*3.3. PAs Disturb Bile Acid Homeostasis Intra- and Extracellular*

The amounts of the primary bile acids cholic acid (CA) and chenodeoxycholic acid (CDCA), as well as the respective glycine- or taurine-conjugated bile acids (GCA, TCA, GCDCA, TCDCA), were measured in the cell lysates and medium supernatants of PAtreated HepaRG cells by UPLC/MS after 48 h of incubation with PAs under serum-free conditions. Due to their formation by gut bacteria, secondary bile acids were not considered in our model.

The detected intra- and extracellular amounts of bile acids are summarized in Figure 4. PAs were sorted according to their potential to induce cytotoxicity in HepaRG cells (cp. Figure 2). The amounts of bile acids in the medium of the treated cells were normalized to the values obtained with medium of the respective solvent control (untreated cells, set to 100%). The amount of bile acids in the cell lysates was first normalized to the total protein content of the cells and then to the corresponding solvent control.

The pattern of the effects on bile acid homeostasis is obviously similar to the effects on gene regulation and cytotoxicity. The higher the cytotoxic potential of the PA, the stronger the effects on bile acid balance. Furthermore, the amount of bile acids in the medium and in the cells decreased significantly with increasing concentration of the respective PA. The effect was much more pronounced in the cell lysates than in the medium. Besides the reduced amounts of most detected bile acids, the conjugated bile acid TCDCA represents an exception due to a slightly increased level in the medium of PA-treated cells. Additionally, a slight but significant increase of intracellular CDCA could be detected after treatment with the non-cytotoxic PAs indicine and heliotridine.

Cyclosporine A, a drug known to induce cholestasis, was used as a positive control [46]. Exposure of HepaRG cells to 20 μM cyclosporine A for 48 h showed similar but much weaker effects than treatment with toxic PAs.

**Figure 4.** Changes in the intra- and extracellular bile acid concentration after PA treatment for 48 h. Differentiated HepaRG cells were treated with PAs in concentrations of 5, 21 and 35 μM under serumfree conditions. The cells of the heat map are colored according to the relative change of the respective bile acid as mean of three replicates compared to untreated cells (solvent control, 0.35% ACN and 1.7% DMSO). An increase is indicated by yellow color and blue filling indicates a decrease compared to the solvent control (100%). Cyclosporine A (20 μM) was used as positive control (PC) for the induction of cholestasis [46]. Bile acid levels below 10% of solvent control are additionally highlighted by white + (+ below 10%; ++ below 5%; +++ below 1%). Statistics: \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 (one-way ANOVA followed by Dunnett's post hoc analysis versus the respective solvent control). Mean values, standard deviations and *p*-values are summarized in the supplemental material (Table S4). Structural characteristics: retronecine (R), heliotrine (H), otonecine (O) or platynecine (P) type; free base (B), monoester (M), open-chained diester (D(o)) or cyclic diester (D(c)). Abbreviations for measured bile acids: cholic acid (CA), glycocholic acid (GCA), taurocholic acid (TCA), chenodeoxycholic acid (CDCA), glycochenodeoxycholic acid (GCDCA), taurochenodeoxycholic acid (TCDCA).

#### **4. Discussion**

Due to their widespread distribution and their strong hepatotoxic properties, 1,2 unsaturated PAs are among the most important naturally occurring toxins. In current risk assessment, the general assumption that all PAs have equipotent hepatotoxic properties regardless of their structural characteristics has been followed [13]. However, many recent studies suggest structure-dependent effects of PAs on various endpoints like the induction of cytotoxicity and apoptosis, the occurrence of DNA double-strand breaks, and the formation of micronuclei and DHP-DNA adducts [26,47–52]. For all studies mentioned, a highly comparable order of PAs, sorted by their respective effect level, could be observed. The group of PAs showing the strongest effects, if used in the respective test set, always included the representatives of the open-chained diesters of the heliotridine type (lasiocarpine and heliosupine), and the cyclic diesters of the retronecine type (senecionine and seneciphylline). Echimidine (open-chained diester, retronecine type) also always showed a clear effect on the endpoint studied, despite some variations. This ranking has been observed independently of the test system or endpoint investigated and is comparable to the observations in this study. As discussed in detail in Glück et al. [25], it is becoming increasingly apparent that the representatives of the open-chained and cyclic diester groups of the heliotrine and retronecine type show the strongest toxic effects, whereas free bases and monoesters show no or only weak effects. This pattern of effect levels was also observed in the present study dealing with the disturbance of bile acid homeostasis.

A balanced bile acid level is important for normal liver function. Changes in the concentration of cytotoxic bile acids in the liver can lead to severe damage to hepatocytes. Jaundice and HSOS are frequently reported symptoms after PA intoxication, and indicate a connection between PA exposure and disturbance in bile flow [31]. Currently, there are very few studies dealing with an association between PA exposure and bile acid or cholesterol imbalance. Yan and Huxtable identified the first evidence for such a relation in the 1990s by studying changes in bile flow and bile composition in rat liver after PA administration [53,54]. They showed that detoxification of PAs occurs via glutathione conjugation. The resulting hydrophilic conjugates are then efficiently secreted via the bile. Taurine from liver cells and bile also appears to reduce PA toxicity [55]. Furthermore, Yan and Huxtable [56] found evidence that different PAs can induce different effects or effect levels, as retrorsine and senecionine stimulated bile flow and monocrotaline and trichodesmine did not.

Xiong et al. [57,58] investigated the effect of the PA senecionine as a single substance, or of extracts from PA-containing plants, on the expression of genes associated with the synthesis and transport of bile acids in two in vivo studies. In addition, the concentration of various bile acids was measured in the serum of rats after PA ingestion. In both studies, an increase in serum concentration was detected for all bile acids analyzed. In addition to the abovementioned studies, Hessel-Pras et al. [41] showed an induction of liver necrosis, inflammation and a disturbance of bile acid homeostasis in mouse liver after the exposure to senecionine, resulting in increased bile acid concentrations in serum.

Effects on bile acid homeostasis have also been described in several in vitro studies. Comparable to the information provided by the abovementioned in vivo studies, Luckert et al. [30] found indirect evidence for effects of PAs on the metabolism and transport of bile acids in primary human hepatocytes. After treatment with four different PAs, downregulation of the transcript level was observed for the liver transport proteins ABCB11 (BSEP), ABCC2, ABCC3, ABCC6, SLC10A1 (NTCP), SLC22A7, SLC22A9, SLCO1B1, SLCO1B3, SLCO2B1. Waizenegger et al. [40] examined the regulation of an extensive set of genes associated with bile acid homeostasis, with an experimental setup comparable to the design of this study. After treatment of HepaRG with four different PAs, the qPCR analyses and bile acid content measurements showed results very similar to the present study, including a strong reduction of gene expression of several enzymes involved in bile acid uptake, synthesis, metabolism, and excretion. Enzymes involved in de novo bile acid synthesis (CYP7A1, CYP8B1, BAAT) show decreased transcript levels in the in vitro experiment of

this study. In rat liver, however, this change is only partially visible and rather weak. After exposure to senecionine, the expression of *Cyp7a1* and *Baat* was slightly decreased, whereas after ingestion of *Senecio vulgaris* extract, the transcript levels of *Cyp8b1* and *Baat* were reduced [57,58].

According to formerly published results from Waizenegger et al. [40], the analysis of the intra- and extracellular bile acid levels in the present study shows a strong reduction of their concentrations. The effects of the cholestasis-inducing drug cyclosporine A [46] were very similar to the PA-induced variations in bile acid balance, indicating a possible cholestasis-inducing potential of PAs in vivo. Overall, the changes of bile acid amounts were intracellularly stronger than extracellularly. This may be due to the fact that the cells are able to secrete the toxic bile acids very efficiently. The concentration changes for the primary bile acids CA and CDCA were much weaker than for their corresponding conjugates. However, since the unconjugated bile acids are rapidly converted into the glycine and taurine conjugates, their intracellular levels and effects on overall bile acid concentrations are very small compared to the conjugated bile acids.

Comparing gene expression data from the two in vivo studies in rat liver and the in vitro study in human HepaRG, many similarities can be found [40,57,58]. The genes of the nuclear receptors FXR and SHP are less expressed in both, in vitro in the human cell system and in vivo in the rat liver, after exposure to senecionine and the extract of *Senecio vulgaris*. The gene expression of the transporters SLC10A1 (NTCP) and SLCO1B1 is also significantly downregulated in all three studies. These transporters are responsible for the uptake of bile acids into hepatocytes from the blood [59–62]. Reduced expression of these transporters could result in lower bile acid concentrations in hepatocytes and thus counteract cholestasis. Differences are apparent in the regulation of the transporters ABCB11 (BSEP), SLC22A7, and ABCC3. While *Abcb11* (*Bsep*) and *Slc22a7* showed hardly any altered expression at the mRNA level in rat liver, their orthologs were significantly downregulated in human HepaRG cells after PA exposure. ABCB11 (BSEP) transports bile acids from the hepatocytes into the bile, whereas SLC22A7 imports organic anions (potentially bile acids) from the basolateral (blood) side into the hepatocytes [63]. *ABCC3*, on the other hand, showed only minimal downregulation in cell culture experiments, whereas gene expression was greatly increased in the rat. ABCC3 is responsible for exporting bile acids to the basolateral side of hepatocytes. This pathway is considered as an alternative route for the secretion of bile acids from cholestatic hepatocytes [64]. Thus, *Abcc3* upregulation in vivo fits very well with the increased serum concentration of bile acids that was also measured. However, this discrepancy of enhanced bile acid amounts in vivo to low levels observed in our in vitro model can probably be associated with the simplified 2D culture of HepaRG cells. Despite differentiation of HepaRG into hepatocyte-like and biliary epithelium-like cells, the culture model does not fully reflect the compartmentalization and polarization of the liver that may strongly affect the regulation of various processes like the efflux of bile acids. Nevertheless, the results of the cell culture experiment clearly show a PA-mediated impairment of bile acid homeostasis in human hepatocyte-like HepaRG cells.

#### **5. Conclusions**

In the present study, we have shown that PAs have a significant structure-dependent effect on bile acid homeostasis. We were able to demonstrate that especially PAs of the diester type caused strongest dysregulation of expression of several genes responsible for bile acid synthesis, uptake and secretion in HepaRG cells. Furthermore, the amounts of intraas well as extracellular bile acids were strongly affected. The dramatic decreases of intraand extracellular bile acid amounts were also predominantly detected for diester-type PAs showing a clear impairment of the bile acid balance, which may contribute to cholestatic liver disease in vivo. Therefore, our in vitro results support in vivo observations [36,53,54] that PAs could stimulate the formation of cholestasis. Our data show very impressively the

structure dependence of this effect, although this correlation must be verified and further investigated in vivo.

Nevertheless, for a reliable risk assessment of PAs, some knowledge gaps need to be filled. Especially with regard to the classification of PAs according to their structural characteristics and the resulting differences in their toxicity, further refining studies are necessary. Important aspects for future studies should be the analysis of toxicokinetics and the associated metabolic reactions of toxification and detoxification. Further in vivo and in vitro studies are therefore essential for a more precise and reliable assessment of the risk to human health from PA contamination in food.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/foods10051114/s1, Table S1: Cytotoxicity MTT, Table S2: PCR Primer Sequences, Table S3: Gene expression PCR, Table S4: Bile acid quant. LC-MS.

**Author Contributions:** Conceptualization, J.G., A.B. and S.H.-P.; methodology, M.H.; investigation, J.G. and M.H.; writing—original draft preparation, J.G.; writing—review and editing, M.H., A.B. and S.H.-P.; visualization, J.G.; supervision, A.B. and S.H.-P.; project administration, S.H.-P.; funding acquisition, S.H.-P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the German Research Foundation (Grant number LA1177/12- 1) and by the German Federal Institute for Risk Assessment (grant numbers 1322-591 and 1322-624).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in the Supplementary Materials.

**Conflicts of Interest:** The authors declare no conflict of interest.
