Altered Levels of mRNAs for Calcium-Binding/Associated Proteins, Annexin A1, S100A4, and TMEM64, in Peripheral Blood Mononuclear Cells Are Associated with Osteoporosis

Background Osteoporosis is the most common metabolic bone disease in the world. Since osteoporosis is clinically symptomless until the first fracture occurs, early diagnosis is critical. Calcium, along with calcium-binding and calcium-associated proteins, plays an important role in homeostasis, maintaining healthy bone metabolism. This study is aimed at investigating the level of calcium-binding/associated proteins, annexin A1, S100A4, and TMEM64, in peripheral blood mononuclear cells associated with osteoporosis and its clinical significance. Methods The levels of mRNAs of annexin A1, S100A4, and TMEM64 in human peripheral blood mononuclear cells were evaluated among 48 osteopenia and 23 osteoporosis patients compared to 17 nonosteoporotic controls. Total RNAs were isolated from clinical samples, and quantitation of mRNA levels was performed using real-time quantitative PCR. Results The levels of mRNAs for calcium-binding proteins, annexin A1 and S100A4, and calcium-associated protein, TMEM64, in human peripheral blood mononuclear cells were significantly reduced in osteopenia and osteoporosis patients compared with nonosteoporotic controls (one-way ANOVA, P < 0.0001, P = 0.039, and P = 0.0195, respectively). Annexin A1 and TMEM64 mRNAs were also significantly reduced in female osteoporosis patients over the age of 50 years compared to nonosteoporotic controls (one-way ANOVA, P = 0.004 and P = 0.0037, respectively). ROC analysis showed that the reduction in the level of mRNA for annexin A1, S100A4, or TMEM64 in the patients' peripheral blood mononuclear cells has a good diagnostic value for osteoporosis. Conclusions The results show for the first time that calcium-binding/associated proteins, annexin A1 and TMEM64, could be future diagnostic biomarkers for osteoporosis.


Introduction
Osteoporosis is an age-related bone disease and has a severe impact on public health and economy worldwide, due to fragility fracture [1]. Since osteoporosis is clinically symptomless until the first fracture, there is a pressing need to identify osteoporosis at an earlier stage. Osteoporosis arises from an imbalance between bone-resorbing osteoclasts and bone-producing osteoblasts [2]. Whilst osteoblasts are derived from bone marrow cells [3], the bone-degrading osteoclasts are derived by cell fusion and differentiation of precursor monocyte blood cells [4], which can be found in the peripheral blood mononuclear cells (PBMCs) of a blood sample [5]. The PBMC fraction could be a source of markers which reflect changes in osteoclast activity in osteoporosis patients and possibly in those with osteopenia, the milder, early form of osteoporosis.
Calcium plays an important role in the development of osteoclasts with the involvement of calcium-binding proteins, such as annexins, S100 proteins, and proteins involved with calcium signaling, such as TMEM64. Three calcium-binding/associated proteins, annexin A1, S100A4, and TMEM64, are expressed in monocyte cells (Gene Expression Atlas: https://www.ebi.ac.uk/gxa/home). The calcium-binding, EF-hand-containing S100 protein, S100A4, has been identified in osteoclasts in developing mouse bone [6]. In mice, knockdown of S100A4 with SiRNA led to a higher bone mass and reduced numbers of osteoclasts [7], suggesting its involvement in some way with bone degradation. Furthermore, it has been suggested that S100A4 protein not only prevents bone excess but also can prevent cortical bone loss in estrogen-deprived mice [8]. Synchronised fusion of mouse osteoclasts has been reported to involve S100A4 and to depend upon a member of the calcium-dependent, phospholipid-binding annexin protein family, annexin A1 [9]. Annexin A1 is functional in the human monocytic cell line, U937 [10], reducing its vascular tethering and transmigration [11]. Furthermore, annexin A1 is an important mediator of the resolution of inflammation [12], in part by enhancing leucocyte apoptosis [13]. Inflammatory diseases have been associated with osteoporotic bone fragility [14].
The transmembrane protein, TMEM64, has been reported to interact with and modulate the activity of the protein, sarcoplasmic endoplasmic reticulum Ca 2+ ATPase 2 (SERCA2), thereby enhancing RANKL-induced internal calcium oscillations that are a part of the pathway of osteoclast generation [15].
Thus, the aim of this study was to find out whether the levels of mRNAs for annexin A1, S100A4, and TMEM64 were significantly changed in the PBMCs of participants with osteopenia or with osteoporosis compared to nonosteoporotic controls.  Table S1).

Isolation of Peripheral Blood Mononuclear Cells.
Peripheral blood mononuclear cells (PBMCs) were isolated from 25 mL blood samples and stored at -80°C, as described previously [16] including the final 14,000 g centrifugation step to yield a tight pellet of cellular material without maintaining cell viability.
2.3. Isolation and Purification of Total RNA. Total RNA from human PBMCs was extracted using a combination of TRIzol reagent and PureLink RNA mini kit (Thermo Fisher, UK), according to the manufacturer's recommendations as described previously [16]. The resulting purified RNAs were eluted in 30 μL of RNase-free water and stored frozen at -80°C until used. RNA concentrations and purity were measured using a Thermo Scientific NanoDrop™ 2000 spectrophotometer.

Quantitation of mRNA Levels Using Real-Time
Quantitative PCR (RT-qPCR). Reverse transcription reactions were carried out using the RT-First Strand kit, RT-qPCR Primer Assays (Qiagen, UK), according to the manufacturer's recommendations and as described previously [16] on 500 ng isolated RNA in a total volume of 20 μL with a no reverse transcriptase, no DNA controls. Reverse transcription reactions were diluted 1 : 20, and 10 μL samples were subjected to RT-qPCR for annexin A1, S100A4, TMEM64, and glyceraldehyde-3-phosphate dehydrogenase (GPDH) cDNAs, using the primers indicated in Table 1, with a SYBR Green PCR kit (Qiagen, UK) in a Roche LightCycler 96 Real-Time PCR system (Roche, UK). Each set of reactions was accompanied by the no-reverse-transcription, no-DNA control sample. The relative levels of each RNA were determined from the Ct value after normalization with control mRNA, GPDH using the 2 -ΔΔCT method [17]. Ct values > 35 obtained from RT-qPCR were considered to be below the level of detection of the methodology.

Statistical
Analyses. Statistical analysis between two groups used Student's t-test and between multiple groups used ANOVA with post hoc Bonferroni correction. Diagnostic values were determined using receiver operator characteristic curves. Statistical analyses were carried out using Stats Direct 3 (Altrincham, Cheshire). P values of <0.05 were considered significant as described previously [18].
3.4. The Levels of mRNAs for Annexin A1, TMEM64, or S100A4 in Peripheral Blood Mononuclear Cells from Participants with or without Chronic Inflammatory Disorders. Osteoporosis has been associated with chronic inflammatory disorders. In order to find out whether the reduction in the levels of mRNAs for annexin A1, TMEM64, or S100A4 in peripheral blood mononuclear cells was associated with the presence of chronic diseases, the participants with osteopenia or osteoporosis were divided into those without and those with reported chronic diseases or undergoing treatment with steroids. The levels of annexin A1, TMEM64, or S100A4 mRNAs in PBMCs were not significantly different between    , and S100A4 (c) proteins were determined in peripheral blood mononuclear cell preparations of nonosteoporotic control subjects and patients suffering from osteopenia or osteoporosis. On each box and whisker plot, the black diamond shows the median value, the cross shows the mean value, and white and black circles denote outliers of 1.5 times and 3 times the interquartile range, respectively.  Figure S1).

Discussion
Three mRNAs, encoding calcium-binding/associated proteins, ANXA1, S100A4, and TMEM64, were found to be at a lower level in the PBMC preparations from osteoporosis patients than nonosteoporotic controls. The reductions of S100A4 mRNA levels were found to be due to participants with other disorders, those receiving existing treatments for osteoporosis, males, and under 50-year olds. In contrast, the observed reductions in annexin A1 or TMEM64 mRNAs were not affected by any of these factors.
Annexin A1 protein and, to a lesser extent, TMEM64 protein are found in the CD14+/CD16-negative monocytes [19] (Gene Expression Atlas: https://www.ebi.ac.uk/gxa/ home). These cells have been shown to be the likely source of osteoclasts by differentiation in vitro of cells from normal human subjects [20]. For TMEM64 mRNA in PBMCs, the observed reduction in the level with osteoporosis in the present experiments seems unexpected in view of previous results with TMEM64 knockout mice, in which reduced levels of TMEM64 mRNA resulted in a phenotype of increased bone mass [21]. Apart from possible species differences between mouse and human, such as observed for annexin A1 [9], the difference between the present results and the TMEM64 mouse knockout results could arise from the cell preparations being examined in the two studies. In the knockout mice, there was a reduction in bone-associated, fully differentiated osteoclasts [21], whereas in the present study, the TMEM64 mRNA levels were examined in the blood PBMC fraction, which contains monocytes as yet undifferentiated into osteoclasts [4].
Annexin A1 has been reported to be present in the human U937 cultured monocyte cell line, where it has been reported to reduce adhesion of U937 cells to bone marrow endothelial cells by directly interacting with α4 integrin and competing with binding of the endothelial α4 integrin receptor, VCAM-1 [11]. A reduction in annexin A1 mRNA, and thus possibly a reduction in annexin A1 protein, in monocytes could therefore result in increased retention of   [22]. TMEM64 mRNA is present at a low level in both T and B lymphocytes and is more abundant in natural killer cells (Gene Expression Atlas: https://www.ebi.ac.uk/gxa/ home). Annexin A1 (lipocortin 1) has been reported to be present at a consistent level in many T lymphocyte subtypes and at a higher level in natural killer cells [19], but not present in B lymphocytes (Gene Expression Atlas: https://www.ebi.ac.uk/gxa/home). Thus, it is possible that the observed reductions in annexin A1 and TMEM64 mRNAs arise in the lymphocyte populations in the PBMC preparations. It is not known whether there are specific differences in the cell types populating PBMCs of nonosteoporotic individuals and osteopenia and osteoporosis patients. However, T lymphocytes [23,24] and natural killer cells [25] have been reported to contribute to osteoporosis by their activation and production of osteoclastogenic cytokines, but the observation that various T cell subtypes can exhibit either osteoclastogenic, e.g., Thy 17, or antiosteoclastogenic activity, e.g., Thy1 [26], suggests that there is a balance between pro-and antiosteoclastic activity in T cell subtypes. Whilst, little, if anything, is known about any possible role of TMEM64 in these processes in T lymphocytes, the presence of annexin A1, on the other hand, has been widely associated with the resolution of inflammatory disorders [12,13,27]. Thus, a change in the balance of T cell osteoclastogenic activity, arising from the reduced annexin   Figure 3: Diagnostic values of annexin A1, TMEM64, and S100A4 mRNAs for nonosteoporotic controls vs osteoporosis. ROC curves between nonosteoporotic controls and osteoporosis sufferers are shown for the mRNAs of annexin A1 (a, d), TMEM64 (b, e), and S100A4 (c, f) for all participants (a, b, c) and female participants over the age of 50 (d, e, f). The circled point indicates the optimum cut-off.

Disease Markers
A1 mRNA levels in PBMCs from osteopenia/osteoporosis patients, could contribute to their osteopenia/osteoporosis. However, this might not be consistent with the observation on the present group of participants that the levels of annexin mRNA in the PBMCs of participants with chronic inflammatory disorders were not significantly different from those without chronic inflammatory disorders, either when all participants or when only the female-over-50 group were considered. ROC curve analysis for the changes in these mRNAs showed moderate accuracy for the areas under the curves [28]. Annexin A1 exhibited the best trade-off of sensitivity against specificity of the three possible markers for all participants when nonosteoporotic controls were compared with osteoporosis (area under the curve, 0.893), with osteopenia (area under the curve, 0.805), or with osteopenia and osteoporosis combined (area under the curve, 0.833) or when only females over the age of 50 were included (areas under the curves, 0.809, 0.747, and 0.767, respectively).
In conclusion, this study demonstrates for the first time that the levels of mRNAs of annexin A1 and TMEM64 in peripheral blood mononuclear cells are reduced in association with osteoporosis, at least in this group of patients with loss of annexin A1 mRNA in peripheral blood mononuclear cells being a good marker for osteopenia/osteoporosis.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Ethical Approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the England Health Research Authority National Research Ethics Service Committee, East of England-Essex (REC reference 15/EE/0051) Ethics Committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Disclosure
The current address of Abdullah Y. Mandourah is Al Hada Armed Forces Hospital, Taif, Saudi Arabia. The current address of Ayed Dera is Department of Clinical Laboratory Sciences, College of Applied Medical Sciences, King Khalid University, Abha, Saudi Arabia.