Correlation Investigation between Pyrrole-DNA and Pyrrole-Protein Adducts in Male ICR Mice Exposed to Retrorsine, a Hepatotoxic Pyrrolizidine Alkaloid

Pyrrolizidine alkaloids (PAs) have been found in over 6000 plants worldwide and represent the most common hepatotoxic phytotoxins. Catalyzed by hepatic cytochrome P450 enzymes, PAs are metabolized into reactive pyrrolic metabolites, which can alkylate cellular proteins and DNA to form pyrrole-protein adducts and pyrrole-DNA adducts, leading to cytotoxicity, genotoxicity, and tumorigenicity. To date, the correlation between these PA-derived pyrrole-protein and pyrrole-DNA adducts has not been well investigated. Retrorsine is a representative hepatotoxic and carcinogenic PA. In the present study, the correlations among the PA-derived liver DNA adducts, liver protein adducts, and serum protein adducts in retrorsine-treated mice under different dosage regimens were studied. The results showed positive correlations among these adducts, in which serum pyrrole-protein adducts were more accessible and present in higher abundance, and thus could be used as a suitable surrogate biomarker for pyrrole-DNA adducts to indicate the genetic or carcinogenic risk posed by retrorsine.


Introduction
Pyrrolizidine alkaloids (PAs) are phytotoxins produced by more than 6000 plant species of 13 distant families [1,2]. Approximately 500-600 potentially poisonous PAs and PA N-oxides have been identified in about 3-5% of the world's floriferous plants [1,3]. Humans are readily exposed to toxic PAs via the consumption of PA-producing plants used in herbal medicines, herbal teas, dietary supplements, and/or PA-contaminated staple foods [4,5]. The transfer of PAs through the food chain from livestock to dietary foodstuffs, such as milk, eggs, meats, and honey, significantly increases the chances of PA exposure in humans [6][7][8][9][10].
PAs are esters formed by a necine base and one or two necic acids. Based on the necine base structure, PAs are generally classified into three types: retronecine (including its 7-α enantiomer, heliotridine), otonecine, and platynecine [1,11]. The retronecine and otonecine types of PAs possessing unsaturated necine bases are hepatotoxic, while platynecine-PAs with saturated necine bases are non-toxic. Retronecine-type PAs are present in many plants, such as various species in genus of Asmachila, Heliotropium, Comfrey, Gynura, Senecio, etc. [1,5,12], and also found to exist in plants in both tertiary base form and N-oxide form [13][14][15]. Catalyzed by hepatic cytochrome P450 monooxygenases, the The PA-derived DNA adducts were firstly characterized in a PA-exposed rat model by Fu et al. in 2010 with full structures elucidated as a pair of epimers of 7-hydroxy-9-(deoxyadenosin-N6-yl) dehydrosupinidine (named DHP-dA-3 and DHP-dA-4) and a pair of epimers of 7-hydroxy-9-(deoxyguanosin-N2-yl)dehydrosupinidine (named DHP-dG-3 and DHP-dG-4) [18] (Figure 1B). These identified pyrrole-DNA adducts, although they have not been detected in biological liver specimens, are found as a mechanism-based common biological biomarker of PA-induced tumorigenicity [19]. On the other hand, DHPAs, as reactive electrophiles, can also bind to nucleophilic sites of proteins to form pyrrole-protein adducts ( Figure 1C). Pyrrole-protein adducts were first detected in the serum of PA-induced liver injury patients who consumed a PA-containing herb by our group in 2011 [12]. Since then, pyrrole-protein adducts were unequivocally detected in the specimens of both liver and serum from PA-treated rats, demonstrating that pyrrole-protein adducts can be used as a biomarker for PA-induced hepatotoxicity [20,21].
Apparently, both PA-derived DNA and protein adducts can serve as useful biomarkers for detecting human exposure to toxic PAs. So far, PA-derived DNA and protein adducts are studied separately [20,[22][23][24], while whether there is a correlation between the formation of these adducts is unknown. Considering the clinical practice, exploration of correlations between PA-derived DNA and protein adducts can potentially develop the use of blood pyrrole-protein adducts as a surrogate biomarker of pyrrole-DNA adducts to correlate with PA-induced tumorigenicity. Retrorsine is a commonly studied hepatotoxic and carcinogenic PA. To pursue this goal, in the present study, we conducted the correlation investigation of retrorsine-derived adducts, including liver pyrrole-DNA adducts, liver pyrrole-protein adducts, and serum pyrrole-protein adducts, formed in retrorsine-treated mice.

Materials and Methods
Procedures involving the care and handling of mice were reviewed and approved by the Animal Experimental Ethics Committee, the Chinese University of Hong Kong (CUHK), under the regulations of the Hong Kong SAR government. Male ICR mice (20-25 g) supplied by the Laboratory Animal Services Center at CUHK were randomly divided into different groups (n = 4 mice per group) for the treatment of retrorsine, a representative toxic PA to induce liver injury in a mouse model [12,[24][25][26] and vehicle control. For the dose-response experiment, mice in different groups were treated orally with a single dose of retrorsine at 10, 20, 40, or 60 mg/kg b.w. and sacrificed at 24 h after the treatment. For the kinetic investigation experiment, mice in different groups were orally administrated with a single dose of 40 mg/kg b.w. of retrorsine and sacrificed at 2, 4, 6, 8, and 12 h, 1, 2, and 3 days, or 1, 2, 4, and 8 weeks after the treatment. For both studies, mice treated in parallel with water served as controls. After the sacrifice by overexposure to carbon dioxide, blood and liver samples were collected. Blood samples were immediately centrifuged at 2000× g for 10 min at 4 • C, and serum samples were harvested and stored at −80 • C for further analysis of serum pyrrole-protein adducts. Liver tissues were stored at −80 • C until the analysis of liver pyrrole-protein and pyrrole-DNA adducts.
Quantitation of liver pyrrole-DNA adducts was conducted according to our previously reported method [18,19]. Liver DNA was extracted using a Blood & Cell Culture DNA Isolation kit (QIAGEN Inc., Valencia, CA, USA) and enzymatically hydrolyzed to nucleosides with micrococcal nuclease, spleen phosphodiesterase, and nuclease P1 as previously reported. The internal standards (DHP- All data are expressed as mean ± SEM. Statistical analyses were performed by Prism 6.0 (GraphPad Software Inc., San Diego, CA, USA).

Results
To investigate the dose-response relationship among different retrorsine-treated groups, mice were orally gavaged with a single dose of vehicle or 10, 20, 40, or 60 mg/kg of retrorsine. Liver injury, manifested by hepatic sinusoidal hemorrhage, was observed in mice dosed with 40 mg/kg and 60 mg/kg of retrorsine [24]. After a 24-h treatment, liver pyrrole-DNA adducts, liver pyrrole-protein adducts, and serum pyrrole-protein adducts were all detected in retrorsine-treated groups with increased formation in a dose-dependent manner but were absent in the vehicle control group (Figure 2A-C, Table S1 and Figures S1 and S2). Pearson's analysis was used to assess correlations among three types of adducts as moderate, strong, or very strong correlations if the Pearson correlation coefficient (r) was between 0.40 and 0.69, 0.70 and 0.89, or 0.90 and 0.99, respectively. Significantly positive correlations have been found in each of the comparisons, indicating that they were closely related over the investigated dose range (Figure 3).  Further, kinetics for the series of retrorsine-derived adducts investigated in the mice model were orally administered with a single dose of 40 mg/kg retrorsine and sacrificed at 2, 4, 6, 8, and 12 h, or 1, 2, and 3 days, or 1, 2, 4, and 8 weeks after dosing. The results demonstrated that both pyrrole-DNA and pyrrole-protein adducts were quickly formed in the liver, and the maximum concentrations (T max ) were reached at about 6 h after the treatment ( Figure 4A,B and Table S2). Afterwards, the decline followed a biphasic elimina-tion pattern, with a fast initial phase followed by an extended slower phase, suggesting sufficient persisting time in the liver. The persistent time for liver pyrrole-DNA adducts and pyrrole-protein adducts has been observed to be 4 and 8 weeks, respectively. While for the pyrrole-protein adducts determined in the serum, without the peak levels observed, the adducts started decreasing initially and remained detectable until 2 weeks post-dosing ( Figure 4C and Table S2). On the other hand, the removal of the serum pyrrole-protein adducts exhibited a one-phase profile, which was synchronous with the fast elimination phase of liver pyrrole-protein adducts. We determined the Pearson's correlation coefficient for each of the comparisons. Similar to the dose-response experiment, there were strong, statistically significant Pearson's correlations among them ( Figure 5).  Taken together, in the present study, the correlation results showed that the binding of DHPA to liver and serum proteins was proportional to its binding to liver DNA. As protein binding is more abundant and can be found in more easily accessible fluids such as blood, it might be possible to estimate DNA binding through the measurement of protein binding.

Discussion
The continuous accumulation of evidence for the harm of PAs to human health necessitates the development and application of robust biomarkers for PA exposure. Characterization of the different types of PA-derived protein and DNA adducts has not only determined the mechanisms of PA-induced toxicities but also potentially provides a means for exposure monitoring and risk assessment. The formation of PA-derived DNA adducts is considered a critical step in the initiation of PA-induced tumorigenicity, which, therefore, could serve as risk-associated biomarkers for early prediction of cancer risk [28,29]. However, the analysis of DNA adducts has limitations in terms of either accessibility or availability of DNA samples from human biological specimens. As a more abundant cellular constituent, protein adducts can thus serve as a useful surrogate biomarker of DNA adducts. Unlike DNA adducts, which can be removed by repair mechanisms, protein adducts are not subject to repair, are apparently stable for the lifetime of different proteins, are more easily found in accessible fluids such as blood, are present in adequate concentrations, and therefore could be considered in the estimation of genetic or carcinogenic risk whenever they can be correlated with DNA binding.
The correlation analysis conducted in the present study showed strong correlations between each pair of the investigated adducts, indicating their molecular events may run in parallel or in a chain of consecutive events. It is known that all toxic PAs undergo metabolic activation to generate DHPAs, which possess the same core structure (pyrrolic moiety) regardless of the original structures of the PAs. DHPAs are chemically and biologically reactive, capable of alkylating cellular proteins and DNAs to form protein adducts and DNA adducts [1,2]. In the present study, extraction of DNA and protein from separate aliquots of liver tissue and comparison of DNA and protein adduct levels did show a very strong correlation, further supporting that they were formed in parallel and not in a competitive process.
For the serum protein adducts, little is known about the mechanisms of pyrroleprotein adducts released into circulation. Some in vitro experiments suggest that this may involve multiple mechanisms, including secretion of adducted proteins and diffusion of DHPAs directly into the plasma. For instance, pyrrole-protein adducts were detected in the medium of human hepatic sinusoidal endothelial cells (HSEC) and human HepaRG hepatocytes cultured in the absence of extracellular protein, suggesting that pyrrole-protein adducts are actively secreted out of hepatocytes into the extracellular space [30][31][32][33]. On the other hand, DHPAs were found to exist in the blood efflux from the hepatic veins of PA-exposed rats, indicating that DHPAs could diffuse out of the hepatocyte and bind covalently to extra-hepatic proteins [34]. Not only detectable in the serum of PA-induced hepatotoxicity patients [35], pyrrole-protein adducts were also detected in the serum of PA treated mice, even with low doses of PAs which do not cause liver injury [16]. Therefore, the measurement of pyrrole-protein adducts in serum can potentially be used as a more practical biomarker for PA intoxication.
In conclusion, by using retrorsine, a representative hepatotoxic and carcinogenic PA, the present study provided the general phenomena on the correlations among the PAderived DNA and protein adducts, which fills a knowledge gap in this vitally important but largely unexplored research field for biomarker studies. With the significantly positive correlation between liver pyrrole-DNA and pyrrole-protein adducts, the pyrrole-protein adducts which appeared in the serum could be considered a useful indicator for PA-induced hepatotoxicity as well as liver cancer risk. Further research is required in order to make progress in understanding the molecular background of the correlations among these biomarkers and to apply this tool successfully for human genotoxic risk prediction.  Table S1: Levels of liver pyrrole-DNA adducts, liver pyrrole-protein adducts, and serum pyrrole-protein adducts formed in male ICR mice at 24 h after a single oral dose of 10, 20, 40, or 60 mg/kg of retrorsine. Table S2: Levels of liver pyrrole-DNA adducts, liver pyrrole-protein adducts, and serum pyrrole-protein adducts formed in male ICR mice at different time points after a single oral dose of 40 mg/kg of retrorsine.

Conflicts of Interest:
All authors declare that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript.