第一部分

营养、氧化应激与衰老标志物在不同糖代谢人群中与NAFLD的关联性

【研究背景】

尽管非酒精性脂肪肝（NAFLD）的具体发病机制不清，营养与氧化应激在其发生和进展中发挥重要的作用。饮食中抗氧化剂摄入可能改善机体的氧化应激状态和NAFLD，但目前的研究结果并不一致，并且针对中国人群的报道相对较少。不同糖代谢状态下维生素E补充对非酒精性脂肪性肝炎(NASH)的治疗效果可能不同，当研究抗氧化剂摄入与NAFLD的关系时，根据糖代谢状态进行亚组分析值得进一步探究。此外，外周血白细胞端粒长度（LTL）和线粒体DNA拷贝数（mtDNAcn）等衰老的标志物也与氧化应激密切相关, 是预测NAFLD的潜在指标，但是关于mtDNAcn、LTL与NAFLD之间关系的研究仍无定论，且在研究上述关系时，大多数研究者很少同时考虑氧化应激等潜在混淆因素的影响。

【研究目的】

分析饮食抗氧化剂摄入、氧化应激指标与NAFLD的内在联系，并按照受试者是否患有糖尿病进行亚组分析，进一步比较不同糖代谢状态下此关系的差异；探讨氧化应激、衰老标志物及NAFLD的内在联系，以助于进一步理解NAFLD的发病机制和寻找简单的NAFLD的血液标志物。

【研究方法】

本横断面研究纳入来自于北京农村社区2型糖尿病管理项目的307名不同糖代谢状态的受试者（糖尿病，n=66；非糖尿病，n=241）。利用腹部超声诊断NAFLD（NAFLD，n=103；无NAFLD，n=204）。采用24小时回顾法收集饮食信息。实时定量PCR法检测mtDNAcn和LTL。ELISA法检测血清超氧化物歧化酶（SOD）、谷胱甘肽还原酶（GR）、8-羟基脱氧鸟苷（8-oxo-dG）等氧化应激指标。将人群按照是否患有糖尿病和NAFLD进行分组分析。使用Logistic回归分析研究饮食抗氧化剂摄入、血清氧化应激指标、衰老标志物与NAFLD的关系。使用中介效应分析研究SOD是否介导抗氧化维生素摄入与NAFLD之间的关系。

【研究结果】

1. 不同糖代谢状态下营养、氧化应激与非酒精性脂肪肝的关联性

(1) 整个人群按照是否患有NAFLD和糖尿病进行分组比较。NAFLD受试者与无NAFLD受试者相比，摄入更少的饮食维生素A（P=0.023），但是血清8-oxo-dG水平更高（P=0.010）。糖尿病受试者与非糖尿病受试者相比，总热量、碳水化合物、维生素C和δ-生育酚摄入增多（P<0.05），血清氧化应激指标两组差异不显著（P>0.05）。Spearman相关性分析提示，整个人群SOD与维生素A（r=0.128，P=0.025）及α-生育酚摄入正相关（r=0.230，P<0.001）。校正传统的NAFLD的危险因素和热量摄入后，Logistic回归分析提示，整个人群饮食α-生育酚摄入与NAFLD的患病风险边缘显著相关(OR=0.905, 95% CI:0.819-1.000, P=0.050)，而血清8-oxo-dG与NAFLD的患病风险正相关(OR=1.607, 95% CI: 1.082-2.386, P=0.019)。 

(2) 在非糖尿病人群, NAFLD受试者饮食维生素A、α-生育酚摄入、血清SOD水平较无NAFLD受试者明显下降（P<0.05），而血清8-oxo-dG明显升高（P<0.05）；Spearman相关性分析提示SOD与维生素A (r=0.173, P=0.007) 及α-生育酚 (r=0.325, P<0.001) 摄入正相关 ；校正传统的NAFLD的危险因素和热量摄入后，Logistic回归分析提示维生素A (OR=0.997, 95% CI:0.994-0.999, P=0.007)、α-生育酚摄入(OR=0.777, 95% CI:0.658-0.919, P=0.003)、血清SOD水平(OR=0.963, 95% CI:0.936-0.992, P=0.012)与NAFLD的患病风险显著负相关；中介分析进一步提示SOD部分介导了维生素A或者α-生育酚摄入与NAFLD之间的关系。

(3) 在糖尿病人群，NAFLD受试者饮食抗氧化剂摄入、血清氧化应激指标与无NAFLD受试者相比无显著统计学差异(P>0.05)；Spearman相关性分析和Logistic回归分析提示饮食抗氧化剂摄入、血清氧化应激指标与NAFLD患病风险无显著相关性(P>0.05)。

(4) 在糖尿病和非糖尿病人群中均未发现维生素C、β-/γ-生育酚、δ-生育酚、锌、硒等抗氧化剂摄入与NAFLD患病风险显著相关(P>0.05)。

2．氧化应激、衰老标志物与非酒精性脂肪肝的内在联系

(1) NAFLD受试者与无NAFLD受试者相比，有更高的mtDNAcn (P=0.032)，但LTL在两组无显著差异（P=0.347）。

(2) 单因素logistic回归分析提示mtDNAcn与NAFLD的患病风险显著正相关(P=0.033)，但是校正8-oxo-dG后发现mtDNAcn与NAFLD的关系不再有统计学意义(P=0.055)，进一步校正糖代谢状态、抗氧化剂摄入和其他的传统的NAFLD的危险因素后这个趋势仍未改变(P>0.05)，而8-oxo-dG是NAFLD的独立危险因素(P<0.05), 这提示氧化应激可能影响了mtDNAcn和NAFLD之间的关系。Spearman相关性分析进一步证实mtDNAcn与8-oxo-dG正相关(r=0.123, P=0.031)，与SOD负相关(r=-0.141, P=0.013)。

【研究结论】

1. 在整个人群中，饮食α-生育酚摄入与NAFLD的患病风险边缘显著相关；在非糖尿病人群中，饮食维生素A和α-生育酚的摄入与NAFLD患病风险显著负相关，并且这种关系部分由SOD中介；在糖尿病人群中，抗氧化剂摄入与NAFLD的患病风险的关系不显著。因此，我们推测不同糖代谢状态人群饮食抗氧化剂摄入与NAFLD的患病风险的关系可能不同。然而，我们的样本量相对较小，需要更大规模的长期随访研究和相关的机制探索进一步证实这一观点。

2. 在一个包含不同糖代谢状态的中国人群，血清8-oxo-dG而不是外周血白细胞mtDNAcn是NAFLD的独立危险因素，mtDNAcn升高的部分原因可能是由于机体对氧化应激的代偿，氧化应激可能影响了外周血白细胞mtDNAcn和NAFLD之间的关系。因此，氧化应激可能是NAFLD的重要驱动力。

第二部分

抗氧化剂对脂肪变性HepG2细胞系的保护作用及机制探索

【研究背景】

NAFLD的发生和发展机制十分复杂，肝脏脂质代谢异常、氧化应激、细胞凋亡等多种因素都可能参与其中。我们发现在非糖尿病人群中，饮食摄入α-生育酚和维生素A与NAFLD的患病风险负相关，需要进一步的基础实验来验证。此外，关于α-生育酚对NAFLD的具体保护机制尚不明确；而关于维生素A与NAFLD发生风险的关系的研究很不一致，需要更加深入的研究来明确。

【研究目的】

探讨α-生育酚和维生素A的代谢产物（视黄酸）对棕榈酸诱导HepG2细胞脂质沉积、脂质合成、脂肪酸β氧化、氧化应激和凋亡的影响及相关机制。

【研究方法】

1. 不同浓度棕榈酸培养HepG2细胞系，根据CCK-8检测的细胞活性和油红O染色检测的细胞脂质沉积情况，选择合适的棕榈酸浓度。

2. α-生育酚或视黄酸联合棕榈酸培养细胞，根据CCK-8细胞活性检测结果选择合适的α-生育酚和视黄酸浓度。

3. α-生育酚或视黄酸联合棕榈酸培养细胞，分为对照组、棕榈酸组和棕榈酸联合α-生育酚（视黄酸）组，使用油红O染色检测细胞脂滴沉积情况，使用试剂盒检测HepG2细胞甘油三酯（TG）、丙二醛（MDA）、超氧化物歧化酶（SOD）水平。

4. Western blot检测不同处理组细胞脂质合成（Srebp-1、ACCα）、脂肪酸β氧化（CPT1）、SOD和细胞凋亡（BCL-2、Bax、Cleaved Caspase-3）相关蛋白的表达。

【研究结果】

1.随着棕榈酸浓度升高，细胞活性呈现下降趋势，细胞内脂滴沉积呈增加趋势，棕榈酸浓度大于等于0.4mmol/L时，细胞活性出现下降且细胞内脂滴形成明显，故选择棕榈酸浓度为0.4mmol/L。

2. CCK-8检测结果提示相对于棕榈酸培养组，α-生育酚或视黄酸联合棕榈酸培养可增加细胞活性，α-生育酚和视黄酸浓度分别为100umol/L和1umol/L显著增加细胞活性，故选择这两个浓度培养细胞。

3. α-生育酚联合棕榈酸培养与单独棕榈酸培养HepG2细胞相比，可显著减少细胞脂质沉积，降低TG、MDA（P<0.05），升高SOD (P<0.05); 视黄酸培养也可显著减少脂质沉积，降低TG（P<0.05）。

4. Western blot提示α-生育酚联合棕榈酸培养与棕榈酸培养HepG2细胞相比，CPT1、SOD蛋白及抗凋亡蛋白BCL-2表达显著上调（P<0.05），而脂质合成（Srebp-1、ACCα）、凋亡相关（Bax、Cleaved Caspase-3）蛋白显著下调(P<0.05); 视黄酸联合棕榈酸培养细胞CPT1、SOD蛋白表达与棕榈酸培养相比显著上调, 而Bax蛋白表达显著下调（P<0.05）。

【研究结论】

α-生育酚和视黄酸对棕榈酸诱导的HepG2脂肪变性细胞有一定的保护作用。α-生育酚和视黄酸不只是抗氧化剂，发挥抗氧化应激的作用，它们的保护作用还可能是通过改善细胞脂质代谢和凋亡实现的，但是具体的信号通路仍需要进一步深入的研究来探索。

Part 1

Associations of nutrition, oxidative stress, aging markers with non-alcoholic fatty liver disease in population with different glucose metabolism status

Backgrounds:

Although the detail mechanisms of non-alcoholic fatty liver disease (NAFLD) remain to be well clarified, nutrition intake and oxidative stress is involved in the pathogenesis of NAFLD.While dietary antioxidants intake may improve oxidative stress and alleviate NAFLD, previous studies have reported inconsistent results and epidemiological research on those associations based on Chinese population is limited. The therapeutic effect of vitamin E on non-alcoholic steatohepatitis may be different under different glucose metabolism status，therefore when studying the relationship between antioxidants intake and NAFLD, subgroup analysis based on different glucose metabolism status should be further explored. In addition, cellular aging markers leukocyte telomere legth (LTL) and mitochondrial DNA copy number (mtDNAcn) are closely associated with oxidative stress and also potential biomarkers of NAFLD. However, the association between aging markers and NAFLD was not consistent, and when studying the association between aging markers and NAFLD, the effect of potential confusing factors such as oxidative stress markers was seldom taken into account simultaneously by most researchers.

Objectives:

We aimed to explore the association between dietary antioxidants intake, oxidative stress and NAFLD, and to perform subgroup analysis according to whether the participants had diabetes，and further to compare the differences in that association; to explore the internal relationship between oxidative stress, aging markers and NAFLD to deepen our understanding of the mechanism of NAFLD and to find simple blood markers of NAFLD.

Methods:

A total of 307 participants with different glucose metabolism status(diabetes，n=66; non-diabetes, n=241)from a diabetes project in ChangPing district of Beijing were included in the cross-sectional analysis. NAFLD was diagnosed by abdominal ultrasound (NAFLD, n =103; Non-NAFLD, n=204). Dietary intake was assessed by a 24-h food recall. The mtDNAcn and LTL were detected using realtime PCR assay. Serum oxidative stress related markers superoxide dismutase (SOD), glutathione reductase (GR) and 8-oxo-2’-deoxyguanosine(8-oxo-dG) were measured by ELISA. Participants were divided into different groups according to whether they had diabetes and NAFLD. Logistic regression analysis was applied to explore the association of antioxidants intake, oxidative stress related markers, aging markers with NAFLD. Mediation model analysis was used to investigate if SOD was involved in the association between antioxidant vitamins intake and NAFLD.

Results:

1. Associations of nutrition and oxidative stress with non-alcoholic fatty liver disease in different glucose metabolism status

(1) The whole population was divided into different groups according to whether they had NAFLD or diabetes. The consumption of dietary vitamin A (P=0.023) was lower while serum 8-oxo-dG (P=0.010) was higher in the NAFLD group compared to the non-NAFLD group. Total caloric, carbohydrate, δ-tocopherol and vitamin C intake were higher in diabetic population compared to those without diabetes (P<0.05), while there were no significant differences in serum oxidative stress markers between groups. Spearman correlation analysis indicated that dietary vitamin A (r=0.128, P=0.025) and α-tocopherol intake (r=0.230, P<0.001) were positively correlated with SOD in whole population. Multiple logistic regression analysis founded that 8-oxo-dG (OR=1.607, 95%CI:1.082-2.386, P=0.019) was positively while α-tocopherol (OR=0.905, 95%CI:0.819-1.000, P=0.050) was marginally associated with NAFLD after adjusting for conventional NAFLD risk factors and caloric intake in whole population.

(2) In non-diabetic population, we found that α-tocopherol and vitamin A intake and serum SOD were lower while 8-oxo-dG was higher in NAFLD group compared to those of the non-NAFLD group（P<0.05）. Spearman correlation analysis indicated that dietary vitamin A (r=0.173, P=0.007) and α-tocopherol (r=0.325, P<0.001) intake were positively correlated with SOD. Multiple logistic regression analysis found dietary vitamin A (OR=0.997, 95%CI: 0.994-0.999, P=0.007), α-tocopherol intake (OR=0.777, 95%CI:0.658-0.919, P=0.003) and serum SOD (OR=0.963, 95%CI:0.936-0.992, P=0.012) were inversely associated with NAFLD after adjusting for conventional NAFLD risk factors and caloric intake. Mediation analysis indicated that SOD significantly mediated the indirect effect of dietary vitamin A or α-tocopherol intake on NAFLD.

(3) In diabetic population, there were no statistical differences between NAFLD group and non-NAFLD group regarding dietary antioxidants intake and serum oxidative stress markers. Moreover, Spearman correlation analysis and logistic regression analysis indicated that there were no significant associations between antioxidants intake, oxidative stress markers and NAFLD.

（4）No significant associations between vitamin C, β-/γ-tocopherol, δ-tocopherol, zinc, and selenium intake and NAFLD were observed in population with diabetes or without diabetes.

2. Associations between oxidative stress, aging markers and non-alcoholic fatty liver disease

(1) Participants with NAFLD had higher mtDNAcn compared to those without NAFLD (P=0.032), while there was no significant difference in LTL between groups (P=0.347).

(2) Univariate logistic regression analysis revealed that mtDNAcn was positively associated with NAFLD (P=0.033). However, following further adjustment for 8-oxodG, the association between mtDNAcn and NAFLD was not significant (P=0.055). With further adjustment for glucose tolerance status, antioxidants intake and other conventional NAFLD risk factors, this trend did not change (P>0.05), while 8-oxo-dG was independently associated with NAFLD (P<0.05), suggesting that oxidative stress may affect the association between mtDNAcn and NAFLD. Spearman correlation analysis further confirmed that leukocyte mtDNAcn was positively correlated with 8-oxo-dG (r=0.123, P=0.031) and negatively correlated with SOD (r=-0.141, P=0.013).

Conclusions:

1. Dietary α-tocopherol intake was marginally associated with NAFLD in whole population, and dietary vitamin A and α-tocopherol intake were negatively associated with NAFLD in non-diabetic population, and this relationship was partly mediated by SOD, while no significant associations between antioxidants intake and the risk of NAFLD were observed in diabetic population. Thus, the relationship between antioxidants intake and NAFLD may be different in population with different glucose metabolism status. However, the sample size was relatively small, and therefore large-scale longitudinal studies and related mechanism studies are needed to further confirm this view.

2. In a Chinese population with different glucose metabolism status，serum 8-oxo-dG, rather than leukocyte mtDNAcn, was an independent risk factor for NAFLD. The relationship between leukocyte mtDNAcn and NAFLD could be affected by oxidative stress, and the higher mtDNAcn may be partly due to a compensatory mechanism in response to elevated oxidative stress. Therefore, oxidative stress may be an important driver of NAFLD.

Part 2

Study on the protective effect and related mechanism of antioxidants on HepG2 cells with steatosis

Backgrounds:

The mechanisms of occurcence and development of NAFLD are very complex, and many factors such as abnormal lipid metabolism, oxidative stress and apoptosis may be involved. We found that in non-diabetic population，dietary α-tocopherol and vitamin A intake were negatively associated with NAFLD，and basic studies are needed to confirm these findings. In addition, the detailed protective mechanism of α-tocopherol on NAFLD is still unclear, and previous studies have reported inconsistent results on the relationship between vitamin A and NAFLD, and therefore more in-depth research is needed to clarify.

Objectives:

We aimed to explore the effects of α-tocopherol (α-T) and a bioactive metabolite of vitamin A [retinoic acid, (RA)] on lipid deposition, lipid synthesis, fatty acid β-oxidation, oxidative stress and apoptosis in palmitic acid (PA)-induced HepG2 cells.

Methods:

1.HepG2 cells were treated with different concentrations of PA for 24 hours, and the appropriate concentration of PA was selected according to the cell viability detected by CCK-8 and the lipid deposition detected by oil red O staining.

2. HepG2 cells were treated with α-T or RA combined with PA, and the appropriate concentrations of α-T and RA were selected according to the cell viability detected by CCK-8.

3. The cells were divided into control group, PA group and PA combined with α-T (RA) group. The lipid droplet deposition was detected by oil red O staining. The intracellular triglyceride (TG), malondialdehyde (MDA) and superoxide dismutase (SOD) of HepG2 cells were detected by kit.

4.Western blot was used to assayed the protein expression of lipogenesis related enzymes (Srebp-1 , ACCα ), CPT1, SOD and apoptosis related factors (BCL-2, Bax, Cleaved Caspase-3) in different drug treated groups.

Results:

1. As the increasement of PA concentration, the cell viability was decreased while the lipid deposition was increased in HepG2 cell compared to that of control group. The cell viability decreased significantly and the lipid deposition increased greatly when HepG2 cells were treated by PA with the concentration greater than or equal to 0.4 mmol/L. Thus, 0.4mmol/L PA was used for subsequent experiments.

2.The results of CCK-8 test showed that α-T or RA together with PA treatment could increase cell viability compared to that of PA treatment. The concentration of α-T and RA were determined to be 100umol/L and 1umol/L according to the cell viability, respectively.

3. α-T together with PA treatment markedly decreased intracellular lipid droplets accumulation, intracellular TG and MDA (P < 0.05), and increased the level of SOD (P < 0.05) compared to that of PA treatment; RA together with PA also reduced lipid deposition and intracellular TG significantly compared to that of PA treatment (P < 0.05).

4. Western blot showed that α-T together with PA treatment significantly increased the protein expression of CPT1, SOD and BCL-2 (P < 0.05) but significantly decreased the protein expression of lipogenesis related enzymes (Srebp-1, ACCα) and apoptosis related factors (Bax, Cleaved Caspase-3) (P < 0.05) compared to PA treatment. In terms of RA toghther with PA treatment, the protein expressions of CPT1 and SOD were significantly up-regulated while Bax was down-regulated compared to PA treatment (P < 0.05).

Conclusition:

α-T and RA treatment can protect PA-induced HepG2 cell to some extent. α-T and RA are not just antioxidants with antioxidant effect，their protective effects may also be achieved by improving cell lipid metabolism and reducing cell apoptosis, but the specific signaling pathways still need to be further explored.