![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Cellular, Molecular, and Exploratory Toxicology, Abbott Laboratories, Abbott Park, Illinois
(Received July 12, 2007; accepted October 25, 2007)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
). They are essential for the classification and elucidation of molecular transport activity, especially as they relate to new pharmaceutical agents in development (Pauli-Magnus and Meier, 2006). They have been applied to predict and understand mechanisms of hepatotoxicity (Dambach et al., 2005). Furthermore, they have been used in the evaluation of normal liver function and in the study of hepatic disease (Elaut et al., 2006).
Whereas PHH systems are clearly advantageous in the evaluation of human toxicological and ADME phenomena, their use is not without limitations. For example, PHH tend to dedifferentiate over time, eventually losing their drug-metabolizing capability (LeCluyse, 2001; Boess et al., 2003). Moreover, human hepatocytes are subject to various stimuli such as exposure to a wide assortment of drugs, environmental influences, disease, diet, natural hormones, alcohol, tobacco, trauma, and genetic variation, among other factors that contribute to experimental variability. Indeed, differences in gene expression have been observed across donors of human hepatocytes (LeCluyse, 2001; Liguori et al., 2005). Lastly, the sporadic availability and significant cost associated with the use of human hepatocytes must be considered.
Because of the technical and financial difficulties associated with the use of isolated human hepatocytes, liver cell lines are often used for experimental assessment at the in vitro level. Of these, the most prominently used is the human hepatocarcinoma-derived HepG2 cell line, which was first established in 1979 (Aden et al., 1979). Since this time, the cell line has become a system of choice for the study of liver-like responses in vitro (Washington and Tchounwou, 2004). However, of particular importance to toxicology and ADME studies, this cell line possesses lower levels of the major phase I drugmetabolizing enzymes then liver, most predominately the cytochrome P450 (P450) enzymes (Rodriguez-Antona et al., 2002). These monooxygenases are the central machinery catalyzing the biotransformation of many xenobiotics, including pharmaceuticals. Albeit, some reports have demonstrated the utility of HepG2 cells in the study of drug-metabolizing enzyme regulation (Wilkening et al., 2003).
In some recent studies, investigators sought to find potential causes for differences in cytotoxicity determination between PHH and HepG2 cell culture systems (Westerink and Schoonen, 2007a,b). One possible cause may be a difference in metabolism between these cell types, and, thus, the expression of phase I and II enzymes was extensively evaluated. It was found that the basal level of many P450s was substantially lower in HepG2 cells, which could account for the differences in cytotoxicity. Furthermore, mRNAs for several UDP-glucuronosyltransferases were expressed at varying levels between the two cell types, which also may contribute to the differences in cytotoxicity.
Although the HepG2 cell line is used rather extensively in some types of ADME and toxicology research, only a handful of reports have focused on the global genomic expression patterns of HepG2 cells (Morgan et al., 2002; Sonna et al., 2003). Even fewer reports have focused on the transcriptomic diversity between human hepatocytes and HepG2 cells. In one study, microarray analysis was performed on two hepatic cell lines, cultured primary hepatocytes and liver slices. Substantial gene expression differences between these systems were observed (Boess et al., 2003). In another example, Harris et al. (2004) compared the basal gene expression profiles of HepG2 cells and human hepatocytes as well as those resulting from exposure to chemical carcinogens. Approximately 31% of the genes expressed basally by HepG2 cells were unique to the cell line, and many of these coded for transcription factors, signal transduction proteins, growth factors, and transporters. Not surprisingly, several transcripts differentially modulated in HepG2 cells were involved in carcinogenesis. Hence, genes that are associated with critical pathways of toxic response were transcribed at disparate levels between the transformed HepG2 cells and PHH, with potential implications for some toxicological and ADME research.
Olsavsky et al. (2007) recently examined the transcriptomic variation of PHH in Matrigel sandwich cultures versus liver cell lines including HepG2 cells. Whereas the PHH sandwich cultures showed basal and postxenobiotic exposure expression profiles similar to those from liver, the cell lines evaluated (HepG2 and Huh7) failed to elicit a comparable response.
Here, we sought to compare the transcriptomes of HepG2 cells and PHH before and after exposure to a panel of fluoroquinolones, a class of bacterial DNA gyrase and DNA topoisomerase IV inhibitors. The quinolones as a class are generally safe and effective with few examples of liver toxicity. One exception is trovafloxacin, the use of which has been restricted because of hepatic idiosyncratic drug reaction (IDR) issues (Bertino and Fish, 2000). Our laboratory has previously demonstrated that microarray analysis can distinguish trovafloxacin from other quinolones not associated with hepatic toxicity in PHH (Liguori et al., 2005). Since then, other examples of unique gene expression patterns in PHH or other culture systems from IDR-causing drugs relative to pharmacologically related compounds have been reported (Kier et al., 2004; Guo et al., 2006). In these examples, several endpoints including gene expression patterns induced by the IDR-causing drug, troglitazone, could be distinguished from those for other thiazolidinediones in several cell types, including PHH, rat hepatocytes, or HepG2 cells.
Thus, a central question of this work was to determine how the overall gene expression profiles of HepG2 cells and PHH compared in response to TVX exposure. At present, these types of questions have become increasingly relevant because of a variety of efforts, including the use of stem cell technology, to establish an immortalized human hepatocyte cell line with a phenotype more similar to its primary counterparts (Hengstler et al., 2005) for potential use in ADME or toxicity studies. Data such as these may be valuable for interpretation of results arising from drug evaluation and biotransformation studies using the HepG2 cell line and other novel hepatocyte-like cell lines.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Cytotoxicity Assay and Cell Treatment. A 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay was conducted on HepG2 cells to establish the cytotoxic range of the fluoroquinolones evaluated and to establish doses for the genomic experiments. The assay was performed according to the manufacturer's instructions (Promega, Madison, WI). HepG2 cells were cultured in 96-flat well plates and were subsequently exposed to the drugs for 24 h. At this time, the MTS reagent was added to each well, and the cells were incubated for 2 to 3 h. After incubation, the absorbance readings were determined using a SpectraMax Plus plate reader (Molecular Devices, Sunnyvale, CA) using a visible wavelength of 490 nm.
For the genomic experiments, quinolone compounds, dissolved in 0.1 N KOH (Sigma-Aldrich, St. Louis, MO), were added to the wells with fresh media at levels of 100 µM (HepG2) or 400 µM (primary human hepatocytes) for 24 h using at least two technical replicates. Trovafloxacin was dosed using two separate preparations of HepG2 cells and two separate donors of human hepatocytes. Vehicle control cells were dosed with an equivalent volume of 0.1 N KOH as the experimental samples. For intracellular comparison (HepG2 cells versus PHH), naive cells were harvested using TRIzol reagent (Invitrogen Life Technologies).
RNA Isolation and cRNA Sample Preparation. Total RNA was isolated from the TRIzol extracts using the standard procedure from the manufacturer. The optical density at 260 nm was used to determine RNA concentrations. RNA quality was accessed using a bioanalyzer (Agilent Technologies, Palo Alto, CA) before proceeding to microarray sample preparation. Microarray analysis was performed using the standard protocol provided by Affymetrix (Santa Clara, CA) and as previously described, starting with 5 µg of total RNA (Richert et al., 2006).
Microarray Analysis. Fragmented, labeled cRNA was hybridized to an Affymetrix human genome U133A array, which contains sequences corresponding to roughly 22,200 transcripts, at 45°C overnight. The arrays were washed, developed, and scanned. The microarray scanned image and intensity files were imported in Rosetta Resolver gene expression analysis software version 6.0 (Rosetta Inpharmatics, Kirkland, WA). Resolver's Affymetrix error model was applied, and ratios were built for each treatment array versus its respective control. Agglomerative or divisive cluster analysis was performed using the average link heuristic criteria (agglomerative only) and the Euclidean distance metric for similarity measure. Some microarray data were also evaluated using the Panther Classification system (Thomas et al., 2003). The data discussed in this publication have been deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE9166
[NCBI GEO]
.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
).
|
Expression Profiles of Naive HepG2 Cells Versus PHH. The expression profiles of untreated HepG2 cells and PHH were examined to establish the overall basal similarities and differences between the two cell types. Figure 1B is a plot of gene expression intensity from one donor of PHH versus that of HepG2 cells. Of the 4306 probe sets (p 
0.01) that were differentially regulated between the two cell types, 2004 of these were up-regulated in the HepG2 cells relative to PHH, and 2302 were down-regulated in the HepG2 cells relative to PHH. Many of these probe sets were modulated >2-fold between the cell types. The expression levels of several major cytochrome P450 monooxygenases were lower in HepG2 cells relative to PHH, consistent with prior observations (Rodriguez-Antona et al., 2002; Wilkening et al., 2003; Westerink and Schoonen, 2007a). These included CYP1A1 (reduced 22.3-fold), CYP3A5 (reduced 78.9-fold), CYP2E1 (reduced 27.6-fold), and CYP2C9 (reduced 21.8-fold). Additionally, several phase II metabolizing enzymes, including UGT1A6, UGT2B15, and UGT1A6 were markedly down-regulated whereas others, especially members of the sulfotransferase family, had height-ened levels of expression in HepG2 cells.
Other transcripts that code for proteins critical for proper maintenance of the cell cycle were differentially expressed between the cell types. For example, cyclin E1 (induced 31.7-fold), cell division cycle 2 (induced 48-fold), cell division cycle associated 8 (induced 25.7-fold), and cyclin A2 (induced 16.6-fold) were all expressed at a greater level in HepG2 cells versus PHH. The mRNAs from several histones, which are essential to proper chromatin organization, were greatly reduced in HepG2 cells, including Histone 1, H2bk (reduced 100-fold), H2B histone family, member S (reduced 32.2-fold), and Histone 2, H2aa3 (reduced 6.1-fold). A complete list of transcripts differentially regulated between HepG2 cells and PHH as shown in Fig. 1B is included in Supplementary Table 1.
These data suggest that the global transcriptome of the HepG2 cells varies considerably from that of primary hepatocytes. Transcripts present in overabundance in HepG2 cells included many associated with cell cycle regulation and checkpoint control, regulation of transcription, DNA, RNA, and nucleotide metabolism, transport, and cell surface-linked signal transduction. Those present in diminished levels included many genes associated with cell death, lipid metabolism, and xenobiotic metabolism.
Gene Expression Changes in HepG2 Cells Induced by Quinolone Treatment. Clearly, much transcriptomic variation exists between HepG2 cells and isolated human hepatocytes, but it also would be beneficial to determine the effects of drug treatment on gene expression for each cell type. Therefore, we sought to identify gene expression alterations after exposure to a panel of fluoroquinolones. The effects of this class of antibiotics on human hepatocytes have been studied previously in our laboratory because of the link between one member of the class (trovafloxacin) and idiosyncratic hepatotoxicity (Bertino and Fish, 2000; Liguori et al., 2005). TVX has previously been shown to induce a mRNA profile unique from that of other quinolones in PHH, stemming from a significantly larger number and magnitude of gene expression changes.
To evaluate whether TVX may have a similar effect on HepG2 cells, we exposed these cells to TVX and five related quinolones for 24 h and then analyzed the resulting genomic profiles of treated cells relative to their respective vehicle controls (Fig. 2A). Agglomerative cluster analysis revealed distinct expression profiles for the TVX-treated HepG2 cells compared with the other quinolones examined, similar to the prior result observed with PHH. Likewise, the treatment with other fluoroquinolones resulted in only mild alterations in gene expression as previously observed with PHH. Of these, only grepafloxacin treatment showed any notable gene expression changes, which is also in agreement with the observations with PHH.
|
). Of the 33 cell death-related genes that were differentially expressed by troglitazone versus relevant negative control compounds, slightly more than half (17) of these genes were also uniquely regulated by TVX (as listed in Table 1).
|
Comparison of HepG2 Cells and PHH Quinolone-Induced Gene Expression Changes. Using agglomerative cluster analysis, we next compared the quinolone-induced gene expression profiles from HepG2 cells with those from PHH (Fig. 2B). Two distinct major clusters were apparent for the two cell types for most of the compounds examined. However, TVX-treated HepG2 cell expression profiles clustered closer to the TVX-treated PHH and away from HepG2 cells treated with other quinolones. Irrespective of cell type, TVX treatment resulted in unique mRNA profiles compared with the other quinolone compounds, which are associated with a low risk of hepatotoxicity in humans. Whereas the majority of probe sets were regulated differently, small areas of similarity in gene expression were also observed between TVX-treated HepG2 cells and PHH.
The differences in expression were more remarkable when a cluster analysis was performed focusing only on TVX-treated HepG2 cells and PHH from two different donors (Fig. 2C). With this type of analysis, two major clusters were formed, which were linked to the underlying cell type. Together encompassing HepG2 cells and PHH, TVX treatment resulted in almost 5900 gene expression changes of the approximately 22,000 transcripts on the array. These perturbations were reproducible between donors of PHH and between preparations of HepG2 cells.
To determine whether the TVX-induced gene expression changes were truly distinct from those of the other quinolones evaluated, rather than due to artifacts from dosing, we next elevated the doses of ciprofloxacin, gatifloxacin, and levofloxacin to between 500 and 800 µM in HepG2 cells. Figure 3A is a cluster analysis using a multidimensional scaling algorithm in Rosetta Resolver showing that the TVX-gene expression profiles in both cell types are distinct from those induced by high doses of ciprofloxacin, gatifloxacin, and levofloxacin. Also, apparent from this analysis is the fact that TVX-treated HepG2 cells and PHH clustered closely together and away from the other quinolones evaluated. The analysis also was completed using an agglomerative hierarchical clustering algorithm as depicted in Fig. 3B. In this manner, the HepG2 expression level of individual genes induced by the high-dose quinolones could be compared directly with those from the TVX-treated HepG2 cells. The resulting profiles associated with the high doses of ciprofloxacin, gatifloxacin, and levofloxacin were clearly distinct from the TVX-induced profiles.
|
0.0001 for the ANOVA to find probe sets of interest) designated 2025 probe sets that were differentially modulated between the TVX-treated HepG2 cell group and the PHH group (Fig. 4A). Some of the differences were due to an anticorrelation in expression between groups, whereas a greater number were due to genes that were regulated solely in one group. Again, these ANOVA-identified probe sets were regulated similarly despite the differences in donor for the PHH or the preparation of HepG2 cells. Some major biological function gene ontologies and pathways for these ANOVA-identified probe sets are listed in Table 2.
|
|
Some individual transcripts that were regulated dissimilarly by TVX between the two cell types included helicase, lymphoid-specific (repressed in HepG2 cells but unchanged in PHH), which is thought to be involved with cellular proliferation (Lee et al., 2000), and Cyr61 (cysteine-rich 61) (induced by TVX 20–100-fold in HepG2 cells but unchanged in PHH), a heparin-binding protein that oversees a diverse range of critical cellular functions including cell survival, adhesion, and differentiation (Wang et al., 2005). GADD45B (growth arrest and DNA damage-inducible), a stress response factor and cell death regulator (Major and Jones, 2004), was induced, on average, >12.6-fold in HepG2 cells versus 2.5-fold in PHH. The proinflammatory cytokine, IL-8, a critical factor of hepatocyte repair subsequent to injury (Gómez-Quiroz et al., 2005), was induced 9.4-fold by TVX in HepG2 cells. However, its expression was repressed in TVX-treated PHH.
Although differences in TVX-induced gene expression profiles were predominant for HepG2 cells and PHH, 929 probe sets were found that were transcribed similarly between HepG2 cells and PHH (Fig. 4B), which were associated with diverse gene ontologies and pathways (Table 3). Topoisomerase 1, which is responsible for adjusting the torsional state of DNA and is critical for the balance between the life and death of the cell (Kang et al., 2004), is an example of a gene that was potentiated comparably (induced 
2-fold) between these two cell types. Metallothionein 2A, 1X, and 1H, members of a family of proteins whose overexpression has been linked to increased cell death (Li et al., 2005), were induced by TVX treatment in both PHH and HepG2 cells. BCL2 (B-cell leukemia/lymphoma 2)-associated transcription factor 1, which promotes activation of apoptotic activity (Kasof et al., 1999), was also overexpressed in both types of cells.
|
Comparison of a TVX-Specific Gene Expression Profile in HepG2 Cells Versus PHH. In our previous investigation of the TVX-induced effects on PHH, we identified 142 probe sets whose expression modulation was unique to TVX treatment and was not associated with treatment by any other quinolones (Liguori et al., 2005). Thus, it was considered plausible that some of these probe sets ultimately could be linked with the toxic outcome. Here, we questioned whether the HepG2 cell line would also uniquely regulate this set of sequences upon TVX treatment. Figure 5 shows that the HepG2 cell line regulated this set of genes analogously to its PHH counterparts, which may provide further insight into the mechanism of TVX-induced hepatotoxicity. Furthermore, the other quinolones tested in HepG2 cells failed to regulate the 142-member set, again consistent with the observation made using PHH. Additionally, TVX treatment regulated key pathways, such as RNA processing and mitochondrial apoptosis, similarly between HepG2 cells and PHH. Thus, despite the numerous differences in TVX-induced expression profiles between the cell types, it was clear that the regulation of a critical set of genes was similar.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
). Other notable gene categories that were significantly different between these cell types were involved in xenobiotic metabolism, cell death, transport, and lipid metabolism. Viewing gene expression intensities between the two cell types, it was evident that there were many differences between HepG2 cells and PHH, a factor that may assist in the interpretation of data from other HepG2 studies.
Another characteristic of the HepG2 cell line is the lowered level of phase I drug-metabolizing enzymes such as the P450s. The results presented here, in agreement with other studies, showed lowered levels of cytochrome P450 transcripts in HepG2 cells compared with human hepatocytes (Rodriguez-Antona et al., 2002). Our results provide additional evidence that the lowered level of P450s in HepG2 cells is most likely due to decreased transcription of P450 mRNA, rather than to protein or mRNA degradation. Thus, the lowered levels of P450s could be linked to changes in levels or to biochemical modification of transcription factors that govern the expression of the P450s. Although the mechanism for the reduced transcription of P450s remains unclear, it is thought that it may be associated with alterations in the levels of hepatocyte nuclear factor-1, 3, and 4, and C/EBP (CCAAT/enhancer binding protein) and other key liver-enriched transcription factors (Cereghini, 1996; Jover et al., 2001; Rodriguez-Antona et al., 2002). Indeed, C/EBPβ, 
, and 
were all underexpressed in HepG2 cells relative to PHH, and, therefore, consistent with the prior hypotheses.
Other authors have described differences from PHH and HepG2 cells in gene expression responses to chemicals such as styrene, styrene-7, 8-oxide, and aflatoxin B1 (Harris et al., 2004; Diodovich et al., 2006). Given the difference in effects, we focused on a series of fluoroquinolone compounds that had already been studied comprehensively in PHH (Liguori et al., 2005) and sought to identify the effects of the compounds on HepG2 cells. One member of this class, trovafloxacin, has been associated with idiosyncratic hepatotoxicity (Bertino and Fish, 2000). In PHH, TVX was clearly shown to affect the genomic profiles more extensively than any other of the quinolones tested, which may partially relate to its hepatotoxicity potential in humans.
Here, the TVX effect on HepG2 cells was reminiscent of that seen with PHH, yielding unique gene expression profiles compared with those for the other quinolones examined. The modulation in expression for the other quinolones (at 100 µM levels) examined was relatively mild, implying that these compounds have a lessened impact on hepatocyte and, possibly, intact liver physiology. Overall the gene expression profiles for TVX were consistent between different preparations, but some differences were also apparent. These differences were relatively small in number. An ANOVA (p < 0.0001) was conducted to quantify the differences between the TVX HepG2 cell preparations, and of the 4344 total expression changes (with fold change at least ±2.0 and p 0.01 in at least one sample), only 357 probe sets (8.2%) were considered differentially regulated. Although the differences were relatively minor, these results stress the importance of designing biological replicates into these types of experiments to monitor any potential variation. Furthermore, it would be interesting in future studies to determine the degree of variation in the TVX-induced gene expression observed as HepG2 cells age in culture and after increased time of exposure (>24 h) to TVX.
Significantly, the TVX-induced gene expression profiles in HepG2 cells continued to cluster independently of those from other less hepatotoxic fluoroquinolones dosed at higher levels (500 µM ciprofloxacin and 800 µM levofloxacin and gatifloxacin). This finding suggests that the TVX-induced gene expression profiles were truly distinctive to TVX treatment and to the cell line and not due to dosing effects. This notion was verified by comparing the expression of individual probe sets from the high-dose non-IDR-inducing fluoroquinolones with those from TVX. Whereas a sparse number of genes were expressed similarly between these treatment groups, the overall profiles are quite distinct. To compare the gene expression overlap of these treatments, a Venn diagram was examined, consisting of the genes regulated consistently by TVX in Fig. 4B and the genes regulated consistently (in two of three treatments) by the high-dose non-IDR-inducingquinolones (data not shown). Of the 929 probe sets from Fig. 4B, only 118 (12.7%) overlap between these treatment groups, which is more evidence of a unique TVX effect on these cells.
Another hepatic IDR-causing drug, troglitazone, has similarly been shown to induce a unique set of gene expression changes in rat hepatocytes and to induce a greater number of gene expression changes than a set of negative control compounds (Guo et al., 2006). In an attempt to identify potential commonalities in the effects of IDR-inducing drugs on liver cells, it may prove worthwhile to inspect the similarities between distinct IDR-causing drugs, and therefore, potentially identify and characterize concordances in mechanism(s) (if any). Indeed, many of the transcripts identified as being regulated uniquely by troglitazone versus its negative controls (Guo et al., 2006) were also regulated uniquely by TVX in HepG2 cells. Hence, these results point to the utility of HepG2 cells as one potential model in the study of hepatic IDRs, despite the basal transcriptomic variation.
It was of interest to characterize some of the TVX-induced expression differences between PHH and HepG2 cells. We found >2000 probe sets that were divergently modulated between the HepG2 cell line and PHH. As shown in Table 2, the most significant of these were associated with cell cycle control, including cyclins, cyclin-dependent kinases, minichromosome maintenance-deficient homologs, aurora kinases, and protein phosphatases. Many of these transcripts were repressed by TVX in HepG2 cells but were not significantly perturbed in PHH. Thus, genes that are regulated distinctively by TVX between cell types may be inherently related to the phenotypical variations associated with cell immortalization. Alternatively, these observations may be related to the finding that TVX can repress cellular proliferation in some cell types, and the effects on cell cycle control expression may reflect this condition (Holtom et al., 2000; Zakeri et al., 2000). Because PHH do not replicate in culture, these cells may be affected less by TVX at the level of cell cycle control.
Levels of inflammation-associated genes, such as IL-8, were also identified as being disparate between the cell types. Inflammation has been previously connected to the hepatic IDR-causing potential of TVX in vivo in the rat (Liguori and Waring, 2006; Waring et al., 2006). IL-8 is an important response factor in hepatic injury, and its expression is typically stimulated upon insult (Gómez-Quiroz et al., 2005). For the PHH donors evaluated here, IL-8 expression was down-regulated almost 2-fold, on average, in response to TVX. For the HepG2 cells, it was overexpressed approximately 9.4-fold. In vivo and in isolated peripheral blood mononuclear cells, TVX has diminished IL-8 levels (Purswani et al., 2000). The reason for the discrepancy in IL-8 production upon TVX exposure to these cell types is unclear. The observation is further complicated by the suggestion that TVX can diminish IL-8 transcription due to its cross-reactivity to and inhibition of mammalian topoisomerases (Purswani et al., 2000). Perhaps an inflammation-mediated reaction is regulated by distinct factors for these two cell types, and such a reaction may be complicated by the fact that inflammatory responses are usually governed by an intact immune system, which is absent in these in vitro cell systems. Another possibility stems from the actuality that PHH were still in the presence of an intact immune system several days before isolation, whereas the HepG2 cells were not, which may also influence differential inflammatory-mediated gene expression between the cell types.
Despite the expression differences between TVX-dosed PHH and HepG2 cells, a critical set of genes, previously identified as potentially being associated with TVX hepatotoxicity, displayed similar expression profiles between the cell types. Hierarchical cluster analysis reveals that, with this select set of 142 genes, TVX-treated HepG2 clustered with TVX-treated PHH and away from other quinolone-treated PHH and HepG2 cells (data not shown). This observation provides additional evidence that these genes may have a role in TVX-specific effects and also shows that HepG2 cells can respond comparably to normal hepatocytes, and for some instances can serve as a useful cell system to understand certain toxicological effects.
In this TVX-specific sequence set, several genes were associated with critical cellular processes such as RNA processing and mitochondrial function and were thus discussed as being involved in the hepatotoxic response from TVX. Superimposing the gene expression of the RNA processing pathway between the two cell types yielded a pattern of noticeable concordance, suggesting that TVX treatment may result in potentially deleterious effects on RNA processing in both types of cells. Interestingly, for the genes regulated similarly by TVX between PHH and HepG2 cells, the most significant gene ontologies are those associated with RNA processing and metabolism (Table 3). In our previous studies with PHH, RNA metabolism was also identified as one of the most significantly affected pathways upon TVX treatment (Liguori et al., 2005).
TVX regulated genes associated with mitochondrial function similarly between PHH and HepG2 cells. One example was mitofusin 1, which is localized to the mitochondrial membrane where it partakes in regulation of proper mitochondrial fusion events and is critical to the lasting viability of the cell (Rojo et al., 2002). Mice with knocked down levels of mitofusins failed to progress in development past the embryo stage (Chen et al., 2003). TVX treatment has been shown to initiate oxidative stress and also can subdue the expression of mitofusin 1, both of which may result in the eventual dysfunction of the mitochondria (Liguori et al., 2005). Similarly, the level of mitofusin 1 was reduced in both PHH and in some samples of HepG2 cells (p < 0.06), implying a similar TVX-related effect on mitochondria. A more global view of TVX effects on mitochondrial-related genes identified a large number that are analogously regulated between HepG2 cells and PHH, including ACADVL (acyl-coenzyme A dehydrogenase, very long chain), TFAM (transcription factor A, mitochondrial), and MYCBP (c-myc binding protein), a finding that is again consistent with a related mitochondrial impact.
In this article, we have extensively compared the genomic profiles of HepG2 cells and PHH at the basal level and after exposure to a panel of fluoroquinolones. The results revealed an extensive network of genes, which were regulated differentially in the native state of each cell. TVX treatment of each type of cell analogously resulted in an enhanced effect on the transcriptome relative to a panel of pharmacologically and structurally related compounds. A set of genes, previously identified as potentially related to the hepatotoxic mechanism of TVX, were modulated similarly between the cell types, which may suggest some similarities in TVX-related effects. Overall, these results could aide in the interpretation of some types of ADME, toxicity, and other data derived from HepG2 cells. As technology in ADME progresses toward the generation of hepatic-like cells from stem cell progenitors or in immortalization of hepatocytes into cell lines (Hengstler et al., 2005), a comparison of transcriptomes both before and after xenobiotic exposure may assist in the translation of effects between these new cell types and the original hepatocyte or liver.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
ABBREVIATIONS: ADME, absorption, distribution, metabolism, and excretion; PHH, primary human hepatocyte(s); P450, cytochrome P-450; IDR, idiosyncratic drug reaction; TVX, trovafloxacin; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; NCBI, National Center for Biotechnology Information; UGT, UDP-glucuronosyltransferase; ANOVA, analysis of variance; IL-8, interleukin 8.
The online version of this article (available at http://dmd.aspetjournals.org) contains supplemental material. 
Address correspondence to: Michael J. Liguori, Abbott Laboratories, D-R463 AP9A, 100 Abbott Park Rd., Abbott Park, IL 60064. E-mail: michael.liguori{at}abbott.com
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
agonists on rat primary hepatocytes and human HepG2 cells. Mol Divers 10: 349–360.[CrossRef][Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
