Subclinically low BMD in young men is associated with compromised bone microarchitecture and lower lean mass

Abstract

Context

Optimizing bone mass accrual early in life is a strategy for mitigating fracture risk in older age.

Objective

This study aimed to identify factors associated with lower than expected bone mass in healthy young adult men.

Methods

Data from 39 male participants (21 areal bone mineral density [aBMD] Z scores <−1.0 at either the total hip, femoral neck, or lumbar spine [Lower]; 18 Z scores ≥1.0, but not <−1.0 at any site [controls; Con]) were analyzed from a larger study. aBMD and body composition were assessed by dual-energy x-ray absorptiometry (DXA), and bone quality was evaluated at the tibia using high resolution–peripheral quantitative computed tomography (HR-pQCT). Fasted concentrations of bone turnover markers, testosterone, estradiol, parathyroid hormone (PTH), insulin-like growth factor-1 (IGF-1), and insulin-like growth factor-1 binding protein 5 (IGFBP-5) were measured via enzyme-linked immunosorbent assays.

Results

The Lower group was older (age in years: Lower 25 [6]; Con 21 [5]; P = .037), shorter (m; Lower 1.74 [0.10]; Con 1.79 [0.10]; P = .024), and had less lean mass than Con (kg; Lower 54.2 [6.2]; Con 60.3 [6.9]; P = .018). Tibial bone microarchitecture (trabecular number 1/mm; Lower 1.7 [0.2]; Con 1.9 [0.2]; P = .004; trabecular separation mm; Lower 0.6 [0.1]; Con 0.5 [0.1]; P = .002) and strength (stiffness kN/mm Lower 207.7 [43.3]; Con 274.7 [39.7] and failure load kN Lower 11.3 [2.2]; Con 14.8 [2.0]; P < .001) were compromised in Lower. Alkaline phosphatase was greater (Lower 1528.2 [933.8]; Con 605.4 [624.1]; P = .003), and IGF-1 (Lower 265.5 [93.6]; Con 340.4 [106.5]; P = .027) was lower in Lower. Osteoblast activity (P ≥ .361) and gene expression were similar between groups (P ≥ .451).

Conclusion

Men with relatively lower than expected aBMD demonstrated compromised bone microarchitecture and lower lean mass, indicating the importance of targeted intervention for lean mass accrual during young adulthood.

Osteoporosis is an age-related disorder in which bone mineral density (BMD) decreases and bone microarchitecture changes, primarily at the axial skeleton (1, 2), leading to weakened bone and increased fracture risk (3). Osteoporosis, and its precursor osteopenia (low BMD that has not met the threshold for osteoporosis) (4), can take years to manifest and is often discovered much later in life (5). By later adulthood, the hormonal, metabolic, and mechanical environments of the body are less likely to protect against bone loss (6). Ensuring maximum bone accrual early in life is an important protective factor against osteoporosis and osteopenia (7). Although nonmodifiable factors, including race and genetics, account for approximately 60% to 80% of variance in BMD in adults, modifiable factors, like physical activity, nutrition, and body composition, may contribute to the remaining variance and could be optimized to promote the accrual of bone mass (8).

Low bone mass is more common in women than men, with the greatest prevalence in postmenopausal women. While men exhibit lower incidence of low bone mass or osteoporosis than women, the associated mortality following an osteoporotic fracture is actually greater in men (9). Furthermore, a treatment gap is prevalent, which was highlighted in an assessment of US Medicare claims data (n = 35 774) wherein only 5.7% of men were tested/treated for osteoporosis within 6 months following a fracture compared to 12.1% of women, and receiving testing or treatment was associated with a reduction in mortality (10). Current recommendations are for women to be screened for osteoporosis at age 65 using dual-energy x-ray absorptiometry (DXA), but no current clinical standards exist for men (11). However, a recent report from the US Preventive Services Task Force states that more research is needed to support screening for low bone mass in men in the absence of known risk factors (12). Since most fragility fractures occur in those without osteoporosis, identifying risk factors in individuals who have not met the threshold for osteoporosis or osteopenia diagnoses is crucial for fracture mitigation (13). Additionally, identifying precursors of osteoporosis via DXA alone can be insufficient to fully capture bone quality (structure plus strength), underscoring the need for alternative indicators for early bone loss, such as advanced imaging and biochemical assessments.

Since the amount of bone accrued during adolescence and young adulthood has been shown to be approximately equal to the amount lost throughout the remainder of life (14), and a 10% improvement in bone mass can reduce osteoporotic fracture risk by 50% (15), identifying those with relatively lower than expected areal BMD (aBMD) and intervening before the window of bone accrual ends is crucial to long-term bone health. The purpose of this study was to determine if factors associated with relatively lower than expected aBMD in young adult men differed from those with healthy bone mass using assessments of tibial bone quality, body composition, circulating biomarkers, and osteogenic milieu.

Materials and methods

Participants

These data are baseline measures collected from a parent study designed to investigate skeletal adaptations to exercise training. Participants were aged 18 to 40 years, generally healthy, recreationally active, nonsmokers, not on medication, and free from current or previous (within 6 months) injury (full inclusion and exclusion criteria in Supplementary Table 1 (16)). Participants provided written informed consent and were recruited between August 2022 and January 2024. Data collection was approved by the University of Pittsburgh Institutional Review Board (IRB No. 21020044) and US Army Medical Research and Development Command (E03018.1a), and was completed in accordance with the Declaration of Helsinki. After baseline data for all participants screened from August 2022 to January 2024 was reviewed, participants included in this analysis were male participants who were identified to have an aBMD Z score less than −1.0 at either the total hip, femoral neck, or L1 to L4 vertebrae by DXA (Lower). Male participants were selected as normal aBMD controls (Con) if they had an aBMD Z score of 1.0 or greater at any measured site and did not have a Z score less than −1.0 at any site.

Demographics and anthropometrics

Participants completed a custom survey to report demographics, general medical history, and calcium intake. This survey included the validated Eating Pathology Symptoms Inventory (EPSI) (17) for assessing disordered eating and the Aging Males’ Symptoms (AMS) scale (18), which has been used as a surrogate measure for energy deficiency in young adult men (19). Body mass (clothed) was assessed by a digital platform scale (Health o Meter); height was determined using a wall-mounted stadiometer (Seca 216 Accu-Hite). Body mass index (BMI) was calculated as body mass in kilograms divided by height in meters squared.

Areal bone mineral density and body composition

aBMD of the total body, unilateral hip, and lumbar spine (L1-L4), and measures of body composition were obtained using DXA (Lunar iDXA, GE Healthcare) in accordance with the International Society for Clinical Densitometry Guidance best practices guidelines. Participants were positioned in accordance with the manufacturer's guidelines by a trained operator, ensuring proper segmentation of regions of interest. The acquired scans were processed using enCORE Software version 15 (Lunar, GE Healthcare). Total body aBMD, total hip aBMD, femoral neck aBMD, L1 to L4 aBMD, total and regional lean mass, and total fat mass were determined from the relevant scans. For our laboratory, coefficients of variation (CVs) for total body aBMD is 1.9%, 1.3% for total lean mass, and 3.4% for total fat mass.

Tibial volumetric bone mineral density, geometry, microarchitecture, and estimated bone strength

Tibial measures of volumetric bone mineral density (vBMD), geometry, microarchitecture, and estimated bone strength were acquired with HR-pQCT (XtremeCT II, version 6.1, ScanCo Medical, AG) at the tibial distal metaphysis (4% of tibial length measured from distal end plate) and mid-diaphysis (30% of tibial length). Scans were performed on the nondominant leg unless history of fracture or metal hardware was present in that limb. Tibia length (to the nearest millimeter) was measured using anthropometric tape from the tibial plateau to the medial malleolus. Participants were seated with the determined leg extended through the gantry and instructed to remain still. Each scan comprises 168 total slices (10.2 mm stack of bone length), with an isotropic voxel size of 60.7 μm. All scans were visually inspected at time of scan for motion artifacts and if significant artifacts were detected, scans were retaken up to 3 times. Quality assurance checks were performed prior to scanning each day and scan images were assessed following testing. Microfinite element analysis (μFEA) was conducted to determine whole-bone stiffness and estimate failure (yield) load as previously outlined (20) using native software (XtremeCT II, version 1.13, ScanCo Medical, AG). For our laboratory, CVs for density measurements are 0.9% or less, 1.1% or less for cortical microarchitecture, and 3.6% or less for trabecular microarchitecture.

Blood biomarkers

Resting, morning, fasted blood samples were collected from an upper extremity vein. Plasma was collected in EDTA tubes and immediately centrifuged (1500g, 15 minutes). Serum samples clotted for 30 minutes before centrifugation (1500g, 15 minutes). Samples were stored at −80 °C. Commercially available enzyme-linked immunosorbent assays were used to determine circulating concentrations of serum procollagen type I N-terminal propeptide (PINP; Novus catalog No. NBP2-76465, RRID:AB_3720306, 1:200 dilution, CV 5.3%, sensitivity 9.4 pg/mL), alkaline phosphatase (ALP; Novus catalog No. NBP2-68197, RRID:AB_3720307, CV 6.0%, sensitivity 46.9 pg/mL), tartrate-resistant acid phosphatase 5b (TRAP5b; Immunodiagnostic Systems catalog No. SB-TR201A, RRID:AB_3095995, CV 4.7%, sensitivity 0.4 U/L), sclerostin (Biomedica catalog No. BI-20492, RRID:AB_2894889, CV 6.4%, sensitivity 3.2 pmol/L), testosterone (Alpco Diagnostics catalog No. 11-TESHU-E01, RRID:AB_3720308, CV 4.2%, sensitivity 0.02 ng/mL), estradiol (Alpco Diagnostics catalog No. 11-ESTHU-E01, RRID:AB_2756385, CV 7.6%, sensitivity 10 pg/mL), insulin like growth factor-1 (IGF-1; Alpco Diagnostics catalog No. 22-IGFHU-E01, RRID:AB_2923276, 1:20 dilution, average CV 3.3%, assay sensitivity 9.4 pg/mL), IGF binding protein-5 (IGFBP-5; R and D Systems catalog No. DY875, RRID:AB_3720311, 3× dilution, CV 2.7%), parathyroid hormone (PTH; (Alpco Diagnostics catalog No. 21-IPTHU-E01, RRID:AB_2943035), CV 7.3%, sensitivity 1.6 pg/mL), and plasma β C-terminal cross-linked telopeptide of type I collagen (βCTX; Immunodiagnostic Systems catalog No. IS-3000N, RRID:AB_3720313, CV 6.2%, sensitivity 3.2 pmol/L). Urine specific gravity was assessed prior to all blood collection.

Osteoblast culture conditions

A human fetal osteoblast cell line was grown (hFOB1.19; American Tissue Type Culture) under repository-recommended conditions (21). Cells were cultured in a humidified incubator with atmospheric levels of O₂ and 5% CO₂ and grown in Dulbecco’s modified Eagle’s medium:Ham's F12 culture medium (ThermoFisher Scientific) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin (ThermoFisher) with media changes every other day. For proliferative conditions cells were cultured at 34 °C and on reaching confluence, cells were cultured at 39 °C to induce differentiation. On introduction to differentiation conditions, 10% blood plasma from participants was added to the osteoblast medium for treatments, supplemented with 0.5% heparin to prevent plasma from congealing.

Osteoblast assays

Osteoblasts were assayed for mineralization deposits and ALP activity after 3, 7, and 10 days of plasma treatments. Prior to assays, cells were fixed in 10% neutral buffered formalin (3.7% formalin, 4 g/L Na2HPO4 in H₂O) for 10 minutes, then washed 3 times with distilled water. Alizarin red S (10 g/L in distilled H₂O, filtered, pH 4.2; Sigma Aldrich) solution was added to wells and incubated in the dark for 15 minutes to analyze calcium deposits. After careful removal of the staining solution, cells were washed 3 times with distilled water. Ten percent acetic acid was added, and cells were transferred to tubes. Tubes were shaken and incubated at 75 °C until dissolved before being centrifuged at 10 000g for 10 minutes. Supernatant (200 µL) was transferred to a new tube and mixed with 75-µL ammonium hydroxide, then 100 µL of the supernatant mixture was added to a 96-well plate and the absorbance was read at 405 nm in a microplate reader (Bio-Tek Instruments).

ALP staining was performed using an ALP staining kit (Abcam) according to the manufacturer's instructions. Briefly, following fixation cells were detached from the plate, resuspended in assay buffer, homogenized, and centrifuged at 10 000g to remove insoluble debris. Sample and standard wells were set up in 96-well plates and pnPP, an ALP substrate, was added to each well and incubated at room temperature for 60 minutes in the dark. The reaction was stopped, absorbance was read at 405 nm using a microplate reader (Bio-Tek Instruments), and ALP activity of samples was calculated using the standard curve.

RNA isolation, reverse transcription, and real-time quantitative polymerase chain reaction

Total RNA was extracted from osteoblasts using Trizol, following the manufacturer's instructions (ThermoFisher), following 3 days of plasma treatment to analyze expression of key osteoblast genes. RNA concentration was measured via nanodrop spectrophotometer, and 500 ng of first-strand complementary DNA was generated using iScript Reverse Transcription Supermix (Bio-Rad Laboratories) in a CFX Opus 96 real-time polymerase chain reaction (PCR) system (Bio-Rad). Complementary DNA was diluted 1:4 in nuclease-free water and real-time PCR was performed using SYBR Green-based chemistry in the CFX Opus 96. PCR conditions were set according to the manufacturer's recommendations and consisted of an initial 30-second denaturation step at 95 °C followed by 40 cycles of a 5-second denaturation step at 95 °C, and a 30-second annealing step at 60 °C. Gene expression was determined using 2^−ΔΔct relative quantification and normalized to endoplasmic reticulum membrane protein complex subunit 7 expression (22). Primer sequences are displayed in Supplementary Table 2 (16).

Statistical analyses

All data were screened for normality by the Shapiro-Wilks test. For normally distributed data, independent-sample t tests were used to determine differences between the Lower and Con groups and mean (SD) were reported. If data were not normally distributed, Mann-Whitney U tests were used, and median (interquartile range were reported. Analyses of covariance were used to determine the differences between the Lower and Con groups with age as a covariate for all DXA- and HR-pQCT–dependent variables. Two-way analyses of variance (group × time) were used to test the effect of group (Lower vs Con), time (day 3, 7, and 10), and their interaction in the cell culture model. Assumptions for the analyses of variance and analyses of covariance analyses were checked. Statistical significance was set a priori at α = .05, 2-sided. All statistical analyses were conducted in SPSS version 29.0.2.0 (IBM Corp) and figures were made in GraphPad Prism v10.3.1 (GraphPad Software).

Results

Participant characteristics

Table 1 shows participant characteristics of the Lower (n = 21) and Con (n = 18) groups. The Lower group was significantly older (P = .037) and shorter (P = .024). Groups were not significantly different for any other characteristic (P ≥ .305) or EPSI scores (Supplementary Table 3 (16)); however, the Lower group had significantly lower total body lean mass compared to the Con group (P = .018). Inclusion and exclusion criteria allowed us to exclude common causes of secondary osteoporosis (23). As expected, the Lower group had lower aBMD and Z scores at all sites (P < .001; Table 2).

Table 1.

Participant characteristics

	Lower (n = 21)	Con (n = 18)	P
Age, ya	25 (6)	21 (5)	.037
Height, ma	1.74 (0.10)	1.79 (0.10)	.024
Total mass, kg	74.70 (12.02)	77.46 (9.44)	.362
Total body fat mass, kg	17.87 (8.41)	13.97 (6.36)	.305
Total body lean mass, kg	54.16 (6.18)	60.26 (6.85)	.018
BMI	24.50 (3.46)	23.76 (2.55)	.459
Calcium intake/d, mga	850 (810)	1130 (990)	.335
Race, n (%)			.515
Asian	7 (33%)	4 (22%)
Black or African American	2 (10)	3 (17%)
Hispanic or Latino	4 (19%)	1 (6%)
White	8 (38%)	10 (56%)
Stress fracture history, n (%)			.946
No	18 (86%)	15 (83%)
Yes	3 (14%)	3 (17%)
AMS total scorea	20 (4)	21 (4)	.519

Data are presented as mean (SD). Bolded P values indicate significance, P < .05.

Abbreviations: AMS, Aging Males’ Symptoms; BMI, body mass index; Con, controls; Lower, total hip, femoral neck, or lumbar spine.

^a Median (interquartile range) and count (n, %) as appropriate.

Table 2.

Dual-energy x-ray absorptiometry areal bone mineral density

	Lower (n = 20)	Con (n = 18)	P
Total aBMD	1.18 (0.10)	1.34 (0.08)	<.001
Total Z scorea	−0.20 (0.70)	1.22 (1.05)	<.001
L1-L4 aBMD	1.08 (0.09)	1.30 (0.11)	<.001
L1-L4 Z-score	−1.27 (0.46)	0.49 (0.76)	<.001
Total hip aBMDa	0.94 (0.14)	1.14 (0.15)	<.001
Total hip Z scorea	−1.1 (0.90)	0.25 (1.0)	<.001
Femoral neck aBMD	0.96 (0.13)	1.16 (0.11)	<.001
Femoral neck Z score	−0.99 (0.66)	0.40 (0.82)	<.001

Table 2 depicts the aBMD and Z scores between the 2 groups. Age was included as covariate for these analyses. For the Lower group, n = 1 did not have all 3 sites scanned. Bolded P values indicate significance, P < .05.

Abbreviations: aBMD, areal bone mineral density; Con, controls; Lower, total hip, femoral neck, or lumbar spine.

^a Mann-Whitney U test was used to examine group differences. Data represented as median (interquartile range); otherwise data are presented as mean (SD).

Tibial volumetric bone mineral density, geometry, microarchitecture, and estimated bone strength

Table 3 depicts key bone morphology, densitometry, and strength outcomes as measured by HR-pQCT. At the tibial metaphysis (4%), the Lower group had significantly narrower tibias as evidenced by lower total area (P = .020) and smaller cortical area (P = .004) than the Con group. The Lower group had fewer trabeculae (Tb.N; P = .004) with greater separation (Tb.Sp; P = .002), and lower total (P = .002) and trabecular (P < .001) vBMD than the Con group. Measures of bone stiffness and failure load were significantly lower in the Lower group than the Con group (both P < .001). At the tibial mid-diaphysis (30%), cortical area was lower in the Lower group than the Con group (P < .001) and the Lower group had thinner cortices (P < .001) but similar cortical vBMD (P = .062) than the Con group. Measures of bone stiffness and failure load at this site were also lower in the Lower group than in the Con group (both P < .001).

Table 3.

High resolution–peripheral quantitative computed tomography

	Lower (n = 20)	Con (n = 18)	P	Cohen's D
Metaphysis (4%)
Size/morphology
Tt.Ar (mm2)	1134.7 (152.4)	1245.9 (117.4)	.020	−0.81
Tb.Ar (mm2)	1051.1 (162.0)	1141.0 (117.0)	.056	−0.63
Ct.Ar (mm2)	90.8 (20.6)	112.3 (15.1)	.004	−1.18
Microarchitecture
Tb.N (1/mm)	1.7 (0.2)	1.9 (0.2)	.004	−0.70
Tb.Th (mm)	0.3 (0.0)	0.3 (0.0)	.415	−0.35
Tb.Sp (mm)	0.6 (0.1)	0.5 (0.1)	.002	0.79
Density
Tt.vBMD (mgHA/cm3)a	235.8 (23.1)	282.6 (16.9)	.002	1.15
Tb.vBMD (mgHA/cm3)	202.4 (22.8)	227.7 (17.8)	<.001	−1.23
Ct.vBMD (mgHA/cm3)	758.1 (75.2)	773.2 (38.5)	.899	−0.25
μFEA
Stiffness, kN/mm	207.7 (43.3)	274.7 (39.7)	<.001	−1.61
Failure load, kN	11.3 (2.2)	14.8 (2.0)	<.001	−1.64
Mid-Diaphysis (30%)
Size/morphology
Tt.Ar (mm2)a	374.9 (55.3)	408.9 (98.6)	.024	−0.79
Ct.Ar (mm2)	273.4 (28.8)	323.1 (33.1)	<.001	−1.61
Ct.Pm (mm)	78.4 (4.5)	82.6 (5.7)	.018	−0.82
Microarchitecture
Ct.Po (%)a	0.5 (0.8)	0.7 (0.8)	.666	0.16
Ct.Th (mm)	5.7 (0.8)	6.6 (0.5)	<.001	−1.30
Density
Tt.vBMD (mgHA/cm3)a	761.0 (152.2)	783.4 (71.0)	.066	−0.69
Ct.vBMD (mgHA/cm3)	1029.6 (25.5)	1013.0 (17.9)	.062	0.75
μFEA
Stiffness, kN/mm	292.4 (32.6)	348.1 (35.7)	<.001	−1.63
Failure load, kN	16.5 (1.9)	19.7 (1.9)	<.001	−1.69

HR-pQCT parameters between the 2 groups. Data are presented as mean (SD). Bolded P values indicate significance, P < .05.

Abbreviations: μFEA, microfinite element analysis; Con, controls; HR-pQCT, high resolution–peripheral quantitative computed tomography; Lower, total hip, femoral neck, or lumbar spine; Sp, separation; Tb, trabeculae; vBMD, volumetric bone mineral density.

^a Median (interquartile range) as appropriate. Age was included as covariate for these analyses.

Biomarkers

There were no statistically significant differences between the Lower and Con groups for PINP, TRAP5b, sclerostin, or βCTX (P ≥ .196; age-adjusted P ≥ .343) (Fig. 1A, 1C, 1D, and 1E). However, ALP was significantly higher in the Lower group than in the Con group (P = .003; age-adjusted P = .012) (Fig. 1B). There were no differences in serum testosterone, estradiol, PTH, or IGFBP-5 concentrations (P ≥ .118; age-adjusted P ≥ .240) (Fig. 1F, 1G, 1H, and 1J). The Lower group had lower circulating concentrations of IGF-I than the Con group; however, after adjusting for age these differences were no longer statistically significant (P = .027; age-adjusted P = .170) (Fig. 1I).

Figure 1.

Biomarkers. Fig. 1 depicts all assessed biomarker concentrations. The total hip, femoral neck, or lumbar spine (Lower) group is depicted in black filled circles, while the control (Con) group is in clear squares. A, P1NP (pg/mL); B, alkaline phosphatase (ALP) (pg/mL); C, tartrate-resistant acid phosphatase 5b (TRAP5b) (U/L); D, sclerostin (pmol/L); E, β C-terminal cross-linked telopeptide of type I collagen (βCTX) (pmol/L); F, testosterone (ng/mL); G, estradiol (pg/mL); H, parathyroid hormone (PTH) (pg/mL); I, insulin-like growth factor-1 (IGF-1) (ng/mL); J, insulin-like growth factor-1 binding protein 5 (IGFBP-5) (ng/mL).

Cell culture

For the alizarin red staining, there were no statistically significant time × group interaction effects observed (P = .225), or main effects of group (P = .361). Main effects of time were present, showing that mineralization increased over time in the pooled sample (P < .001; Fig. 2A). The ALP staining demonstrated a similar pattern with no significant time × group interaction effects (P = .581), nor main effects of group (P = .579); but enzymatic activity of early osteoblast differentiation, as evidenced by ALP staining, demonstrated changes over time in the pooled sample (P < .001; Fig. 2B) wherein staining increased from day 3 to 7 before decreasing to day 10.

Figure 2.

Cell culture. Fig. 2 depicts the results of the cell culture experiment over 10 days. The total hip, femoral neck, or lumbar spine (Lower) group is depicted in black filled circles, while the control (Con) group is in clear squares. A, Alizarin red staining; B, alkaline phosphatase staining.

Osteoblast gene expression

Fold change was expressed for the following osteoblast genes: BSP2, BGLAP, OSTERIX, and RUNX2 (Fig. 3). Data were normalized so that the mean of the Con expression for each gene equaled 1. There were no statistically significant differences in the fold change between the groups for any osteoblast gene (P ≥ .451).

Figure 3.

Gene expression. Fig. 3 shows gene expression data from quantitative polymerase chain reaction. The total hip, femoral neck, or lumbar spine (Lower) group is depicted in black filled circles, while the control (Con) group is in clear squares. All data were normalized so that the mean of the CON group expression for each gene equaled 1. A, BSP2 fold change; B, BGLAP fold change; C, OSTERIX fold change; D, RUNX2 fold change.

Discussion

This investigation provides phenotypical insights into the body composition, bone strength, and metabolic bone environment of healthy young adult males with aBMD Z scores less than <−1. With relatively lower than expected aBMD relative to peers, these individuals could be at increased risk of osteoporosis and fracture as they age. We sought to determine if any measured factors differentiated these participants (Lower) from those with relatively higher aBMD (Con), which could help identify targets for intervention in otherwise healthy young adult men with relatively lower than expected aBMD. Notably, participants in the Lower aBMD group had compromised bone density, morphology, and geometry as assessed via HR-pQCT, lower lean mass, and higher circulating concentrations of ALP.

Our findings from HR-pQCT echo the observations from the DXA-derived aBMD; that is, relatively lower than expected 2-dimensional aBMD observed in the Lower group was consistent with unfavorable characteristics relative to the Con group in terms of microstructure, geometry, and strength. Our data demonstrate that decreased bone microarchitectural parameters are evidence of relatively lower than expected aBMD in young, healthy, recreationally active men, suggesting the relatively lower than expected aBMD extends to the microarchitecture and strength (bone stiffness and failure load) in this population, which is consistent with prior work showing statistically significant associations between tibial bone stiffness and aBMD of the lumbar spine, femoral neck, and total hip (24). And while DXA is the clinical standard used to diagnose osteoporosis and osteopenia, it cannot measure vBMD, bone microstructure, or trabecular and cortical compartments of bone that can be measured by HR-pQCT via 3-dimensional imaging (25). Furthermore, tibial microarchitecture assessed by HR-pQCT has been predictive of incident and major osteoporotic fracture risk (of the proximal humerus, distal forearm, clinical vertebral, or hip) in older adults (aged >40 years) even independent of femoral neck aBMD (26), highlighting the ability of HR-pQCT to capture clinically relevant bone characteristics not traditionally assessed via DXA.

Acute measures of bone metabolism offer insights into the response of recent skeletal cellular activity, an advantage over imaging that can take up to 24 months for changes to be evident (27). Additionally, the assessment of bone cell activity through cell culture models may help provide a more robust picture of bone metabolism through examination of the effect of the circulating milieu on in vitro cellular differentiation and function (28). Resting osteogenic circulating factors have previously been shown to predict bone accrual in US Marine officer candidates, further illustrating the utility of assessing the circulating osteogenic environment (29). Bone turnover markers are released during the bone remodeling process and can represent either formation (P1NP, ALP), resorption (TRAP5b, βCTX), or metabolism (sclerostin) (30). Other biomarkers, like sex steroids, PTH and growth factors, such as the IGF family, can provide a broader picture of the systemic bone environment (31). IGF action is mediated by IGFBPs. IGFBP5 is the most abundant IGFBP in bone and has roles in osteoblast differentiation (32); however, IGFBP5 concentrations were similar between groups. In contrast, circulating IGF-1 concentrations were lower in the Lower group. IGF-1 has central roles in regulating bone growth and homeostasis (33) and is the most abundant growth factor in the bone matrix, where it acts on osteoprogenitors to promote periosteal bone expansion (34). Nevertheless, after adjusting for age, differences in IGF-1 concentrations were no longer statistically significant. Therefore, lower circulating concentrations of IGF-1 are not the driving factor behind the relatively lower aBMD observed in the Lower group. IGF-1 plateaus in young adulthood and declines with age during the third decade of life related to reductions in growth hormone secretion (35). Given that the Lower group was approximately 4 years older on average (mean age 25 years) compared to Con group members (mean age 21 years), the confounding effect of age on IGF-1 concentrations is not unexpected. ALP was significantly higher in the Lower group compared to the Con group, even after adjusting for age. Elevated serum concentrations of ALP are associated with increased bone remodeling as ALP is expressed on the cell surface of osteoblasts to promote mineralization (36). Prior work has demonstrated an inverse relationship between bone ALP and total aBMD, and aBMD of the left arm, pelvis, lumbar spine, and trunk among adults aged 20 to 49 years with data extracted from the National Health and Examination Survey (37). In the present analysis, total circulating ALP was assessed, and although approximately 80% of ALP originates from bone and liver tissue, concentrations may reflect ALP derived from multiple tissues (38). Furthermore, in our osteoblast cell culture model, ALP activity did not differ between osteoblasts treated with plasma from the Lower aBMD group vs those treated with plasma from the Con group. These findings indicate that the circulating environment in the Lower aBMD group did not promote increased mineralization and suggest that the elevated circulating ALP may not reflect enhanced osteoblast activity. Collectively, these findings suggest that differences in aBMD between groups are not reflected by differences in circulating bone turnover markers, with the exception of ALP, which may represent activity of nonskeletal sources.

This study has several limitations. It is possible we assessed these participants’ aBMD before their peak bone accrual, so while they are currently relatively lower than peers of the same age, sex, and race, they may not remain relatively lower. Our laboratory does not have CVs for lumbar spine or hip aBMD, but there were no repeated scans, and all participants were assessed in accordance with International Society for Clinical Densitometry Guidance guidelines. No historical nutrition or physical activity data were available, so we are unable to determine whether these past factors may have influenced aBMD in our population. Nonetheless, our study has a number of strengths including the use of novel imaging and comprehensive biochemical analyses.

This cross-sectional, secondary analysis reports key characteristics of otherwise healthy young adult men with relatively lower than expected BMD. Phenotypic descriptions of young adult men with relatively lower than expected BMD in our sample include lower lean mass, compromised tibial bone quality, and higher ALP compared to young adult men with higher than expected BMD. Of these factors, lean mass represents a potential target for improving bone health. Specifically, improving body composition in adolescence and young adulthood may act as a preventive measure against relatively lower bone mass. Future studies should incorporate robust measures of past physical activity and dietary intake to determine their role in relatively lower than expected bone mass in otherwise healthy young adult men.

Abbreviations

βCTX

β C-terminal cross-linked telopeptide of type I collagen

μFEA

microfinite element analysis

aBMD

areal bone mineral density

ALP

alkaline phosphatase

AMS

Aging Males’ Symptoms scale

BMD

bone mineral density

BMI

body mass index

Con

controls

CV

coefficient of variation

DXA

dual-energy x-ray absorptiometry

EPSI

Eating Pathology Symptoms Inventory

HR-pQCT

high resolution–peripheral quantitative computed tomography

IGF-1

insulin-like growth factor-1

IGFBP-5

insulin-like growth factor-1 binding protein 5

Lower

total hip femoral neck or lumbar spine

PCR

polymerase chain reaction

PINP

procollagen type I N-terminal propeptide

PTH

parathyroid hormone

TRAP5b

tartrate-resistant acid phosphatase 5b

vBMD

volumetric bone mineral density

Funding

This work was supported by the United States Department of Defense, Army Medical Research and Development Command (W81XWH-21-1-0542).

Disclosures

The authors have nothing to disclose.

Data availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.