Introduction: Spinal deformities and scoliosis in particular, represent the most prevalent type of orthopaedic deformities in children and adolescents. At present, the most significant problem for clinicians is that there is no proven method or test available to identify children or adolescents at risk of developing AIS or to identify which of the affected individuals are at risk of progression. As a consequence, the application of current treatments, such as bracing or surgical correction, has to be delayed until a significant deformity is detected or until a significant progression is clearly demonstrated, resulting in a delayed and less optimal treatment. Among patients with AIS needing treatment, 80% to 90% will be treated by brace and 10% will need surgery to correct the deformity by spinal instrumentation and fusion of the thoracic and/or lumbar spine. About 15000 such surgeries are done every year in North America, resulting in significant psychological and physical morbidity. Moreover, there is no pharmacotherapy available to either prevent or reduce spinal deformities due mainly to our limited knowledge of AIS aetiopathogenesis. We have recently reconciled the role of melatonin in AIS aetiopathogenesis by demonstrating a melatonin signalling dysfunction occurring in a cell autonomous manner in cells derived from AIS patients exhibiting severe scoliotic deformities. This defect could potentially explain the majority of abnormalities reported in AIS since melatonin receptors and signalling activities are normally found in all tissues and systems affected in AIS, thus offering a very innovative and unifying concept to explain the aetiology of AIS. Moreover, several lines of evidence suggested that inactivation of Gi proteins by an increased phosphorylation of serine residues could be at the source of this signalling defect in AIS. The goals of that study were to assess the possibility to establish a molecular classification of AIS patients and to demonstrate the feasibility to correct this melatonin signalling defect in cells of AIS patients using therapeutic compounds.

Methods: Primary cell cultures were prepared from musculoskeletal tissues of AIS patients (n=150) and age- and gender-matched controls (n=35) obtained intra-operatively. An informed consent was obtained for each subject as approved by our Institutional Ethical Committee. The osteoblasts, the bone-forming cells, were selected to assess whether or not an alteration of melatonin signalling pathway occurs in AIS and accordingly to identify which component of the melatonin transduction machinery could be involved. Co-immunoprecipitation experiments with membrane extracts were performed to identify interacting molecules with key components of melatonin signal transduction machinery. The functionality of melatonin signalling was assessed by investigating the ability of Gi proteins to inhibit stimulated adenyl cyclase activity in osteoblast cultures. Inhibition curves of cAMP production were generated by adding melatonin to the forskolin-containing samples in concentrations ranging from 10-11M to 10-5M in a final volume of 1 ml of _-MEM media containing 0.2% bovine serum albumin (BSA) alone or in presence of 2.5 _M of therapeutic compound A or therapeutic compound B (the nature of both compounds tested cannot be disclosed at this stage). The cAMP content was determined using an enzyme immunoassay kit (Amersham-Pharmacia Biosciences). All assays were performed in duplicate. A non-parametric test, the Wilcoxon matched pairs test was performed to verify the significance between 2 means. Significance was defined as P< 0.05.

Results: Osteoblasts from patients with AIS showed a lack or a markedly reduced inhibition of forskolin-stimulated adenyl cyclase activity by melatonin generating three distinct response-curves corresponding to three functional groups. In order to identify candidate genes involved in AIS aetiopathogenesis, we focused our attention on known kinases and phosphatases modulating Gi protein functions and characterised their interacting partners. Interestingly, PKC_ was initially targeted owing to its property to phosphorylate Gi proteins in vitro. Indeed, in normal osteoblast interactions occurring between MT2 melatonin receptor and RACK1 (a cytosolic protein that bind to and stabilises the actives form of PKC and permits its translocation to different sites within the cells) and PKC_ were detected although those interactions among different AIS patients were altered. Interestingly, treatment with compound A or B rescued melatonin signal defect in cells derived from 36% and 47% of AIS patients respectively. Overall, melatonin signal transduction was restored in cells of 64% of AIS patients (23/36) when treated by one of these therapeutic compounds.

Conclusions: The functional classification of AIS patients is correlated at the molecular level by distinct interactions between key molecules normally involved in melatonin signal transduction in spite that these patients exhibited the same curve type (right thoracic, Lenke type 1). Collectively, these data strongly argue that traditional curve pattern classification is not a relevant stratification of AIS patients to identify its genetic causes. Moreover, using that molecular system we have demonstrated also the possibility to identify therapeutic compounds to rescue the melatonin signalling defect observed in AIS without any prior knowledge of mutations in any defective genes causing AIS because we are measuring a function.

Research project supported by La Fondation Yves Cotrel de l’Institut de France