Effect of hormone replacement therapy on bone formation quality and mineralization regulation mechanisms in early postmenopausal women

Abstract

Post-menopausal osteoporosis is characterized by a negative imbalance between bone formation and bone resorption resulting in a net bone loss, increasing the risk of fracture. One of the earliest interventions to protect against this was hormonal replacement therapy (HRT). Bone strength depends on both the amount and quality of bone, the latter including compositional / material and structural properties. Bone compositional / material properties are greatly dependent on both patient-, and tissue-age. Raman spectroscopy is an analytical tool ideally suited for the determination of bone compositional / material properties as a function of tissue age as it is capable of analyzing areas ~1 × 1 μm² in tetracycline labeled bone forming areas. Using such analysis of humeri from an ovariectomized primate animal model, we reported that loss of estrogen results in alteration in the mineralization regulation mechanisms by osteoid organic matrix attributes at actively forming bone surfaces. In the present work, we used Raman microspectroscopic techniques to compare osteoid and youngest mineralized tissue composition, as well as relationships between osteoid organic matrix quality and quality attributes of the earliest mineralized tissue in paired iliac crest biopsies obtained from early postmenopausal women before and after two years of HRT therapy. Significant correlations between osteoid proteoglycans, sulfated proteoglycans, pyridinoline, and earliest mineralized tissue mineral content were observed, suggesting that in addition to changes in bone turnover rates, HRT affects the osteoid composition, mineralization regulation mechanisms, and potentially fibrillogenesis.

Keywords: Bone formation; Bone quality; Early postmenopause; Hormonal replacement therapy; Organic matrix; Osteoid; Raman spectroscopy.

Fig. 1

Raman spectra obtained in the osteon and trabecular osteoid as well as interstitial bone are shown, with the Raman peak position of the sulfated proteoglycan (sPG) ~ 1062 cm⁻¹, gylcosaminoglycans (GAGs) and the amide III area highlighted. The Raman spectrum from the interstitial bone area of the cortex shows that the sPG area is strongly overlapped by the B-type carbonate Raman band, thus we did not consider sPGs in mineralized bone areas.

Fig. 2

Osteoid GAG, tissue water, and Pyd content in the osteoid of the forming cortical (top row), and forming trabecular (bottom row) osteoids. Data for the PMOP group are plotted as open, and HRT as solid circles. The only significant difference was the Pyd content (elevated in the HRT group) in both cortical and trabecular osteoid (paired t-test, p < 0.05 in both case).

Fig. 3

Correlations between the monitored spectroscopic parameters in the osteoid and the mineral content at TA1 (youngest mineralized tissue formed) for the cortical (top row) and cancellous (bottom row) compartments. Data for the PMOP group are plotted as open, and HRT as solid circles. Osteoid GAG and Pyd content significantly correlated with the mineral content at TA1 at both cortical (osteons; for PMOP: r = 0.215, p = 0.55 and for HRT: r = −0.632, p = 0.048, respectively) and trabecular actively forming surfaces (for PMOP: r = −0.408, p = 0.242; and for HRT: r = −0.801, p = 0.005, respectively), in the HRT but not in the PMOP group. Additionally, the GAG and Pyd osteoid content significantly correlated in both the cortical (for cortical PMOP: r = 0.517, p = 0.126; and for HRT: r = 0.948, p < 0.0001) and trabecular (for PMOP: r = 0.134, p = 0.711; for HRT: r = 0.980, p < 0.0001) osteoid in the HRT group only.

Fig. 4

Comparison of spectroscopically determined parameters at actively forming (based on the presence of double fluorescent labels) cortical (top row) and trabecular (bottom row) surfaces, by ANOVA. Data for the PMOP group are plotted as black bars, and HRT as grey bars. Mean and SD values are shown. No significant differences between the two groups was evident, with the exception of tissue water content which was lower in the HRT group, at the oldest of the three tissue ages considered.

Fig. 5

Comparison of spectroscopically determined parameters at the oldest tissue age considered in the present study for the geometrical centers of cortical (top row) and cancellous compartments. No significant differences (paired t-test) between the two biopsy groups were evident in interstitial bone, in any of the measured parameters.

Fig. 6

In the cortical compartment, sPG/GAG ratio was significantly elevated in the HRT treated group (Wilcoxon matched-pairs signed rank test; p < 0.05), while the ratio of sPG/Amide III although elevated, did not reach significance (Fig. 6a, and c, respectively). The data for the sPG/GAG ratio and mineral content at TA1 could not be fitted by the same non-linear regression model (Fig. 6b). The sPG / Amide III ratio against the mineral content at TA1 for both PMOP and HRT data sets could be fitted by a Gaussian distribution, but the two curves were significantly different (extra sum-of-squares F-test, p = 0.0027; Fig. 6d).

Fig. 7

The sPG/GAG ratio in the cancellous compartment was significantly elevated in the HRT treated group (Wilcoxon matched-pairs signed rank test; p < 0.05) (Fig. 7a, and c, respectively). Both the sPG/PG, and sPG/Amide III ratios considered against mineral content at TA1 were described by a Gaussian model, while the curves were significantly different (extra sum-of-squares F-test, p = 0.0009, Fig. 7b; and p = 0.0002, Fig. 7d, respectively).