Molecular landscape of congenital vertebral malformations: recent discoveries and future directions

Abstract

Vertebral malformations (VMs) pose a significant global health problem, causing chronic pain and disability. Vertebral defects occur as isolated conditions or within the spectrum of various congenital disorders, such as Klippel–Feil syndrome, congenital scoliosis, spondylocostal dysostosis, sacral agenesis, and neural tube defects. Although both genetic abnormalities and environmental factors can contribute to abnormal vertebral development, our knowledge on molecular mechanisms of numerous VMs is still limited. Furthermore, there is a lack of resource that consolidates the current knowledge in this field. In this pioneering review, we provide a comprehensive analysis of the latest research on the molecular basis of VMs and the association of the VMs-related causative genes with bone developmental signaling pathways. Our study identifies 118 genes linked to VMs, with 98 genes involved in biological pathways crucial for the formation of the vertebral column. Overall, the review summarizes the current knowledge on VM genetics, and provides new insights into potential involvement of biological pathways in VM pathogenesis. We also present an overview of available data regarding the role of epigenetic and environmental factors in VMs. We identify areas where knowledge is lacking, such as precise molecular mechanisms in which specific genes contribute to the development of VMs. Finally, we propose future research avenues that could address knowledge gaps.

Background

The segmentally organized human vertebral column is built of 31–33 vertebrae, comprising 7 cervical, 12 thoracic, 5 lumbar, 5 sacral, and 2–4 coccygeal vertebrae fused into one bone (i.e. coccyx), housing neurons, the spinal cord, and blood vessels. Development of the embryonic vertebral column is complex, and deep understanding of this process at a molecular level is critical for grasping the origin of vertebral defects. The notochord and somites are the most important structures responsible for the vertebral column formation. Somites develop from the paraxial mesoderm on either side of the midline, and then differentiate into ventromedial sclerotome and dorsolateral dermomyotome. Sclerotome cells migrate around the notochord and the neural tube, subsequently segregating into two distinct regions: a cranial domain comprising loosely arranged cells and a caudal region characterized by densely packed cells. The process ultimately leads to development of the vertebral bodies, arches, and transverse and spinous processes. The notochord plays a role in establishing the embryo's longitudinal axis, determining the vertebral column orientation, and guiding the formation of the nucleus pulposus of the intervertebral discs. On the other hand, the dermomyotome gives rise to the dermis and skeletal muscles [1,2,3,4] (Fig. 1). Chondrification and ossification are the final steps in the formation of the vertebrae [5]. On the molecular level, vertebral column development depends on the proper action of several signaling pathways, including Wnt, fibroblast growth factor (FGF), Notch, Hedgehog (Hh), retinoic acid (RA), transforming growth factor β (TGF-β), and bone morphogenic protein (BMP) [6,7,8]. The primary function of the vertebral column is to provide structural support for the body.

Fig. 1

Schematic representation of vertebral development in human embryo. NT – neural tube. Created with Biorender.com

Vertebral malformations (VMs) is an umbrella term describing an etiologically heterogeneous group of congenital defects that may be caused by pathogenic variants in the somitogenesis genes, environmental factors, or a combination of both [9,10,11]. The prevalence of VMs is approximately 1–2 per 2000 live births, however, their actual incidence may be higher due to missed or delayed diagnosis [12, 13]. Depending on which process of the vertebral development has failed, VMs have been divided into segmentation, formation, mixed (both segmentation and formation), or other defects [14]. In addition to vertebral defects, fused or missing ribs or their malalignment are often noted [15]. Vertebral defects may be isolated or associated with other congenital anomalies, including congenital kyphosis or scoliosis, VACTERL association, or syndromes such as Klippel–Feil, spondylocostal dysostosis, spondylothoracic dystrophy, Alagille, Gorlin, CHARGE, Jarcho-Levin, Goldenhar or Joubert syndromes [10, 13, 16, 17]. Patients affected by VMs may be either asymptomatic or present with significant disabilities, resulting in body deformations, motor impairment, respiratory distress or chronic pain which seriously reduces their quality of life [10, 18]. Since there is no cure for VMs, treatment focuses on symptoms managed with either lifestyle or surgical interventions. Surgery is indicated mainly in younger patients with thoracolumbar anomalies and particular VMs, i.e., Klippel–Feil syndrome and congenital scoliosis [19,20,21]. The surgical intervention options encompass convex hemiepiphysiodesis, instrumented fusion, osteotomies, vertebrectomies, and utilization of growth-promoting systems [22].

Herein, we present a comprehensive clinical description of rare congenital vertebral column defects, provide an overview of the most relevant and recent findings concerning the molecular and environmental etiology of VMs, and discuss future research directions. In 2009 and 2013, Giampietro et al. released their two review articles in this field, and since then no other comprehensive reviews of the current literature have been published [11, 13]. Our paper attempts to fill the knowledge gap by synthesizing and interpreting the latest literature to offer new insights into the molecular background of VMs.

Classification of VMs

Vertebral anomalies result from formation, segmentation, or simultaneous formation and segmentation defects [14]. Formation failure is due to the absence of vertebral elements occurring in the anterior, anterolateral, posterior, posterolateral, or lateral region and may be complete (hemivertebra, butterfly vertebra, vertebral aplasia) or partial (wedge vertebra). On the other hand, segmentation failure (unilateral unsegmented bar, block vertebra) arises from abnormal embryological segmentation of the vertebral column (Fig. 2).

Fig. 2

Classification of vertebrae malformations based on the segmentation or formation failures. Segmentation defects encompass block vertebra and unilateral unsegmented bar, whereas formation defects include wedge vertebra, hemivertebra, and butterfly vertebra. Hemivertebra is classified into fully segmented, incarcerated, semisegmented, and nonsegmented. Segmentation defects were illustrated using the example of the lumbar spine segment. Created with Biorender.com

Hemivertebra (HV) is one of the most common vertebral anomalies, with an estimated incidence from 1 to 10 per 10,000 live births, and it is mainly detected within the thoracic (Th8) and lumbar spine [23,24,25]. HV occurs when half of the vertebral body fails to develop (unilateral defect), and one pedicle is missing [14]. It has been shown that HV is not a supernumerary vertebra but rather an underdeveloped innate vertebra that originates from asynchronous growth of the hemimetameric pair [26]. Based on the growth pattern and positioning of the HV, the deformity is classified into four subtypes – fully segmented, incarcerated, semi-segmented, and nonsegmented [27]. Importantly, HV represents a common cause of congenital scoliosis [28]. Butterfly vertebra (BV), also termed sagittal cleft vertebra, anterior rachischisis, somatoschisis, or anterior spina bifida, is a rare vertebral malformation of unknown incidence. Due to a lack of midline fusion of two lateral chondrification centers, BV is characterized by two hemivertebrae separated by a cartilaginous septum giving the butterfly appearance on X-ray imaging [29, 30]. The defect occurs primarily in the lumbar spine or less frequently in the thoracic region, and may cause scoliosis or kyphosis [31]. Total aplasia of the vertebral body was proposed to be the consequence of chondrification center defect, and it usually leads to kyphosis. In addition, the presence of the butterfly malformation is associated with various medical conditions, such as Alagille syndrome, Crouzon syndrome, Jarcho-Levin syndrome, and Pfeiffer syndrome [32,33,34,35]. Finally, a wedge vertebra results from a unilateral asymmetry of the vertebral body where two pedicles are present. The anomaly is generally characterized by partial, unilateral chondrification and ossification [14]. Recent findings underscore the role of wedge-shaped vertebrae as a risk factor in the pathogenesis of symptomatic upper lumbar disc herniation [36].

Segmentation failure is usually observed in the cervical and lumbar spine [37]. The most frequent segmentation defect is the unilateral unsegmented bar resulting from a malformation of two or more adjacent vertebrae, leading to the fusion of over three vertebrae. The malformation results in a bony block that involves the disc spaces and facet joints, accompanied by rib fusions on the same side as the bar. A characteristic feature of an unsegmented bar is a lack of growth plates. However, the unaffected side of the vertebral column continues to grow, leading to significant spinal deformities such as congenital scoliosis [21]. The unsegmented bars can occur together with hemivertebrae, which carries a greater risk for the progression of vertebral deformation than each of these defects alone. Block vertebrae are formed due to somite segmentation failure, culminating in partial or complete fusion of the adjacent vertebrae. The morphological features of the condition include a biconcave shape at the fusion site and the presence of residual intervertebral disk material (chorda remnants) in the proximity of the fusion area. Predominantly only two vertebrae within the cervical, thoracic, or lumbar regions of the spine are affected [14]. The most frequent location for the block vertebrae is C2-C3, exhibiting a strong association with Klippel–Feil syndrome [38, 39].

VM genetic etiology

The genetic etiology of VMs remains unexplored in the majority of affected patients. Vertebral defects may accompany the features of various, often rare, congenital syndromes. Based on the Human Phenotype Ontology database, we have listed syndromes characterized by vertebral defects, in which genetic background has been revealed (Table 1). The KIAA1217 gene has not been associated with any syndrome yet. However, very recent investigations suggest its potential involvement in VMs. Rare variants within this gene have been identified in 10 patients with vertebral fusions and other osseous spine abnormalities [40]. In the following chapters of this review, we describe vertebral defects specific to particular segments of the spine currently intensively investigated for their genetic background. Congenital osseous torticollis in the form of Klippel–Feil syndrome was detailed as a cervical spine defect, congenital scoliosis, and spondylocostal dysostoses were depicted as thoracic/lumbar spine defects, developmental spinal stenosis was listed as lumbar spine defect, whereas sacral agenesis as a sacral spine defect. The comprehensive overview of all the genes from our publication is presented in Table 2. Our analysis shows the participation of VM genes in multiple signaling pathways, particularly in Wnt (Wnt/β-catenin, Wnt/PCP), ERK/MAPK, TGF-β, Notch, Hedgehog, BMP, and PI3K/Akt.

Table 1 Genes associated with pathogenesis of some VMs syndromes [40, 155,156,157,158,159,160,161,162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,177,178,179,180,181, 224]. C–cervical, Th–thoracic, L–lumbar, SD–skeletal deformities, N/A–not applicable, ND–not determined, VBs–vertebral bodies, VMs–vertebral malformations

Table 2 Characterization of gene variants associated with vertebral malformations. Bial–biallelic, Comp het–compound heterozygous, Hemi–hemizygous, Het–heterozygous, Hom–homozygous, MF–multifactorial, ND–not determined; ^agenes associated with several syndromes

Cervical spine

Congenital osseous torticollis—Klippel–Feil syndrome

Klippel–Feil syndrome (KFS) is a complex skeletal disorder characterized by the fusion of at least two cervical vertebrae, initially reported by Maurice Klippel and Andre Feil [41]. Congenital vertebral fusions may occur at any cervical spine level, although the most often affected vertebrae are C2-C3 and C5-C6 [42]. Since the first description of this syndrome, three morphological subtypes of the disorder have been identified: type I, characterized by the fusion of cervical and upper thoracic vertebrae, type II, with only one or two pairs of fused cervical vertebrae (Fig. 3), and type III, with the fusion of cervical vertebrae combined with the fusion of lower thoracic or lumbar vertebrae [43]. KFS is reported in 1 of 40,000 to 42,000 newborns worldwide. However, the incidence of this syndrome remains underreported due to a lack of population screening studies and frequent asymptomatic occurrence. Studies involving 2917 patients at the emergency department and 131 patients with cervical spondylotic myelopathy, who underwent spine imaging, revealed the prevalence of KFS to be 0.58% and 3.82%, respectively [42, 44]. A diagnosis of KFS is based on the clinical triad, which includes a short neck, low-set posterior hairline, and limited head and neck movements. Notably, only 34–74% of the affected individuals manifest all three symptoms [45]. KFS can be isolated or associated with numerous abnormalities, including scoliosis, Sprengel deformity, spina bifida occulta, renal abnormalities, vision and hearing impairment, congenital heart defects, and neurological anomalies [46,47,48].

Fig. 3

Anteroposterior (A) and lateral (B) cervical spine radiographs showing vertebrae fusion at C6-C7 in a patient with Klippel–Feil syndrome

There are four genetic forms of KFS with dominant and recessive inheritance: KFS1, KFS2, KFS3, and KFS4 (Table 3). In KFS patients, many chromosomal abnormalities have been reported, i.e., inv(8)(q22.2q22.3); t(5;17)(q11.2;q23); inv(2)(p12q34) or t(5;8)(q35.1;p21.1) [49,50,51,52]. Furthermore, according to Online Mendelian Inheritance in Man (OMIM), pathogenic variants in different genes are associated with autosomal dominant KFS, i.e., GDF6 (MIM: 601147), GDF3 (MIM: 606522), and autosomal recessive KFS, i.e., MEOX1 (MIM: 600147), and MYO18B (MIM: 607295). The GDF3 and GDF6 genes are members of the TGF-β/BMP family, and their protein products are essential for forming and developing bones and joints. The MEOX1 gene encodes a homeobox protein MOX-1, a transcription factor expressed in somites. MOX-1 regulates separation of vertebrae from one another during early development. Despite the clinical heterogeneity of KFS, the patients harboring pathogenic variants in the MEOX1 gene display multiple common features, i.e., Sprengel’s deformity, congenital scoliosis, and an ectopic omovertebral bone [53, 54]. The MYO18B gene encodes an unconventional class XVIII myosin, mainly expressed in human cardiac and skeletal muscle. The protein plays a potential role in cellular processes and transcriptional regulation of muscle-specific genes [55]. A null variant in MYO18B was linked to a novel developmental disorder that combines KFS and myopathy. Noteworthy, only a small subset of KFS cases could be explained by pathogenic variants in one of the four mentioned genes [56].

Table 3 Genetic classification of Klippel–Feil syndrome. MIM–Mendelian Inheritance in Men

Multiple genes have been proposed as potential candidates responsible for KFS. A homozygous frameshift variant in RIPPLY2 was identified in a patient suffering from KFS with heterotaxy. Studies indicated that variants in RIPPLY2 could be responsible for a new type of KFS. However, further research is required to verify this possible link [57, 58]. Mouse models also identified some variants in the PAX gene family and the Notch signaling pathway as potential genetic cause of the described disorder [59]. Abnormalities in PAX1 have been identified in 8 out of 63 patients with KFS [60]. Furthermore, researchers found out that among five new candidate genes (BAZ1B, FREM2, VANGL1, SUFU, and KMT2D), the variants in BAZ1B had the strongest association with KFS [61]. On the other hand, a study by Li et al. revealed 11 pathogenic missense variants in eight KFS patients, including COL6A1, COL6A2, CDAN1, CHRNG, FLNB, GLI3, MYH3, POR, and TNXB, but none within KFS-related genes – GDF6, GDF3, MEOX1, and MYO18B [62].

Thoracic/lumbar spine

Congenital scoliosis

Congenital scoliosis (CS) is a spinal deformity resulting from the abnormal shape of vertebrae (hemivertebrae, butterfly vertebrae, wedge vertebrae), segmentation failure, or a combination of both [63, 64]. Hemivertebrae are the most common cause of CS. Many CS patients also have defects in other organs, particularly in the heart and the genitourinary system [65]. This condition is estimated to occur in 1 per 2000 live births and manifests as a lateral curvature of the spine (Cobb angle) exceeding 10 degrees. The indication for CS surgery depends on the degree of CS at the time of diagnosis and the disease progression.

The genetic basis of CS is only partially explained. Approximately 10% of the patients harbor heterozygous TBX6 loss-of-function variants or a deletion copy-number variant (CNV) within chromosome 16p11.2, including the TBX6 gene [66,67,68]. Wu et al. reported that CS patients with TBX6 loss-of-function variants carry an additional hypomorphic variant on the second TBX6 allele, which is a specific haplotype corresponding to one of the following common SNVs: rs2289292, rs3809624, and rs3809627 [68]. In two subsequent studies, researchers found these variations in TBX6 in about 9.6% and 7.14% of CS patients, respectively [69, 70]. TBX6 belongs to the T-box family and encodes a transcription factor controlling presomitic mesoderm segmentation and differentiation during development [71, 72]. In 2019, Liu et al. defined TBX6-associated congenital scoliosis (TACS) as a unique clinically recognizable subtype of CS [73, 74].

In addition to 16p11.2 deletion, involving the TBX6 gene, a recent study revealed novel CNVs carried by CS individuals [75]. Lai et al. identified recurrent CNVs encompassing three scoliosis-related genes, including NOTCH2, DSCAM, and SNTG1 and four genes (DHX40, NBPF20, RASA2, and MYSM1) possibly linked to skeletal abnormalities [75].

New CS candidate genes have also been proposed, i.e., TBXT, FBN1, PTK7, SOX9, and Dstyk [76,77,78,79,80,81]. Similarly to TBX6, TBXT (also known as Brachyury or T), a member of the T‐box family, is highly expressed in the notochord and is involved in mesoderm formation and axial elongation [82]. According to some studies, FBN1 may trigger CS by upregulating TGF-β signaling, which is essential for skeletal development [78, 83]. The third candidate gene, PTK7, plays a crucial role in canonical and non-canonical Wnt signaling, whereas the fourth CS candidate gene, SOX9, is involved in chondrocyte differentiation, notochord maintenance, and demarcation of intervertebral disc compartments [84,85,86]. Finally, variants of Dstyk may result in CS-like VMs in zebrafish due to disrupting the formation of the notochord vacuole through the mTORC1/TFEB pathway [81].

Spondylocostal dysostosis

Another congenital spinal disorder, spondylocostal dysostosis (SCD), shares a similar phenotype with CS. SCD is a rare genetic defect characterized by malformations of the ribs and vertebrae (hemivertebrae, butterfly vertebrae, fusion, block, or mixed abnormalities). SCD patients often present with a short neck, short trunk, and scoliosis [17, 87]. To date, SCD has been classified into seven subtypes based on their phenotypes and disease genes: SCD1 with pathogenic variants in DLL3, SCD2 with pathogenic variants in MESP2, SCD3 with pathogenic variants in LFNG, SCD4 with pathogenic variants in HES7, SCD5 with pathogenic variants in TBX6, SCD6 with pathogenic variants in RIPPLY2, and SCD7 with pathogenic variants in DLL1. All these disorders are inherited in an autosomal recessive manner. However, SCD5, in addition to autosomal recessive transmission may also present autosomal dominant inheritance pattern [68, 88,89,90,91,92]. It has been shown that SCD may co-occur with additional cervical and sacral spine malformations or costovertebral malformations. In such phenotypes, pathogenic variants are identified in LFNG or DRMT2, respectively [91, 93, 94]. The results of a functional analysis of the missense LFNG variant (p.Phe188Leu) showed no difference in protein expression between the mutant and wild-type mice [91]. In contrast, the Dmrt2 knock-out mice displayed a similar phenotype to a human neonate with SCD, indicating that pathogenic variants in DMRT2 may be related to a new subtype of SCD [93].

Lumbar spine

Developmental spinal stenosis

Developmental spinal stenosis (DSS), also known as congenital lumbar spinal stenosis, is likely caused by fetal and postnatal abnormal development of the posterior spinal elements [95, 96]. The most common clinical features of DSS include a narrow spinal canal, enlarged lamina, and short pedicles [97]. In some cases, the lumbar vertebrae give the spinal canal a trefoil appearance that leads to lumbar and sacral nerve compression [98]. Genetic predisposition to DSS differs between the upper (L1-L4) and the lower (L5-S1) lumbar spine levels. Genome-Wide Association Study showed that L4 and L5 vertebrae DSS-associated SNVs were located within the ZNF704, and DCC genes, respectively. In addition, three candidate genes, i.e., LRP5, COX2, and VDR can contribute to DSS [99]. DSS is often associated with achondroplasia, a type of skeletal dysplasia resulting from specific FGFR3 activating alterations. Such a complication leads to neurologic symptoms in affected individuals and thus requires surgical interventions [100,101,102]. Sporadically, congenital thoracolumbar stenosis is also noted in alkaptonuria, as described recently [103].

Sacral spine

Sacral agenesis

Sacral agenesis is a congenital absence of the entire sacrum. The classic form of sacral agenesis is autosomal dominant Currarino syndrome (MIM: 176450), in which partial agenesis, i.e., hemisacrum, within S2-S5 vertebrae occurs. In addition, patients present with anorectal malformations, a presacral mass (anterior meningocele, enteric cyst, or presacral teratoma), and urogenital anomalies [104]. Over twenty years ago, a causative gene for this syndrome was found, i.e., MNX1, also known as HLXB9 [105]. Recently, whole exome sequencing studies of 6 patients with Currarino syndrome revealed 7 variants that might be linked to the disorder, i.e., a de novo variant in ETV3L (p.Val126Ile), a de novo variant in NCAPD3, a variant in ARID5A (p.Arg55Leu), a missense variant in CDH2 (p.Arg151Ser), a variant in ITIH2 (p.Ile541Ilefs12), a variant in HOXB4 (p.Lys16Asn), and variant in TLE4 (p.Ser650Leu) [106, 107].

The role of environmental factors and epigenetics in congenital spinal deformities

The role of environmental factors

Neural tube defects

Neural tube defects (NTDs) represent a group of congenital anomalies characterized by incomplete neural tube closure during embryonic development. The defects result from a complex interplay of genetic and environmental factors. NTDs encompass a heterogeneous spectrum of congenital anomalies, including anencephaly, spina bifida (SB), encephalocele, and craniorachischisis [108]. Genetic factors play a key role in the etiology of NTDs, with intragenic susceptibility variants identified in multiple genes, including CCL2 (MIM: 158105), FUZ (MIM: 610622), VANGL1 (MIM: 610132), VANGL2 (MIM: 600533), and TBXT (MIM: 601397) [109,110,111,112,113]. The pathogenic variant in the CCL2 gene predisposes to the development of SB. Notably, the CCL2 gene regulates the export level of monocyte chemotactic protein-1 following treatment with interleukin-1-β in vitro [114]. Research has shown that maternal hyperthermia in the first trimester of pregnancy is associated with a twofold increase in the incidence of SB [115]. Hence, inflammation and increased body temperature, mediated by chemokines, may be contributing factors in the pathogenesis of SB. Jensen et al. linked the CCL2A(-2518)G promoter polymorphism with SB, as the allele could attenuate the response to infection [110]. Another predisposing gene in NTDs, expressed in the emerging neural tube, is the FUZ gene. Seo et al. found 5 missense heterozygous pathogenic substitutions in FUZ in an Italian cohort, i.e., p.Pro39Ser, p.Asp354Tyr, p.Arg404Glu, p.Gly140Glu, and p.Ser142Thr. The variants disrupt primary cilia formation and affect directional cell movement, which are crucial processes in developing the spinal neural tube [113]. Furthermore, several heterozygous missense pathogenic variants within the VANGL1 and VANGL2 genes have been associated with a subset of human NTDs. Merello et al. suggested a correlation between three heterozygous missense variants of VANGL1, p.Ala187Val, p.Asp389His, and p.Arg517His, and the occurrence of NTDs [116]. Interestingly, another research group has indicated a predisposition of pathogenic variants in VANGL2 (p.Ser84Phe, p.Arg353Cys, and p.Phe437Ser) to an increased risk of cranial NTDs in human fetuses [109]. Finally, researchers have identified a pathogenic variant in the TBXT gene, TIVS7-2, in individuals suffering from meningomyelocele. The variant has been concomitantly correlated with elevated predisposition to SB [117]. Numerous studies have also identified other risk-candidate genes such as AMOT, ARHGAP36, CELSR1, COL15A1, DACT1, DISP2, DLC1, DTX1, FREM2, FZD6, GPR50, GRHL3, ITGB1, MTHFR, MYO1E, NKRF, PAX3, PRICKLE1, PTK7, RXRγ, SCRIB, SHROOM3, and TKTL1 [118,119,120,121,122,123,124,125,126]. Despite identifying susceptibility variants responsible for NTDs, recent studies have revealed a significant role of environmental factors in the etiology of NTDs. A prospective study has demonstrated that fever during the first month of pregnancy increases the risk of NTDs [115]. Furthermore, a systematic review and meta-analysis conducted in 2005 confirmed that hyperthermia in early pregnancy is a risk factor for NTDs [127]. Other significant factors contributing to the development of NTDs are maternal diabetes and obesity. Specifically, teratogenic implications of hyperglycemia and hyperinsulinemia increase cellular apoptosis within the developing embryonic neural plate. Women diagnosed with diabetes manifest a notable 2- to tenfold escalation in the risk of NTDs, whereas women affected by obesity demonstrate a 1.5- to 3.5-fold increase, with the severity of risk correlating with maternal body mass index [128,129,130]. Thirdly, inadequate maternal nutritional status during pregnancy, i.e., deficiencies in folate, zinc, and B12, is a factor in the increased risk of NTDs. Notably, research strongly supports the association between folate deficiency and NTDs [131, 132]. The recommended folic acid dosage for women with a previous NTD-complicated pregnancy is 4 mg/day [133]. Among antiepileptic drugs, valproic acid is the most widely recognized teratogenic drug associated with NTDs. The risk of NTDs related to valproate exposure appears to be dose-dependent, necessitating cautionary measures to avoid its use or to limit the dosage [134]. Finally, alcohol and caffeine consumption and maternal exposure to passive smoking are potential risk factors, however, more studies are needed [135,136,137].

Caudal dysgenesis syndrome

Caudal dysgenesis syndrome (CDS; MIM: 600145), also classified as neural tube defect, is a form of sacral agenesis, in which various heterogeneous constellations of symptoms are observed. The CDS phenotype encompasses defects of caudal derivatives, such as anomalies affecting the caudal spine, the spinal cord, the hindgut, the urogenital system, and sporadically the lower extremities (sirenomelia) [138, 139]. Amongst CDS causes, one may list maternal insulin-dependent diabetes during pregnancy (detected in 15–25% of mothers who gave birth to affected children) and pathogenic variants within the VANGL1 or CELSR1 genes [112, 140, 141]. Furthermore, the influence of exogenous substances on the fetus, including retinoic acid and insulin, is also a potential risk factor [142].

The role of epigenetics

Epigenetic factors represent another potential mechanism that may be involved in the pathogenesis of VMs. The epigenetic genes involved in the etiology of vertebral defects are summarized in Table 4. Recent studies showed that aberrant DNA methylation might be linked with the pathogenesis of CS. As compared with healthy individuals, CS patients showed hypermethylation in KAT6B, TNS3, IGHG1, IGHM, IGHG3, RNF213, and GSE1, and hypomethylation in SORCS2, COL5A1, GRID1, RGS3, and ROBO2 [143,144,145]. Moreover, DNA methylation is a critical mechanism in the process of genomic imprinting, an epigenetic mode of inheritance in which genes are expressed exclusively from one parental chromosome, depending on their parental origin. These epigenetic modifications during gametogenesis have been implicated in the etiology of several congenital imprinting disorders (IDs), which present with different clinical features. Silver–Russell syndrome (SRS) and Beckwith–Wiedemann syndrome (BWS) represent examples of imprinting disorders associated with VMs [146]. SRS is characterized by growth retardation, macrocephaly at birth, and dysmorphic facial features (triangular face, prominent forehead). Symptoms associated with VMs include scoliosis, kyphosis, kypho-lordosis, lumbar hypomobility, lumbar hypolordosis with lumbar hypomobility, and abnormally high lumbar vertebrae [147,148,149]. Hypomethylation at the imprinting control region 1 (ICR1) located on chromosome 11p15.5, resulting from the loss of paternal methylation, constitutes a primary cause of SRS. This epigenetic aberration affects the expression of growth-regulatory genes, i.e., IGF2 and H19. Furthermore, patients with SRS carry maternal uniparental disomy of chromosomes 7, 14, 16, and 20, aberrant methylation of 14q32.2, maternal gain-of-function variants in CDKN1C, and paternal loss-of-function variants in IGF2 [150]. BWS manifests clinical features, including macrosomia, macroglossia, abdominal wall defects, and elevated risk for embryonal tumors [151]. Additionally, a recent study identified painful scoliosis with lateralized overgrowth as one of the consequences of BWS [152]. Analogously to SRS, most BWS cases exhibit DNA methylation alterations at the chromosomal locus 11p15.5-11p15.4. In contrast to SRS, BWS is typified by hypermethylation at the ICR1 and hypomethylation at the ICR2, which result in dysregulation of three imprinted genes shared with SRS, namely IGF2, H19, and CDKN1C, and the KCNQ1OT gene [151].

Table 4 Description of epigenetic genes associated with vertebral malformations pathogenesis. BWS–Beckwith–Wiedemann syndrome, CS–Congenital scoliosis, ICR1–Imprinting control region 1, ICR2–Imprinting control region 2

Future perspectives and conclusions

Studies regarding the genetic background of VMs are ongoing worldwide. However, their main limitations remain the rare occurrence of VMs, clinical heterogeneity of these defects, and the economic barrier that all impede performing large cohort research screening using advanced technologies, including whole-genome sequencing, transcriptome profiling via RNA-seq, third-generation sequencing, single-cell sequencing, and other more sophisticated functional studies.

Given the phenotypic heterogeneity of VMs, the application of exact classification systems appears critical for clinical recognition and, next, molecular background research. Studies of clinically homogenous groups of VMs patients are highly needed for identifying the causative genetic lesions underlying vertebral defects and closing the knowledge gap in this area. Simultaneously, exploring the potential contribution of epigenetic factors to the development of vertebral disorders is an interesting avenue for future research. While studies into the epigenetics of CS and IDs have yielded promising results in recent years, there is a knowledge gap in the potential role of epigenetics in other described syndromes. Recent studies on rare diseases such as chromatinopathies and Kabuki syndrome have underscored the crucial role of genome-wide DNA methylation analysis in establishing definitive molecular diagnoses, particularly in the cases where initial genetic screenings yield negative results. Simultaneously, integrating genotype, phenotype, and epigenetic factors has been proposed as a promising approach to unraveling the molecular basis of rare diseases [153, 154]. So far, only one promising study has explored the global genome-wide methylation profile in CS patients, albeit with a small sample size of n = 4 [145]. To expand the scope of methylation investigations in CS and initiate studies in other described VMs disorders, novel methods such as comprehensive whole-genome bisulfite sequencing and methylome arrays covering approximately 850,000 loci could be used. We assume that integrative analyses incorporating multi-omics data, encompassing (epi-)genomic, transcriptomic, and chromatin studies, hold significant promise in providing a comprehensive molecular picture of VMs. Furthermore, to our knowledge, there are no cis-regulatory variants in the non-coding DNA described so far in the medical literature that are causative for VMs. Thus, pathogenic variants located in the regulatory elements of the genes involved in embryonic vertebral development represent another putative disease mechanism. Such causative changes can be identified via array comparative genomic hybridization and whole-genome sequencing analyses.

Importantly, the complexity of VMs etiology cannot be excluded. The involvement of external environmental causes such as maternal drug intake, maternal diseases during pregnancy, or other yet unidentified environmental factors affecting the developing fetus or possibly parents before pregnancy, should also be considered. In VMs disorders influenced by environmental factors, the range of structural abnormalities can differ significantly based on the timing of exposure to these factors during embryonic development and the intensity of their impact. As a result, the affected individuals may display a variety of anomalies, with differences in the type and severity of malformations. Conversely, genetic disorders show a more consistent pattern of inheritance and recurrence within families.

In conclusion, the described heterogeneity of VMs highlights the need for interdisciplinary research approaches that integrate genetics, environmental factors, and epigenetic mechanisms.