Cellular and molecular effects of thiazolidinediones on bone cells: a review

Introduction

Diabetes mellitus is a chronic disease that occurs either when the pancreas does not produce enough insulin (type 1 diabetes mellitus) or when the body cannot effectively use the insulin it produces (type 2 diabetes mellitus). Worldwide it is estimated that 220 million people suffer of diabetes mellitus and among all the diabetic patients, almost 90% suffer of type 2 diabetes. Type 2 diabetes mellitus (T2DM) is largely the results of excess body weight and physical inactivity and is characterized by insulin and glucose intolerance associated with hyperglycemia and hyperinsulinemia leading to serious damage of the heart, blood vessels, eyes, kidneys, and nerves. First line therapy include life style intervention and when this action is not enough to lower sufficiently blood glucose levels, additional therapies including insulin, insulin secretagogues (sulphonylurea, incretins) or blood glucose lowering molecules (biguanides, -glucosidase inhibitors, thiazolidin-ediones, dipeptidyl peptidase inhibitors, glinides and amylin agonists) are required [1]. Among all these molecules, only thiazolidinediones (TZDs) have insulin-sensitizing properties and are the most efficient in controlling durably T2DM [2, 3]. However, there are accumulating evidences that treatment with thiazolidinediones cause bone loss and increase fracture risk in humans, specifically in women although several studies reported the same effects in men [4, 5].

Molecules of the thiazolidinedione family

Several molecules, marketed and non-marketed, belong to the thiazolidinedione family (Figure 1). All members have in common a 5-benzyl-1, 3-thiazolidine-2, 4-dione group and differ by their side chain. Among all thiazolidinedione members, pioglitazone is currently the only molecule marketed in Europe but there is a debate concerning its possibility to increase the frequency of bladder cancer [6]. Rosiglitazone was withdrawn from the european market in October 2010 due to cardio-vascular adverse events and troglitazone was withdrawn worldwide in 2000 due to severe liver problems. The vast majority of thiazolidinedione members have never been marketed but they have all in common to bind and activate partially or fully the nuclear receptor peroxisome proliferator-activated receptor gamma (PPARγ).

PPARγ and its ligands

PPARγ is a member of the PPAR subfamily of nuclear receptor. In humans, the pparγ gene is located on chromosome 3 at the 3p25 locus. Differential promoter usage and alternative splicing produce four variants, including two major forms of the protein, PPARγ 1 and PPARγ 2. PPARγ 1 is expressed in a large range of tissues, including the liver, skeletal muscle, adipose tissue and hematopoietic cells. The expression of PPARγ 2, that contains 30 additional amino acids in its N-terminus extremity, is restricted primarily to adipocytes, stromal cells and osteoblasts [7]. Upon activation, PPARγ heterodimerize with a member of the retinoid X receptor (RXR) family and the heterodimer PPAR γ/RXR activates the transcription of target genes [8]. PPARγ acts primarily as a master gene in metabolic regulation through stimulating insulin sensitivity, glucose-lowering and lipid uptake and storage in peripheral organs such as skeletal muscle, liver and adipose tissue. In particular, PPARγ is the key regulator of adipogenesis and adipocyte metabolism and also exerts anti-inflammatory and anti-atherosclerotic actions. PPARγ activity can be modulated directly or indirectly at different levels of the activation cascade. This topic has been extensively studied and is reviewed in Luconi et al [9].

Fatty acids and eicosanoids derivatives bind and activate PPAR-γ at micromolar concentrations and represent natural ligands for this receptor. Clearly, PPAR-γ prefers polyunsaturated fatty acids including linoleic acid, linolenic acid, arachidonic acid and eicosapentaenoic acid [10]. The micromolar affinity of these metabolites is in range with their serum concentrations. However, their intracellular concentration ranges are unknown. Conversion of linoleic acid to 9-HODE and 13-HODE by 15-lipoxygenase can provide additional micromolar PPAR-γ agonists [11]. A prostaglandin D2-derivative, 15-deoxy-Δ ^12,14-prostaglandin J2 (15d-PGJ2), was demonstrated to be a relatively weak (2-5(μM) PPAR-γ ligand and agonist [12, 13], although the physiological relevance of this ligand is unclear because cellular concentrations cannot be accurately determined. An oxidized phospholipid, hexadecyl azelaoyl phosphatidylcholine, was shown to bind PPAR-γ at nanomolar concentrations [14].

The first association of thiazolidinediones with PPARγ was reported by Lehmann et al. in 1995 [15]. These authors highlighted that TZDs bind and activate PPARγ in a dose-dependent manner. Among all developed TZDs, rosiglitazone has the highest binding affinity at a nanomolar concentration range (Table 1).

Bone loss and increased skeletal fragility in TZD -treated patients

Thiazolidinediones are used in patients with diabetes mellitus and as such it is a prerequisite to appreciate the effects of diabetes mellitus on bone to understand how thiazolidinediones may favor a skeletal fragility. It is well admitted that type 1 diabetes induce skeletal fragility and osteopenia. Far less is known for type 2 diabetes. Although the bone mineral density is similar or higher than non-diabetic volunteers, it appears that T2DM patients present an increased skeletal fragility. This topic has already been reviewed by Hofbauer et al. [16] and will not be detailed herein.

TZDs have been the centre of investigation during the past years and several randomized clinical trials have been undertaken comparing TZDs with placebo. All started with the results of the ADOPT study that investigated the effects of rosiglitazone over metformin or glyburide on metabolic parameters. Results from the ADOPT study highlighted the existence of an increased risk of developing bone fractures in women, but not men, treated with rosiglitazone compared with women treated in the other arms of this study [5]. Similar results with pioglitazone were also released in a note by Eli Lilly Canada Inc [17]. Schwartz et al. examined the association between TZD use (pioglitazone, rosiglitazone and troglitazone) and bone loss of elderly Americans [18]. Sixty-nine patients reported TZD use during a 4-year period of observation. Bone loss was accelerated, by 0.6-1.2% per year, at the trochanter, whole body, and lumbar spine in diabetic women who reported any TZD use compared with those who did not. Longer duration of TZDs and higher compliance with therapy were associated with more rapid rates of bone loss. However, none of the 32 male included in this trial who reported TZD use presented difference in bone mass. Grey et al. reported the results of a 14-week randomized clinical trial comparing rosiglitazone (8 mg/day) with placebo in 50 post-menopausal women who did not have diabetes mellitus or osteoporosis at the beginning of the study [19]. In the rosiglitazone group, each of two specific markers of bone formation (osteocalcin and procollagen type 1 N -telopeptide) declined by 10-12% compared with the placebo group while there was no change in serum β-CTX, the marker of bone resorption. These changes in bone turnover were accompanied by a significant 2% decline in total BMD in the rosiglitazone-treated group, even with the short time frame of this trial. Another study enrolled 56 post-menopausal women with T2DM who were randomized to rosiglitazone (4 mg/day) or dietetary advice for 12 weeks [20]. Bone specific alkaline phosphatase declined by 21% while urinary deoxypyridinoline did not change. These results demonstrate the alterations of bone metabolism (decreased bone formation whilst bone resorption is unchanged) associated with the use of clinically relevant dose of rosiglitazone in a relatively short time. These results or somehow similar to those induced by treatment with glucocorticoids [21].

Thiazolidinediones and bone cells

Three different categories of cells participate to the physiological bone remodeling, osteoclasts, the bone-resorbing cells; osteoblasts, the bone-forming cells and osteocytes, cells embedded in the bone tissue that control bone resorption and bone formation.

Conclusion

In conclusion, TZDs participate to increase adipogenesis at the expense of osteoblastogenesis and as such probably contribute to the reduction of bone mass observed with TZD treatment. TZDs also contribute to the death of osteocytes and the increased expression of sclerostin that in return certainly could emphasize and aggravate the reduction in bone formation. On the other hand, effects of these drugs on osteoclasts seem more controversial and surely require more studies before totally elucidating in which proportion these cells participate to the skeletal phenotype observed with TZD treatment.