Integrated liver-secreted and plasma proteomics approaches identify a predictive model that stratifies MASH in bariatric patients

Background & Aims Obesity is a major risk factor for metabolic associated steatotic liver disease (MASLD) which can progress from metabolic associated steatotic liver (MASL) to metabolic associated steatohepatitis (MASH). There are currently no effective and validated screening tools to stratify obese patients with a greater risk for MASH, independent of liver fibrosis, at a population level. We aimed to characterise the highly abundant and small protein plasma proteomes of worsening MASLD and overlay the liver-secreted proteome to generate a predictive model to stratify patients with and without MASH. Methods Venous blood and liver wedge biopsies were taken from 160 patients undergoing bariatric surgery. MASLD severity was assessed histologically. Liver biopsies from a subset of 96 patients were precision-cut and cultured to assess liver-secreted proteins. Proteomic analysis was performed using liquid chromatography-tandem mass spectrometry on the plasma and the incubation medium cutoff the liver slices. Results Current non-invasive scores failed to stratify MASH in our cohort. The top200 plasma proteome exhibited mild changes in patients with MASH compared to those with No pathology, while the SPEA approach identified substantial differences in plasma proteins of patients with MASH compared to those without MASH. Liver-secreted proteins were remodelled in MASH compared to MASL and individuals with No pathology. There were no significant changes in the liver-secreted proteins and plasma proteome when comparing MASL to those with No pathology. The APASHA model, comprised of APOF, PCSK9, AFM, and S100A6, HbA1c % and AZGP1 stratified MASH in the discovery (AUROC: 0.887, p<0.0001) and validation cohorts (AUROC:0.7673, p=0.0002) and outcompeted other non-invasive scores. Conclusions These proteomic investigations provide a detailed description of liver-secreted and plasma proteome with worsening MASLD. MASH remodels plasma and liver-secreted proteins. Plasma proteomics generated the APASHA model validated in two Australian bariatric cohorts. Further investigation is warranted to interrogate the utility of the APASHA model as a non-invasive risk prediction model in additional cohorts. Lay Summary: Metabolic-associated steatohepatitis (MASH), a more advanced form of metabolic-associated steatotic liver disease (MASLD), alters the levels of many proteins secreted by the liver and in the blood and those. The APASHA model, which is based on proteins that change in the blood or are secreted by the liver in cases of MASH, could potentially be developed into a simple blood test to predict MASH in high-risk groups.

背景与目的：肥胖是代谢相关脂肪性肝病（Metabolic Associated Steatotic Liver Disease, MASLD）的主要危险因素，该病可由代谢相关脂肪变肝（Metabolic Associated Steatotic Liver, MASL）进展为代谢相关脂肪性肝炎（Metabolic Associated Steatohepatitis, MASH）。当前尚无经验证的有效筛查工具，可在人群层面独立于肝纤维化之外，对罹患更高MASH风险的肥胖患者进行分层。本研究旨在表征病情进展的MASLD患者体内丰度较高的小型血浆蛋白质组，并叠加肝脏分泌蛋白质组数据以构建预测模型，从而对合并与未合并MASH的患者进行分层。

方法：本研究从160名接受减重手术的患者中采集静脉血及肝脏楔形活检组织。通过组织病理学评估MASLD严重程度。对其中96名患者的肝脏活检组织进行精准切片并体外培养，以检测肝脏分泌的蛋白质。采用液相色谱-串联质谱法对血浆及肝切片孵育液进行蛋白质组学分析。

结果：现有非侵入性评分无法在本研究队列中对MASH进行有效分层。与无病理改变的受试者相比，合并MASH患者的前200位血浆蛋白质仅呈现轻度变化；而SPEA法发现合并与未合并MASH患者的血浆蛋白质存在显著差异。与MASL患者及无病理改变的受试者相比，合并MASH患者的肝脏分泌蛋白质组发生重塑。将MASL患者与无病理改变的受试者相比，其肝脏分泌蛋白质组及血浆蛋白质组均无显著变化。由APOF、PCSK9、AFM、S100A6、糖化血红蛋白百分比（HbA1c %）及AZGP1构成的APASHA模型，在发现队列（受试者工作特征曲线下面积AUROC: 0.887, p<0.0001）及验证队列（AUROC:0.7673, p=0.0002）中均可有效分层MASH，且性能优于其他非侵入性评分模型。

结论：本蛋白质组学研究详细阐明了病情进展的MASLD患者体内肝脏分泌蛋白质组及血浆蛋白质组的特征。MASH可重塑血浆及肝脏分泌蛋白质组。血浆蛋白质组学构建的APASHA模型已在两个澳大利亚减重手术队列中得到验证。未来需开展进一步研究，以验证APASHA模型作为非侵入性风险预测模型在其他队列中的应用价值。

通俗总结：代谢相关脂肪性肝炎（Metabolic Associated Steatohepatitis, MASH）是代谢相关脂肪性肝病（Metabolic Associated Steatotic Liver Disease, MASLD）的更晚期亚型，可改变肝脏分泌及血液中多种蛋白质的水平。基于MASH患者血液中变化的蛋白质或肝脏分泌的蛋白质构建的APASHA模型，有望开发为一种简易血液检测手段，用于在高危人群中预测MASH的发生。