In the present study, we found that urinary iron was positively associated with serum ALT level as well as increased risks of prevalent and incident hyperALT. Besides, urinary iron was positively associated with urinary 8-iso-PGF2α and 8-OHdG, and urinary 8-iso-PGF2α further mediated the positive association of urinary iron with ALT.
Serum ALT as a marker of liver injury was used in our study. Since liver biopsy is almost impossible to obtain in a general population-based epidemiological study, circulating biomarkers are widely used to identify liver injury. Serum ALT, aspartate aminotransferase, alkaline phosphatase, and bilirubin are all markers of liver injury, of which ALT is a more specific marker of hepatic injury than the others (Kwo et al. 2017). Cave et al used unexplained ALT elevation as the marker of nonalcoholic fatty liver disease among adults without viral hepatitis, hemochromatosis, or alcoholic liver disease (Cave et al. 2010). The study by Ruhl and Everhart also determined liver injury using serum hyperALT among participants without other common causes of liver disease (Ruhl &Everhart 2003). The criteria of abnormal serum ALT level in previous studies were inconsistent. Definitions of hyperALT used in this study (ALT > 30 IU/L for men, or ALT > 19 IU/L for women) has been shown superior sensitivity in patients with minimal to mild histologic lesions of the liver (Prati et al. 2002).
The median concentration of urinary iron was 3.83 µg/mmol Cr in the present study. It has been reported that the urinary concentration of iron among 8 samples from able-bodied persons in America (Denver) was 0.72 µg/mmol Cr (mean) (Howard et al. 1991), and among 64 lead battery repair workers in Ghana was 0.68 µg/mmol Cr (geometric mean) (Dartey et al. 2017). In other studies of Chinese participants, the median concentrations of iron were 7.37 and 5.51 µg/mmol Cr among 122 coke oven workers and 238 controls, respectively (Deng et al. 2019), 7.91 and 5.59 µg/mmol Cr among 275 hypertension and 548 non-hypertension groups, respectively (Wu et al. 2018), and 4.66 µg/mmol Cr among 118 general male participants (Zeng et al. 2013). The level of urinary iron in our study was higher than those in other countries, leading us to speculate that the iron load of the general population in our study might be already beyond the normal range required by the body thus triggered toxic effects. Besides, the concentration of urinary iron of our study participants was lower than that of the other studies in China. These results suggested that the variability of life styles, health condition, environmental pollution, and genetic inheritance might affect the level of urinary iron of study population. Therefore, it is great importance to evaluate the association of urinary iron level with liver ALT among different populations.
In in vitro and in vivo models, the effect of iron overload on liver injury has been documented. Animal models have shown that excessive iron intake through diet could induce liver injury via enhanced oxidative stress in rats and mice (Handa et al. 2016, Kumfu et al. 2016, Mori et al. 2020). However, epidemiological evidence is limited. A clinical study based on 202 pediatric living donor liver transplantation in Japan found that iron overload of the donor is an independent risk factor for liver injury after liver transplantation (Yamada et al. 2020). Our present study clearly demonstrated a significantly dose-response relationships of increased iron load (urinary iron concentration) with increased serum ALT level as well as with increased risks of prevalent and incident hyperALT based on a relatively large general community population.
Studies have shown that excessive iron from iron metabolism disorder are the main incentive factor of ROS production and oxidative stress, inducing lipid peroxidation and oxidative DNA damage (Abalea et al. 1998, Mou et al. 2019, Tang &Kroemer 2020). Bacon et al found that chronic iron overload could induce hepatic mitochondrial and microsomal lipid peroxidation in rat models (Bacon et al. 1983). In in vitro models, Toyokuni and Sagripanti indicated that iron could induce oxidative DNA damage by mediating the production of single and double strand breaks in supercoiled DNA (Toyokuni &Sagripanti 1992). In our current epidemiological study, we found that urinary iron level was significantly positively associated with urinary 8-iso-PGF2α and 8-OHdG, which are biomarkers of lipid peroxidation and oxidative DNA damage, respectively. Furthermore, the association of urinary iron with urinary 8-OHdG was modified by gender, and such an association was stronger in females than in males, suggesting that females might be more susceptible to the effects of iron overload on oxidative DNA damage.
Ferroptosis characterized by iron overload is an iron-dependent novel form of cell death driven by lipid peroxidation, which is morphologically, biochemically, and genetically distinct from apoptosis, necroptosis, and autophagy (Cao &Dixon 2016, Dixon et al. 2012, Xie et al. 2016). Since first proposed by Dixon et al in 2012, ferroptosis has attracted wide attention and been found to play a vital role in the occurrence and progress of many diseases including hepatopathy (Yu &Wang 2021, Yu et al. 2020). The prevailing view of ferroptosis is that small molecule induces depletion of glutathione and inactivation of glutathione peroxidase 4, an enzyme with physiological function of catalyzing the reduction of lipid peroxides (Yang et al. 2014, Yang &Stockwell 2016). Iron overload induced by a dysregulation in iron homeostasis can lead to ferroptosis and further cause acute or chronic diseases (Ng et al. 2019, Sumneang et al. 2020), and the classical lipid peroxidation pathway is an important factor in this process. Fletcher et al found that iron deposited in the parenchymal cells of the liver might change the free radical antioxidant system and further lead to lipid peroxidation, resulting in hepatotoxicity in the rat model (Fletcher et al. 1989). A cross-sectional epidemiological study conducted on 13605 adult participants from the third U.S. National Health and Nutrition Examination Survey indicated that liver injury was associated with increased iron and decreased antioxidants (including carotenoids, vitamin C, and lutein/zeaxanthin) (Ruhl &Everhart 2003), whereas that study did not explore the potential mechanistic roles of antioxidants in the association between iron overload and liver injury. In the present study, we observed that urinary 8-iso-PGF2α rather than 8-OHdG significantly mediated the association of urinary iron with serum ALT. Our results indicated that lipid peroxidation might be one of the potential mechanisms underlying the association of iron overload with liver injury, which might provide an epidemiological evidence for the involvement of ferroptosis in liver injury. The unobserved mediating effect of 8-OHdG might be explained by its newfangled biological effect of anti-inflammation (Choi et al. 2007, Kim et al. 2006), and further research on the potential mechanisms was needed.
This study has several strengths. First, our study was conducted in a general adult population with relatively large sample size. Second, repeated measurements of urinary iron could better reflect the load of systemic iron. Third, the associations of urinary iron with serum ALT level as well as prevalence and incidence of hyperALT were evaluated in a longitudinal study, which could provide causal evidence for iron overload inducing liver injury. Fourth, we further assessed the mediating role of oxidative DNA damage and lipid peroxidation and found that lipid peroxidation significantly mediated the positive association between urinary iron and serum ALT, which could provide epidemiological evidence for the potential mechanism of iron-induced liver injury. Nevertheless, this study still has several limitations worth noting. First, spot morning urine sample rather than 24-hour urine sample or blood sample was used to evaluate the iron level, which might not be an accurate enough indicator of systemic iron load. It is almost impossible to collect 24-hour urinary sample in a population study with such a large sample size. Spot urine sample was widely used in epidemiological studies of large population, with the advantages of non-invasive, convenient collection, and low cost. Furthermore, urinary iron has been reported as a substitute for labile plasma iron (Munoz et al. 2019). Second, some covariates in the present study were collected by self-report. However, data from the Wuhan-Zhuhai cohort were collected by trained staffs following standardized and rigorous protocols to minimize the recall bias. Third, other unmeasured confounders (e.g., other persistent or non-persistent pollutants) that had potential impact on liver injury were not considered, and further studies were imperative.