中文摘要

对于Cobb角＞45°的青少年特发性脊柱侧凸（Adolescent idiopathic scoliosis，AIS）患者，脊柱融合手术是主要治疗手段。然而，脊柱融合术可导致融合节段活动度丢失。对于需要进行腰弯融合的患者，尽可能多的保留腰椎活动节段，对维持患者生活质量、降低腰痛及远端椎间盘退变风险有重要意义。目前，远端融合椎（Lowest instrumented vertebra，LIV）的选择主要依靠站立位X片。而卧位成像手段，如卧位CT，是否有助于AIS患者LIV的选择仍不明了。

对于Cobb角25-40°的AIS患者，支具治疗是主要治疗手段。然而，支具仅可阻止部分患者畸形进展。支具及手术治疗的缺点迫使脊柱外科医生寻求早期干预手段。当前特发性脊柱侧凸（Idiopathic scoliosis，IS）的病因及发病机制仍未阐明。因此，针对IS病程早期的干预手段仍然缺乏。IS患者普遍存在椎旁肌不平衡。而椎旁肌不平衡究竟是导致IS脊柱畸形的原因，还是脊柱畸形引起的继发性改变，目前仍不清楚。探究椎旁肌不平衡和IS的因果关系，可能有助于开发靶向椎旁肌不平衡的治疗手段，从而延缓甚至阻止IS畸形进展。

本研究主要包括两个部分。第一部分，以Lenke-5型AIS患者为研究对象，探究卧位成像是否有助于AIS患者LIV的选择。第二部分，通过多组学手段比较IS及先天性脊柱侧凸椎旁肌不平衡，明确椎旁肌不平衡与IS的因果关系，并探究IS椎旁肌不平衡的关键基因（蛋白）及关键通路。

第一部分 卧位成像在Lenke-5型青少年特发性脊柱侧凸远端融合椎选择中的应用

目的：LIV的选择是Lenke-5型AIS手术治疗的核心问题之一。最后触及椎（Last touched vertebra，LTV）这一概念的提出为LIV的选择带来了便利。然而，相当一部分Lenke-5型AIS患者远端融合至LTV近端的第一个椎体（The first vertebra proximal to LTV，LTV-1）也表现出满意的效果，这促使脊柱外科医生探究更精准的LIV选择策略。在部分Lenke-5型AIS患者中，站立位X片上LTV-1能被卧位CT上的骶骨中垂线（Central sacral vertebral line，CSVL）所触及。这一特性是否有助于Lenke-5型AIS患者LIV的选择仍未知。本研究的目的是制定并评估一项卧位CT辅助的Lenke-5型AIS患者LIV选择策略。

方法：回顾性分析于本中心行后路脊柱融合手术的53例Lenke-5型AIS患者。依据站立位X片上LTV-1是否被卧位CT上的CSVL触及，将患者分为LTV-1组（n=22）及LTV组（n=31）。LTV-1组LIV选择在LTV-1；而LTV组远端融合至LTV。收集人口统计学数据和手术信息。测量术前、术后和末次随访时的影像学参数。应用脊柱侧凸研究学会22项问卷量表（Scoliosis Research Society-22 questionnaire，SRS-22）进行生活质量评估。

结果：两组在性别、手术年龄、手术时长、失血量、异体输血量上无统计学差异。LTV-1组融合范围较LTV组短。术前影像学参数上，LTV-1组Bending相上主弯Cobb角、顶椎偏距、LTV倾斜角、LTV偏距、LTV-1倾斜角、LTV-1偏距、LTV-1旋转度等均小于LTV组。除顶椎偏距外，两组在术后影像学参数上无显著差异。末次随访时，LTV-1组LIV偏距、LIV下位椎间盘开角、冠状位骨盆倾斜角及T2-T12胸后凸大于LTV组。两组在并发症发生率，翻修手术率及末次随访时SRS-22评分上无显著差异。

结论：对于站立位X片上LTV-1可被卧位CT上的CSVL触及的Lenke-5型AIS患者，LIV可选择在LTV-1。

第二部分 多组学比较特发性脊柱侧凸及先天性脊柱侧凸椎旁肌不平衡

目的：既往研究已通过影像学、组织学及肌电图等多种手段证实并解析IS双侧椎旁肌不平衡。然而，椎旁肌不平衡究竟是IS的原因还是结果仍有争议。本研究拟通过多组学手段比较IS与先天性脊柱侧凸（Congenital scoliosis，CS）椎旁肌不平衡，以阐明IS脊柱畸形与椎旁肌不平衡的因果关系，并探究IS椎旁肌不平衡的关键基因（蛋白）及关键通路。

方法：收集5对IS及CS患者凹凸侧椎旁肌。通过转录组测序鉴定IS及CS凹凸侧椎旁肌中的差异表达基因。应用液相色谱串联质谱技术筛选出IS 及CS凹凸侧椎旁肌中的差异表达蛋白。通过液相色谱质谱联用技术检测IS 及CS凹凸侧椎旁肌中的差异代谢物。对差异基因、差异蛋白及差异代谢物进行生物信息学分析。并比较IS及CS中的差异基因、差异蛋白及差异代谢物。通过对差异基因、差异蛋白及差异代谢物的多组学整合分析，鉴定IS及CS凹凸侧椎旁肌中的关键差异基因（蛋白）及差异通路。

结果：

（1）IS中共鉴定到370个差异基因。IS中差异基因主要富集于对外源性刺激的反应、系统过程、细胞-细胞信号系统、信号受体结合、分子功能调节因子等基因本体论（Gene ontology，GO）条目。京都基因和基因组百科全书（Kyoto Encyclopedia of Genes and Genomes，KEGG）分析显示IS中差异基因主要富集于神经活性配体受体互作、细胞因子-细胞因子受体互作、钙离子信号等通路。CS中共鉴定到380个差异基因。CS中差异基因主要富集于多细胞生物过程、多细胞生物过程的调控、系统过程、受体配体活性、受体调节子活性、信号受体结合等GO条目。CS中差异基因主要富集于神经活性配体受体互作、细胞因子-细胞因子受体互作、紧密连接等KEGG信号通路。比较IS及CS中的差异基因仅鉴定出59个呈相同差异趋势的共有差异基因，占IS差异基因的15.9%。

（2）IS中共鉴定到105个差异蛋白。 IS中差异蛋白主要富集于蛋白质糖基化、金属离子反应、ATP依赖性染色质重塑、钙离子结合、DNA结合、蛋白质复合物结合等GO条目。IS中差异蛋白主要富集于糖酵解/糖异生、嘌呤代谢、肥厚型心肌病等KEGG信号通路。CS中共鉴定到48个差异表达蛋白。CS中差异蛋白主要富集于免疫反应、钠离子转运、细胞生长调控、受体活性、磷酸甘油酸激酶活性、内向整流钾通道活性等GO条目。CS中差异蛋白主要富集于糖酵解/糖异生、细胞衰老、Kaposi's肉瘤相关疱疹病毒感染等KEGG信号通路。比较IS及CS中的差异蛋白鉴定出8个呈相同差异趋势的共有差异蛋白，占IS差异蛋白的7.6%。

（3）IS中共鉴定到51个差异代谢物。IS中差异代谢物主要富集于赖氨酸降解、血小板活化等KEGG信号通路。CS中仅鉴定到7个差异代谢物。CS中差异代谢物主要富集于硫胺素代谢，维生素的消化和吸收，缬氨酸、亮氨酸和异亮氨酸降解及血小板活化等KEGG通路。前列腺素G2 是IS和CS仅有的一个共有差异代谢物。

（4）IS凹凸侧椎旁肌中特有的转录组及蛋白组水平均存在差异的基因（蛋白）包括辅肌动蛋白α3和水通道蛋白4等。多组学整合分析显示IS凹凸侧椎旁肌在脂肪细胞脂解的调节和醛固酮的合成和分泌等信号通路存在显著差异。而CS凹凸侧椎旁肌未见转录组、蛋白质组及代谢组差异均富集的信号通路。

结论：IS凹凸侧椎旁肌存在不同于CS的多组学不对称，提示IS中可能存在原发性椎旁肌不平衡。因此，椎旁肌不平衡可能是IS脊柱畸形的原因。辅肌动蛋白α3和水通道蛋白4等基因（蛋白）及脂肪细胞脂解的调节和醛固酮的合成和分泌等通路可能IS椎旁肌不平衡及脊柱畸形的发生、发展中发挥重要作用。

Abstract

Spinal fusion is the mainstay treatment for adolescent idiopathic scoliosis (AIS) patients with Cobb angle exceeding 45 degrees. However, the motion range of fused segments will be lost. For patients in whom the fusion of lumbar curvature is required, preserving mobile segments of the lumbar spine as many as possible is of great significance to maintain the quality of life, reduce the risks of low back pain and distal disc degeneration. Currently, the selection of lowest instrumented vertebra (LIV) mainly relies on standing radiographs. Whether recumbent imaging means, such as recumbent CT, helps for the selection of LIV in AIS patients, is not clear.

Bracing is the primary treatment for AIS patients with Cobb angle of 25 to 40 degrees. However, brace therapy can only prevent the progression of the deformities in a certain proportion of patients. The shortcomings of bracing and surgical treatment urge spine surgeons to seek early interventions. Yet interventions for the early stage of idiopathic scoliosis (IS) is lacking, which may be attributed to the unclarified etiology and pathogenesis. Paraspinal muscle imbalance is prevalent in IS patients. However, it is still unclear whether paraspinal muscle imbalance is the cause of spinal deformities in IS or the change secondary to the deformities. Exploring the causal relationship between paraspinal muscle imbalance and IS may help to develop treatment targeting paraspinal muscle imbalance to delay or prevent the progression of IS.

This study consists of two parts. In part one, Lenke-5 AIS patients were taken as the object to explore whether recumbent imaging is helpful for the selection of LIV in AIS. In part two, congenital scoliosis (CS) was taken as the control. Through multi-omics comparison of paraspinal muscle imbalance between IS and CS, we aimed to clarify the causal relationship between paraspinal muscle imbalance and IS and explore key genes (proteins) and pathways for paraspinal muscle imbalance in IS.

Part I. Application of recumbent imaging for the selection of lowest instrumented vertebra in Lenke-5 AIS

Objective: The selection of LIV is one of the core issues during the surgical treatment of patients with Lenke-5 AIS. Although the proposal of last touched vertebra (LTV) has brought great convenience to the selection of LIV, a considerable number of Lenke-5 AIS patients with the first vertebra proximal to LTV (LTV-1) as LIV displayed satisfactory outcomes. This phenomenon urges spine surgeons to explore more precise strategies for LIV selection. In some Lenke-5 AIS patients, LTV-1 on standing radiographs is touched by central sacral vertebral line (CSVL) on recumbent CT. Whether this characteristic contributes to the selection of LIV in Lenke-5 AIS patients remains unclarified. This study aimed to develop and evaluate a recumbent CT-assisted LIV selection strategy for Lenke-5 AIS.

Methods: 53 Lenke-5 AIS patients who underwent posterior spinal fusion in our center were retrospectively analyzed. Based on whether LTV-1 on standing films was touched by CSVL on recumbent CT or not, patients were divided into the LTV-1 group (n=22) and the LTV group (n=31). Distal fusion to LTV-1 was undertaken in the LTV-1 group, whereas LTV was selected as LIV in the LTV group. Demographic data and surgical information were collected. Radiographic parameters were measured at pre-operation, post-operation and the last follow-up. Quality of life was evaluated using Scoliosis Research Society-22 (SRS-22) questionnaire.

Results: No statistical differences regarding sex, age at surgery, operative time, blood loss and allogeneic transfusion volume, were observed between groups. Shorter fusion range was observed in the LTV-1 group. Smaller preoperative radiographic parameters, including Cobb angle on bending films, apical vertebral translation, LTV tilt, LTV translation, LTV-1 tilt, LTV-1 translation and LTV-1 rotation were observed in the LTV-1 group. Postoperative radiographic parameters were similar between groups except for apical vertebral translation. At final follow-up, greater LIV translation, LIV subjacent disc angle, pelvic obliquity and T2-T12 thoracic kyphosis were observed in the LTV-1 group. There were no significant differences in complication rate, revision rate and SRS-22 questionnaire scores at final follow-up, between groups.

Conclusions: For Lenke-5 AIS patients in whom LTV-1 on standing films is touched by CSVL on recumbent CT, distal fusion to LTV-1 is acceptable.

Part II Multi-omics comparison of paraspinal muscle imbalance between idiopathic scoliosis and congenital scoliosis

Objective: Previous studies have validated and analyzed paraspinal muscle imbalance in IS with multiple approaches, such as imaging, histological methods and electromyography. However, it still remains controversial whether paraspinal muscle imbalance is the cause or the consequence of IS. This study aimed to compare paraspinal muscle imbalance between IS and congenital scoliosis (CS), to clarify the causal relationship between spinal deformities and paraspinal muscle imbalance in IS and explore key differential genes (proteins) and pathways for paraspinal muscle imbalance in IS.

Methods: Bilateral paraspinal muscles of 5 pairs of IS and CS patients were collected. Transcriptome sequencing was performed to identify differentially expressed genes (DEGs) between muscles on the concavity and that on the convexity in IS and CS. Differentially expressed proteins (DEPs) between muscles on the concavity and that on the convexity in IS and CS, were screened out using liquid chromatography coupled to tandem mass spectrometry. Differential metabolites (DMs) between muscles on the concavity and that on the convexity in IS and CS, were detected with liquid chromatography-mass spectrometry. Bioinformatic analysis of DEGs, DEPs and DMs were conducted. Comparison of DEGs, DEPs and DMs between IS and CS was also implemented. Multi-omic analysis of DEGs, DEPs and DMs was conducted to identify pivotal differential genes (proteins) and pathways in paraspinal muscles between the concavity and convexity in both IS and CS.

Results:

(1) 370 DEGs were identified in IS. DEGs in IS were enriched in Gene ontology (GO) terms, including response to external stimulus, system process, cell-cell signaling, signaling receptor binding, molecular function regulator and etc. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed the enrichment of DEGs in IS in neuroactive ligand−receptor interaction, cytokine−cytokine receptor interaction and calcium signaling pathway and etc. 380 DEGs were identified in CS. DEGs in CS were enriched in GO terms, including multicellular organismal process, regulation of multicellular organismal process, system process, receptor ligand activity, receptor regulator activity, signaling receptor binding and etc. KEGG analysis revealed the enrichment of DEGs in CS in neuroactive ligand−receptor interaction, cytokine−cytokine receptor interaction and tight junction. 59 DEGs with the same trend of difference in bilateral paraspinal muscles shared by IS and CS were identified, which account for 15.9% of DEGs in IS.

(2) 105 DEPs were identified in IS. DEPs in IS were enriched in GO terms, such as protein glycosylation, response to metal ion, ATP−dependent chromatin remodeling, calcium ion binding, DNA binding, protein complex binding and etc. KEGG analysis revealed the enrichment of DEPs in IS in pathways, including glycolysis/gluconeogenesis, purine metabolism and hypertrophic cardiomyopathy and etc. 48 DEPs were identified in CS. DEPs in CS were enriched in GO terms, such as immune response, potassium ion transport and regulation of cell growth, receptor activity, phosphoglycerate kinase activity and inward rectifier potassium channel activity and etc. KEGG analysis revealed the enrichment of DEPs in CS in pathways, including glycolysis/gluconeogenesis, cellular senescence, Kaposi's sarcoma-associated herpesvirus infection and etc. 8 DEPs with the same trend of difference in bilateral paraspinal muscles shared by IS and CS were identified, which account for 7.6% of DEPs in IS.

(3) 51 DMs were identified in IS. DMs in IS were enriched in KEGG pathways, including lysine degradation and platelet activation. Only 7 DMs were identified in CS. DMs in CS were enriched in KEGG pathways, including thiamine metabolism, vitamin digestion and absorption, valine, leucine and isoleucine degradation, and platelet activation. Prostaglandin G2 is the only DM shared by IS and CS.

(4) Differential expressed genes (proteins) specific for IS at both the level of transcriptome and proteome include ACTN3 and AQP4. Multi-omics integration analysis showed differential activation of pathways including regulation of lipolysis in adipocytes and secretion of aldosterone in IS. No enriched pathways were found based on DEGs, DEPs and DMs in CS.

Conclusions: Multi-omics asymmetry in IS is different from that in CS, indicating that there might be some primary imbalance in bilateral paraspinal muscles of IS. Therefore, paraspinal muscle imbalance tends to be the cause of spinal deformities in IS. Differential expressed genes, including ACTN3 and AQP4, and pathways, including regulation of lipolysis in adipocytes and synthesis and secretion of aldosterone, may play important roles in the occurrence and development of paraspinal muscle imbalance and spinal deformities in IS.