【作者】 许平波；

【导师】 邓小明； 张卫东； 李金宝； 严诗凯；

【作者基本信息】 第二军医大学 ， 麻醉学， 2008， 博士

【摘要】 【目的】脓毒症预后的早期预测有助于采取更为恰当的治疗措施以改善脓毒症患者的预后,但目前仍缺乏有效的预测手段。本文应用代谢组学方法构建一个能早期、快速、有效预测脓毒症预后的判别模型,筛选与脓毒症预后相关的生物标志物,以便早期采取积极措施以改善脓毒症预后。【方法】75只SD大鼠随机分为盲肠结扎穿孔组(CLP,n=50)和假手术组(n=25),术后12 h在乙醚麻醉下采用断尾法采集外周血1.5ml,室温下静置120min后,3000 g离心20 min,收集上清液,-80℃冻存。继续观察CLP大鼠生存状况6日,然后根据其预后不同,将其分为生存组(存活期＞6日,n=23)、死亡组(存活期24 h～6日,n=22)和废弃标本(存活期＜24 h,n=5)。将三组大鼠血清标本经不同的预处理后,分别行高效液相色谱/质谱仪(HPLC/MS)和核磁共振波谱仪(NMR)分析,获得HPLC/MS谱和~1H NMR谱。对HPLC/MS谱和~1H NMR谱进行数据处理和多元统计分析后,分别获得与各组区分的标志物,然后采用径向基函数神经网络算法(RBFNN)分别构建脓毒症的预后判别模型,并对其预测效能进行评估。【结果】基于HPLC/MS谱图信息,主成分分析法(PCA)可完全区分三组大鼠的生理特征。代谢物鉴定发现,与假手术组相比,CLP组,尤其是死亡组大鼠,其血清中多种游离脂肪酸如棕榈油酸、棕榈酸、亚油酸、软脂酸和硬脂酸的相对浓度显著降低(P＜0.05或0.01)。这表明,脓毒症后外周血中用于氧化供能的多种游离脂肪酸被大量消耗,以弥补机体能量生成的不足,其降低程度与预后相关;与上述5个游离脂肪酸不同,CLP组大鼠尤其是生存组大鼠血清中多个多不饱和脂肪酸,如亚麻酸、二十二碳六烯酸和二十二碳五烯醇相对浓度显著升高(P＜0.05或0.01),这表明脓毒症后,外周血中具有抗炎作用的多不饱和脂肪大量释放,以抑制过度激活的炎症反应,其外周血浓度越高,预后越好。采用RBFNN构建的脓毒症预后判别模型准确预测的敏感性为(96.1±3.6)%,特异性为(91.0±4.3)%。利用~1H NMR谱图信息,正交信号校正偏最小二乘-判别分析法(OPLS-DA)也可较好区分三组大鼠生理特征。代谢物鉴定发现,与假手术组相比,CLP组大鼠外周血羟丁酸、乳酸、丙氨酸、醋酸和乙酰乙酸水平显著增加,而甲酸水平显著下降(P＜0.05),尤其是死亡组大鼠,其变化程度更为显著,这表明脓毒症后糖、蛋白质、脂肪与核酸的代谢明显异常,且其程度与脓毒症预后相关。采用RBFNN构建的脓毒症预后判别模型准确预测的敏感性为(91.0±5.9)%(n=10),特异性为(87.9±1.1)%(n=10)。【结论】脓毒症后,糖、蛋白质、脂肪和核酸的代谢明显异常,应用基于HPLC/MS和NMR的代谢组学方法可全面了解上述应激代谢的变化规律,采用代谢信息构建的模型可早期、快速、有效地预测脓毒症大鼠的预后,有望用于临床。HPLC/MS与NMR侧重点不同,前者侧重于大分子物质,如脂肪酸代谢,而后者侧重于小分子物质,如糖和氨基酸代谢,综合两者信息有利于全面理解脓毒症后物质代谢的变化规律,进而判断脓毒症的严重程度。更多还原

【Abstract】 ObjectiveEarly prognostic evaluation of sepsis is an attractive strategy to decrease the mortality of septic patients,but presently there are no satisfactory approaches.This study is to establish an early,rapid and efficient metabonomic approach for prognostic evaluation of sepsis,and then to explore some potential biomarkers related with the outcome.MethodsSeventy-five rats were randomly allocated to sham-operated group(n=25) and CLP group(n=50).According to the difference in survival duration during the following six days,CLP rats were divided into surviving(n=23,survival duration exceeding 6 d) and nonsurviving group(n=22,survival duration between 24 h and 6 d),excluding five undesirable ones(dead during the first day).At 12 h after surgery, approximately 1.5 ml blood was drawn using tail shearing method in these groups, then left to clot for 2 hours at room tempreture.The serum was separated by centrifugation at 3000 g for 20 min,and the aliquot was stored at -80℃until metabonomic analysis.After special pretreatments,serum samples were analyzed to acquire metabolic profiles using HPLC/MS and NMR spectroscopy respectively. Principle component analysis(PCA) and orthogonal single collection partial least squares discriminant analysis(OPLS-DA) were employed to visualize the changes in the metabolic profiles of sera from surviving,nonsurviving and sham-operated rats, then to discover potential biomarkers related with the otcome of septic rats.Radial basis function neural network(RBFNN) was employed to build predictive model for prognostic evaluation of sepsisResults Based on HPLC/MS data,PCA allows a clear discrimination of the pathologic characteristics among rats from three groups.The results from metabolin identifiction that the levels of palmitoleic acid,palmitic acid,linoleic acid,oleic acid and stearic acid are lower in septic rats,especially in nonsurvivors.It suggests that these important free fatty acids mainly used for energy metabolism are consumed greatly to raise energy supply for organism and their levers may be related with the prognosis of septic rats.Moreover,the levers of another three identified metabolin,including acid and docosapentaenoic acid are higher in septic rats,especially in survivors.It suggests that the levers of these polyunsaturated fatty acids(PUFA) with anti-inflammatory effect increase greatly against the excessive inflammatory reaction and might be related with the prognosis of septic rats.A RBFNN model for outcome predication was built upon the metabolic profile data from rat sera with the sensitivity of(96.1±3.6)%(n=10) and specificity of(91.0±4.3)%(n=10).Based on NMR spectra data,OPLS-DA also allows a clear discrimination of the pathologic characteristics among rats from three groups.Compared with sham-operated group,marked decrease in formate lever and increases in hydroxybutyrate,lactate,acetate,acetoacetate and alanine levers,were observed in septic rats,especially in nonsurvivors,which might due to the enhanced catabolisms of carbohydrate,protein,fatty acid and nucleic acid.Based on the integrity metabolic profile data from ~1H NMR data,another RBFNN model for outcome predication was constructed with the sensitivity of(91.1±5.6)%(n=10) and specificity of(87.9±2.1)%(n=10).ConclusionsAfter sepsis,the metabolism of carbohydrate,protein,fat and nucleinic acid changes markedly.Both HPLC/MS-based and NMR-based metabonomic approachs combined with pattern recognition permit the overall monitoring for the stress metabolism,and allow accurate outcome prediction of septic rats in the early stage. The proposed approach has advantages of rapid,low-cost and efficiency,and is expected to be applied in clinical prognostic evaluation of septic patients. The analytic range of HPLC/MS is different with NMR spectroscopy.The former lays particular emphasis on those metabolins with molecular mass＞100,such as free fatty acid.The latter emphasizes those metabolins with molecular mass＜100, such as carbohydrate and amino acid.Therefore,HPLC/MS coupled with NMR spectroscopy will allow overall metabolin information related with sepsis,resulting in more accurate evaluation on illness severity of sepsis.更多还原