21 (21)	2 (3)
Age (yr)	21.5 ± 3.3a	21.5 ± 3.1b
Race
White/Caucasian	82 (84)	50 (82)
Black/African American	9 (9)	6 (10)
Other	7 (7)	5 (8)
Ethnicity
Hispanic/Latino	15 (15)	8 (13)
Non-Hispanic/Latino	83 (85)	53 (87)
Married	15 (15)	9 (15)
Physical activity before enlistment
Average daily activity (min·d−1)	119 ± 134	113 ± 114
Average vigorous activity (min·d−1)	59 ± 72a	55 ± 67b
Average moderate activity (min·d−1)	62 ± 80	59 ± 69
HEI-2015 total score	58.4 ± 10.7	57.9 ± 10.6c

Values are means ± SD or n (%).

^an = 97.

^bn = 60.

^cn = 31.

Salivary viral DNA

Saliva samples were collected at PRE and POST using passive drool collection to assess latent reactivation of herpes viruses. After saliva collection, samples were frozen, shipped overnight on dry ice to Johnson Space Center (JSC, Houston, TX), and stored at −80°C until analysis. For analysis, saliva was centrifuged at 16,000g for 20 min and the cell pellet was re-suspended in 200 μL residual supernatant. DNA was extracted from the cell pellet suspension using a QIA-Amp DNA kit (Qiagen, Valencia, CA). Real-time polymerase chain reaction was performed to quantify reactivation of Epstein–Barr virus (EBV), herpes simplex virus 1 (HSV1), and varicella-zoster virus (VZV) DNA using QuantStudio 3 (Applied Biosystems, Inc, Life technologies, Foster City, CA). Viral DNA standards were included in all reactions, and the primers and probes used for EBV, VZV, and HSV1 have been published elsewhere (16,27). Reactivation of herpes viruses was defined as the detection of DNA from that virus in a sample.

Salivary IgA

Saliva samples were collected at PRE and POST to determine SIgA concentrations. Participants placed a cotton absorbent swab in their mouth for 3 min to collect a sample. Samples were immediately frozen and shipped overnight on dry ice to the US Army Research Institute of Environmental Medicine (USARIEM) and stored at −80°C until analysis. Salivary IgA concentrations were quantified using the Salimetrics Salivary Secretory IgA enzyme immunoassay kit (Salimetrics, State College, PA) according to the manufacturer’s instructions.

Blood sample collection

Fasted blood samples were collected at PRE and POST using venipuncture of the antecubital vein approximately 1 h after waking between 4:30 am and 5:30 am. Serum was frozen, shipped overnight on dry ice, and stored at −80°C until analysis. Serum cytokine concentrations (interferon gamma [IFNγ], interleukin-6 [IL-6], IL-10, IL-2, IL-6, IL-8, tumor necrosis factor alpha [TNFα]) were measured using a 6-plex magnetic bead multiplex immunoassay (EMD Millipore, Burlington, MA) according to the manufacturer’s instructions in a 96-well plate. Serum cortisol and high sensitivity C-reactive protein were measured by immunoassay (Immulite 2000; Siemens, Malvern, PA). Whole blood was collected into a 6.0-mL heparin anticoagulant tube and shipped overnight at ambient temperature to JSC for flow cytometry and mitogen stimulation as described below.

Peripheral leukocyte distribution (flow cytometry)

Peripheral immunophenotype was assessed by multiparameter flow cytometry analysis. The fluorescent antibodies used, the fluorophores, cellular targets, and manufacturer of each antibody is delineated in Supplemental Digital Content 1 (see Table, Supplemental Digital Content 1, Single analysis tube fluorescent antibody panel for phenotype analysis, http://links.lww.com/MSS/C738). Thresholds for positive were established using unstained internal controls, as well as isotype-matched cytometry controls. Settings for compensation, to eliminate spectral overlap for the various fluorochromes, were established by running each individual antibody as a single stain, in series, and in relevant pairs, before any multicolor analysis. The gating strategy used to enumerate each subpopulation is shown in Supplemental Digital Content 2 (see Figure, Supplemental Digital Content 2, Bivariate flow cytometry dot plots demonstrating resolution of cellular subpopulations for enumeration and gating strategy, http://links.lww.com/MSS/C739). Populations assessed included leukocyte differential, lymphocyte subsets, cluster of differentiation (CD)4+ and CD8+, memory/naive T-cell subsets, cytotoxic T cells, central memory (CM) T cells, natural killer (NK) cells, B cells, and monocyte subsets. Cell surface markers were stained by combining 100 μL of whole blood and the appropriate amount of each labeled monoclonal antibody as determined by titration to maximize signal, followed by a 20-min incubation at room temperature. Red blood cells were lysed using Beckman-Coulter Optilyse (Beckman Coulter, Brea, CA). Stained leukocytes were then fixed in 1.0% paraformaldehyde in PBS for 10 min and analyzed on a Beckman-Coulter “Gallios” flow cytometer. FCS Express version 6 software was used to complete the analysis. For each individual subject sample analysis 100,000 total events were collected.

Mitogen-stimulated cytokine profiles

Cytokine secretion after stimulation with three separate mitogens was assessed using 150 μL whole blood cultured in premixed aliquots of 850 μL Roswell Park Memorial Institute media with the appropriate mitogen added. Mitogenic stimulation was performed with 0.125 μg anti-CD3 and 0.25 μg anti-CD28 (Beckman Coulter, Brea, CA [to activate T cells only via the T cell receptor]), 10 ng·mL⁻¹ of phorbol myristate acetate (PMA) (Sigma Aldrich, St. Louis, MO) and 2 μg·mL⁻¹ of ionomycin (Sigma Aldrich [as a broader pharmacologic stimulus]), or 20 μg·mL⁻¹ of lipopolysaccharide ([LPS] Sigma Aldrich [monocyte activation]). Cultures were incubated for 48 h and supernatants were removed and frozen until analysis. A 13-plex magnetic bead multiplex immunoassay (EMD Millipore, Burlington, MA) was used to assess the cytokine concentrations. Measured cytokines included IL-1β, TNF-α, IL-6, IL-8, IL-2, IL-7, IFNγ, IL-12, IL-4, IL-5, IL-10, IL-13 and granulocyte-macrophage colony-stimulating factor (GM-CSF). For cytokine analysis, samples were processed according to the manufacturer’s instructions in a 96-well plate. Briefly, diluted supernatant was incubated with magnetic beads coated with capture antibody. The bead cytokine concentrations were then washed and incubated with a biotinylated detection antibody, specific for each cytokine. This mixture was then incubated with Streptavidin PE conjugate, which is the reporter molecule. All samples were batch-analyzed to control inter-assay variability using a Luminex MAGPIX (Luminex Corp, Austin, TX). Data were recorded as mean fluorescence intensity to detect subject-relative cytokine production alterations throughout the study period, which was then converted to relative pg·mL⁻¹ concentration.

Diet quality

Diet quality over the past 3 months was assessed at PRE and POST using the Healthy Eating Index-2015 (HEI-2015) score derived from a self-administered paper and pencil version of the 2014 Block Food Frequency Questionnaire (FFQ) (28). The Block FFQ is validated for use in the general U.S. population (29) and has frequently been used for dietary intake and diet quality assessment in military populations (30). Participants were instructed to select the frequency at which they consumed food and beverage items listed on the FFQ and the usual portion size consumed. Pictures of standard portion sizes were provided to assist in the selection of portion sizes. The HEI-2015 scores (range 0–100) were calculated by Nutrition Quest according to standards for 13 components that reflect adherence to the Dietary Guidelines for Americans (31). For components in which higher intakes are desirable, higher scores indicate higher intake, whereas higher scores in component groups that should be moderately consumed are indicative of lower intake (31). The sum of each component group together equates to the total HEI-2015 score. Volunteers who had complete data for both time points and met the criterion of reporting plausible energy intake (32) at PRE and POST (n = 31) were included in the diet analysis. The remainder of volunteers were considered implausible reporters or did not have complete data for both time points and were excluded from the analyses that included dietary data.

Statistical analysis

Normal distribution of data was assessed using the Shapiro–Wilk test. Comparisons of study outcomes at POST relative to PRE were conducted by paired t tests for data that were normally distributed before or after log₁₀ transformation. If data were not normally distributed after log₁₀ transformation, differences between time points were examined using the Wilcoxon Signed Ranks Test on the untransformed data. Prevalence of herpes virus reactivation at PRE and POST was compared using McNemar’s test. Associations between changes (POST minus PRE) in HEI scores, physical activity levels, and markers of immune function were evaluated using Spearman’s or Pearson’s correlations as appropriate. Statistical analyses were performed using SPSS v.26 (IBM Inc., Armonk, NY). Data are presented as mean ± SD or median [IQR] unless otherwise noted. The α level for significance was set at P < 0.05.

RESULTS

A total of 105 recruits (76% male) were enrolled in the study and 61 (97% male) completed data collection PRE and POST IMT. Of the 105 recruits who were enrolled, 7 withdrew consent before beginning data collection. Of the 98 recruits who participated in PRE data collection, 37 did not complete POST data collection for the following reasons: withdrew consent (n = 6), did not complete OSUT training/withdrew from the Army (n = 16), were in quarantine or at sick call at the time of POST data collection (n = 15). Recruits reported sleeping 315 min·d⁻¹ (120 min·d⁻¹) and 420 min·d⁻¹ (60 min·d⁻¹) the week before PRE and POST, respectively (P = 0.28). The amount of time spent engaged in vigorous physical activities demonstrated a tendency to be lower during the week before PRE 34 min·d⁻¹ (79 min·d⁻¹) relative to the week before POST 60 min·d⁻¹ (69 min·d⁻¹; P = 0.08), whereas time spent engaged in moderate physical activities did not differ (PRE: 34 min·d⁻¹ [49 min·d⁻¹], POST: 43 min·d⁻¹ [39 min·d⁻¹]; P = 0.28). Average perceived stress was higher over the month before POST (16 ± 6) compared with the month before PRE (13 ± 5; P < 0.001). After excluding participants with incomplete or implausible data, no difference in overall HEI-2015 scores over the 3 months before PRE (57.9 ± 10.6) compared with the 3 months before POST (61.8 ± 7.3; P = 0.74) training were observed (see Table, Supplemental Digital Content 3, Dietary intake and Healthy Eating Index scores, http://links.lww.com/MSS/C740).

Salivary markers

The frequency of EBV DNA shedding increased from PRE to POST (21 of 61 volunteers at PRE, vs 31 of 61 volunteers at POST; P = 0.04) (Fig. 1). The median [IQR] EBV viral load did not differ between timepoints (PRE: 0 [2520] copies per mL vs POST 325 [12272] copies per mL; P = 0.31). Neither VZV nor HSV1 DNA were detected at either time point. Salivary IgA secretion rates increased by 60% (P < 0.001) from PRE to POST (Fig. 2).

FIGURE 1.

Incidence of EBV reactivation before PRE and POST IMT. †Different than PRE, P = 0.04, χ² test.

FIGURE 2.

Rate of SIgA secretion PRE and POST IMT. **Different than PRE, P < 0.001. Values are mean ± SD. n = 32.

Peripheral leukocyte distribution

Leukocyte distribution data are presented as relative percent, subset to each parent population (Table 2). Distribution was characterized by a relative increase in lymphocyte percentage and decreases in both granulocytes and monocytes from PRE to POST. This was driven by an increase in the relative percentage of NK cells concomitant with decreases in the relative percentages of B cells and T cells. Within T-cell subsets, effector memory (EM) CD4+ cells, EM CD8+ cells, and CM CD8+ cells all decreased (P < 0.05), whereas naive CD8+, terminally differentiated (TD) CD4+ and TD CD8+ cells increased (P < 0.05) from PRE to POST.

TABLE 2.

Changes in leukocytes pre to post OSUT (n = 56).

	PRE (Relative %)	POST (Relative %)	P
Lymphocytes	39.8 ± 13.0	51.0 ± 10.7	<0.001
Granulocytes	48.3 ± 14.0	41.1 ± 10.1	0.001
Monocytes	5.6 ± 1.8	4.8 ± 1.1	0.002
NK cellsa	6.4 (4.3)	9.2 (6.7)	<0.001
T-NK cellsa	1.3 (1.5)	1.4 (1.8)	0.706
B cellsa	11.3 (5.5)	8.8 (4.8)	<0.001
T cellsb	57.7 (9.5)	54.8 (22.2)	0.007
CD4+b	62.8 (10.6)	57.6 (10.6)	<0.001
CD8+	27.5 ± 6.2	25.7 ± 7.4	0.009
TD CD4+a	0.4 (0.4)	0.5 (0.4)	0.014
TD CD8+a	7.3 (7.1)	10.3 (12.5)	<0.001
Naive CD4+b	50.2 (15.9)	51.3 (17.2)	0.063
Naive CD8+b	56.8 (22.1)	68.7 (32.5)	0.03
EM CD4+b	13.2 (5.5)	11.6 (5.3)	0.023
EM CD8+b	24.0 (12.9)	11.0 (13.2)	<0.001
CM CD4+a	34.5 (12.8)	35.8 (15.5)	0.145
CM CD8+a	10.0 (6.1)	6.8 (5.8)	<0.001

Values are mean ± standard deviation or median (interquartile range). Leukocyte distribution data are presented as relative percent, subset to each parent population. For example, granulocytes, monocytes, and lymphocytes total ~100% of the leukocyte population. T, NK, and B cells total ~100% of the lymphocyte population.

^aLog transformed for analysis.

^bWilcoxon signed ranks test.

Mitogen-stimulated cytokine profiles

Concentrations of all measured cytokines except IL-5 increased from PRE to POST (P ≤ 0.004 for all cytokines) after anti-CD3/CD28 stimulation (Fig. 3). Concentrations of GM-CSF, IL-1β, IL-2, IL-4, IL-6, IL-12, and TNF-α increased from PRE to POST (P ≤ 0.001) after PMA and ionomycin stimulation (see Figure, Supplemental Digital Content 4, Cytokine response to PMA stimulation, http://links.lww.com/MSS/C741). In contrast, IFNγ, IL-1β, IL-6, IL-7, IL-10, IL-12 concentrations all decreased (P ≤ 0.001) from PRE to POST after LPS stimulation, whereas GM-CSF and IL-2 concentrations increased (P ≤ 0.001). No other measured cytokines showed significant differences from PRE to POST after PMA or LPS stimulation (see Figure, Supplemental Digital Content 5, Cytokine response to LPS stimulation, http://links.lww.com/MSS/C742).

FIGURE 3.

Anti-CD3 and anti-CD28 stimulated cytokines PRE and POST IMT. Boxes represent median (IQR), whiskers extend to minimum/maximum values. Unless otherwise noted, data analyzed with Wilcoxon sign ranked tests. #Log10 transformed and analyzed with paired samples t-test. n = 53. *Different than PRE, P < 0.01, **different than PRE, P < 0.001.

Serum markers of inflammation

Circulating serum concentrations of cortisol increased (P ≤ 0.001) from PRE to POST, whereas circulating C-reactive protein concentrations decreased (P ≤ 0.001) from PRE to POST. Circulating serum concentrations of IL-6 (P = 0.003) and IL-8 (P = 0.02) both increased from PRE to POST, but differences were not clinically meaningful (See Table, Supplemental Digital Content 6, Serum markers of stress and inflammation, http://links.lww.com/MSS/C743). There were no significant differences in circulating concentrations of IFNγ, IL-2, IL-10, or TNF-α from PRE to POST.

Diet quality and immune function

Changes in total HEI scores were positively associated with changes in the relative percentages of CD8+ T-cells and EM CD8+ T-cells, and inversely correlated with changes in relative percentage of naïve CD8+ T-cells (Figs. 4A–C). Changes in HEI were negatively associated with secreted IL-1β, IL-6, and IL-12 after anti-CD3/CD28 stimulation (Figs. 4D–F). Change in HEI was positively associated with change in concentration of GM-CSF after LPS stimulation (Fig. 4G). Changes in physical activity were not correlated with changes in immune outcomes (data not shown).

FIGURE 4.

Change in T-cell subset (A–C), anti-CD3 and anti-CD28 stimulated cytokines (D–F), and LPS stimulated cytokine (G) association with change in diet quality from PRE to POST IMT. #Change score assessed with log transformed data. Associations assessed with Pearson correlation (A–F). Association assessed with Spearman correlation (G). Diet quality analyzed for plausible reporters only (n = 28 A–C, G; n = 27 D–F).

DISCUSSION

Few studies have measured changes in immune function during IMT, a transitionary period from civilian to military life wherein infection risk is heightened likely due to a combination of physical and psychological stress, communal living, and sleep deprivation. Primary findings from the current study demonstrate evidence of an increased incidence of viral reactivation, which is consistent with immunosuppression during the stressful IMT environment. However, increases in salivary SIgA secretion, lymphocytosis, maturation of CD4+ and CD8+ T cell subsets, heightened T cell reactivity, and no differences in EBV viral load from PRE to POST IMT indicate that immunocompetence was maintained after IMT, if not improved. Notably, increases in diet quality were associated with CD8+ T-cell maturation and a blunted pro-inflammatory response to pathogen stimulation, indicating a possible role for improved nutrition in supporting favorable changes in immune function.

Initial military training resulted in increased reactivation of EBV, but no reactivation of VZV or HSV1. Previous studies from our group conducted in both terrestrial space analog and space environments indicate that there are different “thresholds” of stress required to trigger reactivation and shedding for each virus (33,34). The pattern of reactivation in this study is similar to what we have observed during terrestrial analogs of spaceflight, specifically winterover at both McMurdo and Palmer Stations in Antarctica. Reactivation of EBV results, in part, from stress-associated decrements in cell-mediated immunity (35,36), with increases in glucocorticoid and catecholamine secretion, shifts in cytokine profiles, and declines in function of leukocyte and lymphocyte subsets all contributing (35). The source and duration of stress also impact viral reactivation (37). In support, 6 wk of physical training in military cadets had no effect on HSV1, human herpesvirus 6 (HHV-6), or EBV reactivation, but EBV reactivation was observed during final examinations, a period that often includes both psychological stress and sleep deprivation as potential sources of immune-suppression (37). Reactivation of EBV and HSV-1 in medical students has also been observed and attributed to psychological stress (38). Conversely, elite athletes who underwent 10 months of intense physical training and competition had no evidence of increased viral reactivation (39), whereas in astronauts, the prevalence of VZV and EBV reactivation is higher on long duration (~26 wk) (16) relative to short duration (~2 wk) space flight missions (27), and mission duration is positively correlated with EBV reactivation (40).

Seemingly at a disconnect with the virus incidence data, is the fact that some of the adaptive immune cellular function findings were elevated, and EBV viral load was unchanged from PRE to POST IMT. Generally, reductions in cytotoxic lymphocyte function precede, or cause latent virus reactivation. However, we have observed elevated EBV shedding while “immune function” was also elevated in one other terrestrial analog, winterover at Concordia Station, Antarctica, which was ascribed to the confounding factor of persistent hypobaric hypoxia at that location. For the military recruits studied herein, several variables might explain increased virus reactivation within the context of increases in T-cell function. These factors include diet (detailed below) and the beneficial effects of physical training on immune function versus any immunosuppressive effects of sleep deprivation or excessive exercise. Indeed, sleep deprivation did not appear to be a major factor in this study as recruits reported sleeping 7 h per night before the POST data collection time point, which is consistent with Army regulations. Further, despite the lack of correlations between changes in physical activity and immune markers, time spent engaged in physical activity trended toward increasing by approximately twofold from PRE to POST OSUT. Other explanatory factors include EBV infection being more common among adults and the ability of EBV to partially reactivate (38).

As viral reactivation is often observed during stress exposure and associated with immunosuppression (41), results from the current study suggest that immunosuppression occurred at some point during IMT but was trending toward recovery and adaptation upon completion of IMT. In support, lymphocytosis was observed from PRE to POST IMT as indicated by relative increases in NK cell populations and increased proportions of terminally differentiated T cells relative to memory T cells, reflecting maturation of T cell subsets (42). Further, T-cell activity was heightened POST IMT as evidenced by higher cytokine concentrations after anti-CD3/CD28 and PMA stimulation relative to PRE. Finally, SIgA secretion, a marker of mucosal immunity that is inversely correlated with infection risk in athlete and military populations (3,11), increased from PRE to POST IMT. These observations are consistent with a more active adaptive and mucosal immune response that may reflect a culmination of immune adaptation to pathogen exposures, stress, and physical training during IMT. Studies conducted during shorter, more intense military trainings have reported decreases in lymphocyte populations during training, which have been interpreted as evidence of immune-suppression (14,43). In contrast, studies conducted during military trainings similar in duration to IMT have reported minimal or only transient effects on monocytes, granulocytes and lymphocytes (17,44). Although differences in study populations (new recruits versus tenured military personnel), training loads, living environments, and other factors complicate comparisons between studies, these data collectively suggest that the initial days or weeks of training may be when trainees are at greatest risk of infection.

The observed leukocyte distribution and mitogen stimulated cytokine profiles measured at PRE may support that hypothesis. Specifically, the presence of an innate and early adaptive immune response at PRE was evidenced by increased relative proportions of granulocytes, monocytes, B-cells and memory T-cells in combination with higher cytokine responses to LPS-stimulation compared with POST. Further, CD28/CD3- and PMA-stimulated release of most cytokines, to include both Th1 and Th2 cytokines were lower at PRE, indicating reduced relative suppression of both cellular and humoral immunity (45–47). Collectively, these findings may indicate either early stages of infection or a suppression of the adaptive immune system at PRE. This may suggest that the beginning of IMT, or even before arrival at IMT, is the optimal time for introducing immune-enhancing interventions in military recruits. However, results should be interpreted cautiously. For example, effects of vaccinations received during the initial days of IMT on the immune markers observed herein cannot be ruled out given that recruits had recently received mandatory vaccinations before PRE. As such, studies that measure changes in immune function relative to vaccinations and the development of illness and infection, and that include more frequent assessments than those included herein, especially during the early weeks of IMT, are necessary to more clearly define the time course of greatest infection risk and to determine optimal time points for introducing immune-enhancing interventions.

Nutrition has a well-established role in supporting immune function (48). In this study, average diet quality as measured by the Healthy Eating Index did not change during 22 wk of IMT, and both PRE and POST average total HEI scores translated to a diet that “needs improvement” (49). However, at the individual level, increases in HEI scores correlated with CD8+ cell maturation, indicating a potential influence of improved diet quality on the adaptive immune system and T cell–mediated cytotoxicity. In addition, an inverse association between changes in HEI scores and changes in mitogen-stimulated release of multiple pro-inflammatory cytokines (IL-1, IL-6, IL-12) was observed. These results may suggest that improvements in diet quality had a protective role in regulating, but not preventing, the pro-inflammatory response to antigens. Although these associations are consistent with an effect of nutrition on adaptive immunity, the association between nutrition and innate immunity is less clear. No associations between HEI and granulocytes, monocytes, or NK cells were observed which suggests diet quality may have had less impact on the innate immune responses. Conversely, the positive correlation between changes in HEI scores and changes in LPS-stimulated GM-CSF support a link between diet quality and innate immunity. Lipopolysaccharide is synthesized by gram negative bacteria and activates innate immune cells (15,50). Given that GM-CSF is an important immunomodulator triggered by pathogenic activity, a heightened GM-CSF response to LPS is considered positive (46). To the best of our knowledge these findings are the first to link diet quality to biomarkers of innate and adaptive immune function in the IMT environment. Although correlational, the results do support the need for future research aiming to determine the effects of improved diet quality on immune function during IMT.

Strengths of this study include the longitudinal measurement of multiple markers of innate and adaptive immune function within a logistically challenging field environment. One study limitation is the observational design, which cannot demonstrate cause and effect relationships. Additional limitations include not obtaining more frequent measurements and unmeasured effects resulting from infection mitigation strategies implemented in response to the COVID-19 pandemic. Both of these limitations are a direct result of unanticipated challenges resulting from the COVID-19 pandemic and prevented assessment of immune function and potential contributors to changes in immune function such as physical activity levels, sleep and dietary intake during and after the most novel and strenuous phases of the training period as originally planned. In addition, some aspects of training were altered to protect against COVID-19 transmission. Behavioral modifications included mask wearing and social distancing, which likely reduced pathogen exposures and subsequent immune adaptations experienced during training, and may reduce the generalizability of study results.

An additional limitation was that the dietary intake analysis was restricted to a subset of volunteers due to implausible and incomplete data. This reduced power for detecting associations between dietary and immune variables. Also, measurement of physical activity and sleep levels, both factors that impact immune function, were self-reported and only asked participants to consider the week before data collection. As a result, changes in those outcomes throughout the training period may not have been fully captured. The high attrition rate and the fact that 15 recruits could not complete POST data collection due to quarantine or possible illness are also limitations that may bias results toward more favorable immune profiles at the POST time point. Finally, despite efforts to enroll women in the study, the final cohort was primarily male, which prevented assessment of sex differences. Future studies should aim to include a larger sample size and enroll more women to increase generalizability and determine sex differences.

CONCLUSIONS

In conclusion, we suggest that recruits undergoing IMT represent a unique interplay of age, sleep, crowded living conditions, exercise training, and stress. This is not dis-similar to astronauts during spaceflight who are influenced by stress, isolation, microgravity, and other stressors although recruits appear to be lower on the continuum of stress and immune compromise (33). Study findings nonetheless suggest that immunosuppression likely occurred during the 22 wk of IMT. However, importantly, immunocompetence was maintained at the end of the training period. Although causality could not be determined, diet quality may be a contributing factor to changes in cell-mediated immunity during military training. Findings highlight the need for additional research utilizing a larger sample size, more frequent outcome assessments, especially during the initial weeks of training, and an increased proportion of women to further characterize changes in immune function during IMT and identify factors, such as nutrition, which modulate that response.