Oestrogen-deficiency induces bone loss by modulating CD14⁺ monocyte and CD4⁺ T cell DR3 expression and serum TL1A levels

Osteoporosis is characterized by micro-architectural deterioration of bone tissue and low bone mass that consequently results in increased bone fragility and susceptibility to fracture [1]. In the United Kingdom more than 3 million people are estimated to have osteoporosis with 500,000 osteoporotic fractures every year; costing an estimated £1.8 billion in 2000 with a potential increase to £2.2 billion by 2025 [2, 3]. Osteoporosis can be characterized into two main forms: primary osteoporosis which occurs as part of aging and secondary osteoporosis, when bone loss is driven by a medical condition / disease or treatment [4]. The onset of menopause in females is a major factor in the development of primary osteoporosis. Loss of oestrogen results in two stages of bone loss: an early rapid loss of trabecular and cortical bone due to increased osteoclast activity and decreased osteoclast apoptosis, and a second slower prolonged loss due to decreased osteoblast activity [4, 5]. In contrast, bone loss due to secondary osteoporosis is caused by factors including but not limited to hyperparathyroidism, inflammatory bowel disease (IBD), type 1 diabetes (T1D), arthritis and glucocorticoid treatment. Several mechanisms are known to contribute to the pathology of menopause-induced primary osteoporosis such as increased expression of tumour necrosis factor (TNF) superfamily members TNFα (TNFSF2) and receptor activator of nuclear factor kappa-B ligand (RANKL; TNFSF11) [6‐11]. It is currently unknown however, what role other members of the TNFSF play in this pathological bone loss. This study focused on the TNFSF members TNF-like protein 1A (TL1A, TNFSF15) and its only confirmed trans-membrane receptor, death receptor 3 (DR3; TNFRSF25) [12].

DR3 and its ligand TL1A have been implicated in the pathogenesis of numerous inflammatory conditions associated with secondary osteoporosis including: IBD and rheumatoid arthritis (RA) [13‐15]. While the majority of DR3’s function has been attributed to its expression on T cells and the ability of TL1A to drive the proliferation of effector T cell subsets [16, 17], studies have identified expression of DR3 on the surface of osteoclast precursors and osteoblasts [14, 18, 19]. In vitro studies using circulating CD14⁺ monocytes, osteoclast precursors which migrate to bone to undergo differentiation into osteoclasts [20], identified cell surface expression of DR3 and that addition of TL1A to these cells significantly enhanced osteoclast proliferation and resorptive activity; increasing expression of the chemokine CCL3 and the enzyme matrix metallopepetidase 9 (MMP9) [14]. Furthermore, expression of DR3 has also been identified on human osteoblasts (OB) where, in vitro, it mediated apoptosis under narrowly regulated conditions [21]. These results suggest that signalling by TL1A through DR3 on CD14⁺ osteoclast precursors and osteoblasts could have an important direct role in the pathogenesis of menopause-induced primary osteoporosis; increasing bone resorption and decreasing bone formation. This is further supported by in vivo studies in the murine collagen-induced arthritis (CIA) model where ablation of DR3 was shown to protect against secondary osteoporosis at sites distal from the small joints [14]. Furthermore, CD4⁺ T cells have been identified as being significant drivers of bone loss following oestrogen deficiency [22], leading to the possibility that changes in DR3 / TL1A expression on these cells could potentially have an indirect effect on bone loss.

While the current data suggests that DR3 and TL1A are implicated in adverse bone loss associated with secondary osteoporosis, their complicity in the pathology of menopause-induced primary osteoporosis remains unknown. In the present study we investigated serum levels of TL1A and the expression of DR3 on peripheral blood CD14⁺ cells isolated from pre-menopausal, post-menopausal and late post-menopausal osteoporotic females to determine whether changes in oestrogen status result in significant modulation of these TNFSF members. We demonstrate for the first time that, in contrast to pre-menopausal females, post-menopausal serum levels of TL1A are not significantly elevated and that DR3 expression is not induced on CD14⁺ monocytes. However, utilizing the murine ovariectomy (OVX) model of oestrogen-deficiency we reveal that early post-OVX expression of DR3 on CD4⁺ T cells is significantly elevated suggesting that DR3 and TL1A could play a potentially indirect role in early post-menopausal bone loss.

The pathology of osteoporosis is complex with many factors contributing to the increased bone loss and decreased bone formation. Of these factors, members of the TNFSF such as TNFα and RANKL are known to play critical roles. While roles for DR3 and its ligand TL1A have been demonstrated in numerous conditions that are associated with pathological bone loss [13‐15], it is currently unknown whether DR3 and / or TL1A are affected by loss of oestrogen and contribute to the progression of menopause-induced primary osteoporosis. In the present study we reveal for the first time in humans that loss of oestrogen does not significantly affect serum levels of TL1A and that late post-menopause CD14⁺ monocyte osteoclast precursors are unable to induce DR3 expression. Furthermore, using the murine OVX model of oestrogen-deficiency, we demonstrate that early post-OVX DR3 expression is upregulated on CD4⁺ T cells but returns to normal levels at later time points, when compared to the sham control (mice that have the same surgery but whose ovaries remain intact). These data suggest that DR3 and TL1A could have a potential indirect role in increasing osteoclast formation and bone loss in the early stages of post-menopause.

Oestrogen deficiency associated with menopause has been linked to elevated levels of interleukin (IL)- 1β, IL-8 and TNFα in the serum [6] and increased expression of RANKL and TNFα in ex vivo cultures of PBMCs [7, 8], circulating monocytes [9] and bone marrow macrophages [10, 11]. In the present study, serum levels of TL1A were decreased in post-menopausal females but comparable in the osteoporotic cohort when compared to the pre-menopausal controls; suggesting that unlike TNFα, elevated expression of TL1A does not play a role in post-menopausal osteoporotic bone loss. However, it is important to note that the increased cytokine expression described in previous studies [6‐11] were reported in samples isolated from early post-menopausal females. Interestingly, in the study by Bismar et al. [10] cytokine levels in late post-menopausal (70 ± 6 years old) bone marrow cultures were significantly lower than those from early post-menopausal women (51 ± 5 years old). These observations raise a couple of intriguing possibilities: firstly, that TL1A may be increased during the early stages of post-menopause when bone loss occurs but missed during this study and secondly; that due to physiological changes post-menopause, pre-menopausal levels of TL1A become pathological and contribute to bone loss. Having shown that serum levels of TL1A were not elevated in late post-menopausal osteoporotic patients we next investigated the effect on DR3 to determine if increased receptor expression could contribute to the increased bone loss associated with osteoporosis.

Previously we reported that circulating CD14⁺ monocytes, isolated from pre-menopausal females, express DR3 when cultured in the presence of MCSF on ivory discs and that treatment of these cells with TL1A resulted in significantly enhanced osteoclast formation [14]. Interestingly however, in the present study DR3 expression was not induced on CD14⁺ monocytes isolated from post-menopausal females, even when cultured for an extended period. The difference in DR3 expression between pre- and post-menopausal derived CD14⁺ monocytes suggests a critical change in these cells caused by menopause. The effects of aging and menopause on immune cell DR3 expression has not been extensively studied. In the only other study to investigate the effect of aging on DR3 expression, Slebioda et al. [25] noted differences in CD4⁺, CD8⁺ and CD20⁺ DR3 expression between healthy children (9.8 ± 4.3 years) and healthy adults (45.3 ± 10.5 years); suggesting that aging can have a significant effect on a cell’s DR3 phenotype. While in the Slebioda et al. [25] study aging did not affect CD14⁺ and CD11c⁺ DR3 expression, it does not rule out a change in DR3 phenotype on these cells due to menopause or as a person moves from adult (20–59 years) to elderly (60+ years); changes in monocyte and macrophage functions have been documented as impaired in aged animal models [26] and elderly individuals [27]. Furthermore, work by Sadeghi et al. [28] demonstrated that monocytic CD14 expression alters during aging; cells change from a CD14^bright/CD16^dim phenotype in young individuals (30.5 ± 13.5 years) to a CD14^dim/CD16^bright phenotype in the elderly (87.6 ± 14 years) demonstrating that these cells undergo a significant phenotypic change with age.

To determine whether the lack of DR3 expression on CD14⁺ monocytes affected osteoclast formation, we performed osteoclastogenesis assays. No difference in osteoclast potential was observed between the post-menopausal and osteoporotic cultures. However, levels of osteoclast formation were lower than we have previously reported for pre-menopausal controls under the same conditions [14]. This is in contrast to work published by D’Amelio et al. [7] who demonstrated significantly higher levels of ‘spontaneous’ osteoclast formation from osteoporotic patient PBMC cultures compared to pre-menopausal controls. The apparent discrepancy between the two studies however, could be explained by differences in methods used. In the D’Amelio et al. [7] study whole PBMC cultures were utilized as compared to CD14⁺ monocyte cultures used in this study. Increased production of pro- osteoclastogenic cytokines RANKL, MCSF and TNFα by PBMCs post-menopause has been demonstrated by a number of groups [7‐9]. Elevated levels of these cytokines would have a significant effect on osteoclast formation in these cultures. Of the PBMCs, T cells are believed to play a particularly critical role in bone loss associated with oestrogen-deficiency [4]; loss of oestrogen has been observed to increase T cell expression of TNFα, RANKL and IL-17 [8, 29, 30]. Importantly, signalling by TL1A through DR3 on CD4⁺ T cells has been shown to induce expression of TNFα and IL-17 [17, 31, 32]; suggesting that changes in DR3 and TL1A signalling on these cells could indirectly affect osteoclast formation by modulating T cell cytokine expression. To investigate this we performed a murine model of ovariectomy and investigated splenic CD4⁺ T cell DR3 expression 1 and 8 weeks post-surgery as indicators of early and late post-menopause. Interestingly, DR3 expression on CD4⁺ T cells was significantly elevated at 1 week but not 8 weeks post-surgery. This raises the very significant possibility that elevated CD4⁺ T cell DR3 expression may contribute to the increased bone loss observed in early menopause.

In conclusion, we have demonstrated for the first time that loss of oestrogen has a significant effect on DR3 and TL1A expression. We reveal that post-menopause the ability to induce DR3 expression is lost from circulating CD14⁺ monocytes and serum TL1A levels trend downwards compared to pre-menopausal females. In contrast however, serum TL1A levels in post-menopausal osteoporotic females are comparable to pre-menopausal. However, due to the low numbers of patients recruited to the study, further work is required before the exact effect of menopause on DR3 and TL1A expression can be revealed. In addition, we further identify in a murine ovariectomy model of oestrogen loss that CD4⁺ T cell DR3 expression is significantly upregulated early on, suggesting that DR3 / TL1A signalling could indirectly contribute to the increased bone loss observed post-menopause through modulation of T cell activity.