Epigenetic and Genetic Factors Related to Curve Progression in Adolescent Idiopathic Scoliosis: A Systematic Scoping Review of the Current Literature

Adolescent idiopathic scoliosis (AIS) is a progressive deformity of the spine. Scoliotic curves progress until skeletal maturity leading, in rare cases, to a severe deformity. While the Cobb angle is a straightforward tool in initial curve magnitude measurement, assessing the risk of curve progression at the time of diagnosis may be more challenging. Epigenetic and genetic markers are potential prognostic tools to predict curve progression. The aim of this study is to review the available literature regarding the epigenetic and genetic factors associated with the risk of AIS curve progression. This review was carried out in accordance with Preferential Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. The search was carried out in January 2022. Only peer-reviewed articles were considered for inclusion. Forty studies were included; fifteen genes were reported as having SNPs with significant association with progressive AIS, but none showed sufficient power to sustain clinical applications. In contrast, nine studies reporting epigenetic modifications showed promising results in terms of reliable markers. Prognostic testing for AIS has the potential to significantly modify disease management. Most recent evidence suggests epigenetics as a more promising field for the identification of factors associated with AIS progression, offering a rationale for further investigation in this field.


Introduction
Adolescent Idiopathic Scoliosis (AIS) is a complex three-dimensional deformity of the spine, with a different grade of involvement of the frontal, sagittal, and axial planes [1]. It affects 2-3% of the adolescent population [2]; females are more often involved than males [3].
The diagnosis of scoliosis is based on patient clinical examination and radiographical evaluation [4]. After AIS is diagnosed, patients need different management (ranging from observation alone to orthotic treatment and surgical correction) according to curve magnitude at the time of diagnosis and curve progression potential.
Scoliotic curves progress until skeletal maturity, causing important aesthetic problems, such as humps, with psychological problems and loss of self-esteem, coronal, and/or sagittal imbalance and muscle fatigue [5]. In rare cases, the curve progression can lead to a severe deformity with the occurrence of a lung restrictive disease, a consequent increase in right atrial and ventricular pressure, alongside neurological impairment [6]. a severe deformity with the occurrence of a lung restrictive disease, a consequent increase in right atrial and ventricular pressure, alongside neurological impairment [6].
While the Cobb angle is a straightforward tool in initial curve magnitude measurement, assessing the risk of curve progression for each patient at the time of diagnosis may be more challenging.
At the same time, identifying predictors of curve progression is still fundamental to avoid erroneous clinical management depriving patients of adequate treatment or exposing others to unnecessary one. For this purpose, many clinical parameters are widely accepted as predictors of scoliosis progression: curve location, age at diagnosis (<12 years), pre-menarche status, low Tanner stage, and peak height velocity [4,7]. Moreover, some radiographic parameters are currently considered by clinicians, such as curve magnitude at the time of diagnosis (>25°), Risser stage (0-1), open triradiate cartilage, and demonstration of significant curve progression between serial radiographs [6,7]. Figure 1 represents the parameters related to scoliosis progression. Epidemiological and genetic studies indicated AIS as a polygenic disease, and several studies investigated genetic and epigenetic factors associated with an increased risk of the onset of the scoliotic curve [8][9][10][11]. Several loci associated with AIS susceptibility were identified and evaluated in different ethnic groups, even if the value of AIS susceptibility in clinical practice is limited. Less information is available regarding candidate genetic Epidemiological and genetic studies indicated AIS as a polygenic disease, and several studies investigated genetic and epigenetic factors associated with an increased risk of the onset of the scoliotic curve [8][9][10][11]. Several loci associated with AIS susceptibility were identified and evaluated in different ethnic groups, even if the value of AIS susceptibility in clinical practice is limited. Less information is available regarding candidate genetic and epigenetic factors related to scoliotic curve progression and its prediction, which would be a key tool for disease management.
A review of the literature was carried out following the Preferential Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [12].
The Oxford level of evidence scale [13] was used to assess the level of evidence of the included studies. The full version was used to assess randomized and non-randomized clinical trials, whereas the modified version was used to assess all other studies.
Inclusion criteria considered papers describing genetic and epigenetic factors associated with AIS curve progression published in English peer-reviewed journals. Isolated case reports/series with less than 5 patients, literature reviews, and meta-analyses were excluded. The included articles met the PICO criteria for systematic reviews (Population, Intervention, Comparison, and Outcomes). Different types of studies were considered for inclusion: case series, case-control, cohort studies, comparative studies, genome-wide association studies, and case-only studies. These studies were conducted either retrospectively or prospectively.

Search Strategy
Pubmed-MEDLINE, The Cochrane Central Registry of Controlled Trials, Google Scholar, and the Embase Biomedical Database were searched over the years 1990-2022 to identify eligible studies in the English literature describing the genetic factors associated with AIS curve progression. The online literature search was conducted in January 2022 by three reviewers (MM, FB, and GV). The authors stated the following research question: "Are there genetic and epigenetic factors correlated with scoliotic curve progression in adolescent idiopathic scoliosis patients?". This research question matched all four PICO concepts. Subsequently, the following key concepts were formulated "Adolescent Idiopathic Scoliosis", "curve progression", "curve severity" and "genetic variants", "epigenetic variants", and "polymorphism", and various alternative terms were considered for each key concept to include the maximum number of articles available in the literature pertaining to the research question. Details on the search strategy are summarized in Supplementary Table S1.

Study Selection
After screening the titles and abstracts, the full-text articles were obtained and reviewed. A manual search of the bibliography of each of the relevant articles was also performed to identify potentially missed eligible papers. Duplicates were removed. The study selection process carried out in accordance with the PRISMA flowchart is shown in Figure 2. The present systematic review was accepted for registration in the PROSPERO database for systematic reviews [14] (ID: CRD42022322089). Figure 2. The present systematic review was accepted for registration in the PROSPERO database for systematic reviews [14] (ID: CRD42022322089).

Data Extraction
Two reviewers (MM and SN) extracted the data through a standardized data collection form. Three reviewers (MM, SN, and AR) checked the data for accuracy, and inconsistent results were analyzed for discussion. The extracted data concerning the study design (with the level of evidence), number of patients, demographics of patients, curve progression definition, biological sample, gene/s involved, mutation/s, and results are summarized in Table 1. The following outcomes were considered for analysis: curve severity defined as the Cobb angle; curve progression measured as the increase in the Cobb angle from the initial evaluation; epigenetic or genetic factors associated with curve progression; and clinical features of curve progression: curve location, age at diagnosis (<12 years), premenarche status, low Tanner stage, and peak height velocity time. Moreover, we considered some radiographic parameters currently considered by clinicians, such as the curve magnitude at the time of diagnosis (>25°), Risser stage (0-1), and open triradiate cartilage.

Data Extraction
Two reviewers (MM and SN) extracted the data through a standardized data collection form. Three reviewers (MM, SN, and AR) checked the data for accuracy, and inconsistent results were analyzed for discussion. The extracted data concerning the study design (with the level of evidence), number of patients, demographics of patients, curve progression definition, biological sample, gene/s involved, mutation/s, and results are summarized in Table 1. The following outcomes were considered for analysis: curve severity defined as the Cobb angle; curve progression measured as the increase in the Cobb angle from the initial evaluation; epigenetic or genetic factors associated with curve progression; and clinical features of curve progression: curve location, age at diagnosis (<12 years), pre-menarche status, low Tanner stage, and peak height velocity time. Moreover, we considered some radiographic parameters currently considered by clinicians, such as the curve magnitude at the time of diagnosis (>25 • ), Risser stage (0-1), and open triradiate cartilage.

Methodological Quality Assessment of Included Studies
The assessment of the methodological quality of the studies was performed using checklist criteria. The quality assessment tool adopted from the National Institutes of Health/National Heart, Lung, and Blood Institute was used [15]. After answering a series of multiple-choice questions, the quality of each study was reported as poor, fair, or good. All details are summarized in Supplementary Table S2.
The studies analyzed both small and large-sized populations (n = 16 to 2645), describing the association between genetic and epigenetic factors involved in AIS curve progression.
The included studies are heterogeneous (or lacking data) in ethnicity, spine deformity, gender, and curve progression definition (Table 1).

Genetic Factors Associated with Disease Progression
Genetic factors possibly influencing the progression of adolescent idiopathic scoliotic curves were analyzed on genomic DNA prevalently obtained from peripheral blood, or alternatively, from saliva [25,27,43,52].
Numerous polymorphisms were described as associated at different levels with scoliosis curve progression (Table 1), and related genes were hypothesized for their possible involvement in disease development.
In more detail, Ward et al. [52] investigated the predictive value of the Scoliscore in Caucasian AIS patients, suggesting that a risk model of patients' natural history could be possible by extracting SNPs from patients' DNA. The prognostic score, ranging from 1 to 200, was applied to three different cohorts with known AIS outcomes (low-risk females, high-risk females, and high-risk males, where high scores corresponded to a higher risk of curve progression and vice versa). Indeed, low-risk scores (<41) had a negative predictive value close to 100% for each of the three cohorts studied.
The promising "Scoliscore" results were not entirely replicated in the Chinese population by Xu et al. [29], with only two SNPs (rs9945359 and rs17044552) found to be associated with curve progression and severity [29]. The authors stated that, despite the existing ethnic differences (Caucasian vs. Chinese), AIS patients could share two SNPs as common traits in the pathogenesis of curve progression, but the Scoliscore was not reliable in the Chinese Han population. Similar results were obtained by three other independent studies [25][26][27] analyzing the validity of the Scoliscore in Caucasian [25,27] and French-Canadian populations [26].
Putting all these findings together, it may be hypothesized that ethnic differences between Asian and Caucasian populations could yield great divergence regarding the prognostic power of "Scoliscore". Moreover, the result was not replicated in studies with the same Caucasian population.
Insulin-like growth factor 1 (IGF-1) has an important role in skeletal growth [57], representing a good candidate to play a role in AIS curve progression. Yeung et al. [28] first reported a weak association (p = 0.04) between the IGF-1 polymorphism and a higher Cobb angle in Chinese AIS patients, suggesting IGF-1 as a disease-modifying gene rather than an AIS-onset gene per se.
This result was not replicated in the Japanese population [58], but an association (p = 0.01) was described between the rs5742612 polymorphism in the upstream region of the IGF-1 gene and disease risk, with a significantly different distribution of IGF-1 genotypes in low-and high-risk groups in the Korean population [21].
The estrogen receptor (ER) gene has been shown to be expressed in both human osteoclasts and osteoblasts and plays a critical role in cellular proliferation in bone tissue [59]. Based on the assumption that the estrogen reaction to skeletal and sexual growth is genetically determined by ER gene polymorphism, Inoue et al. [16,17] and Zhao et al. [51] found ER1 gene polymorphism (Xbal site) to be related to curve progression. However, in Tang et al.'s [39] study, a subgroup of Chinese skeletally immature patients was followed until skeletal maturity at age 16, and the abovementioned hypothesis was not confirmed.
Other successfully replicated genetic factors are FBN1 and FBN2 variants. The FBN1/2 genes encode fibrillin, a glycoprotein of the extracellular matrix, and mutations in these genes have been reported in a variety of fibrillin-related disorders (i.e., Marfan syndrome [60]).
To determine whether FBN1 and FBN2 variants were associated with AIS curve progression, Buchan et al. [23] and Sheng et al. [37] found that rare mutations in FBN1 and 2 were particularly present in severe AIS cases when compared to non-severe cases or healthy controls.
Most of the previously reported associations between genetic markers and AIS curve progression were not replicated in other independent studies. Therefore, Ogura et al. [30] and Wang et al. [47] explored the functional role of the rs35333564 variant located in the MIR4300HG gene in different ethnic populations (Japanese and Chinese). Both studies confirmed that the MIR4300HG functional variant could significantly add risk of curve progression with similar odds ratios and p-values. Moreover, Wang's study [47] evaluated the relative expression of MIR4300 in paraspinal muscles among surgical patients carrying different MIR4300 genotypes, discovering that the GG genotype showed remarkably lower tissue expression than the AA genotype. Interestingly, and for the first time, the tissue expression level of MIR4300 was significantly correlated with curve severity. To the authors' best knowledge, there are no studies that contradict the abovementioned association.
Altogether, available data on genetic factors correlated with AIS evolution do not allow the prediction of disease progression based only on genetic information. Table 2 summarizes the findings concerning genetic factors associated with AIS progression, statistical significance, and the sensitivity/specificity of each variant.

Rs1800469 Rs1800471
TGFβ-1 protein triggers chemical signals that regulate various cell activities inside the cell, including the growth and division (proliferation) of cells, the maturation of cells to carry out specific functions (differentiation), cell movement (motility), and controlled cell death (apoptosis) OR = 3.78 within 95% Confidence interval (CI): 1.42-10.05 0.038 Kruskal-Wallis analysis of variance revealed the relationship between the SNP C-509T of the TGFB1 gene and the curve severity in females with AIS (Kruskal-Wallis statistic = 6.50) I. Ryzhkov

Epigenetic Factors Associated with Disease Progression
In eukaryotes, gene expression is dynamically regulated at the chromatin level by epigenetics, defined as heritable and reversible changes in gene expression without alterations of the underlying DNA nucleotide sequence [61]. Epigenetic marks principally include DNA methylation (the addition/removal of methyl groups to/from cytosines within CpG dinucleotides) and histone post-translational modifications (such as methylation, acetylation, phosphorylation, ubiquitination, and sumoylation). These modifications give rise to local chromatin remodeling that, in turn, modifies the accessibility of regulatory elements to genes. Regulation by non-coding RNAs such as microRNAs is also part of epigenetics. Epigenetic mechanisms regulate cell differentiation and development and are involved in human disease [62].
To date, few studies concerning epigenetic factors involved in AIS progression have been published, but literature data strongly encourage further research in this field.
Meng et al. [34], for the first time, reported a large-scale genome-wide analysis to establish a prognostic model based on methylation status. They analyzed peripheral blood cell DNA of two monozygotic twin pairs discordant for disease progression and validated the results in additional samples. They found a positive correlation between cg01374129 site demethylation and AIS progression (AUC value of 0.805 in the ROC analysis), suggesting epigenetic regulation. Since this site is near the HAS2 gene (hyaluronan synthase 2), playing a critical role in vertebral and intervertebral disc development, they speculated cg01374129 hypomethylation deregulates HAS2 expression, impairing normal spine development and causing scoliosis progression.
Another study [48] used a genome-wide methylation approach to test the influence of DNA methylation status on curve severity, by studying DNA from peripheral blood cells of eight monozygotic twin pairs. The authors found four probes (cg02477677, cg12922161, cg16382077, and cg08826461) where increasing curve severity was associated with hypomethylation. Candidate genes affected by differential methylation include the WNT signaling pathway and neuropeptide Y.
Mao et al. [36] investigated promoter methylation of the COMP gene, encoding the cartilage oligomeric matrix protein as a target gene for AIS curve progression. COMP promoter methylation, associated with low gene expression, was found to directly correlate with AIS curve severity (high Cobb angle of the main curve).
PITX1 (pituitary homeobox 1, a member of the RIEG/PITX homeobox transcription factors) gene promoter hypermethylation in peripheral blood cells of AIS patients is significantly associated with the Cobb angle of the main curve, suggesting a relationship with disease progression [38]. Similarly, average protochaderin 10 (PCDH10) promoter methylation was higher and gene expression was lower in AIS patients compared to controls. Moreover, high PCDH10 promoter methylation was associated with the Cobb angle of major curves in AIS patients [44]. Furthermore, in this case, data were obtained by analysis of DNA from peripheral blood cells.
In paravertebral muscles, H19 and ADIPOQ genes have been shown to be expressed inconsistently [40], with lower H19 levels and higher ADIPOQ levels in concave-sided muscle tissues compared to convex-sided ones. These data positively correlated with the spinal curve and age at initiation [40], suggesting an important role of H19 and ADIPOQ not only in the onset but also in the progression of AIS.
On the contrary, the methylation status of estrogen receptor 2 (ESR2) in deep paravertebral muscles was found to be associated with the occurrence but not progression of AIS [63].
In another study, the methylation status of tissue-dependent and differentially methylated regions (T-DMRs) of the ESR1 estrogen receptor was analyzed in superficial and deep paraspinal muscles to explore the association with AIS progression. The authors found suggestive evidence that methylation status might be associated with disease severity [49].
MicroRNAs are small noncoding RNAs that also participate in the regulation of bone metabolism, osteoclast, and osteoblast function. These molecules are epigenetic factors involved in the control of specific molecular pathways in bone-related disorders.
By performing miRNA expression profile analysis on plasma samples from severe and mild AIS patients and controls, Wang et al. [45] suggested miR-151a-3p as a putative biomarker of severe AIS since it was overexpressed in severe but not mild AIS patients. MiR-151a-3p may contribute to scoliosis progression through the inhibition of GREM1 gene expression in osteoblasts interrupting bone homeostasis.
Via microarray analysis, miRNA-145-5p (miR-145) and β-catenin mRNA (CTNNB1) were found to be overexpressed in AIS bone tissue and primary osteoblasts compared to controls. Significant negative correlations between circulating miR-145 and serum sclerostin, osteopontin, and osteoprotegerin were noted in patients with AIS. The observed aberrant miRNA expression inhibited osteocyte function via Wnt/β-catenin signaling, appearing dysregulated in AIS. MiR-145 was therefore suggested as a prognostic AIS biomarker [35].
In summary, the hypomethylation of some DNA regions, the hypermethylation of some gene promoters (COMP, PITX1, PDCH10), and the overexpression of some miRNAs (miR-145, miR-151a-3p) were associated with AIS progression. Table 3 summarizes the available data on epigenetic factors associated with AIS progression, the techniques used, the tissues analyzed, and the statistical evidence. The methylation level of five CpGs in the COMP promoter was significantly correlated with Cobb angle of the main curve and chronological age (p < 0.0001). AIS patients were associated with high Higher DNA methylation level and low gene expression of PCDH10 gene rather than normal controls. The high methylation level indicated high Cobb angle of major curves in AIS. The abnormal DNA methylation may widely exist and serve as a potential mechanism for AIS progression. The average methylation level was 4.32 ± 0.73 in AIS patients and 3.14 ± 0.97 in healthy controls (p < 0.001). Besides, the PCDH10 gene expression was 0.23 ± 0.04 in AIS patients and 0.36 ± 0.08 in normal controls (p < 0.001).

Discussion
Adolescent idiopathic scoliosis (AIS) is the most common type of scoliosis, a complex phenotype resulting from the interaction of multiple genetic loci with each other and the environment [53].
AIS is a progressive musculoskeletal disease that may result in cosmetic deformity, back pain and functional deficits, psychological problems, and impaired social interactions [64,65]. Among patients initially diagnosed with AIS, curve progression before skeletal maturity occurs in approximately two-thirds of cases, and in 10% of patients, it progresses to severe scoliosis (Cobb angle >40 • ) in the following years [6,66]. Although X-ray exams and clinical examinations are currently considered the gold standard for AIS follow-up, they have limited sensitivity and specificity values and provide limited information on curve progression risk [5]. Serial radiographs can result in relatively high cumulative radiation doses, leading to stochastic effects with long-term increased cancer and mortality risks [67]. A recent AIS cohort study stated an overall cancer rate (mostly breast and endometrial) that was five times higher in AIS patients followed up with X-ray exams than the general population [68]. Surgical intervention is currently the ultimate solution established for patients with a severe curve or with conservative treatment failure [69]. It can achieve powerful curve correction but is characterized by high morbidity and intra and/or post-operative complications [70,71].
The control of curve progression is therefore a crucial clinical task, but its etiology is still largely unknown; therefore, new biomarkers are needed to facilitate early detection and accurate curve progression risk assessment. The identification of such biomarkers has the potential to improve patient management, minimize unnecessary orthopedic intervention, define the best applicative protocol for orthopedic treatment, and identify the subpopulation of patients in which early surgery, even with non-severe curves, can avoid operating on severe curves with worse outcomes and more risks. Since clinical features do not adequately predict disease progression, more reliable prognostic factors need to be identified to increase the accuracy of the predictive model, and genetic/epigenetic markers might represent ideal candidates for AIS management. Although the role of genetic factors in AIS development is widely accepted, their role in disease progression is still under study.
In the present work, we systematically reviewed the available literature from 1990 to the present date, concerning genetic and epigenetic factors associated with AIS progression.
Forty papers met the inclusion criteria of the present review, with fifteen genes reported as having SNPs with a significant association with progressive AIS [25][26][27]29,52]. We also considered the development of a predictive algorithm based on a panel of 53 SNPs associated with AIS curve progression, the so-called "Scoliscore", whose ability to discriminate between patients with a low or high risk of progression failed to be replicated in some populations [25][26][27]29,52].
Available data concerning genetic factors suggest a relatively low association and, if present, an association with low predictive capacity (Tables 1 and 2), low odd risk values, and low level of evidence (III or IV). Moreover, the low replicability in different ethnicities confirms the extreme variability of the genetic influence on curve progression, suggesting its multifactorial nature, as is the case for AIS onset. Of the 15 genes reported as having SNPs with a significant association with progressive AIS, none showed sufficient power to sustain clinical applications.
Discordant AIS progression described in monozygotic twins [37] suggested the involvement of nongenetic factors and epigenetic processes are emerging as the best candidates [37], with a series of genes whose methylation was correlated with AIS curve severity [34,36,38]. Nine studies reporting epigenetic modifications showed promising results in terms of reliable markers suggesting epigenetics as the more promising field for the identification of factors associated with AIS progression, offering a rationale for further investigation in this field.
To the best of our knowledge, this is the first systematic scoping review where the available evidence evaluating the genetic and epigenetic factors influencing AIS curve progression was analyzed and, if necessary, integrated with additional calculations. Moreover, this work included an analysis of epigenetic factors, focusing not only on hereditable factors but also on the importance of environmental influences and tissue-related genetic expression on the AIS phenotype.
The main limitation of the present review is the presence of high heterogeneity among the included studies in terms of a lack of homogeneous study design and prospective comparative studies with high values of associations and predictive capacity, possibly representing the principal selection bias of the present work. Moreover, the absence of a clear, internationally recognized definition of progression of the curve and the low replicability of association between SNP and AIS progression in different populations generate non-reliably comparable conclusions and represent a confounding factor. The number of published papers on genetic and epigenetic factors related to AIS progression is noteworthy and surprising but without a final international consensus. Defining the factors related to AIS curve progression has the potential to completely renew the clinical management of such a frequent disease.
On the other hand, as more AIS progression-associated variants are identified, they could be incorporated into a "risk of progression scoring system" that can predict the risk of progression. Artificial intelligence may be used for this purpose, thanks to the development of algorithms based on deep learning and machine learning, employing data from spine radiographs, clinical patients' features, and genetic/epigenetic factors to create a complete "tailored" diagnostic tool. Although this approach is fascinating, no clinical studies have attempted this approach.
Therefore, in the forthcoming years, different new biomarkers could be combined with clinical and radiographic parameters, hopefully for the development of new therapeutic strategies based on genetic factors and epigenetic modulators. In line with this mission, further prospective comparative studies with homogeneous architecture and cohorts are needed.

Conclusions
In conclusion, prognostic testing for AIS has the potential to significantly modify disease management. This will be achieved only after the identification of reliable markers and an understanding of the underlying biologic pathways. Genetic studies identified a series of loci associated with disease progression, whose power appears, however, insufficient to guide clinical choices. More recent evidence suggests epigenetics as a more promising field for the identification of factors associated with AIS progression, offering a rationale for further investigation in this field. More data are needed, and studies on tissues involved in the pathology, rather than peripheral blood, are necessary.