Load carriage aerobic exercise stimulates a transient rise in biochemical markers of bone formation and resorption

Abstract

Exercise can be both anabolic and catabolic for bone tissue. The temporal response of both bone formation and resorption following an acute bout of exercise is not well described. We assayed biochemical markers of bone and calcium metabolism for up to 3 days after military-relevant exercise. In randomized order, male (n = 18) and female (n = 2) Soldiers (means ± SD; 21.2 ± 4.1 years) performed a 60-min bout of load carriage (30% body mass; 22.4 ± 3.7 kg) treadmill exercise (EXER) or a resting control trial (REST). Blood samples were collected following provision of a standardized breakfast before (PRE), after (POST) exercise/rest, 1 h, 2 h, and 4 h into recovery. Fasted samples were also collected at 0630 on EXER and REST and for the next three mornings after EXER. Parathyroid hormone and phosphorus were elevated (208% and 128% of PRE, respectively, P < 0.05), and ionized calcium reduced (88% of PRE, P < 0.05) after EXER. N-terminal propeptide of type 1 collagen was elevated at POST (111% of PRE, P < 0.05), and the resorption marker, C-terminal propeptide of type 1 collagen was elevated at 1 h (153% of PRE, P < 0.05). Osteocalcin was higher than PRE at 1 through 4 h post EXER (119%–120% of PRE, P < 0.05). Sclerostin and Dickkopf-related protein-1 were elevated only at POST (132% and 121% of PRE, respectively, P < 0.05) during EXER. Trivial changes in biomarkers during successive recovery days were observed. These results suggest that 60 min of load carriage exercise elicits transient increases in bone formation and resorption that return to pre-exercise concentrations within 24 h post-exercise.

NEW & NOTEWORTHY In this study, we demonstrated evidence for increases in both bone formation and resorption in the first 4 h after a bout of load carriage exercise. However, these changes largely disappear by 24 h after exercise. Acute formation and resorption of bone following exercise may reflect distinct physiological mechanoadaptive responses. Future work is needed to identify ways to promote acute post-exercise bone formation and minimize post-exercise resorption to optimize bone adaptation to exercise.

INTRODUCTION

Repeated physical activity increases bone strength by inducing anabolic bone formation. Studies in military populations who engage in physical training regimes have reported anabolic bone formation at the tibia, as evidenced by increases in volumetric bone mineral density of 2%–4% following 8 to 44 wk of military training in young adults (1–4). Exercise also stimulates bone catabolism in response to declining concentrations of serum-ionized calcium (5) and for replacing fatigue damage that can occur with repetitive loading during prolonged physical activity (6, 7). Over time, exercise-induced bone resorption can offset gains in bone strength derived from exercise-induced formation. Accordingly, optimizing the processes of anabolic bone formation while minimizing bone resorption is needed to fully realize the benefits of exercise. Capturing circulating biochemical markers of bone cellular activity in response to exercise under controlled laboratory conditions can provide the evidence of the many potential factors that influence the effects of exercise on bone (8). Although numerous bone biomarker studies have been conducted (9, 10), the study methods vary greatly, including in exercise duration and intensity, selected biochemical markers of bone activity, and in the time course of biological sample collection (11). One consistent finding is that for activity to illicit anabolic effects on bone, it must have sufficient loading or impact in multiple planes of direction (10).

Ideally, biomarker and exercise studies should include markers of osteocyte, osteoblast, and osteoclast activity, and allow for enough time following exercise to capture the initial phases of targeted bone remodeling. Several days of post-exercise biological sample collection may be required to capture biochemical markers of osteoblast and osteoclast activity that would follow differentiation, recruitment, and initiation of anabolic and catabolic activity by these cells. Furthermore, initiation of bone resorption for targeted remodeling and repair of areas of microdamage occurs between 24 h and 10 days following mechanical loading (6). Finally, another important methodological consideration is that there are nonmechanical stimuli that may affect circulating biochemical markers of bone cellular activity, such as macronutrient and micronutrient intake and diurnal factors (12–14). These factors must be controlled to delineate the effects of the exercise on bone cellular activity.

Accordingly, we evaluated the responses of serum biochemical markers of osteocyte, osteoblast, and osteoclast activity, as well as ionized calcium and parathyroid hormone (PTH), to a self-paced aerobic exercise bout in young, healthy men and women. We evaluated bone biochemical markers before exercise, at several time points after exercise, and throughout 3 days following exercise.

MATERIALS AND METHODS

Design

A block-randomized, crossover design was used to compare the effects of a load carriage aerobic exercise on circulating markers of bone formation and resorption. Participants completed two experimental trials, separated by 14.5 ± 5.4 days. One trial included 60 min of load carriage exercise (EXER) and the other comprised 60 min of rest (REST). We chose the load carriage exercise because it is an activity familiar to Soldiers and was expected to induce a high degree of mechanical loading. The resting trial was implemented to control for known circadian patterns in markers of bone metabolism. Blood samples collected before and after exercise in EXER were collected at the same time during the REST trial. Ad libitum breakfast was provided and was matched between trials in terms of content and timing. Blood samples were collected for three follow-up mornings after EXER only.

Participants

Twenty Soldiers (18 male and 2 female) participated in the study. Participant demographics can be found in Table 1. Potential participants were included in the study if they were active-duty Soldiers between the ages of 18 and 42 yr, with a current passing score on their most recent physical fitness test. Volunteers had less than 1 yr of military service. Participants had to be healthy without history of endocrine, cardiovascular, renal or bone-modifying disorders, or an injury that could be aggravated by load carriage exercise. Females who were pregnant or lactating in the previous 6 mo, and anyone donating blood in the previous 8 wk were excluded. Female participants were studied within their follicular menstrual cycle phase. Before enrollment, participants were briefed on the study procedures in the presence of an ombudsperson and provided their written informed consent. The study was approved by the Institutional Review Board at the US Army Medical Research and Development Command. The investigators have adhered to the policies for the protection of human subjects as prescribed in the Department of Defense instruction 3216.02 and the research was conducted in adherence with the provisions of 32 Code of Federal Regulations Part 219.

Table 1.

Participant demographics

Characteristic
Sex, n (%)
Male	18 (90)
Female	2 (10)
Race, n (%)
White/Caucasian	12 (60)
Black/African American	7 (35)
Other	1 (5)
Age, yr	21.2 ± 4.1
Body weight, kg	74.7 ± 12.3
Height, cm	171.2 ± 12.6
V̇o2peak, mL·kg−1·min−1	50.9 ± 5.3

All values are means ± SD.

Load Carriage Protocol

Participants performed a week of baseline testing that included an incremental treadmill V̇o_2peak test on one occasion and three 10- to 15-min practice sessions for the load carriage task on the treadmill. On the trial day (Day 0), participants arrived at the laboratory before 0600, after an overnight fast, and had an indwelling venous (IV) catheter inserted into a vein in the antecubital space. Participants were block-randomized to either complete their EXER or REST trial first. For the EXER trial, a 60-min load carriage time trial commenced at 0900. Participants wore a weight vest loaded with 30% of their body weight during the time trial. The treadmill was set at a constant 1% grade and the participants were instructed to cover as much distance as possible in the 60-min time period by self-selecting and adjusting the speed. Participants were blinded to the treadmill speed and grade during the test, and the only feedback given was the amount of time remaining. The load carriage exercise with a weighted vest was chosen because of its ease of distributing, balancing, and adjusting the load. Similar self-paced treadmill tests have been used successfully in previous studies (15, 16). We chose to use 60-min duration versus a fixed distance so we could precisely match sample collection time points to the actual time in both EXER and REST. Study staff recorded distance every 5 min and heart rate (HR) and rating of perceived exertion (RPE) (17) every 15 min until the end of the task. Exercise intensity was estimated by the percentage of actual maximal heart rate. Water was available ad libitum. During the same time period in the REST trial, participants remained in the laboratory and engaged in activities such as watching movies or reading. Blood samples, feeding, and all other tests during a participant’s second trial (REST or EXER) were time-matched to the first trial to within ±5 min. Participants were instructed to restrict their physical activity to light activities only for 2 days before their Day 0 (for both EXER and REST) and for the three follow-up days after the EXER trial.

Pre-Exercise Feeding

Breakfast was provided at 0700 during each trial, and participants were instructed to choose foods that would not cause gastrointestinal upset during their treadmill test. Breakfast options included pre-packaged breakfast sandwiches, oatmeal, bakery items, and sports drinks. Participants consumed identical breakfast meals during REST and EXER trials. Water was provided ad libitum throughout the testing day. On average, the breakfast meal consisted of 820 ± 243 kcal, 278.5 ± 98.1 mg calcium, and was 39.6 ± 5.5% fat, 11.7 ± 3.8% protein, and 50.0 ± 11.7% carbohydrate. Dietary intake data were calculated using ESHA Food Processor v. 11.6.522 (Salem, OR).

Blood Sampling

To assess the short-term biomarker exercise response, blood was collected from the IV catheter at the following time points on Day 0, immediately before the time trial (PRE), immediately upon completion of the time trial (POST), and at 1 h, 2 h, and 4 h post-EXER or REST. To assess a longer-term biomarker response, fasted AM samples were collected by venipuncture at 0630 on REST and EXER days and the three successive days following the EXER trial only (Day +1, Day +2, and Day +3).

Biochemical Analyses

Whole blood was immediately assayed for hemoglobin and hematocrit using a Beckman Coulter Ac2Diff hematology analyzer (Beckman Coulter, Brea, CA), and for ionized calcium (iCa) on an Abbott ISTAT analyzer (Abbott Laboratories, Chicago, IL). Serum was assayed for markers of calcium metabolism, including intact parathyroid hormone (PTH), on a Siemens Immulite 2000 immunoassay system. Serum phosphorus (PHOS) was assayed on a Siemens Dimension clinical chemistry system (Siemens Healthcare Diagnostics). Serum markers of osteoclast activity, including C-terminal propeptide of type 1 collagen (CTX, Immunodiagnostic Systems Inc., Gaithersburg, MD) and tartrate-resistant acid phosphatase (TRAP5b, Quidel Corp., San Diego, CA) were assayed by enzyme-linked immunoassay (ELISA). Serum markers of osteoblast activity included N-terminal propeptide of type 1 collagen (P1NP) radioimmunoassay (Immunodiagnostic Systems Inc.), bone alkaline phosphatase ELISA (BAP, Quidel Corp.), and total osteocalcin (OCN) were assayed via multiplexing kit (MilliporeSigma, Burlington, MA). Osteocyte activity was assessed via multiplexing (MilliporeSigma) for osteoprotegerin (OPG), sclerostin (SCL), Dickkopf-related protein (DKK1), and fibroblast-related protein 23 (FGF23). All samples were assayed in duplicate and re-assayed in the event of poor replication of the duplicates. Interassay coefficients of variation (CVs) were as follows: 7.8% for PTH, 4.3% for PHOS, 1.2% for iCa, 4.1% for BAP, 6.3% for TRAP5b, 6.9% for CTX, 3.8% for P1NP, 9.5% for OCN, 7.9% for OPG, 12.6% for SCL, 10.8% for DKK1, and 9.1% for FGF23.

Statistical Analysis

Sample size was calculated using G*Power software to have 80% power with an α = 0.05 for a repeated measures design. Our means and standard deviations for the biomarkers were calculated in preliminary unpublished data collected in our laboratory. It was estimated to detect a 20% unadjusted, absolute biomarker change in bone formation markers induced by exercise within individuals, a minimum of 20 participants was required. Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS v. 26, Chicago, IL) and Prism GraphPad (v. 8.2.1, San Diego, CA) software. Two-way repeated measures analysis of variance (or mixed-model in a case with a missing data point) with Bonferroni’s post hoc test was used when there was a significant interaction effect to determine changes in biomarkers over time and between trials. Effects of time were compared with the PRE time point. To account for changes, over and above that of the circadian pattern, the effects of condition (EXER vs. REST) were compared at each sample time point. For both EXER and REST, acute phase biomarker concentrations are expressed as % change from PRE and have been adjusted for changes in plasma volume using hemoglobin and hematocrit (18), unless otherwise noted. The fasted AM samples on EXER and REST and Days +1, +2, and +3 samples were not adjusted and are presented as biomarker concentrations. Significance was set at P < 0.05. Data presented are means ± SD.

RESULTS

Self-Paced Load Carriage Exercise

Table 2 presents the physical performance data from the 60-min load carriage bout. The participants covered 6.15 ± 0.75 km (range 5.1–7.5 km), at an average RPE of 13.6 ± 2.9 (on a scale of 6–20 where 6 is resting and 20 is all out exertion), and percentage of maximal heart rate of 84.4 ± 8.5%. During the last minute of the exercise bout, RPE was 16.2 ± 2.0, and the percentage of maximal heart rate was 91.6 ± 6.0%. All participants adjusted pace accordingly and many alternated fast walking with periods of slow running throughout the 60-min test.

Table 2.

Load carriage exercise performance data

	15 min	30 min	45 min	60 min
Cumulative distance, km	1.52 ± 0.28	3.03 ± 0.46	4.57 ± 0.63	6.15 ± 0.75
Split pace, km/h	6.09 ± 1.14	6.04 ± 0.79	6.13 ± 0.85	6.32 ± 0.64
% Maximum heart rate	80.8 ± 9.2	81.8 ± 7.0	83.3 ± 7.3	91.6 ± 6.0
RPE	10.1 ± 2.0	13.4 ± 1.6	14.6 ± 1.8	16.2 ± 2.0

All values are means ± SD. RPE, rating of perceived exertion (6–20 scale).

Short-Term Response (0–4 h after Exercise)

There were no differences (P > 0.05) in the biomarkers assessed at the PRE time point between EXER and REST conditions except for TRAP5b and BAP, which were both higher (P < 0.05) at PRE during EXER than REST (Supplemental Table S1, https://doi.org/10.6084/m9.figshare.21440616).

Calcium and phosphorus metabolism.

During EXER, effects of time were observed as PTH (Fig. 1A) was 208% higher than PRE at POST, 2 h (173%), and 4 h (225%, all P < 0.05). The 2-h time point was not different from REST (P > 0.05). During EXER, iCa (Fig. 1B) was 88% of PRE at the POST timepoint, returned to be equal to PRE at 1 h and then higher at 2 h and 4 h (105% and 104%, respectively, P < 0.05) compared with PRE. iCa was unchanged during REST. PHOS (Fig. 1C) was 128% higher than PRE at POST, and again at 4 h (141%, P < 0.05).

Figure 1.

Biomarkers of calcium metabolism in response to 60-min load carriage exercise (EXER) or rest (REST). Data are means ± SD, adjusted for plasma volume changes from PRE, and expressed as % change from pre-exercise. Pre, immediately pre-exercise; Post, immediately post-exercise; and 1 h, 2 h, and 4 h after exercise. A: PTH, parathyroid hormone; B: iCa, ionized calcium; C: PHOS, phosphorus. *Significantly different from Pre, P < 0.05. #Significantly different from respective time point in REST, P < 0.05.

Markers of bone resorption.

CTX (Fig. 2A) was higher than PRE during EXER 1 h after exercise (153% of PRE, P < 0.05) and was not different from corresponding REST at subsequent time points (P > 0.05). TRAP5b (Fig. 2B) was higher than REST at 4 h after exercise (P < 0.05). OPG (Fig. 2C) was increased from PRE during EXER at every post-exercise time point (113%–119%, P < 0.05). OPG was unchanged during REST.

Figure 2.

Biomarkers of bone resorption in response to 60-min load carriage exercise (EXER) or rest (REST). Data are means ± SD, adjusted for plasma volume changes from PRE, and expressed as % change from pre-exercise. Pre, immediately pre-exercise; Post, immediately post-exercise; and 1 h, 2 h, and 4 h after exercise. A: CTX, C-terminal propeptide of type 1 collagen; B: TRAP5b, tartrate-resistant acid phosphatase; C: OPG: osteoprotegerin. *Significantly different from Pre, P < 0.05. #Significantly different from respective time point in REST, P < 0.05.

Markers of bone formation.

P1NP (Fig. 3A) was higher than PRE at POST (111%, P < 0.05) during the EXER trial. Although P1NP increased 1 h through 4 h in both trials but were not different between EXER and REST (P > 0.05). BAP (Fig. 3B) was unchanged in both trials. OCN (Fig. 3C) during EXER was higher than PRE at 2 h and 4 h (120% and 119%, respectively, P < 0.05).

Figure 3.

Biomarkers of bone formation in response to 60-min load carriage exercise (EXER) or rest (REST). Data are means ± SD, adjusted for plasma volume changes from PRE, and expressed as % change from pre-exercise. Pre, immediately pre-exercise; Post, immediately post-exercise; and 1 h, 2 h, and 4 h after exercise. A: P1NP, N-terminal propeptide of type 1 collagen; B: BAP bone alkaline phosphatase; C: OCN total osteocalcin. *Significantly different from Pre, P < 0.05. #Significantly different from respective time point in REST, P < 0.05.

Markers of osteocyte activity.

SCL (Fig. 4A) was higher than PRE during EXER at POST (132% of PRE, P < 0.05) and again at 4 h (125% of PRE, p < 0.05). Although SCL was elevated over PRE at 1 h and 2 h during EXER, this was not different from the corresponding time points during REST (P > 0.05). DKK1 (Fig. 4B) was higher than PRE during EXER at POST only (121% of PRE, P > 0.05). Small differences were found in FGF23 between REST and EXER at 1 h and 4 h (Fig. 4C).

Figure 4.

Biomarkers of osteocyte activity in response to 60-min load carriage exercise (EXER) or rest (REST). Data are means ± SD, adjusted for plasma volume changes from PRE, and expressed as % change from pre-exercise. Pre, immediately pre-exercise; Post, immediately post-exercise; and 1 h, 2 h, and 4 h after exercise. A: SCL, sclerostin; B: DKK1, Dickkopf-related protein 1; C: FGF23, fibroblast growth factor 23. *Significantly different from Pre, P < 0.05. #Significantly different from respective time point in REST, P < 0.05.

Follow-up Day Response

Calcium and phosphorus metabolism.

There were no differences in any of the fasted, absolute concentrations in PTH and PHOS (P > 0.05). However, iCa was higher (P < 0.05) on the morning of Day 2 than on the morning of REST or EXER (Table 3).

Table 3.

Fasted, morning biomarker concentrations for resting trial (REST), exercise trial (EXER), and three follow-up days after EXER (Day 1, Day 2, Day 3)

Biomarker	REST	EXER	Day +1	Day +2	Day +3
PTH, pg/mL	45.4 ± 27.8	45.8 ± 22.6	41.9 ± 25.7	47.4 ± 30.2	51.9 ± 30.9
iCa, mg/dL	4.77 ± 0.18	4.78 ± 0.17	4.83 ± 0.17	4.88 ± 0.14*#	4.82 ± 0.18
PHOS, mg/dL	4.25 ± 0.62	4.22 ± 0.56	4.30 ± 0.63	4.40 ± 0.72	4.45 ± 0.70
CTX (ng/mL)	1.06 ± 0.59	1.12 ± 0.69	1.11 ± 0.70	1.12 ± 0.78	1.11 ± 0.76
TRAP5b, U/L	3.66 ± 1.47	3.69 ± 1.40	3.71 ± 1.41	3.79 ± 1.55	3.84 ± 1.51
P1NP, ng/mL	94.8 ± 53.0	108.4 ± 66.0	99.3 ± 44.8	110.0 ± 59.2	105.3 ± 53.8
OPG, pg/mL	319.2 ± 70.8	320.9 ± 58.0	318.9 ± 58.0	325.3 ± 54.3	321.6 ± 54.2
BAP, U/L	24.2 ± 10.9	24.6 ± 10.2	24.7 ± 11.1	24.5 ± 10.2	24.9 ± 10.4
OCN, ng/mL	25.7 ± 11.0	25.2 ± 10.8	24.1 ± 10.8	23.4 ± 10.9*#	23.4 ± 10.2*#
SCL, pg/mL	1833 ± 817	1830 ± 776	1721 ± 752	1684 ± 655	1670 ± 701
DKK1, pg/mL	1666 ± 482	1822 ± 524	1784 ± 487	1884 ± 449	1769 ± 512
FGF23, pg/mL	32.3 ± 10.1	33.8 ± 13.4	32.4 ± 12.2	31.7 ± 9.5	32.0 ± 9.7

All values are means ± SD. *different from REST, P < 0.05, #different from EXER, P < 0.05. BAP, bone alkaline phosphatase; CTX, C-terminal propeptide of type 1 collagen; DKK1, Dickkopf-related protein 1; FGF23, fibroblast-related protein 23; iCa, ionized calcium; OCN, osteocalcin; OPG, osteoprotegerin; PHOS, phosphorus; PTH, parathyroid hormone; P1NP, N-terminal propeptide of type 1 collagen; SCL, sclerostin; TRAP5b, tartrate resistant acid phosphatase.

Markers of bone resorption.

There were no changes among absolute concentrations of CTX, TRAP5b, or OPG during any of the AM samples (P < 0.05).

Markers of bone formation.

The fasted morning OCN on Day +2 (23.4 ± 10.9 ng/mL) and Day +3 (23.4 ± 10.2 ng/mL) was lower than both REST (25.7 ± 11.0 ng/mL) and EXER (25.2 ± 10.8 ng/mL) AM samples (P < 0.05).

Markers of osteocyte activity.

There were no differences in the AM samples on any study morning for SCL, DKK1, and FGF23 (P > 0.05).

DISCUSSION

This study was undertaken to evaluate the temporal responses of bone to aerobic exercise using circulating biochemical markers of calcium regulation, osteocyte, osteoblast, and osteoclast activity. We found that biochemical markers of calcium regulation, bone formation, and bone resorption were elevated over baseline in the first 4 h after a single bout of aerobic exercise. However, these changes in bone metabolism did not persist into successive days of recovery. These observations suggest that the bone catabolic and anabolic responses to a single bout of exercise are initiated in the hours directly after exercise, and studies designed to investigate factors that may influence these adaptive responses may not gain added value from measuring serum bone biomarkers in successive days after exercise.

Calcium and Phosphorus Metabolism

Immediately after exercise, we found a decrease in circulating iCa concentrations, an increase in circulating PHOS concentrations, and an increase in PTH (Fig. 1), which are responses similar in magnitude to other studies using 60 min of treadmill running at similar intensities (79%–87% maximum heart rate) (19, 20) or 60 min of cycling at 80% maximum heart rate (5). Exercise has been shown to rapidly decrease circulating iCa and result in a homeostatic increase in PTH (5, 21), which may be dependent on exercise intensity(22). As PHOS is coupled with calcium in hydroxyapatite crystals in bone, the observed increase in circulating phosphorous may be reflective of the bone resorption that accompanies increases in PTH (21). The mechanisms initiating the decline in iCa with exercise have not been fully elucidated; however, calcium-containing sweat loss during prolonged exercise and calcium uptake into skeletal muscle are postulated as contributing factors (23). Supplementing with dietary calcium before and during exercise or use of a calcium clamp technique (5) have all been shown to attenuate declines in iCa (24, 25), and/or increases in PTH (24, 26, 27), and the biochemical marker of bone resorption, CTX, in response to exercise (24, 27). These studies suggest that attenuating decreases in circulating calcium is a potential means for minimizing bone catabolism during exercise.

Markers of Bone Resorption

The ∼53% increase in CTX at 1 h after exercise and modest increase in TRAP5b at 4 h (Fig. 2) both indicate that the exercise protocol used in this study did induce bone catabolism in the first 4 h of recovery. As a biomarker of bone resorption, post-exercise increases in CTX have been the most commonly reported response to prolonged, strenuous exercise (9, 28–30) and have been shown to stay elevated for as long as 3 h after exercise (31). The increases in TRAP5b 4 h after exercise in our study is a much less common finding. TRAP5b has been reported to increase after 8–12 wk of military training (4, 32, 33), potentially because TRAP5b is considered an indicator of osteoclast number rather than activity (34), and the time it would take for the mechanical stimulation during exercise to translate to an increase in osteoclast differentiation and recruitment, is likely outside the window of the acute phase after exercise (34).

Another mode of osteoclast resorption following exercise is the process of targeted remodeling, in which fatigue damage generated during exercise is targeted for osteoclastic removal and subsequent deposition of new bone by osteoblasts within the same location, with the resorption phase of targeted remodeling is reported to being with 1 to 14 days (6) Although we collected samples for 3 days after the exercise bout, we did not observe elevations in CTX or TRAP5b in this time period. It is possible that the elevation of TRAP at +4 h may be indicative of increases in osteoclast number in support of this process, and that resorption of the fatigue-damaged bone occurred in the days following the sampling. It is also possible that the exercise was not strenuous enough to induce fatigue damage in our relatively fit population, or that numerous bouts of strenuous activity are needed to stimulate this physiological process.

We observed a post-exercise increase in OPG of ∼20%, which remained elevated throughout the 4-h recovery period and returned to baseline for the three follow-up days. This observation of an acute increase following exercise is similar to responses seen in other studies (30, 35–37). However, unlike our study, OPG has been reported to be elevated over baseline the day after exhaustive treadmill running (36) and a marathon (38), and 3 days after an ultramarathon (30). OPG, which is secreted from the osteoblasts, functions as a decoy receptor for RANKL and inhibits osteoclastogenesis. OPG is often reported in ratio format with RANKL; however, RANKL is often difficult to interpret in the circulation because it has been found to be undetectable in 50%–70% of the samples, as was the case in our study (data not reported) (39, 40). The increase we observed in OPG suggests that our exercise protocol resulted in antiresorptive activity only during the immediate recovery period following exercise.

Markers of Bone Formation

We saw an increase in the bone formation marker P1NP immediately post-exercise (Fig. 3), which returned to the equivalent concentration of the rest condition by 1 h and for the remainder of the recovery period and follow-up days. This observation is congruent with the concept that exercise leads to bone anabolism, or adaptive bone formation (8). Others have reported transient P1NP changes after moderate to high-intensity exercise. For example, treadmill running at intensities of 55%–75% V̇o_2max to exhaustion resulted in transient 10%–33% increases in P1NP over baseline (19, 22, 28) and a study using bilateral jumping reported a positive correlation between P1NP response and exercise loading parameters (41). As in our data, P1NP was found to be unchanged during successive days of recovery after treadmill running in active and endurance-trained men (36). However, other studies report no post-exercise increase in P1NP after cycling exercise in adolescents (42). Although P1NP is considered one of the most sensitive markers of bone formation, it is important to note that increases in P1NP may be related to collagen synthesis in tendons and muscles (43, 44).

Besides P1NP, we assessed two additional markers of bone formation, OCN and BAP. OCN is one of the most abundant proteins in bone, is a marker of osteoblast activity (45), and its uncarboxylated form may have a role in energy metabolism (46). In our study, OCN was increased by 115% at 1 h and remained elevated at 119% at 4 h after exercise. Like P1NP, these elevations in OCN likely represent bone anabolism following exercise. It has recently been suggested that OCN may best respond to protocol using weight-bearing activities compared to activities with low-impact forces (11). In our data, we did observe that OCN was lower at Day 2 and Day 3 than the morning fasted samples of both REST and EXER. Our participants were Active-Duty Soldiers who engage in regular physical exercise. In our study, they were instructed to remain inactive during the 3 days of recovery (via being placed on restricted physical activity with their leadership), and thus, inactivity may have suppressed normal osteoblast activity.

The final biochemical marker of bone formation we assayed was BAP, a protein produced during the mineralization of bone that has been demonstrated to increase with military training (4, 47, 48). We did not find any change in BAP in the first 4 hours after exercise or in subsequent days of recovery. In other acute exercise studies, small post-exercise increases in BAP have been reported (26, 37, 49), which in one study was abolished after adjusting for changes in plasma volume as was done in this study (26).

Although we only observed increases in P1NP and OCN in the hours following exercise, no biochemical marker of bone formation was elevated 1 through 3 days after exercise. These observations are consistent with other studies reporting no changes in the biomarkers examined from 24 h to 48 h after exercise (20, 28, 50, 51). Unlike these aforementioned studies, reduced (52) and increased (22) markers of bone formation have been reported after a single bout of running during 1 to 4 days of recovery. Collectively, our study and others suggest that increased bone formation that can follow exercise may be transient, and that changes in bone density and morphology that accompany chronic physical activity may be the result of cumulative effects of repeated bouts of physical activity.

Markers of Osteocyte Activity

Immediately after exercise, we found that SCL, an inhibitor of bone formation secreted by osteocytes, increased to 131.6% of the pre-exercise concentrations of SCL, which returned to resting levels at 1 h and 2 h, but then was again elevated at 4 h of recovery. The immediate post-exercise increase was similar in timing and magnitude to other reports (50, 53–55). This increase in SCL following exercise that we and others have observed is perplexing, given that suppression of SCL with exercise in mechanically loaded osteocytes is thought to be one of the primary mechanisms of anabolic bone formation (56). Other studies that have reported transient increases in SCL after exercise suggest the increases may be a result of previously synthesized SCL released into the circulation or even a decrease in kidney clearance (55).

We also assayed for other markers of osteocyte activity, DKK1 and FGF23. DKK1, like SCL, is a negative regulator of bone formation secreted from osteocytes. We observed no changes in the DKK1 response after the exercise. FGF23, a known regulator of phosphorus metabolism, was elevated over REST at 1 h and 4 h post-exercise. FGF23 promotes phosphaturia, which may be a compensatory response to our observed post-exercise increases in circulating PHOS. In support of this interpretation, FGF23 has been reported to increase in runners immediately following an ultramarathon (57). Collectively, our results suggest that a 60-min bout of moderate-intensity aerobic exercise can stimulate osteoclast, osteoblast, and osteocyte activity indicative of calcium regulation and bone anabolism and catabolism.

Practical Implications

The observations that both formation and resorption occur in response to exercise have important implications for military training and exercise interventions—specifically, that to optimize the bone anabolic effects of exercise, certain environmental and lifestyle practices should be considered. These factors include maintaining adequate energy availability, macro- and micronutrient intake, and sleep, as well as abstaining from analgesic drugs that may inhibit anabolic bone formation (9, 12, 13). There are also practical countermeasures that may prevent bone resorption with exercise, such as limiting the repetitive loading that can generate fatigue damage (8) and supplementing calcium intake during exercise (5).

Besides observing osteoblast and osteoclast activity following exercise, we also observed increased osteocyte activity, as well as indications of calcium and phosphorus regulation. These observations likely reflect the many physiological responses and of the bone cellular activity that accomplishes tissue- and organ-level bone functional adaptation to exercise (8, 9). As this study and others(10) continue to delineate the magnitude and time course of these physiological responses, means of promoting bone anabolism and preventing catabolism could help optimize the bone response to exercise and military training.

Strength and Limitations

Our study had several limitations. First, the mode exercise as a self-paced, loaded walking protocol may not be as applicable as other exercise tests, such as steady-state exercise where exercise intensity would equal among participants. We only collected follow-up samples after the EXER trial only, and we only had two women participants in our study. Another limitation, which is consistent with most studies of circulating biochemical markers of bone metabolism, is that some of the markers are not secreted solely by bone cells. Therefore, we cannot rule out that changes in circulating concentrations of some biomarkers are not from other tissues. Finally, another limitation of studies like ours that investigate changes in biochemical markers of bone metabolism in humans is that without bone histomorphometry, biomarkers can only be suggestive of cellular activity and actual tissue anabolism or catabolism. The strengths of our study include the degree of control of variables known to affect bone biomarkers such as matching meal timing and composition, matching for potential circadian patterns in biomarkers, and controlling physical activity outside of study participation.

Conclusions

We found that a 60-min bout of self-paced load carriage aerobic exercise induced transient increases in circulating markers of calcium regulation, osteocyte activity, and bone resorption and formation, which returned to baseline by 24 h of recovery. Increased formation and resorption markers after exercise indicate that initiation of anabolic and catabolic responses of bone to exercise occurs in the initial hours after exercise rather than in subsequent days of recovery. These observations may inform studies aimed at investigating mechanisms for optimizing bone anabolism with exercise while preventing bone catabolism.

DATA AVAILABILITY

The data from the current study are only available with an approved data sharing agreement requested through the corresponding author.

SUPPLEMENTAL DATA

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.21440616.

GRANTS

This study was supported by the Military Operational Medicine Program and US Army Medical Research and Development Command, Fort Detrick, Maryland.

DISCLAIMERS

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Army or the Department of Defense.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.S.S., L.J.L., S.A.F., E.G.-S., and J.M.H. conceived and designed research; J.S.S., L.J.L., S.A.F., and E.G.-S. performed experiments; J.S.S. and L.J.L. analyzed data; J.S.S., L.J.L., S.A.F., E.G.-S., and J.M.H. interpreted results of experiments; J.S.S. prepared figures; J.S.S. and J.M.H. drafted manuscript; J.S.S., L.J.L., S.A.F., E.G.-S., and J.M.H. edited and revised manuscript; J.S.S., L.J.L., S.A.F., E.G.-S., and J.M.H. approved final version of manuscript.