Most experts agree that the causes of adolescent idiopathic scoliosis (AIS) are multifactorial with no generally accepted theory of pathogenesis (Appendix 1) 14 51 57 58 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 . This reflects shortcomings in our understanding of the complex biological and biomechanical multifactorial processes involved in AIS pathogenesis which needs innovative thinking 73 , to which we add new findings not explained by prevailing theories. One recent review suggests that genetics and the unique mechanics of the fully upright human spine play a decisive role in AIS pathogenesis 75 . A genome-wide association study revealed 30 markers identified as the most useful prognostically 56 .

A commonly held pathogenetic theory is that initiating changes in the spine of unknown origin lead to biomechanical spinal growth modulation causing curve progression 80 81 82 107 . Brace treatment is based on this view of pathogenesis.

Studies over many years in AIS subjects have shown abnormalities of visual, vestibular, proprioceptive and postural control 67 69 70 94 95 96 99 100 101 102 103 104 involving the brain stem 69 95 97 98 99 , cerebral hemispheres and corpus callosum 69 95 104 111 112 113 114 115 , though not without controversy. Lowe et al 67 suggested that the pathogenesis of adolescent idiopathic scoliosis (AIS) results from a primary pathology in the hind brain causing a defect of central control, or processing in the central nervous system (CNS) that affects a normal growing spine 116 . Neurological abnormalities with AIS have been explained by four fairly comprehensive concepts for pathogenesis:

(1) visuo-spatial perceptual impairment producing a motor control problem 104 ;

(2) body-spatial orientation concept 69 ;

(3) neurodevelopmental concept 105 106 ; and

(4) sensory integration disorder 102 .

Abnormal asymmetries of brain structure and function are found in AIS girls for each of cerebral hemispheres 112 113 114 115 , dichotic listening 112 , brain stem 97 98 99 and, in preliminary research for left thoracic AIS, on MR brain scans, reduced white matter density in the left internal capsule and corpus callosum 114 115 .

Summarizing concepts of AIS pathogenesis in 2008 51 , we suggested a novel neuro-osseous escalator concept for AIS in girls (Figures 1, 2 and 3). This involves interaction between the growing skeleton and postural mechanisms of the maturing somatic nervous system. The dependence of AIS progression on growth is attributed not to growth\velocity, but to rapid skeletal enlargement hormonally-induced, producing skeletal sizes for age beyond the capacity of postural mechanisms of the somatic nervous system to control the initiating deformity.

Later in 2008, from analyses of anthropometric data of adolescent girls - normal, screened and preoperative, we reported that relatively higher and lower subsets of body mass index (BMI) reveal different features of skeletal maturation 46 47 117 118 119 and asymmeties of spinal deformity and upper arm lengths 46 120 121 (Figure 1). Subsequently, skeletal overgrowth patterns for age were found in preoperative AIS girls compared with normal girls when analysed separately by higher and lower BMI subsets 29 122 . Then, in normal girls and boys, an excess of severe back humps was found to be associated with lower BMI subsets 123 124 125 . These and other findings were not explained by any of the theories surveyed (Appendix 1, items 1-15). A more comprehensive hypothesis for AIS pathogenesis in girls was needed incorporating energy homeostasis (bioenergetics) and the hypothalamus in a disorder presenting as abnormalities of trunk growth with axial and appendicular skeletal asymmetries and systemic skeletal features in preoperative girls. The components included in the new formulation are white adipose tissue, leptin, hypothalamus and sympathetic nervous system (LHS concept). Together with the escalator concept, they form the double neuro-osseous theory (Figure 1). It has common ground with the thoracospinal concept 59 60 61 62 63 . These findings for AIS girls and the severe trunk asymmetry of healthy adolescents 123 124 125 are consistent with the hypothesis that the control mechanisms of bioenergetics have relevance to the etiopathogenesis of such shape deformities/distortions.

Sporns and Edelman 126 wrote:

"There is overwhelming evidence that the emergence of coordinated movements is intimately tied to both the growth of musculoskeletal system and to the development of brain. The neural development and learning cannot be considered outside of their biomechanical context. A key theoretical issue is how the changes in brain circuitry controlling muscles and joints become matched to simultaneously occurring developmental changes at the periphery."

The CNS body schema in adults is defined as a ".....system of sensory-motor processes that continually regulate posture and movement - processes that function without reflective awareness or the necessity of perpetual monitoring." 127 . This control involves the posterior parietal cortex which participates in the dynamic representation of the body schema integrated with other cortical areas 127 128 129 130 .

We postulate that during normal growth and maturation, a physiological balance is continuously renewed between two synchronous polarized processes we term neuro-osseous timing of maturation (NOTOM) escalators (Figure 2) 24 51 111 , namely:

(1) Osseous escalator. Increasing skeletal size, changing skeletal shape and relative mass of the different body segments which, through posture and motion of the body by producing developmental biomechanical and kinematic changes at the periphery, create developmentally-altering proprioceptive and visuo-spatial inputs to the neural escalator in the brain.

(2) Neural escalator and postural control. The brain and CNS body schema are recalibrated as they continuously adjust to skeletal enlargement, shape and relative mass changes to enable them to coordinate motor actions. The posterior parietal cortex (Area 7) in human clinical and experimental studies has been shown to participate in the dynamic representation of the CNS body schema (Figure 2) 127 128 129 130 . Leptin functionally enhances NMDA receptors which are critically involved in most models of learning and memory 131 132 . Increased circulating leptin levels may explain the reduced grey matter of certain brain areas in obese subjects 133 .

The term escalators are applicable only during growth. Muscles are not included in this terminology because they do not primarily drive skeletal growth, but have key roles in sensory and motor function and contribute to segmental masses. Similar mechanisms are being evaluated in robotics and specifically the learning in, and from, brain-based devices 134 .

Figures 1 and 3 provide an outline of the escalator concept for AIS pathogenesis in girls. Putative abnormalities of the two polarized components of the escalators - with asynchrony and asymmetry(ies) - provide the mechanisms of the escalator concept for AIS pathogenesis before and during the curve acceleration phase 5 in:

(1) spine growing rapidly with asymmetry(ies), and

(2) brain and CNS body schema with -

a) postural maturational delay, and/or

b) brain asymmetry(ies) 113 114 115 .

Postural maturational delay in the CNS may be relative to earlier skeletal maturation 135 136 137 138 139 140 141 , or absolute arising from an abnormality in afferent 100 101 102 103 142 143 144 145 , central 104 113 , or motor mechanisms 104 146 . A study of stroke subjects suggests that in axial postural control, the right hemisphere undertakes higher-order spatial processing than the left hemisphere [ 147 , see 148 ].

The fate of early AIS - to progress, become static or resolve (rarely) according to the double neuro-osseous theory generally depends on the relative contribution and outcome of the disharmony (Figure 1) between:

a) vertebral growth plate asymmetries in up to three dimensions arising wholly or in part from dysfunction in the autonomic nervous system 25 26 27 28 29 ;

b) postural control, with or without asymmetries, of a rapidly enlarging and actively moving 52 71 149 adolescent spine; and

c) postural maturity (see Discussion, Explanations for undisputed facts about AIS, (2) Predilection for females b)).

Bipedal mice and the protection by melatonin. Machida et al 150 suggested that the scoliosis development in bipedal melatonin-deficient mice and the protection from scoliosis by restoring melatonin levels, are crucial influences for a postural mechanism and bipedalism in scoliosis development. Deficiency of osteopontin or CD44 receptor also protect transgenic melatonin-deficient C57Bl/6J mice from scoliosis 19 20 . Later, we examine whether the scoliosis of these three mouse models may be markers of stress reactions involving the hypothalamus rather than crucial influences for scoliosis development (see Scientific basis of leptin-hypothalamic-sympathetic nervous system (LHS) concept, items 11 & 12).

Little discussed features of AIS pathogenesis are:

• Prescoliotics of both sexes show body height, sitting height, and growth of sitting height greater than in non-scoliotic children 135 136 .

• Early radiological maturation at 11-12 years of age in AIS subjects 137 .

• Early adolescent skeletal growth attained for age by AIS girls 38 39 41 121 135 136 137 138 139 140 141 . In the preoperative AIS girls of the relatively higher BMI subset, all the skeletal parameters we measured when plotted as standard deviation scores against age, showed negative regressions - several statistically significant, but not for the lower BMI subset of preoperative AIS girls (unpublished observations).

Together, these observations suggest that, collectively, AIS girls have a growth pattern different from normal, involving growth factors connected to the disease 137 151 , confirmed in subsequent research 64 65 89 90 152 .

Periapical ribs longer on the concavity of right thoracic AIS in elderly scoliosis cadavers were found 30 and given pathogenetic significance, but the finding is controversial 31 63 . In thoracic idiopathic scoliosis, upper arm length asymmetry (relatively longer on convexity) is significantly associated with each of apical vertebral rotation (AVR) and Cobb angle 32 . Also in scoliosis subjects but with lower spine scoliosis (thoracolumbar and lumbar), iliac height asymmetry (relatively taller on concavity) is associated with Cobb angle and apical vertebral rotation 34 , confirming an observation for subjects with lumbar scoliosis 33 .

It is unknown whether these asymmetries of upper arm, iliac height and also femoral anteversion 38 39 are pathogenetically-related to any local asymmetry in the AIS spine. We speculate that they are 24 25 32 34 35 36 37 40 41 105 106 120 121 . In this connection we outlined evidence supporting a common pathogenesis of upper arm length asymmetry and thoracic AIS spinal deformity 32 . In a similar way that the extraspinal general skeletal overgrowth for age in AIS girls is associated with the relative anterior spinal overgrowth (RASO) 64 65 89 90 giving it pathogenetic significance, we view the abnormal asymmetry of paired bones as sentinels of vertebral and/or rib growth plate asymmetries and having pathogenetic significance. There is some evidence of a primary vertebral growth plate disorder in AIS (Figure 1) 43 44 65 90 . Extra-spinal skeletal length asymmetry is also found in ilio-femoral lengths 35 . More such asymmetries need to be sought in other bilateral bones of AIS girls - sacral alae 153 154 155 , clavicles and scapulae.

In girls with AIS and young adults with scoliosis, lower body mass index 157 158 159 160 161 162 163 164 165 has been found by most but not by all workers 46 135 166 167 These findings have implications for body development, abnormal spinal development, or nutrition of patients with AIS 165 . There is some evidence of disordered eating behavior 159 168 169 , but the low body-mass index of girls with AIS is said not to be the result of the eating disorder 168 .

There is a trend towards increasing numbers of adolescents with AIS in the overweight category 170 171 . The hypothesis that increased BMI can influence scoliosis presentation was tested in 427 adolescents with idiopathic scoliosis 170 . Female subjects who presented with larger curves (>50 degrees) were older and had a greater BMI than those with curves less than 50 degrees (p = 0.0557). Possible curve detection difficulties, endocrine factors and an earlier puberty with increased fat mass were suggested for the association of the larger curves with obesity.

In humans, common variants at only two loci, FTO and MC4R (melanocortin-4 receptor) have been reproducibly associated with body mass index (BMI) 172 173 . Mutations of MC4R are the leading cause of severe childhood-onset obesity 172 . A meta-analysis of 15 genome-wide association studies for BMI identified six additional loci, including SH2B1 173 . Several of the likely causal genes are expressed, or known to act in the central nervous system 172 173 174 . Different versions of the human gene FTO strongly correlate with BMI 174 ; the FTO gene with significant polymorphic variation has been identified in several papers as a candidate gene predisposing to obesity. In rats, fto is significantly up-regulated (41%) after food deprivation 174 . In humans, fat mass, and genetic markers for obesity genes MC4R and FTO, are strongly related to bone mineral content, total body and regional, measured by DXA 175 .

SH2B1 is a strong prior candidate for regulating body weight; it is implicated in leptin signaling; Sh2b1-null mice are obese; and the evidence suggests that the effects of this gene on obesity are mediated through the central nervous system 173 (see Leptin, hypothalamus and AIS).

Overall, these findings support the view that fat mass is on the causal pathway for bone mass in normal children 175 .

In mice, loss of the Fto gene leads to postnatal growth retardation, reduction of adipose tissue and serum leptin, increased energy expenditure, enhanced circulating levels of adrenaline and noradrenaline; these changes are attributed to sympathetic system activation (sympathoactivation) controlling energy expenditure through mitochondria and fatty acids/triglycerides 176 177 . In Fto-deficient mice the sympathoactivation associated with decreased circulating leptin levels is similar to the hypothalamic up-regulation and sympathoactivation we postulate for AIS girls, but without the skeletal overgrowth for age (see Autonomic nervous system - leptin-hypothalamic-sympathetic nervous system (LHS)-driven mechanism in health and LHS concept in AIS).

Most previous research on AIS has evaluated BMI as a sole parameter, or in relation to a few skeletal features 163 164 167 . The genetic aspects of BMI for AIS have not been reported but it may be difficult in such research to disentangle the contributions of lower BMI from that of the AIS.

Our recent findings for AIS girls show that higher and lower BMI subsets relative to median BMI values for age have different patterns by each of (1) skeletal sizes for age, (2) bilateral skeletal length asymmetries, and (3) skeletal overgrowth for age in preoperative AIS compared with normal girls, which is systemically distributed suggesting hormonal effects.

In leptin-deficient mice (ob/ob) altered leptin signaling has significantly different effects on bone growth in the axial and appendicular skeletons 179 . Compared with normal mice, leptin-deficient mice have significantly shorter femora, and significantly increased vertebral lengths, a trend confirmed in subsequent research 180 . Suggested reasons for this axial/appendicular skeletal growth difference in mice include: (1) decreased thigh muscle mass as a factor for the femoral shortening through mechanotransduction pathways 179 ; and (2) vertebral growth plates respond to absent leptin signals in a fundamentally different manner from long bone growth plates 180 . The latter interpretation is consistent with the view that leptin-deficient mice have energy priority of vertebral linear growth relative to limb bones, in contrast to the energy priority of trunk width growth in girls (Figure 4). This apparent human/mouse difference is consistent with an evolutionary change to the trunk broadening of hominins (Figure 5) (hominins include living humans and fossil species that are ancestral to living humans, see Evolutionary Origins).

In the lower BMI subset, mean upper arm length asymmetry (7.0 mm, right minus left)) is significantly greater preoperative than in screened (-0.8 mm) and normal girls (2.1 mm)(each p < 0.001 with statistically significant variance ratios). In the higher BMI subset, mean upper arm length asymmetries are respectively 3.7 mm, 1.1 mm, and 2.4 mm, greater in preoperative than screened girls (p = 0.031) (analyses of variance correcting for age) 46 .

Figure 6 181 shows that apical vertebral rotation is significantly associated with upper arm length asymmetry for the lower, but not higher BMI subset, also for Cobb angle (p < 0.001, r = 0.510) 46 120 121 . These findings suggest that the abnormal upper arm length asymmetry of thoracic AIS 32 is not secondary to the spinal deformity but has a pathogenesis common to the spinal deformity 32 .

In girls with right thoracic AIS, mean upper arm length asymmetry is significantly greater than normal girls (5.6/2.2. mm, p < 0.001). The asymmetry is similar at 11-12 years of age in both higher and lower subsets. It negatively regresses on age in the higher BMI subset (p < 0.001, r = -0.486) but not significantly in the lower BMI subset (p = 0.125, r = -0.212, variance ratio of lower to higher BMI subset = 2.05, p < 0.01); and menarcheal age negatively regresses on upper arm length asymmetry in the higher BMI subset (p = 0.027, r = -0.325). This 'transient' asynchronous upper arm length growth detected with abnormal systemic earlier skeletal overgrowth for age as in some younger preoperative girls (Figure 7), suggests a relation to pathogenesis. There were insufficient girls with left thoracic AIS for separate analyses (n = 12 46 ) (see Discussion, Upper arm length asymmetry and the higher BMI subset of right thoracic AIS, and Skeletal asymmetries and lower BMI subsets).

Figure 7 182 shows that with relatively higher BMIs, the younger AIS girls, have larger corrected stature for age than do the normal girls, becoming normal sizes by 16 years of age (p < 0.001, ANOVA with age correction). This pattern is found in each of 11 skeletal segments, four of them in bilateral limb segments suggesting a systemic response. Mean menarcheal ages are not significantly different. The skeletal pattern for age suggests earlier skeletal maturation with overgrowth in these younger girls probably from circulating hormones ? GH/IGF-I and possibly estrogen 29 122 . The AIS girls with relatively lower BMIs show a more complex pattern with two growth phases: earlier phase similar to normals, and later phase in most skeletal segments, mainly postmenarcheal, with larger overall skeletal growth attained for age in preoperatives relative to normals, ? estrogen effect 29 122 . The similar mean Cobb angle and apical vertebral rotation show that while curve severity at the time of surgery appears independent from (1) skeletal growth patterns, and (2) BMI subsets, we suggest that common factors in different proportions and other common factors, determine the similar curve severities in both subsets (see Discussion Skeletal sizes for age - curve severity, sympathoactivation and hormonal stimulation).

The excess of severe back humps in girls and boys was associated with lower BMI subsets 123 124 125 .

A more comprehensive hypothesis for girls with AIS was required involving energy homeostasis and the hypothalamus in a disorder presenting as abnormalities of trunk growth with axial and appendicular skeletal asymmetries and in preoperative girls with systemic skeletal features.

Right thoracic, but not left thoracic AIS in girls, is considered by Sevastik and colleagues to be initiated by dysfunction of the sympathetic nervous system leading through vascular changes to relative overgrowth of concave periapical rib lengths 59 60 61 62 63 . This section is written in collaboration with Professor JA Sevastik. Compared with right thoracic AIS, the pathogenesis of left thoracic AIS in girls remains relatively unexplored [ 114 115 , see DISCUSSION (6)]. The thoracospinal concept of pathogenesis was established from anatomical and clinical evidence including left-right asymmetries of thoracic skin temperature, breast size and vascularity, and periapical rib length asymmetry 30 . Subsequent experimental studies 61 provided evidence for the correction of experimentally-induced scoliosis consistent with the pathogenetic conclusions. The thoracospinal concept is supported by recent studies on breast size 183 , vascular 184 185 and peripheral nerve 186 findings. It does not encompass evidence relating to the new neuroskeletal biology, energy homeostasis, or white adipose tissue which is central to the regulation of energy balance by adipokines, particularly leptin, hormones of the digestive system and metabolites, particularly glucose (Figure 8).

Biomechanical mechanisms are thought to be involved in pathogenesis. Evidence 60 showed that gradual elongation of one rib affects the position of the numerically corresponding vertebra in the three cardinal planes in a way similar to the apical vertebra in idiopathic scoliosis. The disc space wedging is explained by the rotational movement of the central vertebra in the frontal plane, and the lordotic tendency of the scoliotic segment is explained by ventral vertebral translation in combination with tilt in the sagittal plane. Curve progression is attributed to biomechanical mechanisms 63 80 81 82 .

In the last decade it was shown initially in mice, that the central nervous system regulates bone remodeling, and more recently longitudinal bone growth via the sympathetic nervous system linking leptin-responsive hypothalamic neurons to bone tissue 187 188 189 190 191 192 193 194 195 196 197 198 . In reviewing this new field of neuroskeletal biology, Patel and Elefteriou 195 summarize long-standing clinical observations relating to bone and the nervous system including reflex sympathetic dystrophy, hyperplastic callus associated with head injury and myelomeningocoele, and osteopenia associated with stroke, spinal cord injury and peripheral neuropathy. Conflicting reports on the effect of β-blockers for risk of fractures are published, and randomized clinical trials are needed 199 . Theoretically, neuroskeletal mechanisms expressed via the sympathetic nervous system through its bilaterality (Figure 5), could create asymmetries, although from animal experiments there is no evidence for or against such asymmetries.

Bodily energy reserves are managed actively by complex systems that regulate food intake, substrate partitioning and energy expenditure thereby regulating long-term adiposity 200 . Energy homeostasis, fat and glucose metabolism are regulated by integratory centers in the central nervous system which receive, and convey information by signals from peripheral organs (such as adipocytes, gut and pancreatic islets - eg insulin and amylin both short-term satiety signals, the latter being a hind brain signal), and which send efferent neural and hormonal signals to peripheral tissues that regulate food intake, energy expenditure, metabolism and behavior (feeding) 200 201 202 203 . The obesity genes MC4R, FTO and SH2B1 may participate in the central control of energy homeostasis 172 173 174 200 203 . A neuroanatomical framework explaining the effects of leptin on neuroendocrine and sympathetic nervous system function has been reported 204 .

Adipose tissue, where fatty acids are stored as triglycerides in lipid droplets, is central to the regulation of energy balance 205 . White adipose tissue constitutes separate depots that contribute with the hypothalamus as the key centre for integration and control of energy balance 200 . Leptin, best known as a satiety hormone, a signal of energy sufficiency and long-term adiposity, is one of several cytokine-like hormones secreted by adipocytes 1 2 200 . In girls there are gradual age- and BMI-related increases in circulating leptin levels 206 . Molna-Carballo et al 12 from a longitudinal study reported that the leptin concentration increases in both sexes with the progression of puberty, this value being 40% greater in girls, which correlates with the increase in body volume and fat accumulation 206 207 . Girls have higher serum leptin levels before, during, and after puberty than boys, even after accounting for the development of greater female adiposity 207 . The sexual dimorphism in leptin concentrations during puberty appears to be partly due to a stimulatory effect of estradiol on fat deposition and leptin concentration in females and a suppressive effect of testosterone on leptin concentration in males 207 . Leptin levels in men are lower than women at all decades of life 208 .

Leptin, the product of the obesity gene (ob) circulates in both free and bound form, and targets neurons including the arcuate nucleus and other nuclei of the hypothalamus 200 . Leptin is a master hormone that acts via a specific receptor (OB-R with six types of receptor, LepRa-LepRf; the longest form, LepRb is the only receptor isoform that contains active intracellular signaling domains). The leptin receptor is present in a number of hypothalamic nuclei, where it exerts its effects. within a complex web of signals with many regulatory functions for food intake, body weight, increasing energy expenditure through sympathoactivation, thermogenesis, other metabolic and endocrine functions, reproduction, immune/inflammatory responses, and wound healing, mainly through signaling to the hypothalamus including 1 2 200 209 :

a) appetite repression and body weight control (anti-obesity, anorexigenic);

b) initiation of puberty in girls as one gate with kisspeptin in a permissive role 1 2 ; genetic variation in LIN28B on chromosome 6 is associated with the timing of puberty 210 ;

c) stimulation of the sympathetic nervous system, more in females than in males, possibly because of their greater fat mass 211 212 ;

d) in bone formation, anti-osteogenic in mice acting centrally through the sympathetic nervous system 187 188 189 190 191 192 194 195 196 197 213 involving the molecular clock and circadian regulation 214 , possibly with an opposite direct effect on bone 190 195 196 198 . Several genes are identified having high levels of expression in the hypothalamus 192 195 196 . Mice lacking β-adrenergic receptors have increased bone mass 215 . In feedback, the skeleton exerts an endocrine regulation of energy metabolism through the Esp gene exclusive to osteoblasts controlling secretion of the hormone-like substance osteocalcin 216 217 218 (Figure 8).

Animal experimentation suggests a two-way interaction between leptin and the sympathetic nervous system, with leptin causing sympathoactivation, and the sympathetic nervous system exercising regulatory feedback inhibition over leptin release 219 .

Leptin stimulates longitudinal bone growth in leptin-deficient (ob/ob) and leptin-receptor deficient (db/db) mice 180 194 220 221 222 , and growth plates in culture 180 222 223 224 being chondro-osteogenic and angio-genic 190 . The leptin appears to act centrally through the sympathetic nervous system (Figure 8) 190 194 213 , growth hormone stimulation 180 190 220 222 , and peripherally 190 222 with a direct effect on growth plate chondrocytes by its signaling receptor 180 220 222 , regulating IGF-I receptor expression 190 223 , and by other mechanisms (Figure 9) 180 . There is evidence for mice, that vertebral body growth plates may respond to leptin differently from long bone growth plates 179 180 . Iwaniec et al 194 propose that hypothalamic leptin plays a role in coupling energy homeostasis and bone growth, acting as an important permissive factor for normal bone growth. Leptin appeared in evolution with the bony skeleton 216 .

Maor et al 223 reviewed clinical evidence that after craniopharyngioma surgery in children, circulating leptin may contribute to bone growth including normal height velocity 225 . Children with exogenous obesity usually show increased height velocity 226 , and their serum leptin levels are approximately five times that of normal children 227 , with obese children being taller than average from 6-9 years 225 , showing more advanced bone age/chronological age 227 , earlier puberty and menarche 226 and no significant correlation of leptin and estradiol levels 228 .

Montague et al 229 reported two severely obese consanguinous children with congenital leptin deficiency, the findings of which strongly suggested that leptin critically influences energy balance in prepubertal humans. One child developed abnormalities of growth in long bones of her legs treated by corrective surgery, an abnormality attributed to growth plate fragility 180 . Subsequently, in three children who were congenitally deficient in leptin and morbidly obese, Farooqi et al 230 reported radiological skeletal maturation was increased by 2.1 years, and that leptin therapy produced beneficial effects on the skeleton.

Severe dietary restriction, a common cause of leptin insufficiency and growth/length restriction in humans 194 , is probably associated with, and explained by, decreased GH and IGF-I receptors in growth plates 231 .

Qiu and colleagues 163 164 reported a marked decrease in circulating leptin in AIS girls compared with controls, confirmed by Dr A Moreau (personal communication). Positive correlations were found between leptin and each of age, menstrual status, weight, corrected height, BMI, Risser sign, bone mineral content and bone mineral density (lumbar spine and femoral neck) but not Cobb angle, suggesting that leptin may play an important role in the lower BMI of AIS girls 164 . Longitudinal studies are needed.

Central leptin resistance is defined as reduced ability of circulating leptin to suppress appetite and weight gain and to promote energy expenditure 232 .

In obesity. Central leptin resistance isconsidered to be one of the main causes of obesity 232 233 . It is thought to result mostly from a state of diminished hypothalamic responsiveness to increased levels of circulating leptin 200 which may be selective 232 233 234 235 236 .

In healthy females: normal juvenile girls and somatotropic axis. Central leptin resistance may occur normally in girls 227 , and in pregnancy thereby permitting the accumulation of adipose tissue stores necessary for growth, reproduction and lactation 227 237 ; leptin sensitivity returns, possibly by signaling mechanisms 232 , or by altering the leptin dose-response curves 223 238 . There is preliminary evidence 50 suggesting that the hypothalamus of some normal juvenile girls, but not boys, functions with central leptin resistance of the somatotropic (growth hormone/IGF) axis. This putative mechanism, is interpreted as limiting energy invested in female skeletal growth thereby conserving energy for reproductive development 50 . It may be related to the female predisposition to AIS.

Several mechanisms have been revealed to explain central leptin resistance in obesity 232 , namely:

(1) Impaired leptin transport across the blood-brain barrier e.g. triglycerides 238 239 240 .

(2) Serum leptin interacting proteins (SLIPS) such as C-reactive protein [ 241 , but see 200 ].

(3) Inflammation 239 242 .

(4) Intracelluar inhibitory molecules (negative regulators) of leptin signaling including -

a) the suppressors of the cytokine signaling (SOCS) family 200 243 244 ,

b) protein-tyrosine phosphatases (PTPs) 200 245 246 ; and

c) OB-R gene related protein (OB-RGRP) 247 248 .

a) Suppressors of the cytokine signaling (SOCS). Howard et al 243 and Mori et al 244 noted that the leptin receptor is highly expressed in the hypothalamus and belongs to the cytokine-receptor superfamily that activates the Janus tyrosine kinase-signal transducers and the activators of transcription (JAK/STAT) pathway to modulate cellular responses in a negative feedback loop [ 249 250 , for detail and other pathways see 232 ]. They report evidence for mice that SOCS-3 neuronal deletion enhances leptin sensitivity 244 250 as does haploinsuffiency of SOCS-3 243 . SOCS-3 is also a human gene. SOCS-2, a genetic determinant of height growth in normal children, is involved in the regulation of IGF-I signaling 251 .

b) Protein-tyrosine phosphatases (PTPs). PTP-1B also contributes to leptin resistance by inhibiting intracellular leptin receptor signaling by inhibiting JAK2 activation 232 240 252 . PTP-1B deficient mice by knockout and by an antisense (anti-DNA) oligonucleotide designed to blunt the expression of PTP-1B, showed improved leptin and insulin action 252 . PTP-1B is a major regulator of energy balance, insulin sensitivity, and body fat stores 246 . PTP-1B is also a human gene.

c) OB-R gene related protein (OB-RGRP). Couturier and colleagues 247 248 253 report that OB-RGRP negatively regulates the specific leptin receptor OB-R in the hypothalamus of mice. They comment that if the results obtained in the diet-induced obesity mouse model are transposable to humans, targeting the regulator of the leptin receptor rather than the receptor itself (either by RNA interference or by pharmacological antagonists), could be a more appropriate basis for identifying potential new therapeutic targets for a variety of diseases, including obesity.

(5) Intracelluar stimulatory molecules (positive regulators) of leptin signaling. According to Morris and Rui 232 , SH2B1 enhances leptin signaling. It appears to be required for the maintenance of leptin sensitivity, energy balance and body weight, ultimately through activation of the PI 3 kinase pathway. The ability of SH2B1 to enhance leptin sensitivity may be modulated by other members of the SH2B family. Cellular leptin sensitivity may be determined, at least in part, by a balance between positive (e.g. SH2B1) and negative (e.g. SOCS3 and PTP-1B) regulators.

(6) Chronic endoplasmic reticulum (ER) stress, mediated through protein tyrosine phosphatase 1B and not through suppressors of cytokine signaling-3 233 , contributes to leptin resistance and obesity, presumably by activating various unfolding protein response signaling pathways, 232 . Inhibition of ER stress in the hypothalamus by either genetic or pharmacological means markedly improves leptin sensitivity and decreases food intake and body weight in mice 232 .

(7) Defects in neural circuitry including impairment of MC4R signaling in the paraventricular nucleus, induce leptin resistance, hyperphagia and obesity, with genetic and environmental factors modulating the synaptic remodeling and rewiring of this circuitry 232 .

The challenge is to develop diagnostic approaches for the different forms of central leptin resistance and design personalized healthcare programs to treat obesity 232 .

Lowe et al 21 22 suggested that altered paraspinal muscle activity explained the relationship between platelet calmodulin level changes and Cobb angle changes in AIS with calmodulin acting as a systemic mediator of tissues having a contractile system (actin and myosin). An alternative speculative concept to explain the findings of Lowe is that in predisposed subjects, platelet activation with calmodulin changes occurs within dilated vessels of deforming vertebral bodies 107 . The activated platelets in juxta-physeal vessels release growth factors which, after extravasation, abet the hormone-driven growth of the already mechanically-compromised vertebral endplate physes to promote the relative anterior spinal overgrowth (RASO) and curve progression of AIS.

Moreau et al 19 20 found all transgenic melatonin-deficient C57Bl/6J mice 150 devoid of OPN or CD44 receptor were protected against scoliosis, contrasting with wild-type ones. May this be, not because OPN is essential for scoliosis pathogenesis, but because OPN deficiency reduces stress reactions in mice 260 ?

For, in mice, circulating OPN plays a significant role in the body's reaction to stress by regulating hormones of the hypothalamic-pituitary-adrenal axis (HPA) 260 modulated by leptin which activates the JAK/STAT pathway. Stressors cause less up-regulation of the stress hormone corticosterone in OPN-deficient mice 260 . This may be tested in the model used for mice: (1) rendered bipedal at 3 weeks of age, and (2) kept in tall cages to make them reach up increasingly for food and water 150 . The developmental stress hypothesis 261 , if confirmed, suggests that OPN deficiency through reduced corticosterone up-regulation causes less stress-reaction damage to the neural development of posture and so protects against the scoliosis. If so, these transgenic mice findings 19 20 may not be relevant to AIS pathogenesis.

Osteopontin, a major non-collagenous bone matrix glycoprotein originally isolated from bone - sialic acid rich, phosphorylated and inhibitor of calcification - has a critical role in bone remodeling which in OPN-knockout mice was suppressed 262 . Hence, the interpretation under item 11. above, and the evidence from Fujihara et al 262 , together raise caution about attributing a causal, rather than a consequential, role to increased plasma OPN in AIS pathogenesis.

Promoter polymorphisms of the gene for melatonin receptor 1B (MT1B) are associated with the occurrence of AIS, but not directly with curve severity; this supports the hypothesis of a MLT-signaling pathway dysfunction in AIS 263 . There is a lack of association between promoter polymorphism of the MTNR1A gene and AIS 264 . Genome-wide association studies have shown that melatonin receptor 1B variation is also associated with insulin and glucose concentrations; the risk genotype of this SNP predicts future type 2 diabetes suggesting that blocking the melatonin ligand-receptor system in the endocrine pancreas could be a therapeutic avenue for type 2 diabetes 265 266 . These genetic findings:

• are consistent with hormone receptors having a variety of parallel but independent downstream effects; and

• raise the question: Do post-operative AIS girls after 60 years of age have a lower prevalence of type 2 diabetes, because they are protected by being leaner and using their energy in a different way with a more efficient burn within their systemic disorder?

We postulate that in normal girls, trunk widening of the pelvis, ribcage and shoulder girdle, characteristic of humans, is contributed to by a leptin-hypothalamic-sympathetic nervous system (LHS)-driven mechanism acting bilaterally (Figure 5). Differential sympathetic innervation between axial and appendicular bones may be present 196 . The pattern of skeletal sizes for age 47 48 49 suggests that any differential innervation by the sympathetic nervous system may differ between girls and boys.

In normal human growth, biacromial broadening reflects widening mainly of the underlying upper thorax (Figures 10 and 11) 149 267 268 269 , and pelvic broadening reflects iliac flaring and widening mainly of the sacral alae (Figure 12); the latter reaches its maximum in hominins to provide a firm base of support for the trunk during bipedal posture and locomotion (Figures 13, 14) 153 267 269 270 271 . Hominid lumbar vertebrae also exhibit a caudally progressive widening of their laminae and of the space separating their articular processes 270 . Pelvic inlet width is a predictor of pediatric chest width 272 .

The evidence suggests that pelvic widening in the frontal plane 267 (which varies with climatic conditions), together with pelvic incidence in the sagittal plane 273 274 , provided hominins with conservation of energy 273 through biomechanical economy enabling -

• bipedalism with upright posture 75 153 ,

• modified spinal movements 275 , and in the last 3 million years -

• increasing fetal brain size 270 271 276 277 with sagittal expansion of birth canal (Figure 12) 149 270 271 , possibly with the bigger brain, from (1) a bigger baby,. (2) longer lumbar region, and (3) ability to conceive of tool construction and usage 276 .

The evidence suggests that the medio-lateral dimension of the birth canal has been relatively (but not absolutely) ample since the australopithecine stage about 3 million years ago (mya = megaannum) with a funnel-shaped upper thorax (Figure 11) 269 , as in the contemporary chimpanzee (Figure 13). A more ovoid pelvic shape with increase particularly of the sagittal dimension, then evolved in response to increasing brain size particularly from about 0.5 mya (Figure 12) 270 271 (see Evolutionary Origins).

AIS in girls from the standpoint of the autonomic nervous system is viewed as expressing increased central leptin sensitivity of hypothalamic sympathetic functions and, in some girls, of the somatotropic axis, which subsequently develop an inverse relationship. We speculate that AIS arises from dysfunction of the normal LHS-driven mechanism (Figure 5) by genetically-determined and selectively increased hypothalamic sensitivity (up-regulation from mutations) to circulating leptin leading to hypothalamic asymmetry. The asymmetry is viewed as an adverse response to stress 25 36 , with asymmetric activity mediated via the sympathetic nervous system bilaterally to vertebrae and/or ribs (Figures 1 and 5), to upper arm lengths in thoracic AIS, and to iliac heights in thoracolumbar and lumbar AIS. The increased sensitivity of the hypothalamus to leptin is viewed as being enhanced by increasing circulating levels of leptin from the fat accumulation of adolescent girls 12 , despite the lower leptin levels of AIS girls 163 164 .

The requirements for the theory are that in dysfunction, the sympathetic nervous system (SNS)-driven effects contribute with neuroendocrine mechanisms to produce 25 :

(1) Earlier skeletal maturation (hormonal).

(2) Sympathoactivation expressed asymmetrically in vertebral growth plates in 1-3 dimensions - left-right, front-back and/or torsionally - and in some paired bones (Figures 5 and 6).

(3) General skeletal overgrowth for age systemically distributed (hormonal)(Figure 7) 152 .

(4) Left-right extra-spinal skeletal length asymmetries (ribs, upper arms and ilia) (Figure 1) with upper arm length asymmetry being a signal of thoracic vertebral and/or rib length asymmetry (Figure 6).

(5) Increased hypothalamic sensitivity to circulating leptin (up-regulation) involves the somatotropic (GH/IGF-I) axis 222 in some younger preoperative AIS girls (Figure 7, see Neuroendocrinology,. Sympathetic nervous system and GH/IGF axis).

(6) Hormonal effects of the GH/IGF axis cause exaggeration of the SNS-induced vertebral/rib length asymmetry contributing to curve progression of preoperative AIS girls in an inverse relationship (Figure 5, see Neuroendocrinology. Sympathetic nervous system and GH/IGF axis).

(7) Relative osteopenia 88 278 279 which results in part from sympathoactivation.

The lower BMI 163 164 and body fat of AIS girls may be determined genetically 172 173 174 and contributed to by sympathoactivation 176 219 from the putative hypothalamic up-regulation to leptin (LHS concept) 25 . Overweight girls with AIS 170 171 probably reflect changes from genetic (leptin resistance in relation to satiety) and societal factors.

The LHS concept for AIS pathogenesis of girls, views the increased hypothalamic sensitivity to leptin as being at the opposite end of the spectrum to the central leptin resistance of obesity. This increased sensitivity to circulating leptin affects the hypothalamic sympathetic nervous system and, in some AIS girls, the somatotropic neuroendocrine axis. The effects produced in growing bones by these neural and endocrine mechanisms are influenced by the availability of energy, allocated by the hypothalamus through hormones and the nervous system, modulated by circulating leptin levels that measure long-term adiposity.

The putative dysfunction of hypothalamic neurons in AIS - increased and asymmetric sensitivity to leptin, may result from an abnormality of a G-protein-coupled receptor, or G protein, to leptin 25 . The melatonin-signaling dysfunction caused by the inactivation of Gi proteins so far detected is peripheral 14 15 16 17 18 19 20 , and it is unknown whether any hypothalamic mechanism of etiopathogenesis is involved [Dr A Moreau personal communication]. Melanocortin-3 (MC3R) and MC4R are G-protein coupled receptors highly expressed in the hypothalamus 232 .

Subject to the caveat expressed for circulating OPN levels having a causal role in AIS, increased levels of circulating OPN 19 20 may act as a gate for AIS in the hypothalamus as does kisspeptin for puberty through its G-protein-coupled membrane receptor GPR54 2 280 281 .

Subject to the demonstration of a significant functional variation in human populations, inhibitory molecules such as SOCS-3 232 243 244 250 , PTB-1B 232 240 252 and possibly the regulator of the leptin receptor (OB-RGRP) 247 248 253 - all as negative regulators of leptin sensitivity, by their decreasing action, are candidates to increase hypothalamic sensitivity to leptin in the LHS-driven concept for AIS pathogenesis.

As positive regulators of leptin sensitivity, members of the SH2B family by their increasing action 232 , are candidates to increase hypothalamic sensitivity to leptin in the LHS-driven concept for AIS pathogenesis.

Hormesis is a bimodal dose response to drugs and toxins, first stimulation and then an adverse response, usually inhibition 282 283 284 . There is evidence that this normal hormetic process applies to leptin 223 . The dose effect will be influenced by the combined effects of 1) increased hypothalamic sensitivity to leptin, and 2) raised circulating leptin levels from adolescent female fat accumulation. We speculate that in the hypothalamus the hormesis of leptin, in adversity leads not to inhibition but to increased sensitivity and asymmetry 36 . The concept is considered plausible by Dr EJ Calabrese [personal communication]. In rats, infused leptin increases sympathetic nervous system activity in a dose-dependent manner suggesting that leptin may act hormetically on the normal rat hypothalamus 285 .

Rett syndrome is a genetic neuro-osseous developmental disorder much more prevalent in girls than boys, characterized by profound and progressive loss of intellectual functioning and growth failure 286 287 . Raised circulating leptin levels and overactivity of the sympathetic nervous system 288 are associated with its pathophysiology 286 287 . The skin sympathetic responses are related to the side of the scoliosis, on the foot ipsilateral to the convex side of the scoliosis where it shows a relatively lower amplitude 286 . These findings are consistent with the view that leptin and sympathetic nervous system dysfunction, under certain conditions, may be associated with scoliosis expression and curve laterality.

PWS, a rare multisystem genetic disorder, is thought to result from a central hypothalamic-pituitary dysfunction 289 290 . It is associated with failure to thrive in infancy and progressive hyperphagia and obesity in childhood; there is short stature with growth hormone (GH) deficiency, obesity, eating disorders, decreased muscle mass, hypotonia, hypogonadism, and a high prevalence of scoliosis in infants, juveniles and adolescents (15-86%) with 67% affected at skeletal maturity 289 291 292 . The pathogenesis of the scoliosis is unknown 293 ; it is unrelated to gender and BMI 292 and may be related to decreased muscle mass, hypotonia, and hypo-excitability of motor cortical areas with defective neurogenesis of cortical tissue 294 . The contribution of the autonomic nervous system, if any, to the scoliosis appears to be unknown. PWS is not accompanied by deranged leptin concentrations and there was no evidence of an interaction of the GH/IGF axis with leptin metabolism in GH-deficient children 295 . While infants with PWS, have higher leptin levels than controls, suggesting a relative excess of fat to lean body mass 296 , adults with PWS have leptin assessment corresponding to their degree of obesity 297 (see Endocrine and Therapeutic Implications, GH treatment and the Prader-Willi syndrome (PWS)).

Explanations of "what makes us human" often include a bridge between culture and biology 51 . Recently, it has been suggested that decreased circulating melatonin levels due to light from campfires extending the day, "changed the timetable of growth, development and reproduction, because sitting by the fire altered the night's flow of melatonin and the cascade of hormones that follow it." 304 .

Power and Schulkin 301 in their book, 'The Evolution of Obesity', outline an evolutionary hypothesis in relation to fat and hominin brain growth 299 300 . The book is one of the first to use an evolutionary framework to analyse a major body of neuroendocrine knowledge about a specific condition 53 . Power and Schulkin write:

"Human beings have evolved to become very good at storing fat; fat appears to have been very important in our evolution. For example, human babies are among the fattest of all mammals... ...The importance of fat, both in our diet and on our bodies, appears to have increased in human beings compared to our nonhuman primate relatives. We suggest that this change in nutritional biology was linked to the seminal evolutionary event in our lineage: our larger brain." 301 .

Nutritionally, human brain growth is said not to be costly 299 , but it does require docosahexaenoic acid (DXA), present in body fat more at birth than at any other time in life 300 . The functioning human brain enlarging particularly in the first two years of postnatal life, imposes a burden on metabolism by -

• increasing energy demands, and

• restricting flexibility in energy allocation when nutritional supply is disrupted - as in the nutritional stresses of weaning and childhood infections 299 301 302 .

The relation of leptin to brain growth is not considered here 133 .

We suggest that another 'seminal evolutionary event' - earlier in our lineage than brain growth, was trunk width growth which has increased more in human beings compared with our nonhuman primate relatives; the latter lack the extended childhood and rapid and large acceleration of growth velocity at adolescence in humans (Figures 11, 12, 13, 14) 153 267 270 271 303 .

• Pelvic width. In hominins, increased pelvic as iliac and sacral width for habitual erect walking was established by about 3 mya (Figure 12).

• Thorax and shoulder gitrdle width. Ribcage widening, particularly of the upper thorax (Figure 11) happened in the last 3 million years. The wide shoulders characteristic of Homo 303 evidently resulted from upper ribcage widening relative to depth (Figures 10 and 11), with clavicular lengthening (Figure 14). This trunk widening at the shoulder girdles is likely to have been selected by:

a) the evolution of upright posture giving an enhanced respiratory importance to the upper thorax [see 268 ]; and

b) counter-rotations of upper thorax and arms (but not the head) providing counter-balancing torques generated by shoulder girdles and arm-swinging needed to oppose torques created by the pelvic rotations of hominin bipedalism 71 149 268 303 .

• Brain and pelvic depth. The large fetal brain size enabling a dramatic jump of adult brain size from about 0.5 mya, was made possible by further expansion of the birth canal, particularly sagittally (pelvic depth) (Figure 12) 75 267 299 300 301 302 303 .

The LHS mechanism suggests that the fatness of hominins, starting over 3 mya, raised circulating leptin levels which, through the hypothalamus and sympathetic nervous system, supplemented the hormonally-driven growth in width of pelvis and ribcage 26 (Figures 10, 11, 12), and not in nonhuman primates 153 267 (Figures 5, 13 and 14). This mechanism, we suggest, provided a process in evolution that contributed to:

• pelvic widening mainly from sacral widening, enabling bipedalism with upright posture, later

• upper thorax with shoulder widening, and still later

• increased pelvic depth of Homo sapiens (Figure 12).

The LHS mechanism is interpreted as being evident today in normal human development as 'energy priority of trunk width growth' in girls (Figures 4 and 5) 47 48 49 .

We speculate:

• In evolution, to reduce toxicity to the hypothalamus of the raised circulating leptin levels - signaling greater adipose tissue stores particularly in females, hypothalamic sensitivity to circulating leptin became diminished (desensitized, or down-regulated, i.e. central leptin resistance), possibly involving increased action of inhibitory molecules such as SOCS-3 and PTP-1B, or decreased action of stimulatory molecules such as SH2B1. It needs to be established whether humans deal with SOCS-3, PTP-1B, and SH2B1 differently from other apes.

• In evolution, the development of human bipedalism and upright posture necessitated adaptations of postural control by the somatic nervous system 51 .

• The putative central leptin resistance in the somatotropic (GH/IGF) axis of normal juvenile girls [ 50 , see 227 237 ] is linked to a greater evolutionary down-regulation to leptin in the female than the male hominin hypothalamus.

The LHS concept for AIS pathogenesis suggests that the putative genetically-determined selectively increased hypothalamic sensitivity (up-regulation from mutations) to leptin leading to hypothalamic sympathetic asymmetry is rooted in the evolutionary origins of hominin fat deposition providing the energy needed for trunk width growth and later, brain growth and metabolism. We posit that increasing levels of circulating leptin associated with fat accumulation of adolescent girls 12 , enhance the putative increased hypothalamic sensitivity (sympathetic and somatotropic) to leptin of AIS girls. This raises the question: Is the societal fat accumulation of normal adolescent girls 156 associated with increasing severity 170 171 and/or prevalence of AIS?

Left-right asymmetries of the neuroendocine system and of hypothalamic structure and sex-linked function are reported in normal animals 305 .

Badman and Flier 200 state that the improvement in central leptin signaling by PTP-1B may provide a target for pharmacological intervention for weight-loss therapies 306 . Similarly, the LHS concept for AIS pathogenesis suggests that impairment of central leptin signaling may ultimately provide a target for pharmacological intervention for progressive AIS in girls, if this can be done selectively.

Other manipulatable causes of AIS pathogenesis are suggested by the melatonin-signaling dysfunction detected in osteoblasts and chondrocytes.

(1) Osteoblasts. In vitro, MLT significantly stimulates osteoblast proliferation, differentiation and mineralization from controls 336 , but not in osteoblasts from AIS subjects 337 338 ; this defect is suggested to play a role in the low bone mineral density of AIS patients and contribute to pathogenesis 338 .

MLT-signaling dysfunction in AIS subjects has been revealed mainly using bone tissue because (a) osteoblasts respond to MLT, and (b) relative osteopenia is often observed in patients with AIS 14 15 88 278 279 . In some girls with AIS, a particular MLT-signaling defect is evident 17 256 258 337 . Correction of this defect in vitro by estradiol suggested that "the lack of estrogen that results in late menarche may be corrected by estrogen agonists having a positive effect on bone tissue remodeling" 256 . Leboeuf et al 258 suggest estrogens as important pharmacological targets to consider in AIS therapy directed to patients selected on their tissue response to MLT. This is in contradistinction to the suggestion of delaying the adolescent growth spurt for subjects in the lower BMI subset using a gonadorhelin analogue 330 331 332 333 (see Sex hormones).

(2) Chondrocytes. In cartilage from controls, MLT significantly inhibits chondrocytes proliferation in vitro but not from AIS subjects 339 . According to Wang and colleagues 339 , the non-responsiveness (i.e. lack of inhibition) of AIS chondrocytes to MLT might play a role in the abnormally increased bone growth of AIS girls from dysfunction of the MLT-signaling pathway. In this connection, there is a decreasing expression of MT1 and MT2 mRNA in chondrocytes from AIS patients which may be related to the molecular pathogenesis of AIS 340 .

Rather than a clinical trial of a somatostatin analogue and β-blockers, we suggest that currently there is a need to evaluate circulating hormones and sympathoactivation in AIS girls by relatively higher and lower BMI subsets.

In addition to using cellular dielectric spectroscopy for AIS diagnosis based on G-protein coupled receptor detection 18 , Moreau et al 19 20 suggest OPN and sCD44 as useful markers for diagnosis and prognosis of idiopathic scoliosis. Subject to further study, as already mentioned, OPN may be a potential target for therapeutic intervention in AIS subjects as suggested for psoriatic patients 341 (see Some melatonin-deficient mouse models of scoliosis - markers of developmental stress?).

The analysis of our skeletal data by relatively higher and lower BMI subsets distinguishes two types of effect: skeletal sizes for age (Figures 4 and 7), and skeletal asymmetries (Figure 6).

Skeletal sizes for age - energy priority of trunk width in girls. The skeletal size for age effect in the girls is shown as differences between -

(1) higher and lower BMI subsets in each of preoperative, screened and normal girls (Figures 4 and 5) restricted mainly to the trunk 46 117 118 119 ; and

(2) preoperative and normal girls in higher and lower BMI subsets (Figure 7).

The trunk width growth priority of girls is seemingly a human characteristic. It is not explained by any of the prevailing theories of AIS pathogenesis (Appendix 1, items 1-15) each of which solely addresses pathogenesis. The trunk width features are accommodated by the LHS mechanism which invokes the sympathetic nervous system and hormones (Figure 5).

Theories about the pathogenesis of AIS have to explain several undisputed facts 91 110 364 .

(1) Dependence of the deformity upon growth and growth rate. The relation of skeletal growth velocity to curve progression in AIS is established 4 5 137 365 366 , but its mechanism of action is unclear - causative, conditional, amplifying, or coincidental 91 . In the escalator concept, the dependence of AIS progression on growth is explained not by velocity of growth, but by rapid spinal lengthening and trunk enlargement beyond the capacity of the postural mechanisms to control the deformity 24 51 111 .

(2) Predilection for females. Two putative mechanisms explain the greater susceptibility of girls than boys to progressive AIS:

a) In the autonomic nervous system, the increased sensitivity (up-regulation) of the hypothalamus (sympathetic NS and somatotropic axis) to leptin by mutations with its asymmetries contributing to AIS, greater in females than in males 25 , is attributed to: i) diminished sensitivity (down-regulation, i.e. resistance) to leptin of the female hypothalamus established by mutations in hominin evolution; and ii) central leptin resistance in the somatotropic axis of normal juvenile girls 50 which, through mutations causing central leptin sensitivity, may predispose some girls to AIS.

b) In the somatic nervous system, girls may enter their adolescent skeletal growth spurt in postural immaturity, compared with boys who may enter their adolescent growth spurt in postural maturity so they are protected from developing a scoliosis curve (Figure 15) 330 331 332 .

(3) Involvement of members in involved families. This is determined by genetic factors operating in the autonomic and somatic nervous systems 56 77 78 79 and other mechanisms.

(4) Curve types and laterality patterns. Biomechanical factors involving ribs 59 60 61 62 63 and/or vertebrae 64 65 91 92 93 and spinal cord 64 65 92 93 , acting during growth may localize AIS to the thoracic spine and cause the sagittal spinal shape alterations 83 84 85 86 87 88 89 90 . The non-random laterality of thoracic AIS curves has been explained by several factors including handedness, aorta, lungs, diaphragm, pre-existing lateral curve, axial rotation and embryology 367 368 369 370 371 . We suggest that the laterality and site of thoracic, thoracolumbar and lumbar curves is determined, in part, by the location of the putative abnormalities of the LHS-driven mechanism in the hypothalamus and sympathetic nervous system.

(5) Varied progression patterns. These are explained by the interaction of autonomic and somatic nervous systems in the spine and trunk compounded by any relative osteopenia of vertebrae 88 278 279 , biomechanical spinal growth modulation 80 81 82 , accelerated disc degeneration 45 342 343 344 345 346 347 348 349 350 351 , and platelet calmodulin dysfunction 21 22 107 . Circulating leptin levels in AIS girls did not correlate significantly with Cobb angle 163 164 . This finding does not preclude circulating leptin levels acting with increased hypothalamic sensitivity to leptin to contribute to the magnitude of the hypothalamic asymmetry, and from that to the sympathetic nervous system-induced skeletal asymmetry(ies).

(6) 3-D rotatory deformity of the spine. In thoracic AIS, Davids et al 372 found that the most valuable single MRI indicator for abnormal central nervous system findings was the absence of an apical segment lordosis. This and other evidence 91 373 suggests that in thoracic AIS, apical lordosis 83 84 85 86 87 is determined by processes either intrinsic to the spine ("primary", i.e. relative anterior spinal overgrowth = RASO 51 76 ), and/or extrinsically by the sympathetic nervous system acting on vertebrae in 1-3D - left-right, front-back, and/or torsionally. Recent evidence shows that while right thoracic AIS has a reduced thoracic kyphosis (T5-12), increased pelvic incidence and sacral slope consistent with the RASO theory of pathogenesis 374 , left thoracic AIS 374 has a normal thoracic kyphosis and pelvic incidence, not consistent with the RASO theory. This may signify that left thoracic AIS has a pathogenesis different from right thoracic AIS 374 , possibly involving reduced white matter density of the central nervous system 114 115 . We suggest that right and left thoracic AIS in girls may be driven separately by the two nervous system components of the double neuro-osseous theory: right thoracic AIS mainly by the autonomic/sympathetic nervous system and left thoracic AIS, mainly by the somatic nervous system.

(7) Vertebral bodies grow faster than the posterior vertebral elements 64 65 83 84 85 86 87 88 89 90 . This is explained in part by a greater enhancing effect of the sympathetic nervous system on vertebral bodies and their growth plates than on posterior vertebral growth leading to asymmetry in the sagittal plane and the relative anterior spinal overgrowth (RASO) of progressive AIS.

(8) AIS is exclusive to humans. We suggest that AIS in girls is a consequence of abnormalities occurring in the putative physiological LHS-driven (Figure 5) and escalator (Figure 2 ) mechanisms of the theory, both of which are unique to humans 24 25 50 and emanating from these and other features of their evolution 298 299 300 301 302 303 .