Stimulatory Effect of Tofacitinib on Bone Marrow Adipocytes Differentiation.

Background: Systemic inflammation is the main factor underlying secondary osteoporosis in patients with rheumatoid arthritis (RA). Janus kinase inhibitors (JAKi), such as tofacitinib (Tofa), can control systemic inflammation and may have beneficial effects on bone in various models. This might be due to direct effects on the bone microenvironment and not exclusively based on their anti-inflammatory function. Bone marrow adipocytes (BMAds) are abundant in the bone microenvironment. The effect of JAKi on BMAds is unknown, but evidence suggests that there is competition between human bone marrow-derived stromal cell (hBMSC) differentiation routes towards BMAds and osteoblasts (Ob) in osteoporosis. Objectives: The aims of the study are to determine whether Tofa influences BMAds and Ob derived from hBMSCs and to investigate the potential effects of Tofa on bone marrow adiposity in RA patients. Methods: To determine the effect of Tofa on cellular commitment, hBMSCs were differentiated to BMAds or OBs for 3 days together with Tofa at 200, 400, or 800 nM and TNFα. This study was also conducted using differentiated BMAds. The impact of Tofa was determined by gene and protein expression analysis and cell density monitoring. In parallel, in a pilot study of 9 RA patients treated with Tofa 5 mg twice a day (NCT04175886), the proton density fat fraction (PDFF) was measured using MRI at the lumbar spine at baseline and at 6 months. Results: In non-inflammatory conditions, the gene expression of Runx2 and Dlx5 decreased in Ob treated with Tofa (p <0.05). The gene expression of PPARγ2, C/EBPα, and Perilipin 1 were increased compared to controls (p <0.05) in BMAds treated with Tofa. Under inflammatory conditions, Tofa did not change the expression profiles of Ob compared to TNFα controls. In contrast, Tofa limited the negative effect of TNFα on BMAd differentiation (p <0.05). An increase in the density of differentiated BMAds treated with Tofa under TNFα was noted (p <0.001). These findings were consolidated by an increase in PDFF at 6 months of treatment with Tofa in RA patients (46.3 ± 7.0% versus 53.2 ± 9.2% p <0.01). Conclusion: Together, these results suggest a stimulatory effect of Tofa on BMAd commitment and differentiation, which does not support a positive effect of Tofa on bone.

An integrative bioinformatics approach to decipher adipocyte-induced transdifferentiation of osteoblast.

In human, bone loss is associated with increased marrow adipose tissue and recent data suggest that medullary adipocytes could play a role in osteoporosis by acting on neighboring bone-forming osteoblasts. Supporting this hypothesis, we previously showed, in a coculture model based on human bone marrow stromal cells, that factors secreted by adipocytes induced the conversion of osteoblasts towards an adipocyte-like phenotype. In this work, we employed an original integrative bioinformatics approach connecting proteomic and transcriptomic data from adipocytes and osteoblasts, respectively, to investigate the mechanisms underlying their crosstalk. Our analysis identified a total of 271 predicted physical interactions between adipocyte-secreted proteins and osteoblast membrane protein coding genes and proposed three pathways for their potential contribution to osteoblast transdifferentiation, the PI3K-AKT, the JAK2-STAT3 and the SMAD pathways. Our findings demonstrated the effectiveness of our integrative omics strategy to decipher cell-cell communication events.

Adipocyte-induced transdifferentiation of osteoblasts and its potential role in age-related bone loss.

Our preliminary findings have lead us to propose bone marrow adipocyte secretions as new contributors to bone loss. Indeed, using a coculture model based on human bone marrow stromal cells, we previously showed that soluble factors secreted by adipocytes induced the conversion of osteoblasts towards an adipocyte-like phenotype. In this study, microarray gene expression profiling showed profound transcriptomic changes in osteoblasts following coculture and confirmed the enrichment of the adipocyte gene signature. Double immunofluorescence microscopic analyses demonstrated the coexpression of adipogenic and osteoblastic specific markers in individual cells, providing evidence for a transdifferentiation event. At the molecular level, this conversion was associated with upregulated expression levels of reprogramming genes and a decrease in the DNA methylation level. In line with these in vitro results, preliminary immunohistochemical analysis of bone sections revealed adipogenic marker expression in osteoblasts from elderly subjects. Altogether, these data suggest that osteoblast transdifferentiation could contribute to decreased bone mass upon ageing.

New insights into the epigenetics of osteoporosis.

Osteoporosis is characterized by reduced bone formation and accumulation of adipocytes in the bone marrow compartment. The decrease in bone mass results from an imbalance between osteoclast-mediated bone resorption and osteoblast-mediated bone formation. The deficiency of bone cells to replace the resorpted bone can be due to a preferential differentiation of bone marrow stromal cells into adipocytes at the expense of osteoblasts. Consequently, the processes that control the differentiation of osteoclasts, osteoblasts and adipocytes play a crucial role in bone metabolism. It is known that epigenetic mechanisms are critical regulator of the differentiation programs for cell fate and moreover are subject to changes during aging. Here, we summarize recent findings on the role of epigenetics in the modulation of mechanisms that may be associated with osteoporosis. In particular, we focus on disturbances in the bone remodeling process described in human studies that address the epigenetic regulation of the osteoblast/adipocyte balance.

Variations of secretome profiles according to conditioned medium preparation: The example of human mesenchymal stem cell-derived adipocytes.

One challenging point in analyzing cellular secretome collected as conditioned medium is cross-contamination by cell culture media components, especially bovine serum proteins. A common approach for serum removal is to wash the cells, an alternative is to grow cells using serum-free conditions. Given that the sample processing may influence the phenotype of cells and thus the secretome, it is important to establish the optimal protocol for each cell type. In this study, we compared two methods for preparing conditioned medium from human adipocytes derived from mesenchymal stem cells. Cells were either washed twice with PBS or cultured the last four days of differentiation in serum-free adipogenic medium. Gene expression of the cells was evaluated by using real-time PCR and 1D LC-MS/MS was used to compare secreted proteins present in the culture supernatants. Surprisingly, results showed significant differences in gene expression patterns of the cells and in protein content of the conditioned media and suggested that PBS washes induced severe modifications of the phenotype of cells and thus changes in protein secretion profiles. These data emphasize the significant variations in protein species related to cell manipulations and underline the importance of procedure optimization prior to any proteomic investigation.

Dexamethasone in osteogenic medium strongly induces adipocyte differentiation of mouse bone marrow stromal cells and increases osteoblast differentiation.

BACKGROUND: Osteoblasts and adipocytes share a common mesenchymal stem cell origin. Therefore, it has been suggested that the accumulation of marrow adipocytes observed in bone loss is caused by a shift in the commitment of mesenchymal stem cells from the osteogenic pathway to the adipogenic pathway. Supporting this hypothesis the competition between adipogenic and osteogenic lineages was widely demonstrated on partially homogeneous cell populations. However, some data from mouse models showed the existence of an independent relationship between bone mineral content and bone marrow adiposity. Therefore, the combination of adipogenesis and osteogenesis in primary culture would be helpful to determine if this competition would be observed on a whole bone marrow stromal cell population in a culture medium allowing both lineages. In this aim, mouse bone marrow stromal cells were cultured in a standard osteogenic medium added with different concentrations of Dexamethasone, known to be an important regulator of mesenchymal progenitor cell differentiation. RESULTS: Gene expression of osteoblast and adipocyte markers, biochemical and physical analyses demonstrated the presence of both cell types when Dexamethasone was used at 100 nM. Overall, our data showed that in this co-differentiation medium both differentiation lineages were enhanced compared to classical adipogenic or osteogenic culture medium. This suggests that in this model, adipocyte phenotype does not seem to increase at the expense of the osteoblast lineage. CONCLUSION: This model appears to be a promising tool to study osteoblast and adipocyte differentiation capabilities and the interactions between these two processes.

Adipogenic RNAs are transferred in osteoblasts via bone marrow adipocytes-derived extracellular vesicles (EVs).

BACKGROUND: In osteoporosis, bone loss is accompanied by increased marrow adiposity. Given their proximity in the bone marrow and their shared origin, a dialogue between adipocytes and osteoblasts could be a factor in the competition between human Mesenchymal Stem Cells (hMSC) differentiation routes, leading to adipocyte differentiation at the expense of osteoblast differentiation. The adipocyte/osteoblast balance is highly regulated at the level of gene transcription. In our work, we focused on PPARgamma, CEBPalpha and CEBPdelta, as these transcription factors are seen as master regulators of adipogenesis and expressed precociously, and on leptin and adiponectin, considered as adipocyte marker genes. In 2010, our group has demonstrated, thanks to a coculture model, that in the presence of hMSC-derived adipocytes (hMSC-Adi), hMSC-derived osteoblasts (hMSC-Ost) express lesser amounts of osteogenic markers but exhibit the expression of typical adipogenic genes. Nevertheless, the mechanisms underlying this modulation of gene expression are not clarified. Recently, adipocytes were described as releasing extracellular vesicles (EVs), containing and transferring adipocyte specific transcripts, like PPARgamma, leptin and adiponectin. Here, we investigated whether EVs could be the way in which adipocytes transfer adipogenic RNAs in our coculture model. RESULTS: We observed in hMSC-Ost incubated in hAdi-CM an increase in the adipogenic PPARγ, leptin, CEBPα and CEBPδ transcripts as well as the anti-osteoblastic miR-138, miR30c, miR125a, miR-125b, miR-31 miRNAs, probably implicated in the observed osteocalcin (OC) and osteopontin (OP) expression decrease. Moreover, EVs were isolated from conditioned media collected from cultures of hMSC at different stages of adipocyte differentiation and these specific adipogenic transcripts were detected inside. Finally, thanks to interspecies conditioned media exposition, we could highlight for the first time a horizontal transfer of adipogenic transcripts from medullary adipocytes to osteoblasts. CONCLUSIONS: Here, we have shown, for the first time, RNA transfer between hMSC-derived adipocytes and osteoblasts through EVs. Additional studies are needed to clarify if this mechanism has a role in the adipocytic switch driven on osteoblasts by adipocytes inside bone marrow and if EVs could be a target component to regulate the competition between osteoblasts and adipocytes in the prevention or in the therapy of osteoporosis and other osteopenia.

Long-term physiological alterations and recovery in a mouse model of separation associated with time-restricted feeding: a tool to study anorexia nervosa related consequences.

BACKGROUND: Anorexia nervosa is a primary psychiatric disorder, with non-negligible rates of mortality and morbidity. Some of the related alterations could participate in a vicious cycle limiting the recovery. Animal models mimicking various physiological alterations related to anorexia nervosa are necessary to provide better strategies of treatment. AIM: To explore physiological alterations and recovery in a long-term mouse model mimicking numerous consequences of severe anorexia nervosa. METHODS: C57Bl/6 female mice were submitted to a separation-based anorexia protocol combining separation and time-restricted feeding for 10 weeks. Thereafter, mice were housed in standard conditions for 10 weeks. Body weight, food intake, body composition, plasma levels of leptin, adiponectin, IGF-1, blood levels of GH, reproductive function and glucose tolerance were followed. Gene expression of several markers of lipid and energy metabolism was assayed in adipose tissues. RESULTS: Mimicking what is observed in anorexia nervosa patients, and despite a food intake close to that of control mice, separation-based anorexia mice displayed marked alterations in body weight, fat mass, lean mass, bone mass acquisition, reproductive function, GH/IGF-1 axis, and leptinemia. mRNA levels of markers of lipogenesis, lipolysis, and the brown-like adipocyte lineage in subcutaneous adipose tissue were also changed. All these alterations were corrected during the recovery phase, except for the hypoleptinemia that persisted despite the full recovery of fat mass. CONCLUSION: This study strongly supports the separation-based anorexia protocol as a valuable model of long-term negative energy balance state that closely mimics various symptoms observed in anorexia nervosa, including metabolic adaptations. Interestingly, during a recovery phase, mice showed a high capacity to normalize these parameters with the exception of plasma leptin levels. It will be interesting therefore to explore further the central and peripheral effects of the uncorrected hypoleptinemia during recovery from separation-based anorexia.

Fibroblast growth factor 2 inhibits up-regulation of bone morphogenic proteins and their receptors during osteoblastic differentiation of human mesenchymal stem cells.

Understanding the interactions between growth factors and bone morphogenic proteins (BMPs) signaling remains a crucial issue to optimize the use of human mesenchymal stem cells (HMSCs) and BMPs in therapeutic perspectives and bone tissue engineering. BMPs are potent inducers of osteoblastic differentiation. They exert their actions via BMP receptors (BMPR), including BMPR1A, BMPR1B and BMPR2. Fibroblast growth factor 2 (FGF2) is expressed by cells of the osteoblastic lineage, increases their proliferation and is secreted during the healing process of fractures or in surgery bone sites. We hypothesized that FGF2 might influence HMSC osteoblastic differentiation by modulating expressions of BMPs and their receptors. BMP2, BMP4, BMPR1A and mainly BMPR1B expressions were up-regulated during this differentiation. FGF2 inhibited HMSCs osteoblastic differentiation and the up-regulation of BMPs and BMPR. This effect was prevented by inhibiting the ERK or JNK mitogen-activated protein kinases which are known to be activated by FGF2. These data provide a mechanism explaining the inhibitory effect of FGF2 on osteoblastic differentiation of HMSCs. These crosstalks between growth and osteogenic factors should be considered in the use of recombinant BMPs in therapeutic purpose of fracture repair or skeletal bioengineering.

Human osteoblasts derived from mesenchymal stem cells express adipogenic markers upon coculture with bone marrow adipocytes.

In osteoporosis, bone loss is accompanied by greater adiposity in the marrow. Given the cellular proximity within the bone marrow, we wondered whether adipocytes might have a paracrine impact on osteoblast differentiation. To test this hypothesis, we cocultured adipocytes with osteoblasts derived from mesenchymal stem cells (MSCs) in the absence of direct cell contact and then analyzed gene expression changes in the osteoblastic population by using real-time reverse transcription polymerase chain reaction. We found that, upon coculture, MSC-derived osteoblasts showed appearance of adipogenic (lipoprotein lipase, leptin) and decrease of osteogenic (osteocalcin) mRNA markers. Our results indicate that in vitro, MSC-derived adipocytes are capable of inducing MSC-derived osteoblasts to differentiate to an adipocyte phenotype. These new data suggest that (i) transdifferentiation of committed osteoblasts into adipocytes may contribute to the increase in marrow fat content at the expense of bone-forming cells and (ii) this switch might be initiated by the adipocytes themselves.

TNF-alpha's effects on proliferation and apoptosis in human mesenchymal stem cells depend on RUNX2 expression.

RUNX2 is a bone-specific transcription factor that plays a critical role in prenatal bone formation and postnatal bone development. It regulates the expression of genes that are important in committing cells into the osteoblast lineage. There is increasing evidence that RUNX2 is involved in osteoblast proliferation. RUNX2 expression increases during osteoblast differentiation, and recent data even suggest that it acts as a proapoptotic factor. The cytokine tumor necrosis factor alpha (TNF-alpha) is known to modulate osteoblast functions in a manner that depends on the differentiation stage. TNF-alpha affects the rate at which mesenchymal precursor cells differentiate into osteoblasts and induces apoptosis in mature osteoblasts. Thus we sought to establish whether or not the effects of TNF-alpha and fetal calf serum on proliferation and apoptosis in human mesenchymal stem cells (hMSCs) were dependent on RUNX2 level and activity. We transfected hMSCs with small interfering RNAs (siRNAs) directed against RUNX2 and found that they proliferated more quickly than control hMSCs transfected with a nonspecific siRNA. This increase in proliferation was accompanied by a rise in cyclin A1, B1, and E1 expression and a decrease in levels of the cyclin inhibitor p21. Moreover, we observed that RUNX2 silencing protected hMSCs from TNF-alpha's antiproliferative and apoptotic effects. This protection was accompanied by the inhibition of caspase-3 activity and Bax expression. Our results confirmed that RUNX2 is a critical link between cell fate, proliferation, and growth control. This study also suggested that, depending on the osteoblasts' differentiation stage, RUNX2 may control cell growth by regulating the expression of elements involved in hormone and cytokine sensitivity.

Leptin receptors and beta2-adrenergic receptor mRNA expression in brain injury-related heterotopic ossification.

Heterotopic ossification (HO) frequently occurs after brain injury. Recently, we found that leptin levels were decreased in the serum of patients with HO. Data suggest two mechanisms mediating leptin effects: a central suppressive mechanism acting via the beta(2)-adrenergic system and a direct stimulatory action starting when leptin binds to its receptors in osteoblastic cells. In this study, we analyzed leptin and beta(2)-adrenergic receptors mRNA expression in osteocytes originated from normal or heterotopic bone biopsies to investigate whether direct or indirect pathway signaling might be implicated in this pathological bone formation. We report for the first time the mRNA expression of the leptin receptor isoforms in osteocytes isolated from all biopsies. Moreover, quantitative reverse transcription-polymerase chain reaction allowed us to measure a significant decrease in the level of beta(2)-adrenergic receptor mRNA in cells isolated from heterotopic bone biopsies. These results could suggest an association between hypothalamic leptin signaling and brain injury-related HO.

Runx2 regulates the expression of GNAS on SaOs-2 cells.

Runx2 is a key regulator of osteoblast-specific gene expression and controls the expression of multiple target genes during osteoblast differentiation. Although some transcriptional targets for Runx2 are known, it is believed that the osteogenic action of Runx2 is mediated by additional target genes, and increasing studies are performed in order to identify such Runx2-responsive genes. To identify genes following the inhibition of Runx2 in osteoblastic cell line, SaOs-2 was stably transfected with a dominant negative mutant of Runx2 (Deltacbfa1) under the control of a strong promoter. Comparison of gene expression patterns by differential display on selected SaOs-2 clones allowed us to observe that GNAS mRNA which encodes for the Gsalpha protein is overexpressed (5 to 8 fold) in cells presenting high levels of Deltacbfa1. This overexpression was also observed at the protein level and seemed to be reflected by an increased basal cAMP level. Gel shift experiments performed in this study indicate that Runx2 is able to bind to the promoter of GNAS, suggesting a direct regulation at the transcriptional level. Well-described GNAS mutations like fibrous dysplasia or Albright hereditary osteodystrophy are linked to abnormality in osteoblast function, and numerous evidences showed that Gsalpha coupled adrenergic receptors increase the expression of osteotrophic factors and regulate bone mass. Regulation of Gsalpha protein by Runx2 seems to be of particular interest considering the increasing evidences on bone metabolism regulation by G proteins.

Identification of CBFA1-regulated genes on SaOs-2 cells.

Current knowledge about mechanisms controlling osteoblast-specific gene expression has led to the identification of Cbfa1 as a key regulator of osteoblast differentiation. Several essential questions about this transcription factor remain to be addressed, e.g., the nature of stimuli that may modulate its own expression, as well as the genetic repercussions following alterations in Cbfa1 levels. To identify such Cbfa1-responsive genes, the SaOs-2 cell line was stably transfected with a dominant negative mutant of Cbfa1 (DeltaCbfa1). Comparison of gene expression patterns by differential display on selected SaOs-2 clones allowed the identification of four new genes that may be under the control of Cbfa1. Three of them, SelM, elF-4AI, and RPS24, seemed to be linked to a global change in cellular metabolism and cell growth. The fourth, the CD99/MIC2 gene, was strongly overexpressed (around tenfold) in cells presenting high levels of Deltacbfa1. This observation adds evidence to show that this marker of Ewing family tumors is linked to the osteoblast lineage. The exact function of CD99 remains largely undefined, and this is the first time that its regulation by an essential transcription factor involved in osteoblast differentiation has been observed.

Gene expression in human osteoblastic cells from normal and heterotopic ossification.

Heterotopic ossification (HO), a possible complication of head injury, develops in sites where it is not normally present like at the vicinity of joints. It may cause pain, decrease motion and in severe cases complete joint ankylosis requiring surgical intervention. To our knowledge, no study has been made to analyze HO at the molecular level on human biopsies, whereas its etiology remains to be determined. We defined a procedure of cell fractionation from bone resections and developed quantitative RT-PCR to compare genetic expression patterns between human normal osteoblasts and heterotopic ossification forming cells. This quantitative study demonstrated a specific and strong overexpression of osteocalcin mRNA in HO-isolated cells associated with a significant upregulation of type 1 collagen and osteonectin mRNA while histological analysis showed only small cellular variations. Our results give a first molecular characterization of heterotopic ossification and we conclude that such overexpressions in HO-isolated cells could be associated with the high activity of this pathological bone.

In vitro control of human bone marrow stromal cells for bone tissue engineering.

For the clinical application of cultured human mesenchymal stem cells (MSCs), cells must have minimal contact with fetal calf serum (FCS) because it might be a potential vector for contamination by adventitious agents. The use of human plasma and serum for clinical applications also continues to give rise to considerable concerns with respect to the transmission of known and unknown human infectious agents. With the objective of clinical applications of cultured human MSCs, we tested the ability of autologous plasma, AB human serum, FCS, and artificial serum substitutes containing animal-derived proteins (Ultroser G) or vegetable-derived proteins (Prolifix S6) to permit their growth and differentiation in vitro. To conserve as much autologous plasma as possible, we attempted to mix it at decreasing concentrations with the serum substitute containing vegetable-derived mitogenic factors. Under control conditions, by day 10 all the fibroblast colony-forming units (CFU-Fs) were alkaline phosphatase (ALP) positive. However, their number and size were highly variable among donors. Better CFU-F formation was obtained with Ultroser G, and with human AB serum and autologous plasma mixed at, respectively, 5 and 1% with Prolifix S6. The effects of these mixtures on CFU-F formation demonstrate synergy, with the human serum or plasma supplying the factors that favor differentiation of MSCs while Prolifix S6 supplies the mitogenic factors. Finally, we demonstrated the possibility of controlling human MSC growth and differentiation in vitro. Notably, by means of a minimal quantity of human serum or human plasma mixed with a new serum substitute containing vegetable-derived proteins, we displayed growth and differentiation of human MSCs comparable to that obtained with FCS or serum substitutes containing animal-derived proteins. These results will have crucial significance for future applications of cultured human MSCs in bone tissue engineering.