应用蛋白质芯片技术从血清中筛选肺癌标志蛋白的研究

An Investigation on Screening of Lung Cancer Marker Proteins from Serum Using Protein Chip Technology

摘要: 目的：基于表面增强激光解吸电离飞行时间质谱(SELDI-TOF-MS)蛋白芯片技术筛选肺癌患者血清差异表达蛋白，构建具备潜在诊断价值的肺癌血清蛋白质指纹图谱候选模型，为肺癌早期无创诊断提供新的研究靶点与技术思路。方法：采用WCX2弱阳离子交换蛋白芯片对80例原发性肺癌患者及80例健康对照者血清样本开展蛋白质谱检测，通过Biomarker Wizard软件筛选组间差异蛋白峰，结合支持向量机(SVM)算法构建肺癌诊断候选模型，并通过肺癌荷瘤小鼠模型对核心差异蛋白进行体内初步验证。结果：在1000~30,000 Da质谱检测范围内共识别113个有效蛋白质峰，筛选获得16例组间差异显著的蛋白峰(P < 0.0001)，其中肺癌组高表达10个、低表达6个。基于差异蛋白峰构建的SVM候选诊断模型内部测试集灵敏度为96.56%，特异度为94.79%，AUC达0.987，展现出良好的内部判别效能。动物实验初步证实，核心候选标志蛋白血清浓度与小鼠肿瘤体积呈显著正相关(r = 0.87, P < 0.01)。结论：SELDI-TOF-MS技术可有效筛选肺癌血清差异蛋白，联合SVM构建的诊断模型为具备潜在应用价值的肺癌候选诊断模型，暂未达到临床成熟应用标准。该模型在肺癌早期诊断、临床鉴别诊断中的有效性及稳定性仍需通过蛋白质谱鉴定、多中心大规模前瞻性队列验证、亚组分层分析等后续研究进一步证实。

Abstract: Objective: To screen differentially expressed proteins in serum of lung cancer patients based on surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) protein chip technology, and to construct a candidate serum protein fingerprint model with potential diagnostic value, so as to provide new research targets and technical ideas for non-invasive early diagnosis of lung cancer. Methods: WCX2 weak cation exchange protein chips were used to detect serum protein profiles of 80 patients with primary lung cancer and 80 healthy controls. Biomarker Wizard software was applied to screen differentially expressed protein peaks between groups, and a candidate lung cancer diagnostic model was constructed combined with support vector machine (SVM) algorithm. A lung cancer tumor-bearing mouse model was established for preliminary in vivo verification of core differential proteins. Results: A total of 113 valid protein peaks were detected in the mass range of 1000~30,000 Da, and 16 significantly differential protein peaks were screened (P < 0.0001), including 10 up-regulated and 6 down-regulated peaks in lung cancer group. The SVM candidate diagnostic model constructed based on differential protein peaks presented good internal discriminative efficiency, with a sensitivity of 96.56%, specificity of 94.79%, and AUC of 0.987 in the internal test set. Animal experiments preliminarily verified that the serum concentration of core candidate marker protein was significantly positively correlated with mouse tumor volume (r = 0.87, P < 0.01). Conclusion: SELDI-TOF-MS technology can effectively screen differential serum proteins of lung cancer. The SVM combined diagnostic model is a potential candidate diagnostic model for lung cancer and has not yet met the standard of mature clinical application. The efficacy and stability of the model in early diagnosis and clinical differential diagnosis of lung cancer need to be further verified by follow-up studies including protein mass spectrometry identification, large-scale multi-center prospective cohort verification and subgroup stratified analysis.