Site-specific changes in bone microarchitecture, mineralization, and stiffness during lactation and after weaning in mice

Authors

  • X Sherry Liu,

    Corresponding author
    1. Division of Endocrinology, Department of Medicine, College of Physicians and Surgeons, Columbia University, New York, NY, USA
    • McKay Orthopedic Research Laboratory, Department of Orthopedic Surgery, University of Pennsylvania, 424 Stemmler Hall, 36th Street and Hamilton Walk, Philadelphia, PA 19104, USA.
    Search for more papers by this author
  • Laleh Ardeshirpour,

    1. Section of Endocrinology and Metabolism, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA
    Search for more papers by this author
  • Joshua N VanHouten,

    1. Section of Endocrinology and Metabolism, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA
    Search for more papers by this author
  • Elizabeth Shane,

    1. Division of Endocrinology, Department of Medicine, College of Physicians and Surgeons, Columbia University, New York, NY, USA
    Search for more papers by this author
  • John J Wysolmerski

    1. Section of Endocrinology and Metabolism, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA
    Search for more papers by this author

Abstract

Despite the dramatic bone loss that occurs during lactation, bone mineral density rapidly recovers after offspring are weaned and milk production stops. The goal of this study is to quantify site-specific changes in bone quantity and quality during and after lactation in a mouse model. We used micro computed tomography (µCT), individual trabecula segmentation (ITS), digital topological analysis (DTA)-based tissue mineral density (TMD) analysis, and micro finite element analysis (µFEA) to quantify the effects of lactation and weaning on bone microarchitecture, mineralization, and stiffness at the spine, tibia, and femur. We found a significant decrease in trabecular plate microarchitecture, tissue mineralization of the trabecular surface, trabecular central skeleton, and intervening envelopes, and whole bone stiffness in lactating versus nulliparous mice at all three sites. In recovered mice, all these different aspects of bone quality were comparable to nulliparous mice at the spine. In contrast, trabecular plate microarchitecture and whole bone stiffness at the tibia and femur in recovered mice were lower than nulliparous mice, as were central trabecular tissue mineralization and cortical structure at the femur. These findings are consistent with clinical observations of partial recovery of femoral bone mineral density BMD after lactation in humans. The observed differences in trabecular surface tissue mineralization in nulliparous, lactating, and recovered mice are consistent with prior observations that maternal bone turnover shifts from resorption to formation at the time of pup weaning. The significant differences in trabecular central tissue mineralization during these three states suggest that osteocytes may contribute to the reversible loss of mineral during and after lactation. Future studies are necessary to determine whether differing functions of various bone cells at individual skeletal sites cause site-specific skeletal changes during and after lactation. © 2012 American Society for Bone and Mineral Research.

Introduction

Lactation induces substantial changes in maternal calcium and bone metabolism, as the mother's skeleton serves as an important source of calcium for milk production.1 During lactation, there is a dramatic increase in the rate of maternal bone resorption and in calcium losses from the maternal skeleton.2–4 Although the rate of bone formation is increased during this time, it is outpaced by the rate of bone resorption, resulting in a rapid net decline in bone mass.2, 5–8 Nursing women lose as much as 5% to 7% of their bone mass over 6 months.5, 6 Rodents lose as much as 20% to 30% of their bone mass over 3 weeks of lactation, possibly because they have many more offspring.7–9

Despite the dramatic bone loss that occurs during lactation, bone mineral density (BMD) recovers rapidly after offspring are weaned and milk production is halted. The recovery process is characterized by a sudden halt of bone resorption and an elevated rate of bone formation.8, 10 Studies on rats and mice have shown partial or full recovery of bone mass and mechanical properties after the end of lactation.3, 7, 9 Several epidemiological studies in humans suggest that the number of pregnancies and the duration of lactation have no long-term deleterious effects on bone mass or on fracture risk.11–14 However, controversy exists in current literature as other studies suggest the duration of breast feeding may correlate with lower subsequent BMD in women.15–18

Alterations in the circulating levels of estrogen, parathyroid hormone-related protein (PTHrP), and calcitonin have been shown to contribute to the regulation of bone loss during lactation in both mice and women.4, 19–21 However, knowledge of the characteristics and mechanisms of maternal bone recovery after lactation are still rather limited. Previous studies of nursing women have reported complete restoration of BMD at the lumbar spine, but incomplete restoration at the hip after they stop lactating.5, 22–24 Rodent studies have also suggested that bone volume and microarchitecture improved to a greater extent at the vertebra than the tibia after pups were weaned.10, 25, 26 These data suggest that in addition to the systemic regulation of mineral metabolism by ovarian hormones and PTHrP, local factors may also affect skeletal recovery after lactation. If this is the case, then prolonged breastfeeding may increase future fracture risk at some skeletal sites but have no influence on other sites.

The primary goal of the current study was to quantify site-specific changes in bone quantity and quality during and after lactation in a mouse model. We used high resolution micro computed tomography (µCT) to image bone microarchitecture at the lumbar vertebra, proximal tibia, distal femur, and femur midshaft.27, 28 We have previously demonstrated that plate- and rod-like trabeculae have distinct roles during the initiation and progression of trabecular bone failure, and that trabecular plates dominate apparent strength of trabecular bone.27–30 Therefore, one objective of the current study was to analyze trabecular bone microarchitecture using individual trabecula segmentation (ITS)-based morphological analysis, a state-of-the-art image analysis technique capable of quantifying the number, orientation, average size, and total volume of plate-like and rod-like trabeculae. A second objective was to quantify trabecular tissue mineralization using hydroxyapatite (HA) density-calibrated µCT imaging and digital topological analysis (DTA).28, 31–34 This technique can separately evaluate tissue mineralization at the surface and central portion of trabeculae. Tissue mineralization at the trabecular bone surface is more likely to be influenced by osteoblast and osteoclast remodeling activities, whereas mineralization within the central portion of the trabeculae may be associated with osteocyte activity. Last, we examined changes in the mechanical competence of each skeletal site during lactation and postweaning using micro finite element analysis (µFEA). This technique noninvasively estimates bone stiffness, which is closely related to bone strength and fracture risk.35–38

Materials and Methods

Experimental animals

Female, 10- to 13-week-old, CD1 mice were purchased from Charles River Laboratories (Wilmington, MA, USA). Mice in lactation (n = 8) and recovered (n = 7) groups were allowed to become pregnant, deliver, and lactate. Litter size was adjusted to 8 to 12 pups to equalize suckling intensity between dams. Pups were removed on the 12th day of lactation to halt milk production and trigger the recovery process. Mice in lactation groups were sacrificed on the 12th day of lactation, and recovered mice were killed on the 28th day after weaning. Age-matched, nulliparous (virgin) CD1 mice were used as controls (n = 5). All experiments were approved by Yale University's Institutional Animal Care and Use Committee.

µCT imaging

Bone specimens from tibia, third lumbar vertebra, and femur were harvested from each mouse, and scanned and analyzed in the Department of Biomedical Engineering, Columbia University, NY, USA, using a µCT imaging system (VivaCT 40; Scanco Medical AG, Bassersdorf, Switzerland) at an isotropic spatial resolution of 10.5 µm. Specimens were stabilized with gauze in a 2-mL centrifuge tube filled with 70% ethanol and fastened in the specimen holder of the µCT scanner. A total of 90 slices corresponding to a 0.95-mm region distal from the growth plate of the proximal tibia, 200 slices corresponding to a 2.1-mm region of the central vertebra L3, and 224 slices corresponding to 2.35-mm region proximal from the growth plate of the distal femur were scanned for trabecular bone analysis. A total of 90 slices corresponding to a 0.95-mm region of the mid femur shaft were scanned to analyze cortical bone structure. A 3D image of each scan was reconstructed and bone voxels were segmented from bone marrow and background using Gaussian filtering and a global threshold corresponding to 30% of maximal grayscale value of the images. Standard microstructural analyses of trabecular bone were performed for trabecular compartments of all three sites and results were reported in Supplemental Material Tables S1–S3.39

Individual trabecula segmentation (ITS)-based morphological analyses

The trabecular compartment was semi-automatically extracted from the cortex for the proximal tibia, L3, and distal femur, and subjected to ITS analyses to evaluate trabecular bone volume, number, thickness, and connectivity separately for trabecular plates and rods27 (Fig. 1). First, a complete volumetric decomposition technique was applied to segment the trabecular network into individual plates and rods27 (Fig. 1, left). Based on the 3D evaluations of each individual trabecular plate and rod, bone volume and plate and rod number were evaluated by plate and rod bone volume fraction (pBV/TV and rBV/TV), as well as plate and rod number densities (pTb.N and rTb.N, 1/mm). Plate-to-rod ratio (P-R ratio), a parameter of plate versus rod characteristics of trabecular bone, was defined as plate bone volume divided by rod bone volume. The average size of plates and rods was quantified by plate and rod thickness (pTb.Th and rTb.Th, mm), plate surface area (pTb.S, mm2), and rod length (rTb.l, mm). Intactness of the trabecular network was characterized by plate–plate, plate–rod, and rod–rod junction density (P-P, P-R, and R-R Junc.D, 1/mm3), calculated as total junctions between trabecular plates and rods normalized by the bulk volume. Orientation of trabecular bone network was characterized by axial bone volume fraction (aBV/TV), defined as axially aligned bone volume divided by the bulk volume. The definition of these ITS measurements can be found in Glossary 1 (Supplemental Material). Detailed methods describing the complete volumetric decomposition technique and ITS-based measurements can be found in our recent publications.27, 28

Figure 1.

Representative 3D cortical and trabecular bone microarchitecture of the mouse lumbar vertebra L3 imaged by µCT. (Left) Trabecular bone was decomposed into individual trabeculae represented by different colors. (Right) The same trabecular bone was also illustrated by blue and pink to represent plate- and rod-like trabeculae, respectively.

Tissue mineralization density (TMD) analysis

TMD reflects the tissue-level characterization of degree of mineralization of bone matrix. To assess TMD, µCT attenuation values were first converted to hydroxyapatite density (mg HA/cm3) using a linear conversion calibrated by phantom. By DTA-based skeletonization, trabecular bone voxels were iteratively peeled off layer by layer until no more bone voxels could be removed without altering the shape or topology of the trabecular microarchitecture.28, 33, 34 The remaining structure, a minimal representation of trabecular network topology and connectivity, was defined as the central trabecular skeleton. The first layer removed was defined as the surface trabecular bone. The remaining bone voxels that occupy the envelope between the surface and central skeleton was defined as the intervening (middle) trabecular bone (Fig. 2). For the proximal tibia, L3, and the distal femur, trabecular bone TMD was calculated for surface (sTMD), central (cTMD), and middle (mTMD) envelopes, as the total bone mineral content divided by the bone volume within the envelope.

Figure 2.

Representative 3D distribution of tissue mineral density (TMD) of surface trabecular bone (sTMD), central trabecular bone (cTMD), and intervening trabecular bone (mTMD) occupying the envelop between surface and central trabecular bone of the mouse lumbar vertebra L3. (Top) demonstrates a cross-sectional reconstruction of L3 viewed from the superior aspect of the vertebral body; (bottom) demonstrates a transverse section through the midportion of L3 viewed from the anterior aspect of the vertebral body. TMD increased from surface to center of trabeculae: mean sTMD (yellow), mTMD (orange), and cTMD (red) of nulliparous mice were 767, 969, and 1086 mgHg/cm3, respectively.

Cortical bone analysis

Cortical TMD, thickness, bone area, medullary area, endosteal perimeter, periosteal perimeter, porosity, and polar moment of inertia (pMOI) were evaluated for the cortical bone of the mid femur shaft by using manufacturer provided software (Scanco Medical AG). In addition, stiffness was evaluated for the mid femur shaft by µFEA. pMOI indicates bone's resistance to bending and stiffness is related to bone's resistance to compression.

µFEA

Based on the µCT images of whole bone segments of the proximal tibia, lumbar vertebra, distal femur, and mid femur shaft, each bone voxel was converted to an eight-node brick element to construct Finite Element (FE) models for bone stiffness measurements.40 Bone tissue was modeled as an isotropic, linear elastic material with a Young's modulus (Es) of 15 GPa and a Poisson's ratio of 0.3.41 A uniaxial compression was applied along the axial direction of the model and the model was subjected to linear analysis to determine stiffness using an element-by-element precondition conjugate gradient solver. Whole bone stiffness, defined as the reaction force divided by the applied displacement, characterizes the mechanical competence of both cortical and trabecular compartments and is closely related to whole bone strength35 and fracture risk.36–38

Statistical analysis

Analyses were conducted using KaleidaGraph 3.6 software (Synergy Software, Reading, PA, USA). All the parameters were presented as mean ± standard deviation (SD). One-way analysis of variance (ANOVA) with the Tukey-Honesty Significant Difference (HSD) post-test were performed to compare microarchitectural, mineralization, and micromechanical parameters between nulliparous, lactating, and recovered groups. Two-sided p values <0.05 were considered to indicate statistical significance.

Results

Trabecular bone microarchitecture

As shown in Figures 3, 4, and 5, and Tables 1, 2, and 3, lactating mice had severely compromised bone microarchitecture. Trabecular bone volume fraction (BV/TV) was markedly lower in lactating than nulliparous mice (by −61% to −75%) at all three trabecular bone sites (L3, proximal tibia, and distal femur). ITS analysis revealed that trabecular plate microarchitecture was significantly disrupted in lactating compared with nulliparous mice, with lower plate bone volume fraction (pBV/TV; by −88% to −91%), plate number density (pTb.N; by −19% to −21%), plate thickness (pTb.Th; by −30% to −31%), plate surface area (pTb.S; by −68% to −72%), plate–rod and plate–plate junction densities (P-R and P-P Junc.D; by −48% to −75% and −66% to −82%), plate-to-rod ratio (P-R Ratio, by −81% to −91%), and axial bone volume fraction (aBV/TV, by −78% to −82%) (all p < 0.05). The differences between lactating and nulliparous mice were similar at all three sites (Tables 1, 2, 3, Fig. 7). Trabecular rod microarchitecture also differed between lactating and nulliparous mice at each skeletal site (Tables 1, 2, 3, Fig. 7). At the proximal tibia and distal femur, lactating mice had lower rod bone volume fraction (rBV/TV, by −51% and −29%) but no difference in rod number density (rTb.N). In contrast, at L3, lactating mice had similar rBV/TV but higher rTb.N (by 16%) compared with nulliparous mice. Rod thickness (rTb.Th) was lower at all three sites (by −16% to −20%), whereas rod length (rTb.l) did not differ at any site in lactating compared with nulliparous mice. Furthermore, rod-rod junction density (R-R Junc.D) was significantly higher in lactating than nulliparous mice at L3 and the proximal tibia, but did not differ at the distal femur (Tables 1, 2, 3, Fig. 7).

Figure 3.

3D reconstruction of representative lumbar spine L3 from (A) nulliparous, (B) lactating, and (C) 28-day recovered mice.

Figure 4.

3D reconstruction of representative proximal tibia from (A) nulliparous, (B) lactating, and (C) 28-day recovered mice.

Figure 5.

3D reconstruction of representative distal femur from (A) nulliparous, (B) lactating, and (C) 28-day recovered mice.

Table 1. Results of ITS-Based Morphological Analysis, DTA-Based TMD Analysis, and µFE Analysis of Trabecular Bone at the Lumbar Vertebra L3
 Groupsp values
Nulliparous (n = 5)Lactating (n = 8)28-day Recovered (n = 7)All groupsNulliparous versus lactLact versus recoveredNulliparous versus recovered
ITS-based microstructural parameters
 BV/TV (%)20.5 ± 4.38.1 ± 2.417.6 ± 4.4<0.001<0.001<0.0010.37
 pBV/TV (%)14.8 ± 4.11.7 ± 0.912.5 ± 5.4<0.001<0.0010.0040.14
 rBV/TV (%)5.8 ± 0.76.84 ± 0.86.13 ± 1.10.52NSNSNS
 pTb.N (1/mm)6.8 ± 0.45.5 ± 0.36.6 ± 0.7<0.001<0.001<0.0010.35
 rTb.N (1/mm)6.2 ± 0.37.3 ± 0.26.3 ± 0.40.0030.0040.030.50
 pTb.Th (µm)44 ± 332 ± 244 ± 2<0.001<0.001<0.0010.72
 rTb.Th (µm)40 ± 234 ± 140 ± 1<0.001<0.001<0.0010.78
 pTb.S (µm2)10234 ± 11213078 ± 9299555 ± 918<0.001<0.001<0.0010.16
 rTb.l (µm)175 ± 3180 ± 4176 ± 40.16NSNSNS
 P-P Junc.D (1/mm3)345 ± 62123 ± 53340 ± 113<0.001<0.001<0.0010.94
 P-R Junc.D (1/mm3)374 ± 54210 ± 77387 ± 109<0.0010.005<0.0010.88
 R-R Junc.D (1/mm3)104 ± 26289 ± 41121 ± 29<0.001<0.0010.0010.48
 P-R Ratio2.6 ± 0.70.2 ± 0.11.7 ± 0.7<0.001<0.001<0.0010.04
 aBV/TV (%)12.5 ± 2.93.0 ± 0.710.5 ± 0.4<0.001<0.001<0.0010.11
Tissue mineralization
 sTMD (mgHg/cm3)767 ± 7733 ± 4760 ± 7<0.001<0.001<0.0010.07
 mTMD (mgHg/cm3)969 ± 15902 ± 10950 ± 16<0.001<0.001<0.0010.06
 cTMD (mgHg/cm3)1086 ± 20999 ± 161057 ± 25<0.001<0.001<0.0010.06
Mechanical parameter
 Whole bone stiffness (kN/mm)3.1 ± 0.40.9 ± 0.52.6 ± 0.3<0.001<0.001<0.0010.13
Table 2. Results of ITS-Based Morphological Analysis, DTA-Based TMD Analysis, and µFE Analysis of Trabecular Bone at the Proximal Tibia
 Groupsp values
Nulliparous (n = 5)Lactating (n = 8)28-day recovered (n = 7)All groupsNulliparous versus lactLact versus recoveredNulliparous versus recovered
ITS-Based Microstructural Parameters
 BV/TV (%)15.8 ± 4.13.9 ± 0.69.2 ± 3.5<0.001<0.0010.0070.003
 pBV/TV (%)9.7 ± 2.80.9 ± 0.44.8 ± 2.1<0.001<0.001<0.0010.002
 rBV/TV (%)6.1 ± 1.43.0 ± 0.34.4 ± 1.4<0.001<0.0010.040.07
 pTb.N (1/mm)5.6 ± 0.54.4 ± 0.24.6 ± 0.6<0.001<0.0010.570.004
 rTb.N (1/mm)5.7 ± 0.55.3 ± 0.25.0 ± 0.60.030.190.430.03
 pTb.Th (µm)50 ± 235 ± 350 ± 1<0.001<0.001<0.0010.99
 rTb.Th (µm)45 ± 135 ± 145 ± 2<0.001<0.001<0.0010.69
 pTb.S (µm2)10794 ± 10003046 ± 10019331 ± 1041<0.001<0.001<0.0010.06
 rTb.l (µm)182 ± 2187 ± 13185 ± 50.53NSNSNS
 P-P Junc.D (1/mm3)181 ± 8233 ± 1296 ± 46<0.001<0.0010.060.02
 P-R Junc.D (1/mm3)232 ± 9658 ± 16139 ± 59<0.001<0.0010.040.04
 R-R Junc.D (1/mm3)89 ± 3086 ± 1560 ± 210.040.070.960.07
 P-R Ratio1.6 ± 0.20.3 ± 0.11.1 ± 0.2<0.001<0.001<0.001<0.001
 aBV/TV (%)9.3 ± 2.21.9 ± 0.45.3 ± 2.0<0.001<0.0010.0020.003
Tissue mineralization
 sTMD (mgHg/cm3)778 ± 4742 ± 4776 ± 6<0.001<0.001<0.0010.82
 mTMD (mgHg/cm3)1001 ± 12926 ± 12988 ± 16<0.001<0.001<0.0010.23
 cTMD (mgHg/cm3)1136 ± 181034 ± 211106 ± 21<0.001<0.001<0.0010.06
Mechanical Parameter
 Whole bone stiffness (kN/mm)9.0 ± 1.72.6 ± 0.86.8 ± 0.4<0.001<0.001<0.0010.003
Table 3. Results of ITS-Based Morphological Analysis, DTA-Based TMD Analysis, and µFE Analysis of Trabecular Bone at the Distal Femur
 Groupsp values
Nulliparous (n = 5)Lactating (n = 8)28-day recovered (n = 7)All groupsNulliparous versus lactLact versus recoveredNulliparous versus recovered
ITS-based microstructural parameters
 BV/TV (%)12.6 ± 3.14.1 ± 0.95.6 ± 2.9<0.001<0.0010.47<0.001
 pBV/TV (%)8.2 ± 2.41.0 ± 0.23.3 ± 2.1<0.001<0.0010.08<0.001
 rBV/TV (%)4.5 ± 0.73.3 ± 0.63.1 ± 1.50.010.050.590.01
 pTb.N (1/mm)5.3 ± 0.34.3 ± 0.44.2 ± 0.90.0020.0080.570.003
 rTb.N (1/mm)5.2 ± 0.25.4 ± 0.44.6 ± 0.70.0060.90.0070.03
 pTb.Th (µm)52 ± 436 ± 347 ± 0<0.001<0.001<0.0010.13
 rTb.Th (µm)44 ± 237 ± 142 ± 1<0.001<0.001<0.0010.37
 pTb.S (µm2)10482 ± 8883696 ± 6668038 ± 1955<0.001<0.001<0.0010.003
 rTb.l (µm)182 ± 3183 ± 2184 ± 30.8NSNSNS
 P-P Junc.D (1/mm3)134 ± 2848 ± 1477 ± 42<0.001<0.0010.360.002
 P-R Junc.D (1/mm3)174 ± 2984 ± 24109 ± 57<0.001<0.0010.70.005
 R-R Junc.D (1/mm3)64 ± 7109 ± 2554 ± 24<0.0010.0089<0.0010.35
 P-R Ratio1.8 ± 0.30.3 ± 0.11.0 ± 0.3<0.001<0.001<0.001<0.001
 aBV/TV (%)7.5 ± 2.01.4 ± 0.32.8 ± 2.0<0.0010.36<0.001<0.001
Tissue mineralization
 sTMD (mgHg/cm3)777 ± 6741 ± 3772 ± 5<0.001<0.001<0.0010.12
 mTMD (mgHg/cm3)1003 ± 14920 ± 9984 ± 10<0.001<0.001<0.0010.02
 cTMD (mgHg/cm3)1137 ± 171025 ± 171098 ± 17<0.001<0.001<0.0010.003
Mechanical parameter
 Whole bone stiffness (kN/mm)7.1 ± 0.92.8 ± 0.95.4 ± 0.5<0.001<0.001<0.0010.008

In mice recovering from lactation (recovered mice), as shown in Figure 3, bone microarchitecture at L3 was comparable to nulliparous mice in all ITS measures except P-R ratio (−35%, Table 1, Fig. 7). At the proximal tibia (Fig. 4), most ITS parameters were significantly better in recovered than lactating mice. Plate-related parameters including pBV/TV, pTb.S, pTb.Th, P-R Junc.D, P-R Ratio, and aBV/TV were greater in recovered than lactating mice while pTb.N and P-P Junc.D were not different between recovered and lactating mice. In addition, rod-related parameters rBV/TV and rTb.Th were greater in recovered than lactating mice. Among the above measurements, which were lower in lactating than in nulliparous mice, only measurements of rBV/TV, pTb.Th, rTb.Th, and pTb.S were equivalent in recovered and nulliparous mice (Table 2, Fig. 7). Furthermore, rTb.N, which was not different between lactating and nulliparous mice, was lower in recovered than in lactating mice. The microstructural difference between recovered and nulliparous mice were most significant at the distal femur (Figs. 5, 7, Table 3). In this regard, plate-related measurements pBV/TV, pTb.N, P-R Junc.D, P-P Junc.D, and aBV/TV and rod-related parameters rBV/TV and rTb.N, were significantly lower in both lactating mice and recovered mice than in nulliparous mice. Furthermore, rTb.N, which was not different between lactating and nulliparous mice, was lower in recovered than in lactating mice. Only pTb.Th, rTb.Th, and R-R Junc.D in recovered mice were comparable to nulliparous mice at the femur (Table 3).

Trabecular bone mineralization

TMD increased from the surface to the center of trabeculae (Fig. 2). cTMD was 11% to 13% higher than mTMD, and mTMD was 23% to 29% higher than sTMD at each site and in all experimental groups. At all three sites, sTMD was 4% to 5% lower in lactating than in nulliparous mice; in contrast, sTMD in recovered mice did not differ from sTMD in nulliparous mice (Tables 1, 2, 3, Fig. 7). In addition, mTMD (7%–8%) and cTMD (8%–10%) were significantly lower in lactating than in nulliparous mice at all three sites. At L3 and the proximal tibia, mTMD and cTMD in recovered mice did not differ from those in nulliparous mice. At the distal femur, mTMD and cTMD in recovered mice were greater than lactating mice, but 3% lower than nulliparous mice (Tables 1, 2, 3, Fig. 7).

Cortical bone microarchitecture and mineralization

At the midfemur shaft, cortical bone area (Ct.Area) and cortical thickness (Ct.Th) were 28% and 30% lower and cortical porosity (Ct.Po) 36% higher in lactating than in nulliparous mice (Fig. 6, Table 4). In recovered mice, femur Ct.Area and Ct.Th were still 18% lower and Ct.Po 16% higher than in nulliparous mice. Cortical tissue mineral density (Ct.TMD) in the femur was 5% lower in lactating than in nulliparous mice, but did not differ between recovered and nulliparous mice (Table 4). Furthermore, medullary area, periosteal perimeter, and endosteal perimeter did not differ in nulliparous, lactating or recovered mice.

Figure 6.

3D reconstruction of representative midfemur shaft from (A) nulliparous, (B) lactating, and (C) 28-day recovered mice.

Table 4. Results of Standard µCT Analysis of Cortical Bone at the Midfemur Shaft
 Cortical bone parametersp values
Nulliparous (n = 5)Lactating (n = 8)28-day recovered (n = 7)All groupsNulliparous versus lactLact versus recoveredNulliparous versus recovered
Bone area (mm2)1.11 ± 0.120.80 ± 0.090.91 ± 0.06<0.001<0.0010.080.007
Medullary area (mm2)2.04 ± 0.352.33 ± 0.282.27 ± 0.420.42NSNSNS
Periosteal perimeter (mm)5.18 ± 0.295.34 ± 0.495.13 ± 0.230.55NSNSNS
Endosteal perimeter (mm)3.28 ± 0.283.43 ± 0.553.35 ± 0.240.82NSNSNS
Cortical thickness (mm)0.26 ± 0.020.18 ± 0.010.22 ± 0.02<0.001<0.0010.003<0.001
Porosity (%)14.7 ± 1.020.0 ± 1.217.1 ± 1.5<0.001<0.0010.0010.02
TMD (mg/cm3HA)1413 ± 191340 ± 471397 ± 290.0060.010.020.77
Moment of inertia (mm4)0.53 ± 0.110.40 ± 0.090.43 ± 0.040.06NSNSNS
Stiffness (kN/mm)16.2 ± 1.810.9 ± 1.512.2 ± 1.3<0.001<0.0010.220.002

Bone stiffness

At L3, whole bone stiffness was 71% lower in lactating mice compared with nulliparous mice, but did not differ in recovered compared with nulliparous mice (Table 1, Fig. 7). At the proximal tibia and distal femur, whole bone stiffness was 71% and 61% lower in lactating than in nulliparous mice. In contrast to L3, the stiffness of the proximal tibia and distal femur were 24% and 25% lower in recovered compared with nulliparous mice (Tables 2, 3, Fig. 7). The cortical bone stiffness of mid femur shaft was 33% and 25% lower in lactating and recovered mice, respectively, compared with nulliparous mice. There were no significant differences in stiffness between lactating and recovered mice. In addition, there was also a nonsignificant trend toward a lower moment of inertia in lactating and recovered mice compared with nulliparous mice (Table 4).

Figure 7.

Percentage differences of microarchitecture, mineralization, and mechanical parameters of trabecular bone at L3, the proximal tibia, and distal femur of lactating and recovered mice compared with nulliparous mice. a,b,c Indicates significant difference compared with nulliparous mice (p < 0.05 by ANOVA) at L3a, proximal tibiab, and distal femurc. All the comparisons are cross-sectional.

Discussion

In this study, we examined the maternal bone microarchitecture, tissue mineralization, and mechanical properties at various skeletal sites during and after lactation in mice. We found that lactating mice had fewer, thinner, and smaller trabecular plates; fewer connections between plates and between plates and rods; less axially aligned trabeculae, and thinner trabecular rods at all three trabecular sites: L3, the proximal tibia, and distal femur. These microarchitectural differences were associated with a 70% reduction in predicted bone stiffness at each site. Additionally, the degree of mineralization at the surface, middle, and center of trabeculae was significantly lower in lactating mice at all sites. After lactation, the microarchitectural alterations persisted to varying degrees at some skeletal sites. Twenty-eight days after pups were weaned, trabecular microarchitecture, tissue mineralization, and stiffness at the lumbar spine were indistinguishable in recovered mice compared with nulliparous mice. In contrast, the microarchitecture of the tibia and femur of recovered mice was significantly worse than the microarchitectural parameters observed in nulliparous mice. Moreover, lactation-associated differences in thickness, bone area, porosity, and stiffness of the femur cortical midshaft were also present in recovered mice. These results suggest that despite remarkable improvements in all aspects of bone quality after lactation, trabecular and cortical microarchitecture, trabecular bone tissue mineralization, and bone stiffness at the tibia and femur may not recover to values comparable to nulliparous mice.

We have applied ITS analysis to µCT and high-resolution peripheral quantitative CT (HR-pQCT) images to trabecular plate and rod microarchitecture in several studies and found that plates play a far more important role than rods in determining the apparent mechanical properties of trabecular bone.27–29, 42 Thus, changes in trabecular plate microarchitecture are likely to be closely associated with bone's resistance to fractures. In this study, we found similar changes in plate bone volume, number, thickness, and junction densities in lactating mice at all three sites, suggesting that lactation is associated with marked deterioration in trabecular plate microarchitecture. However, the changes in rod microarchitecture varied between skeletal sites. At the lumbar spine, trabecular bone loss was associated with an increase in rod number and a steady rod bone volume fraction, suggesting that plates were likely converted into rods. In contrast, at the femur and tibia, there were net reductions in rod number and volume suggesting that both compartments change simultaneously or that the microarchitectural changes were more severe causing conversion of plates to rods and then severing of rods, leading to less residual connectivity in the trabecular network.

In contrast to the relative uniformity of the trabecular bone loss during lactation, the restoration of trabecular bone after the end of lactation differed significantly among the three sites. The thickness of plates and rods recovered fully at all sites, and the lumbar spine demonstrated complete microarchitectural recovery. However, in the tibia and femur, recovered mice still had reduced plate bone volume, fewer plates and rods, fewer junctions between plates and rods, a less axially aligned trabecular network, and a more rod-like trabecular structure than nulliparous mice. These trabecular properties contribute more to mechanical integrity than trabecular thickness suggesting that there may be persistent post-lactational alterations in bone strength at these sites.

We used finite element analysis, a computational technique, to estimate whole bone stiffness of L3, the proximal tibia, and distal femur. Stiffness estimated by µFEA at these three sites includes the mechanical contribution of both the cortical and trabecular bone compartments. We found that stiffness was reduced by more than 50% in lactating mice at all three sites. Moreover, in recovered mice only at the L3 vertebra was stiffness similar to nulliparous mice. A previous study of lactating rats also showed full restoration of stiffness at the lumbar vertebra after lactation by mechanical testing.3 We also measured the bone stiffness of midfemur shaft: a site that only consists of cortical bone. The cortical bone stiffness was lower in lactating than nulliparous mice and remained low in recovered mice. Although differences in geometrical measures of cortex did not reach statistical significance, there was a trend toward a greater resorption in endocortical surface than periosteal formation during lactation, leading to greater medullary area and thinner cortex in lactating and recovered mice than nulliparous mice. To our knowledge, this is the first time that stiffness of the proximal tibia, distal femur, and midfemur shaft have been evaluated in mice associated with lactation, and our results suggest that both the trabecular bone-abundant regions and the cortical regions of these long bones have incomplete recovery of their mechanical competence after lactation.

Our findings in vertebral trabecular bone are consistent with a previous study that reported decreased bone volume fraction, thickness, connectivity, and a more rod-like structure, as indicated by structure model index (SMI) in lactating versus nulliparous mice, with full recovery of these structural deficits after lactation.8 However, no change was reported for trabecular number in that study.8 The overall trabecular number by standard analysis was also measured in our study (results reported in Supplemental Material Tables S1–S3) and no change was found. However, in our ITS results, because plate and rod properties were analyzed separately, we were able to document that plate and rod number at L3 changed in opposite directions during and after lactation, likely explaining the lack of change in overall trabecular number noted in the previous study.8 Also consistent with our findings, 2D histomorphometry studies showed persistent differences in bone volume fraction and connectivity between recovered and nulliparous rats as well as in bone volume fraction, thickness, and trabecular number between recovered and nulliparous mice.10, 25, 26 Our findings in mice, and those of others in rodents, also parallel the clinical observations of BMD changes in women: BMD of the spine is completely restored after lactation, whereas BMD of the hip is incompletely restored.5, 22–24

In addition to the loss of bone matrix, decreases in bone tissue mineralization may also contribute to the decreased BMD and bone strength observed during lactation. Lactating mice had reduced trabecular bone surface TMD, which may be related to increases in undermineralized osteoid caused by accelerated bone turnover. The recovery in surface TMD after lactation may be explained by the documented shift from net maternal bone resorption to net maternal bone formation triggered by weaning.8, 10 It is intriguing that we also observed a 10% loss and recovery in central trabecular tissue TMD in lactating and recovered mice. Emerging data from Qing et al.43, 44 suggests that osteocytes remodel the perilacunar/canalicular bone during reproductive cycles, removing mineral during lactation and then replacing it after milk production ceases. Our observations of reversible changes in central trabecular TMD provide further evidence that osteocytes may be partially responsible for the reduction and restoration of bone tissue mineral by removing and replacing their perilacunar matrix.

An important strength of our study is the use of novel and state-of-the-art technologies to investigate site-specific changes in different aspects of bone qualities in association with lactation and recovery. There are also limitations associated with this study. First, a cross-sectional study design was used. By comparing groups of nulliparous, lactating, and recovered mice, we observed restored plate number, junction densities, and plate–rod ratio. However, how these microarchitectural deficits were reversed was not clear. Second, TMD analysis used in the current study is an emerging technique that needs further validation by higher resolution analysis technique, such as quantitative backscattered electron imaging or synchrotron light source µCT imaging. Because of the relatively low resolution of µCT images, the measurement of TMD, particularly the surface TMD, may be subjected to a partial volume effect. Mineral density of voxels at the bone–marrow boundary can be underestimated because of partially filled voxels. However, it has been recently demonstrated in newborn pig that mineralization distribution across the surface, middle, and central trabeculae detected by µCT imaging technique corresponds to a similar pattern of bone tissue stiffness distribution measured by nanoindentation: lower at trabecular surface and higher in the center.45 Thus, the alterations observed in TMD of lactating and recovered mice may correlate with more significant changes in whole bone stiffness than what has been estimated by µFE analysis, which only accounts for the contributions from bone mass and bone structure but not the intrinsic mechanical properties of bone tissue.

In conclusion, we found highly significant differences between lactating and nulliparous mice in trabecular plate microarchitecture, surface, middle, and central trabeculae tissue mineralization and whole bone stiffness at the lumbar spine, proximal tibia, and distal femur. All aspects of bone quality at the lumbar spine in recovered mice were comparable to those of nulliparous mice. In contrast, trabecular plate microarchitecture and whole bone stiffness at the tibia and femur remained below these parameters in nulliparous mice. Moreover, recovery of central trabecular tissue mineralization and cortical structure at the femur was incomplete. Together, these finding suggest that these skeletal sites may have lasting alterations in both trabecular and cortical structure after lactation. As a working model to explain these differences, we propose that lactational bone loss in the spine occurs primarily through thinning and central perforation of the trabecular plates, converting them into rod-like structures. This leaves the trabecular connectivity relatively intact and allows for rapid and complete recovery of the microarchitecture after lactation by filling in the central perforations. However, in long bones, rods and plates are simultaneously resorbed leading to perforations of plates and breakage of the resultant rods and less connectivity. Having to bridge compromised rods and plates likely impairs the ability to restore microarchitecture fully at these sites. Testing of this hypothesis will require longitudinal µ-CT measurements in vivo in order to trace the fate of individual plates and rods, an important goal of our future studies. Clarification of these issues will also inform future studies to determine how the functions of bone cells during and after lactation might differ at individual skeletal sites in order to cause the site-specific skeletal responses that we have observed.

Disclosure

Dr. Liu is an inventor of the ITS analysis software used in the study. All the other authors have no conflicts of interest.

Acknowledgements

This work was supported by NIH Grants NIH DK077565 and NIH CA153702 and by the Thomas L. Kempner and Katheryn C. Patterson Foundation. We thank Dr. X. Edward Guo for providing the µCT instrument and for helpful comments and discussions.

Author's roles: Study design: XSL and JJW. Study conduct: XSL, LA, and JNV. Data analysis: XSL. Data interpretation: XSL, LA, JNV, ES, and JJW. Drafting manuscript: XSL. Revising manuscript content: LA, JNV, ES, and JJW. Approval final version of manuscript: XSL, LA, JNV, ES, and JJW. XSL takes responsibility for the integrity of the data analysis.

Ancillary