Short-term in vivo testing to discriminate genotoxic carcinogens from non-genotoxic carcinogens and non-carcinogens using next-generation RNA sequencing, DNA microarray, and qPCR
Genes and Environment volume 45, Article number: 7 (2023)
Next-generation RNA sequencing (RNA-Seq) has identified more differentially expressed protein-coding genes (DEGs) and provided a wider quantitative range of expression level changes than conventional DNA microarrays. JEMS·MMS·Toxicogenomics group studied DEGs with targeted RNA-Seq on freshly frozen rat liver tissues and on formalin-fixed paraffin-embedded (FFPE) rat liver tissues after 28 days of treatment with chemicals and quantitative real-time PCR (qPCR) on rat and mouse liver tissues after 4 to 48 h treatment with chemicals and analyzed by principal component analysis (PCA) as statics. Analysis of rat public DNA microarray data (Open TG-GATEs) was also performed. In total, 35 chemicals were analyzed [15 genotoxic hepatocarcinogens (GTHCs), 9 non-genotoxic hepatocarcinogens (NGTHCs), and 11 non-genotoxic non-hepatocarcinogens (NGTNHCs)]. As a result, 12 marker genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Gdf15, Lrp1, Mbd1, Phlda3, Plk2, and Tubb4b) were proposed to discriminate GTHCs from NGTHCs and NGTNHCs. U.S. Environmental Protection Agency studied DEGs induced by 4 known GTHCs in rat liver using DNA microarray and proposed 7 biomarker genes, Bax, Bcmp1, Btg2, Ccng1, Cdkn1a, Cgr19, and Mgmt for GTHCs. Studies involving the use of whole-transcriptome RNA-Seq upon exposure to chemical carcinogens in vivo have also been performed in rodent liver, kidney, lung, colon, and other organs, although discrimination of GTHCs from NGTHCs was not examined. Candidate genes published using RNA-Seq, qPCR, and DNA microarray will be useful for the future development of short-term in vivo studies of environmental carcinogens using RNA-Seq.
Lovett published the article “Toxicogenomics: Toxicologists brace for genomics revolution” in Science in 2000. He described the new approach of toxicogenomics, in which DNA microarrays are used to profile gene expression in cells exposed to test compounds . Quantitative real-time PCR (qPCR) has been used independently or to confirm DNA microarray results [2, 3]. However, RNA-Seq is now an important tool for examining the role of the transcriptome in biological processes , which could surpass DNA microarray and qPCR [5, 6]. Nevertheless, to date, only a small number of studies have been published on the discrimination of GTHCs from NGTHCs and NGTNHCs using RNA-Seq-based toxicogenomics [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31] (File 1).
Carcinogenicity testing plays an essential role in identifying carcinogens in environmental chemistry and pharmaceutical drug development. However, it is a time-consuming and labor-intensive process to evaluate the carcinogenicity with conventional 2-year rodent-based animal studies . There is thus an increased need to develop novel alternative approaches to these rodent bioassays for assessing the carcinogenicity of substances .
Carcinogens have conventionally been divided into two categories according to their presumed mode of action: genotoxic carcinogens (GTCs) and non-genotoxic carcinogens (NGTCs). An OECD expert group defined that a GTC has the potential to induce cancer by interacting directly with DNA and/or the cellular apparatus involved in preserving the integrity of the genome, while an NGTC has the potential to induce cancer without interacting directly with either DNA or the above-mentioned apparatus .
Bevan and Harrison asserted that genotoxic carcinogens are usually identified based on positive results in different in vitro and in vivo test systems, including detection of DNA strand breaks, unscheduled DNA synthesis, sister chromatid exchange, DNA adduct formation, mitotic recombination, and gene mutation. Typical tests of mutagenicity include the Ames test, in vitro metaphase chromosome aberration assay, in vitro micronucleus assay, and mouse lymphoma L5178Y cell Tk (thymidine kinase) gene mutation assay, in vivo micronucleus assay in rodents, and transgenic rodent mutation assay. NGTCs are considered to have a threshold for exerting hazardous effects and guidelines regarding appropriate levels of exposure to them are set by the various authoritative bodies in the same way as for other hazardous substances. Bevan and Harrison recommend that clear differentiation between threshold and non-threshold carcinogens should be made by all expert groups and regulatory bodies dealing with carcinogen classification and risk assessment .
RNA-Seq has identified more DEGs and provided a wider quantitative range of expression level changes than conventional DNA microarrays. Because of its wider dynamic range as well as its ability to identify a larger number of DEGs, RNA-Seq may generate more insight into mechanisms of toxicity and mode of action (MOA) . In this context, the successful development of a short-term in vivo assay in rodents for discriminating GTCs, NGTCs, and non-carcinogens (NCs) using RNA-Seq would be valuable.
Only a few papers have been published on discriminating GTCs from NGTCs using RNA-Seq in vivo [8, 9]. Therefore, this review also includes data on discriminating GTCs, NGTCs, and NHCs using DNA microarray and qPCR [36,37,38,39,40,41,42,43,44,45,46,47], as these data would be helpful in creating a toxicogenomics database. This review also incorporates recent reports on whole-transcriptome RNA-Seq on animals in vivo, in the liver, kidney, and other organs, although reports did not include the discrimination of GTCs from NGTCs [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31].
In this manuscript, we introduce candidate marker genes published using RNA-Seq, qPCR, and DNA microarray to develop RNA-Seq to discriminate GTCs, NGTCs, and NCs among the chemicals to which humans are exposed in daily life.
Discrimination of GTHCs and NGTHCs and/or NGTNHCs using DNA microarray and qPCR in vivo
In the early days of toxicogenomics research, Ellinger-Ziegelbauer et al. reported DEGs in rat liver upon exposure to 4 GTHCs [ 2-nitrofluorene (2NF), dimethylnitrosamine (DMN), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), and aflatoxin B1 (AFB1)] and 4 NGTHCs [methapyrilene (MPy), diethylstilbestrol (DES), Wy-14643, and piperonylbutoxide (PBO)] for 1–14 days using DNA microarray and the support vector machine (SVM) algorithm as a statistical analysis [36,37,38]. They presented marker genes, such as Cdkn1a, Ccng1, and Mgmt for GTHCs and Apex1, Pcna, Cdk1, Ccnb1, Rps27, Hspd1, and Hspa9 for NGTHCs, whose expression was characteristically changed upon exposure to these carcinogens .
In the form of collaborative studies of the Toxicogenomics/The Japanese Environmental Mutagen Society ·Mammalian Mutagenicity Study Group (JEMS·MMS), Furihata et al. conducted research to discriminate GTHCs from NGTHCs and/or NGTNHCs using rodent liver [3, 39,40,41,42,43]. They selected 50 candidate marker genes and Gapdh as a control gene for normalization based on their preliminary results with nine chemicals using an original DNA microarray and Affymetrix GeneChip Mu74AV2. They reported the dose-dependent changes of expression determined by qPCR at 4 h and 28 days for 50 genes in the liver of mice treated with a single dose of two N-nitroso GTHCs, diethylnitrosamine (DEN) and ethylnitrosourea (ENU), as shown in Fig. 1 . Next, they studied the effects of 12 chemicals on mouse liver at 4 and 48 h after their single dosing and successfully discriminated eight GTHCs [2-acetylaminofluorene (2AAF), 2,4-diaminotoluene, diisopropanolnitrosamine, 4-dimethylaminoazobenzene, NNK, N-nitrosomorpholine, quinoline, and urethane] from four NGTHCs [1,4-dichlorobenzene, dichlorodiphenyltrichloroethane, di(2-ethylhexyl)phthalate (DEHP), and furan] using qPCR and PCA, as shown in Fig. 2 . They also identified by qPCR that 4 and 48 h after administration were key time points from the time-dependent changes in gene expression during the acute phase (4 to 48 h) following the administration of chrysene . Additionally, in rat liver, they successfully discriminated two GTHCs (DEN and 2,6-dinitrotoluene) from an NGTHC (DEHP) and an NGTNHC (phenacetin) at 4 and 48 h, as shown in Fig. 3 . They then proposed 12 candidate marker genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Gdf15, Lrp1, Mbd1, Phlda3, Plk2, and Tubb4b) (JEMS/MMS marker genes) to discriminate GTHCs and NGTHCs and/or NGTNHCs. Subsequent gene pathway analysis on these genes by Ingenuity Pathway Analysis indicated that they are particularly involved in the DNA damage response, resulting from the signal transduction of a p53-class mediator leading to the induction of apoptosis. These studies suggest that the application of PCA to the gene expression profile in rodent liver during the acute phase is useful for predicting that a chemical is a GTHC rather than an NGTHC and/or an NGTNHC [41, 43].
U.S. Environmental Protection Agency (EPA) studied DEGs induced by 4 known GTHCs: 2NF, AFB1, NNK, and DMN in rat liver and proposed 7 biomarker genes, Bax, Bcmp1, Btg2, Ccng1, Cdkn1a, Cgr19, and Mgmt for GTHCs . Four genes, Bax, Btg2, Ccng1, and Cdkn1a were also proposed as GTHC-associated DEGs by Furihata et al. [41, 43].
Park et al. studied DEGs induced by 2 GTHCs (2AAF and DEN), 1 GTC, melphalan, and 1 NGTNC, 1-naphthylisothiocynate in rasH2 mouse liver upon repeated administrations for 7- and 91- days using DNA microarray and qPCR and presented the results in a heatmap. They selected 68 significantly deregulated genes that represented a GTHC-specific signature; these genes were commonly deregulated in both the 2AAF- and DEN-treated rasH2 mice, namely, 52 up-regulated genes, including Aen, Bax, Btg2, Ccng1, Cdkn1a, Ddit4l, Plk2, Mdm2, Phlda3, and Tubb4b as also proposed as GTHC-associated DEGs upon exposure to DEN and 2AAF by Furihata et al. [41, 43], and 16 down-regulated genes, .
Kossler et al. examined a total of 13 chemicals, including 3 known GTHCs: (C.I. Direct Black 38, DMN, and 4,4’-methylenedianiline), 3 NGTHCs: (1,4-dichlorobenzene, phenobarbital sodium, and piperonyl butoxide), 4 NHCs (medical drugs;): cefuroxime sodium, nifedipine, prazosin hydrochloride, and propranolol hydrochloride), and 3 chemicals exhibiting ambiguous results in genotoxicity testing: (cyproterone acetate, thioacetamide, and Wy-14643), in CD-1 mouse liver after their oral administration for 3 and 14 days. They proposed 51 marker candidate genes for differentiating GTHCs from NGTHCs and NHCs (Table 1) and 58 marker candidate genes for differentiating NGTHCs from GTHCs and NHCs (Table 2) in mouse liver, as examined with DNA microarray, in the course of the IMI MARCAR (Innovative Medicines Initiative/Biomarkers and molecular tumor classification for non-genotoxic carcinogenesis) project, involving a European consortium of partners in EFPIA “a research-based pharmaceutical industry operation in Europe” and academics . Using two-step heatmaps, they suggested successfully discriminating GTHCs, NTHCs, and NHCs.
Discrimination of GTHCs and NGTHCs and/or NGTNHCs in public DNA microarray data by PCA
Furihata and Suzuki analyzed in vivo rat data from the public DNA microarray data, in Open TG-GATEs [(Database Description—Open TG-GATEs | LSDB Archive (biosciencedbc.jp)] with the 12 mouse marker genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Gdf15, Lrp1, Mbd1, Phlda3, Plk2, and Tubb4b) (JEMS/MMS marker genes) . They analyzed the data associated with exposure to a total of 23 chemicals: 5 typical rat GTHCs (2AAF, AFB1, 2-nitrofluorene, DEN, and N-nitrosomorpholine), 7 typical rat NGTHCs (clofibrate, ethanol, fenofibrate, gemfibrozil, hexachlorobenzene, phenobarbital, and WY-14643), and also 11 NGTNHCs (allyl alcohol, aspirin, caffeine, chlorpheniramine, chlorpropamide, dexamethasone, diazepam, indomethacin, phenylbutazone, theophylline, and tolbutamide) from Open TG-GATEs. The analysis was performed 3, 6, 9, and 24 h after a single administration and 4, 8, 15, and 29 days after repeated administrations. Genes that were differentially expressed in a dose-dependent manner that was specific to GTHCs were observed, and their significance was assessed using the Williams test during 3–24 h and 4–29 days. PCA successfully discriminated GTHCs from NGTHCs and NGTNHCs at 24 h and 29 days, as shown in Fig. 4 . The results demonstrated that 12 previously proposed mouse marker genes (JEMS/MMS marker genes) are useful for discriminating rat GTHCs from NGTHCs and NGTNHCs.
In another study, Kanki et al. studied 13 NGTHCs with various MOA from OPEN TG-GATEs (28 days) and selected 42 genes that were up-regulated and 8 that were down-regulated upon exposure to them . However, none of them coincided with the 55 genes associated with NGTHCs exposure proposed by Kossler et al. . It is considered that the reason for this discrepancy is that NGTHCs were compared only with the control but not with GTHCs in the study .
Discrimination of GTHCs from NGTHCs using RNA-Seq in short-term in vivo test
Furihata et al. used intact RNA derived from freshly frozen rat liver tissues after 4 weeks of the feeding of chemicals in the water or the food . Using targeted RNA-Seq with specific primers for 12 candidate marker genes (JEMS/MMS marker genes) previously proposed by Furihata and Suzuki  and sample-specific sequence tags, they evaluated the rat hepatocarcinogen 1,4-dioxane (DO) with ambiguous genotoxicity compared with typical GTHCs, DEN and 3,3-dimethylbenzidine·2HCl (DMB), and an NGTHC, DEHP. Gene expression profiles of the 12 genes under DO treatment differed significantly from those with DEN and DMB, as well as DEHP. Finally, PCA successfully differentiated GTHCs from DEHP and DO using these 11 genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Lrp1, Mbd1, Phlda3, Plk2, and Tubb4b), as shown in Figs. 5 and 6 . The present results suggest that RNA-Seq and PCA are useful for differentiating typical GTHCs and typical NGTHCs in the rat.
Discrimination of a GTHC from an NGTHC using RNA-Seq with formalin-fixed paraffin-embedded (FFPE) samples
Furihata et al. used RNA-Seq with FFPE samples from rat liver tissues after 4 weeks of the feeding of chemicals in the water or the food . Specifically, targeted RNA-Seq was applied to FFPE samples to analyze 12 genes (JEMS/MMS marker genes) as potential markers for rat responses to GTHCs and NGTHCs, with the comparison between a typical GTHC, 2AAF, and p-cresidine (CRE), the genotoxicity of which is ambiguous. 2AAF induced remarkable differences in the expression of eight genes (Aen, Bax, Btg2, Ccng1, Gdf15, Mbd1, Phlda3, and Tubb4b) from that in the control group, while CRE only induced expression changes in Gdf15, as shown by Tukey’s test. Meanwhile, gene expression profiles for nine genes (Aen, Bax, Btg2, Ccng1, Cdkn1a, Gdf15, Mbd1, Phlda3, and Plk2) differed between samples treated with 2AAF and CRE. Finally, PCA of 12 genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Gdf15, Lrp1, Mbd1, Phlda3, Plk2, and Tubb4b) (JEMS·MMS marker genes) using our previous Open TG-GATE data  plus 2AAF and CRE successfully differentiated 2AAF, as a GTHC, from CRE, as an NGHTC (Fig. 7) . It was thus concluded that targeted RNA-Seq on FFPE samples and PCA are useful for evaluating a typical rat GTHC and an NGTHC.
Recent whole-transcriptome RNA-Seq reports on in vivo analyses in animal liver, kidney, and other organs
Various whole-transcriptome RNA-Seq studies on the effects of hepatocarcinogens in rodent liver have been reported [5, 6, 10,11,12,13,14,15,16,17,18, 23], although they did not examine the discrimination of GTHCs from NGTHCs.
Li et al. examined rat livers treated with a GTHC, AFB1, for 5 days and analyzed the effects using RNA-Seq, TempO-Seq, DNA microarray, and qPCR. They showed that RNA-Seq revealed toxicological insights from pathway enrichment, with overall higher statistical power compared with TempO-seq and DNA microarray. They detected 862 DEGs (491 up-regulated and 371 down-regulated by AFB1) in HiSeq2000 and confirmed 11 up-regulated genes (Ccnb1, Cenpw, G6pd, Nt5dc2, Pttg1, Spp1, Stmn1, Tacc3, Tk1, Ube2c, and Ube2t) by qPCR .
In another study, Nault et al. examined an NGTHC, acetamide, in rat liver after treatment for 7 and 28 days. They showed the DEGs results using heatmaps. They reported 9 up-regulated genes: (E2f4, Ar, Mybl1, Kdm6a, Sox2, Mycn, Sry, Mybl2, and EF1) and 10 down-regulated ones: (Esr1, Rxr, Ppara, LXRalpha, Pparg, Cebpa, Egr1, Cebpb, Foxo1, and Foxp1). Additionally, they wrote complex increase/decrease in the following genes Hebp2, Acot1, Ifit1, Cenpw, Chek2, Parpbp, Cyp17a1, Slc7a1, and Prom1 in the paper .
Elsewhere, Gong et al. reported that the US FDA-led SEQC (i.e., MAQC-III) project conducted a comprehensive study focusing on the transcriptome profiling of rat liver samples treated with 27 chemicals with various MOA for 3 to 7 days to evaluate the utility of RNA-Seq in safety assessment and elucidating the mechanism of toxicity .
Moreover, Bushel et al. examined the effects of treatment with 15 chemicals with various MOA for 3 to 7 days in rat liver and presented the data obtained by DNA microarray, RNA-Seq, and Tempo-Seq in a heatmap .
Li et al. studied the effects of a carcinogenic dose of aristolochic acid for 12 weeks in rat kidney.
Four thousand fifty one up-regulated and 2743 down-regulated mRNAs were observed and 43 up-regulated and 20 down-regulated miRNAs were observed as measured by PCA and hierarchical clustering analysis .
Israel et al. reported DEGs induced by a GTC, 1,3-butadiene, in mouse liver, lung, and kidney for 2 weeks. They performed RNA-Seq, identification of accessible chromatin (ATAC-seq), and characterization of regions with histone modifications associated with active transcription (ChIP-seq for acetylation at histone 3 lysine 27, H3K27ac). Most changes were restricted to lung tissue. The results were shown in heatmaps. They showed that the DEGs were involved in Phase I metabolism (58 Cyp family members and 12 others), Phase II metabolism (58 genes), and IFNγ signaling (75 genes) .
Additionally, Felley-Bosco and Rehrauer reported RNA-Seq data from asbestos-exposed mice. In that study, an asbestos suspension was injected every 3 weeks for eight rounds and an examination was performed 33 weeks after the first injection. They performed data mining of publicly available datasets to evaluate how noncoding RNA contributes to mesothelioma heterogeneity. Nine noncoding RNAs (Fendrr, Gm26902, Gm17501, Meg3, miR 17–92 cluster, Dubr, and Firre) were specifically elevated in mesothelioma tumors and shown to contribute to the heterogeneity of human mesothelioma. Because some of these RNAs have known oncogenic properties, this study supported the concept that noncoding RNAs can act as cancer progenitor genes .
Guo et al. reported mechanisms of mouse colitis-accelerated colon carcinogenesis induced by azoxymethane/dextran sulfate sodium treatment for 22 weeks. The 10 most up-regulated genes in tumors were Alb, Alox15, Clca4, Cxcl6, Lyz, Mmp7, Mmp10, Pnliprp 1, Slc30a2, and Wif1, while the 10 most down-regulated ones were Ca3, Chrna3, Folh1, Nos1, Pln, Retnlb, Sst, Stmn3, Sycn, and Zcchc12 .
Asahina et al. reported that alcohol intake for 5 months induced pancreatic ductal adenocarcinoma in Pdx1 Cre; LSL-Kras G12D mutant mice. Whole RNA-seq analysis revealed that the consumption of alcohol increased the expression of markers for tumors (Epcam, Krt19, Prom1, Wt1, and Wwtr1), stroma (Dcn, Fn1, and Tnc), and cytokines (Tgfb1 and Tnf) and decreased the expression of Fgf21 and Il6 in the pancreatic tumor tissues .
Kinaret et al.  asserted that, although the advent of high-throughput hybridization-based technologies, such as DNA microarrays, significantly boosted the generation of large-scale gene expression profiles, recent advances in sequencing technologies further improved such capability. For instance, RNA-Seq allows the detection of gene expression with an increased dynamic range, solving the problem of probe saturation for highly expressed transcripts. Furthermore, RNA-Seq does not need a priori knowledge about the genomic sequence of the studied organism and does not suffer from the above-mentioned cross-hybridization events, especially in the analysis of complex genomes. As a consequence, RNA-Seq allows de novo transcript discovery to be performed to identify unannotated transcripts and characterize new transcripts generated by alternative splicing. However, an appropriate analytical plan should be made to avoid or mitigate certain biases that could occur during the data management and analysis. For instance, previous works  showed that standard normalization procedures can affect the sensitivity of differential expression analysis, reflecting the behavior of a relatively small number of either high-count or ubiquitous genes. RNA-Seq typically produces larger and more complex data, which require more time and more sophisticated analytical approaches, than in DNA microarray experiments, for example. Although transcriptome profiling is increasingly being employed in toxicogenomic experiments, the analytical pipelines are still far from being standardized. To date, no benchmark of the optimal analytical procedures in transcriptome profiling in toxicogenomic experiments has been formulated. Recently, the reduction of the cost of analyzing a single transcriptome made the accomplishment of large-scale studies possible, which have been carried out by international programs such as CMAP, TOX21, and LINCS1000 .
Comparing RNA-Seq with qPCR and DNA microarray, RNA-Seq is reflecting the absolute amount of RNA expression more directly than others as read counts. The reliability of the results can be confirmed by sequence without a disturbance of mismatch in probes or primers and is applicable for alternative splicing. The qPCR method is easy to perform and does not require advanced experience but is applicable only after the selection of target genes. It is not a comprehensive method compared to total RNA-seq or DNA microarray. The DNA microarray methods require many steps and skills and have more variances among different platforms. The reliability of the results is slightly lower than the other two methods. The major results should often be confirmed by qPCR. From the analysis of previous DNA microarray papers, we have learned that the marker genes differ depending on the type of chemicals studied. The marker genes in previous DNA microarray papers do not always match. It would be useful to examine published DNA microarray papers to identify candidate marker genes, and it would be useful to accumulate RNA-Seq (whole) data, which is more reliable than DNA microarray, to converge the marker genes. This requires easy-to-use bioinformatics.
Kinaret et al.  introduced the following public data.
Connectivity Map (CMAP, Connectivity Map, Broad Institute) ,
LINCS 1000 NIH LINCS Program (lincsproject.org) ,
DrugMatrix (norecopa.no) ,
Open TG-GATEs (LSDB Archive; biosciencedbc.jp) ,
ArrayExpress (EMBL-EBI) , and
The qPCR has been used as an efficient screening method after narrowing down biomarker genes by comprehensive analysis using DNA microarray. Similarly, “targeted” RNA-Seq, in which specific PCR primers are designed to amplify only selected gene transcripts, can be used. In “targeted” RNA-Seq, the unique sequencing tag allows a large number of samples to be mixed and sequenced at the same time, making it a simpler and more cost-effective method than qPCR. To increase the efficiency of the analysis, it is recommended to combine genes with similar expression levels for “targeted” RNA-Seq .
The next newly established technology for RNA-Seq is single-cell RNA-Seq (scRNA-Seq). The scRNA-Seq pipeline has emerged as a valuable tool for uncovering individual cellular functions in thousands to millions of cells, an advancement over the bulk RNA-seq method of averaging gene expression across all cells in a tissue . However, to the best of our knowledge, scRNA-Seq has yet to be applied to toxicogenomics, including to the discrimination of GTCs, NGTCs, and NCs.
When discussing the proposed candidate genes that can act as markers of GTHCs and NGTHCs in RNA-Seq, DNA microarray, and qPCR data on samples from rodent liver, they are not always consistent among different published papers [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31, 36,37,38,39,40,41,42,43,44,45,46,47]. For example, JEMS·MMS·Toxicogenomics group proposed 12 candidate genes (Aen, Bax, Btg2, Ccnf, Ccng1, Cdkn1a, Gdf15, Lrp1, Mbd1, Phlda3, Plk2, and Tubb4b) to discriminate GTHCs from NGTHCs and NGTNHCs by PCA from analyses of mouse liver , rat liver , public DNA microarray data (OPEN TG-GATEs) , RNA-Seq , and RNA-Seq on FFPE samples  upon 4 h to 28 days of treatment with a total of 35 chemicals (15 GTHCs, 9 NGTHCs, and 11 NGTNHCs). Meanwhile, Kossler et al. proposed 51 marker candidate genes that could differentiate GTHCs from NGTHCs and NHCs (Table 1) and 58 marker candidate genes that could differentiate NGTHCs from GTHCs and NHCs (Table 2) in mouse liver examined by DNA microarray. They examined a total of 13 chemicals (3 GTHCs, 6 NGTHCs, and 4 NHCs) in mouse liver after treatment for 3 and 14 days . They proposed 15 genes involved in the DNA damage response, four of which (Bax, Ccng1, Ddit4l, and Phlda3) overlapped with those in the studies of JEMS·MMS·Toxicogenomics group. However, Kossler et al. examined three GTHCs that differed from the 15 GTHCs examined by JEMS·MMS·Toxicogenomics group. Moreover, Park et al. presented significantly deregulated genes in rasH2 mouse liver upon treatment with DEN and 2AAF; there were 47 upregulated genes, including Aen, Bax, Btg2, Ccng1, Cdkn1a, Ddit4l, Plk2, Mdm2, Phlda3, and Tubb4b, which were also proposed by JEMS·MMS·Toxicogenomics group, and 11 downregulated genes . JEMS·MMS·Toxicogenomics group also studied DEN and 2AAF. Furthermore, Jonker et al. reported the discrimination of 2 GTCs, 2NGTCs and 2 NGTNCs in the liver of both wild-type and DNA repair-deficient Xpa2/2/p531/2 (Xpa/p53) mice using DNA microarray and heatmap . However, their candidate genes differed from those in other published papers. Finally, Li et al. examined rat liver upon treatment with a GTHC, AFB1, for 5 days and performed analyses using RNA-Seq, TempO-Seq, DNA microarray, and qPCR. They proposed 11 completely different marker genes in other published papers . Given these conflicting findings, it should be useful to reselect or validate genes from all available databases to discriminate GTCs, NGTCs, and NGTNCs.
In connection with restrictions on animal testing, “OECD Guidelines for the Testing of Chemicals, [Repeated Dose 28-Day Oral Toxicity Study in Rodents (OECD TG 407)] [Test No. 407: Repeated Dose 28-day Oral Toxicity Study in Rodents | READ online (oecd-ilibrary.org)] is still valid for testing chemical toxicity. This assay determines the general toxicity of chemicals in rodents after 28 days of oral dosing (e.g., effects on the liver, kidneys, heart, and lungs). Despite restrictions being placed on animal testing, this test will continue to be applied. We can use the animal organs from the test collaboratively and use the samples, which would reduce the number of experimental animals used.
In toxicogenomic experiments, there are protocol issues to be considered, such as the method and number of administered doses, dose setting, and timing of observation. As yet, no consensus has been reached on the optimal settings for these variables. Therefore, it would be beneficial to adjust the strategy according to each study to find the best protocol, but also to adjust settings to match previous studies, such as using a 28-day repeated dosing test in rats. Regarding the future direction of toxicogenomics concerning the 3Rs concept, we also propose incorporating not only toxicogenomics but also other genotoxicity assays (e.g., micronucleus test, error-corrected sequencing, comet assay, DNA adduct analysis) into 28-day repeated dosing study in rats to enable a reduction in the number of animals used by applying multi-endpoint assays.
Targeted RNA-Seq requires only a few hundred base pairs for sequencing, which enables the use of RNA from FFPE samples. A large number of FFPE samples from pathological examinations in previous studies are available, including those from 2-year rodent bioassays for carcinogenicity. The examination of stored FFPE samples would enable the establishment of substantial expression data with information on toxicological endpoints such as carcinogenicity . The construction of a large database with data on a large set of genotoxic carcinogens would improve the efficiency and reliability of biomarker genes for discriminating such compounds.
There is a growing need to develop alternative in vivo methods to the 2-year rodent bioassay to assess the carcinogenicity of environmental chemicals. Toxicogenomics, including recent RNA-Seq and previous qPCR and DNA microarray, has been studied for its potential as a short-term in vivo alternative to long-term animal studies. RNA-Seq has identified more DEGs and provided a wider quantitative range of expression level changes than conventional DNA microarrays. JEMS·MMS·Toxicogenomics group successfully discriminated GTHCs from NGTHCs and/or NGTNHCs in rat and mouse liver by 12 marker genes using targeted RNA-Seq, RNA-Seq on FFPE samples, qPCR, and DNA microarray with PCA as a statistical approach. The 12 marker genes were re-validated by public DNA microarray data (OPEN TG-GATEs). EPA studied DEGs induced by 4 known GTHCs in rat liver using DNA microarray and proposed 7 biomarker genes, four of which (Bax, Btg2, Ccng1, and Cdkn1a) overlapped with those of JEMS/MMS 12 genes. Candidate genes published using RNA-Seq, qPCR, and DNA microarray will be useful for the future development of short-term in vivo studies of environmental carcinogens using RNA-Seq. In connection with the restrictions on animal testing and the 3Rs concept, it would be beneficial to adjust settings to match a 28-day repeated dosing test in rats rather than seeking the best protocol for toxicogenomics.
Availability of data and materials
Differentially expressed protein-coding genes
European Federation of Pharmaceutical Industries and Associations
Innovative Medicines Initiative
The Japanese Environmental Mutagen Society ·Mammalian Mutagenicity Study Group
Biomarkers and molecular tumor classification for non-genotoxic carcinogenesis
Mode of action
Organization for Economic Co-operation and Development
Quantitative real-time PCR
Principal component analysis
Next-generation RNA sequencing
Support vector machine
- AAF, 2AAF:
- Acot1 :
Acyl-CoA thioesterase 1
- Acot9 :
Acyl-CoA thioesterase 9
- Aen :
Apoptosis enhancing nuclease
- Akap13 :
A-kinase anchoring protein 13
- Akr1d1 :
Aldo-keto reductase family 1 member D1
- Alb :
- Alox15 :
- Apex1 :
Apurinic/apyrimidinic endodeoxyribonuclease 1
- Ar :
- Armt1 :
Acidic residue methyltransferase 1
- Atad2 :
ATPase family AAA domain containing 2
- Atosa :
Atos homolog A
- Atp6v1d :
ATPase, H + transporting lysosomal V1 subunit D
- Atxn10 :
- Bax :
BCL2 associated X, apoptosis regulator
- Bcl2a1 :
BCL2 related protein A1
- Bcor :
BCL6 interacting corepressor
- Btg2 :
BTG anti-proliferation factor 2
- Ca3 :
Carbonic anhydrase 3
- Camkk2 :
Calcium/calmodulin-dependent protein kinase kinase 2
- Ccdc80 :
Coiled-coil domain containing 80
- Ccnb1 :
- Ccnf :
- Ccng1 :
- Ccr2 :
C-C motif chemokine receptor 2
- Cd34 :
- Cdk1 :
Cyclin-dependent kinase 1
- Cdkn1a :
Cyclin-dependent kinase inhibitor 1A
- Cebpa :
CCAAT/enhancer binding protein alpha
- Cebpb :
CCAAT/enhancer binding protein beta
- Chek2 :
Checkpoint kinase 2
- Cenpw :
Centromere protein W
- Ces2a :
- Ces2e :
Cholinergic receptor, nicotinic, alpha polypeptide 3
- Clca4 :
Chloride channel accessory 4
- Coa6 :
Cytochrome c oxidase assembly factor 6
- Col1a2 :
Collagen, type I, alpha 2
- Cox6b2 :
Cytochrome c oxidase subunit 6B2
- Cxcl6 :
C-X-C motif chemokine ligand 6
- Cyp2c65 :
Cytochrome P450, family 2, subfamily c, polypeptide 65
- Cyp17a1 :
Cytochrome P450, family 17, subfamily a, polypeptide 1
- Dcn :
- Ddit4l :
DNA-damage-inducible transcript 4-like
- Dgka :
Diacylglycerol kinase, alpha
- Dleu2 :
Deleted in lymphocytic leukemia, 2
- Dubr :
Dppa2 upstream binding RNA
- E2f4 :
E2F transcription factor 4
- EF1 :
Elongation factor 1-alpha
- Egr1 :
Early growth response 1
- Emp3 :
Epithelial membrane protein 3
- Enc1 :
Ectodermal-neural cortex 1
- Epcam :
Epithelial cell adhesion molecule
- Esr1 :
Estrogen receptor 1
- Exoc4 :
Exocyst complex component 4
- Fam171a1 :
Family with sequence similarity 171, member A1
- Fbn1 :
- Fendrr :
Foxf1 adjacent non-coding developmental regulatory RNA
- Fgf21 :
Fibroblast growth factor 21
- Fgl1 :
Fibrinogen-like protein 1
- Fgl2 :
Fibrinogen-like protein 2
- Firre :
Functional intergenic repeating RNA element
- Fn1 :
Fibronectin type III domain containing 5
- Folh1 :
Folate hydrolase 1
- Foxo1 :
Forkhead box O1
- Foxp1 :
Forkhead box P1
- Fstl1 :
- G6pd :
- G6pdx :
Glucose-6-phosphate dehydrogenase X-linked
- Gapdh :
- Gdf15 :
Growth differentiation factor 15
- Ggta1 :
Glycoprotein alpha-galactosyltransferase 1
- Ginm1 :
Glycoprotein integral membrane 1
- Gm2011 :
Predicted gene 2011
- Gm10419 :
Predicted gene 10419
- Gm17501 :
Predicted gene, 17,501
G protein subunit alpha transducin 1
- Grhl1 :
Grainyhead like transcription factor 1
- Grk3 :
G protein-coupled receptor kinase 3
- Gstm1 :
Glutathione S-transferase, mu 1
- Gstp3 :
Glutathione S-transferase pi 3
- Gtf2b :
General transcription factor IIB
- H2-Dma :
Histocompatibility 2, class II, locus Dma
- H2-DMb2 :
Histocompatibility 2, class II, locus Mb2
- Hebp2 :
Heme binding protein 2
- Hip1r :
Huntingtin interacting protein 1 related
- Hnf4aos :
Hepatic nuclear factor 4 alpha, opposite strand
- Hspd1 :
Heat shock protein family D (Hsp60) member 1
- Hspa9 :
Heat shock protein family A (Hsp70) member 9
- Ifit1 :
Interferon-induced protein with tetratricopeptide repeats 1
- Il6 :
- Iqgap1 :
IQ motif containing GTPase activating protein 1
- Kdm6a :
Lysine demethylase 6A
- Krt19 :
- Loxl2 :
Lysyl oxidase-like 2
- Lck :
Lymphocyte protein tyrosine kinase
- Lrp1 :
LDL receptor related protein 1
- Ltn1 :
Listerin E3 ubiquitin protein ligase 1
- lxr :
LexA regulated function (Escherichia phage P1)
- Lyz1 :
- Map3k20 :
mitogen-activated protein kinase kinase kinase 20
- Mbd1 :
Methyl-CpG binding domain protein 1
- Mbl2 :
Mannose-binding lectin (protein C) 2
- Mcm5 :
Minichromosome maintenance complex component 5
- Meg3 :
Maternally expressed 3
- Mgmt :
- miR 17–92 cluster :
MicroRNA 17–92 cluster
- Mmp7 :
Matrix metallopeptidase 7
- Mmp10 :
Matrix metallopeptidase 10
- Moxd1 :
Monooxygenase, DBH-like 1
MYB proto-oncogene like 1
- Mybl2 :
MYB proto-oncogene like 2
MYCN proto-oncogene, bHLH transcription factor
Non-SMC condensin II complex, subunit G2
- Nebl :
- Nisch :
- Nolc1 :
Nucleolar and coiled-body phosphoprotein 1
- Nos1 :
Nitric oxide synthase 1
- Nr2c2 :
Nuclear receptor subfamily 2, group C, member 2
- Nsl1 :
NSL1, MIS12 kinetochore complex component
- Nt5dc2 :
5'-Nucleotidase domain containing 2
- Parpbp :
PARP1 binding protein
- Pcna :
Proliferating cell nuclear antigen
- Pgap2 :
Post-GPI attachment to proteins 2
- Phlda3 :
Pleckstrin homology-like domain, family A, member 3
- Pierce1 :
Piercer of microtubule wall 1
- Pkp2 :
- Plaat3 :
Phospholipase A and acyltransferase 3
- Plekha2 :
Pleckstrin homology domain-containing, family A (phosphoinositide binding specific) member 2
- Plk2 :
Polo-like kinase 2
- Pln :
- Pnliprp 1 :
Pancreatic lipase related protein 1
- Pola1 :
Polymerase (DNA directed), alpha 1
- Ppara :
Peroxisome proliferator activated receptor alpha
- Pparg :
Peroxisome proliferator-activated receptor gamma
- Pqlc3 :
PQ loop repeat containing
- Prkd2 :
Protein kinase D2
- Prdm15 :
PR domain containing 15
- Prom1 :
- Pttg1 :
PTTG1 regulator of sister chromatid separation, securin
- Rasal2 :
RAS protein activator like 2
- Retnlb :
Resistin like beta
- Rorc :
RAR-related orphan receptor gamma
- Rps27 :
Ribosomal protein S27
- Samd4a :
Sterile alpha motif domain containing 4A
- Siva1 :
SIVA1, apoptosis-inducing factor
- Slc7a1 :
Solute carrier family 7 member 1
- Slc25a32 :
Solute carrier family 25, member 32
- Slc30a2 :
Solute carrier family 30 (zinc transporter), member 2
- Snx6 :
Sorting nexin 6
- Sox2 :
SRY-box transcription factor 2
- Spp1 :
Secreted phosphoprotein 1
- Srprb :
Signal recognition particle receptor, B subunit
- Sry :
Sex determining region Y
- Sst :
- Stmn1 :
- Stmn3 :
- Sycn :
- Tacc3 :
Transforming, acidic coiled-coil containing protein 3
- Tagln2 :
- Tead1 :
TEA domain family member 1
- Tgfb1 :
Transforming growth factor, beta 1
- Tk1 :
Thymidine kinase 1
- Tmem98 :
Transmembrane protein 98
- Tmem181c-ps :
Transmembrane protein 181C, pseudogene
- Tmem268 :
Transmembrane protein 268
- Tmsb10 :
Thymosin, beta 10
- Tnf :
Tumor necrosis factor
- Tnfrsf1b :
Tumor necrosis factor receptor superfamily, member 1b
- Top2a :
Topoisomerase (DNA) II alpha
- Tspan13 :
- Tuba1a :
Tubulin, alpha 1A
- Tubb4b :
Tubulin, beta 4B class Ivb
- Tulp2 :
Tubby-like protein 2
- Ube2c :
Ubiquitin-conjugating enzyme E2C
- Ube2t :
Ubiquitin-conjugating enzyme E2T
- Wif1 :
Wnt inhibitory factor 1
- Wt1 :
WT1 transcription factor
- Wwtr1 :
WW domain containing transcription regulator 1
- Zcchc12 :
Zinc finger, CCHC domain containing 12
- Zdhhc14 :
Zinc finger, DHHC-type palmitoyltransferase 14
- Zeb2 :
Zinc finger E-box binding homeobox 2
- Zfp54 :
Zinc finger protein 54
Zinc finger protein 472
- Zfp750 :
Zinc finger protein 750
- Zfp958 :
Zinc finger protein 958
- Zkscan14 :
Zinc finger with KRAB and SCAN domains 14
Lovett RA. Toxicogenomics. Toxicologists brace for genomics revolution Science. 2000;289(5479):536–7.
Fabian G, Farago N, Feher LZ, Nagy LI, Kulin S, Kitajka K, Bito T, Tubak V, Katona RL, Tiszlavicz L, Puskas LG. High-density real-time PCR-based in vivo toxicogenomic screen to predict organ-specific toxicity. Int J Mol Sci. 2011;12:6116–34.
Furihata C, Watanabe T, Suzuki T, Hamada S, Nakajima M. Collaborative studies in toxicogenomics in rodent liver in JEMS·MMS; a useful application of principal component analysis on toxicogenomics. Genes Environ. 2016;38:15.
Walton K, O’Connor BP. Optimized methodology for the generation of RNA-sequencing libraries from low-input starting material: enabling analysis of specialized cell types and clinical samples. Methods Mol Biol. 2018;1706:175–98.
Wang C, Gong B, Bushel PR, Thierry-Mieg J, Thierry-Mieg D, Xu J, Fang H, Hong H, Shen J, Su Z, Meehan J, Li X, Yang L, Li H, Łabaj PP Toxicol, Kreil DP, Megherbi D, Gaj S, Caiment F, van Delft J, Kleinjans J, Scherer A, Devanarayan V, Wang J, Yang Y, Qian HR, Lancashire LJ, Bessarabova M, Nikolsky Y, Furlanello C, Chierici M, Albanese D, Jurman G, Riccadonna S, Filosi M, Visintainer R, Zhang KK, Li J, Hsieh JH, Svoboda DL, Fuscoe JC, Deng Y, Shi L, Paules RS, Auerbach SS, Tong W. The concordance between RNA-seq and microarray data depends on chemical treatment and transcript abundance. Nat Biotechnol. 2014;32:926–32.
Rao MS, Van Vleet TR, Ciurlionis R, Buck WR, Mittelstadt SW, Blomme EAG, Liguori MJ. Comparison of RNA-seq and microarray gene expression platforms for the toxicogenomic evaluation of liver from short-term rat toxicity studies. Front Genet. 2019;9:636.
McHale CM, Zhang L, Thomas R, Smith MT. Analysis of the transcriptome in molecular epidemiology studies. Environ Mol Mutagen. 2013;54:500–17.
Furihata C, Toyoda T, Ogawa K, Suzuki T. Using RNA-Seq with 11 marker genes to evaluate 1,4-dioxane compared with typical genotoxic and non-genotoxic rat hepatocarcinogens. Mutat Res Genet Toxicol Environ Mutagen. 2018;834:51–5.
Furihata C, You X, Toyoda T, Ogawa K, Suzuki T. Using FFPE RNA-Seq with 12 marker genes to evaluate genotoxic and non-genotoxic rat hepatocarcinogens. Genes Environ. 2020;42:15.
Li D, Gong B, Xu J, Ning B, Tong W. Impact of sequencing depth and library preparation on toxicological interpretation of RNA-seq data in a “three-sample” scenario. Chem Res Toxicol. 2021;34:529–40.
Nault R, Bals B, Teymouri F, Black MB, Andersen ME, McMullen PD, Krishnan S, Kuravadi N, Paul N, Kumar S, Kannan K, Jayachandra KC, Alagappan L, Patel BD, Bogen KT, Gollapudi BB, Klaunig JE, Zacharewski TR, Venkataraman Bringi V. A toxicogenomic approach for the risk assessment of the food contaminant acetamide. Toxicol Appl Pharmacol. 2020;388.
Gong B, Wang C, Su Z, Hong H, Thierry-Mieg J, Thierry-Mieg D, Shi L, Auerbach SS, Tong W, Xu J. Transcriptomic profiling of rat liver samples in a comprehensive study design by RNA-Seq. Sci Data. 2014;1: 140021.
Bushel PR, Paules RS, Auerbach SS. A comparison of the TempO-Seq S1500+ Platform to RNA-Seq and microarray using rat liver mode of action samples. Front Genet. 2018;9:485.
Merrick BA, Phadke DP, Auerbach SS, Mav D, Stiegelmeyer SM, Shah RR, Tice RR. RNA-Seq profiling reveals novel hepatic gene expression pattern in aflatoxin B1 treated rats. PLoS ONE. 2013;8: e61768.
Israel JW, Chappell GA, Simon JM, Pott S, Safi A, Lewis L, Cotney P, Boulos HS, Bodnar W, Lieb JD, Crawford GE, Furey TS, Rusyn I. Tissue- and strain-specific effects of a genotoxic carcinogen 1,3-butadiene on chromatin and transcription. Mamm Genome. 2018;29:153–67.
Zhou D, Hlady RA, Schafer MJ, White TA, Liu C, Choi JH, Miller JD, Roberts LR, LeBrasseur NK, Robertson KD. High fat diet and exercise lead to a disrupted and pathogenic DNA methylome in mouse liver. Epigenetics. 2017;12:55–69.
Schyman P, Printz RL, Estes SK, Boyd KL, Shiota M, Wallqvist A. Identification of the toxicity pathways associated with thioacetamide-induced injuries in rat liver and kidney. Front Pharmacol. 2018;9:1272.
Kurma K, Manches O, Chuffart F, Sturm N, Gharzeddine K, Zhang J, Mercey-Ressejac M, Rousseaux S, Millet A, Lerat H, Marche PN, Macek Jilkova Z, Decaens T. DEN-induced rat model reproduces key features of human hepatocellular carcinoma. Cancers (Basel). 2021;13:4981.
Li Z, Qin T, Wang K, Hackenberg M, Yan J, Gao Y, Yu LR, Shi L, Su Z, Chen T. Integrated microRNA, mRNA, and protein expression profiling reveals microRNA regulatory networks in rat kidney treated with a carcinogenic dose of aristolochic acid. BMC Genomics. 2015;16:365.
Felly-Bosco E, Rehrauer H. Non-coding transcript heterogeneity in mesothelioma: insights from asbestos-exposed mice. Int J Mol Sci. 2018;19:1163.
Guo Y, Wu R, Gaspar JM, Sargsyan D, Su ZY, Zhang C, Gao L, Cheng D, Li W, Wang C, Yin R, Fang M, Verzi MP, Hart RP, Kong Ah-Ng. DNA methylome and transcriptome alterations and cancer prevention by curcumin in colitis-accelerated colon cancer in mice. Carcinogenesis. 2018;39:669–80.
Asahina K, Balog S, Hwang E, Moon E, Wan E, Skrypek K, Chen Y, Fernandez J, Romo J, Yang Q, Lai K, French SW, Tsukamoto H. Moderate alcohol intake promotes pancreatic ductal adenocarcinoma development in mice expressing oncogenic Kras. Am J Physiol Gastrointest Liver Physiol. 2020;318:G265–76.
Merrick BA, Chang JS, Phadke DP, Bostrom MA, Shah RR, Wang X, Gordon O, Wright GM. HAfTs are novel lncRNA transcripts from aflatoxin exposure. PLoS ONE. 2018;13: e0190992.
Schyman P, Printz RL, AbdulHameed MDM, Estes SK, Shiota C, Shiota M, Wallqvist A. A toxicogenomic approach to assess kidney injury induced by mercuric chloride in rats. Toxicology. 2020;442: 152530.
Chikara S, Mamidi S, Sreedasyam A, Chittem K, Pietrofesa R, Zuppa A, Moorthy G, Dyer N, Christofidou-Solomidou M, Reindl KM. Flaxseed consumption inhibits chemically induced lung tumorigenesis and modulates expression of phase II enzymes and inflammatoryc cytokines in A/J mice. Cancer Prev Res (Phila). 2018;11:27–37.
Kim M, Jee SC, Kim S, Hwang KH, Sung JS. Identification and characterization of mRNA biomarkers for sodium cyanide exposure. Toxics. 2021;9:288.
Kawamura T, Yamamoto M, Suzuki K, Suzuki Y, Kamishima M, Sakata M, Kurachi K, Setoh M, Konno H, Takeuchi H. Tenascin-C produced by intestinal myofibroblasts promotes colitis-associated cancer development through angiogenesis. Inflamm Bowel Dis. 2019;25:732–41.
Triff K, McLean MW, Konganti K, Pang J, Callaway E, Zhou B, Ivanov I, Chapkin RS. Assessment of histone tail modifications and transcriptional profiling during colon cancer progression reveals a global decrease in H3K4me3 activity. Biochim Biophys Acta Mol Basis Dis. 2017;1863:1392–402.
Leung YK, Govindarajah V, Cheong A, Veevers J, Song D, Gear R, Zhu X, Ying J, Kendler A, Medvedovic M, Belcher S, Ho SM. Gestational high-fat diet and bisphenol A exposure heightens mammary cancer risk. Endocr Relat Cancer. 2017;24:365–78.
Tang XH, Osei-Sarfo K, Urvalek AM, Zhang T, Scognamiglio T, Gudas LJ. Combination of bexarotene and the retinoid CD1530 reduces murine oral-cavity carcinogenesis induced by the carcinogen 4-nitroquinoline 1-oxide. Proc Natl Acad Sci U S A. 2014;111:8907–12.
Urvalek AM, Osei-Sarfo K, Tang XH, Zhang T, Scognamiglio T, Gudas LJ. Identification of ethanol and 4-nitroquinoline 1-oxide induced epigenetic and oxidative stress markers during oral cavity carcinogenesis. Alcohol Clin Exp Res. 2015;39:1360–72.
Li T, Tong W, Roberts R, Liu Z, Thakkar S. DeepCarc: deep learning-powered carcinogenicity prediction using model-level representation. Front Artif Intell. 2021;4: 757780.
Corvi R, Madia F, Guyton KZ, Kasper P, Rudel R, Colacci A, Kleinjans J, Jennings P. Moving forward in carcinogenicity assessment: report of an EURL ECVAM/ESTIV workshop. Toxicol In Vitro. 2017;45:278–86.
Jacobs MN, Colacci A, Corvi R, Vaccari M, Aguila MC, Corvaro M, Delrue N, Desaulniers D, Ertych N, Jacobs A, Luijten M, Madia F, Nishikawa A, Ogawa K, Ohmori K, Paparella M, Sharma AK, Vasseur P. Chemical carcinogen safety testing: OECD expert group international consensus on the development of an integrated approach for the testing and assessment of chemical non-genotoxic carcinogens. Arch Toxicol. 2020;94:2899–923.
Bevan RJ, Harrison PTC. Threshold and non-threshold chemical carcinogens: a survey of the present regulatory landscape. Regul Toxicol Pharmacol. 2017;88:291–302.
Ellinger-Ziegelbauer H, Stuart B, Wahle B, Bomann W, Ahr HJ. Comparison of the expression profiles induced by genotoxic and nongenotoxic carcinogens in rat liver. Mutat Res. 2005;575:61–84.
Ellinger-Ziegelbauer H, Gmuender H, Bandenburg A, Ahr HJ. Prediction of a carcinogenic potential of rat hepatocarcinogens using toxicogenomics analysis of short-term in vivo studies. Mutat Res. 2008;637:23–39.
Ellinger-Ziegelbauer H, Aubrecht J, Kleinjans JC, Ahr HJ. Application of toxicogenomics to study mechanisms of genotoxicity and carcinogenicity. Toxicol Lett. 2009;186:36–44.
Watanabe T, Tobe K, Nakachi Y, Kondoh Y, Nakajima M, Hamada S, Namiki C, Suzuki T, Madeda S, Tadakuma A, Sakurai M, Arai Y, Hyogo A, Hoshino M, Tashiro T, Ito H, Inazumi H, Sakaki Y, Tashiro H, Futihata C. Differential gene expression induced by two N-nitroso carcinogens, phenobarbital and ethanol in mouse liver examined with oligonucleotide microarray and quantitative real-time PCR. Gene Env. 2007;29:115–27.
Watanabe T, Tanaka G, Hamada S, Namiki C, Suzuki T, Nakajima M, Furihata C. Dose-dependent alterations in gene expression in mouse liver induced by diethylnitrosamine and ethylnitrosourea and determined by quantitative real-time PCR. Mutat Res. 2009;673:9–20.
Watanabe T, Suzuki T, Natsume M, Nakajima M, Narumi K, Hamada S, Sakuma T, Koeda A, Oshida K, Miyamoto Y, Maeda A, Hirayama M, Sanada H, Honda H, Ohyama W, Okada E, Fujiishi Y, Sutou S, Tadakuma A, Ishikawa Y, Kido M, Minamiguchi R, Hanahara I, Furihata C. Discrimination of genotoxic and non-genotoxic hepatocarcinogens by statistical analysis based on gene expression profiling in the mouse liver as determined by quantitative real-time PCR. Mutat Res. 2012;747:164–75.
Sakurai M, Watanabe T, Suzuki T, Furihata C. Time-course comparison of gene expression profiles induced by the genotoxic hepatocarcinogen, chrysene, in the mouse liver. Gene Env. 2014;36:54–64.
Suenaga K, Takasawa H, Watanabe T, Wako Y, Suzuki T, Hamada S, Furihata C. Differential gene expression profiling between genotoxic and non-genotoxic hepatocarcinogens in young rat liver determined by quantitative real-time PCR and principal component analysis. Mutat Res. 2013;751:73–83.
Rooney J, Hill T 3rd, Qin C, Sistare FD, Corton JC. Adverse outcome pathway-driven identification of rat liver tumorigens in short-term assays. Toxicol Appl Pharmacol. 2018;356:99–113.
Park HJ, Oh JH, Park SM, Cho JW, Yum YN, Park SN, Yoon DY, Yoon S. Identification of biomarkers of chemically induced hepatocarcinogenesis in rasH2 mice by toxicogenomic analysis. Arch Toxicol. 2011;85:1627–40.
Kossler N, Matheis KA, Ostenfeldt N, Bach Toft D, Dhalluin S, Deschl U, Kalkuhl A. Identification of specific mRNA signatures as fingerprints for carcinogenesis in mice induced by genotoxic and nongenotoxic hepatocarcinogens. Toxicol Sci. 2015;143:277–95.
Furihata C, Suzuki T. Evaluation of 12 mouse marker genes in rat toxicogenomics public data, Open TG-GATEs: discrimination of genotoxic from non-genotoxic hepatocarcinogens. Mutat Res Genet Toxicol Environ Mutagen. 2019;838:9–15.
Kanki M, Gi M, Fujioka M, Wanibuchi H. Detection of non-genotoxic hepatocarcinogens and prediction of their mechanism of action in rats using gene marker sets. J Toxicol Sci. 2016;41:281–92.
Kinaret PAS, Serra A, Federico A, Kohonen P, Nymark P, Liampa I, Ha MK, Choi JS, Jagiello K, Sanabria N, Melagraki G, Cattelani L, Fratello M, Sarimveis H, Afantitis A, Yoon TH, Gulumian M, Grafström R, Puzyn T, Greco D. Transcriptomics in toxicogenomics, Part I: experimental design, technologies, publicly available data, and regulatory aspects. Nanomaterials (Basel). 2020;10:750.
Waters M, Stasiewicz S, Merrick BA, Tomer K, Bushel P, Paules R, Stegman N, Nehls G, Yost KJ, Johnson CH, Gustafson SF, Xirasagar S, Xiao N, Huang CC, Boyer P, Chan DD, Pan Q, Gong H, Taylor J, Choi D, Rashid A, Ahmed A, Howle R, Selkirk J, Tennant R, Fostel J. CEBS—Chemical effects in biological systems: a public data repository integrating study design and toxicity data with microarray and proteomics data. Nucleic Acids Res. 2008;36(Database issue):D892-900.
Lea IA, Gong H, Paleja A, Rashid A, Fostel J. CEBS: a comprehensive annotated database of toxicological data. Nucleic Acids Res. 2017;45:D964–71.
Lamb J. The Connectivity Map: A new tool for biomedical research. Nat Rev Cancer. 2007;7:54–60.
Subramanian A, Narayan R, Corsello SM, Peck DD, Natoli TE, Lu X, Gould J, Davis JF, Tubelli AA, Asiedu JK, Lahr DL, Hirschman JE, Liu Z, Donahue M, Julian B, Khan M, Wadden D, Smith IC, Lam D, Liberzon A, Toder C, Bagul M, Orzechowski M, Enache OM, Piccioni F, Johnson SA, Lyons NJ, Berger AH, Shamji AF, Brooks AN, Vrcic A, Flynn C, Rosains J, Takeda DY, Hu R, Davison D, Lamb J, Ardlie K, Hogstrom L, Greenside P, Gray NS, Clemons PA, Silver S, Wu X, Zhao WN, Read-Button W, Wu X, Haggarty SJ, Ronco LV, Boehm JS, Schreiber SL, Doench JG, Bittker JA, Root DE, Wong B, Golub TR. A next generation connectivity map: L1000 platform and the first 1,000,000 profiles. Cell. 2017;171:1437–52.
Ganter B, Snyder RD, Halbert DN, Lee MD. Toxicogenomics in drug discovery and development: mechanistic analysis of compound/class-dependent effects using the DrugMatrix database. Pharmacogenomics. 2006;7:1025–44.
Igarashi Y, Nakatsu N, Yamashita T, Ono A, Ohno Y, Urushidani T, Yamada H. Open TG-GATEs: a large-scale toxicogenomics database. Nucleic Acids Res. 2015;43(Database issue):D921-7.
Kolesnikov N, Hastings E, Keays M, Melnichuk O, Tang YA, Williams E, Dylag M, Kurbatova N, Brandizi M, Burdett T, Megy K, Pilicheva E, Rustici G, Tikhonov A, Parkinson H, Petryszak R, Sarkans U, Brazma A. ArrayExpress update—Simplifying data submissions. Nucleic Acids Res. 2015;43(Database issue):D1113-6.
Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002;30:207–10.
Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M, Marshall KA, Phillippy KH, Sherman PM, Holko M, Yefanov A, Lee H, Zhang N, Robertson CL, Serova N, Davis S, Soboleva A. NCBI GEO: Archive for functional genomics data sets—Update. Nucleic Acids Res. 2013;41(Database issue):D991-5.
Haimbaugh A, Meyer D, Akemann C, Gurdziel K, Baker TR. Comparative Toxicotranscriptomics of single cell RNA-seq and conventional RNA-seq in TCDD-exposed testicular tissue. Front Toxicol. 2022;4: 821116.
Jonker MJ, Bruning O, van Iterson M, Schaap MM, van der Hoeven TV, Vrieling H, Beems RB, de Vries A, van Steeg H, Breit TM, Luijten M. Finding transcriptomics biomarkers for in vivo identification of (non-)genotoxic carcinogens using wild-type and Xpa/p53 mutant mouse models. Carcinogenesis. 2009;30:1805–12.
Auerbach SS, Phadke DP, Mav D, Holmgren S, Gao Y, Xie B, Shin JH, Shah RR, Merrick BA, Tice RR. RNA-seq-based toxicogenomic assessment of fresh frozen and formalin-fixed tissues yields similar mechanistic insights. J Appl Toxicol. 2015;35:766–80.
We thank all participants in collaborative studies of JEMS·MMS·Toxicogenomics group, especially Drs. Takashi Watanabe, Shuich Hamada, and Madoka Nakajima. We also thank Drs. Kumiko Ogawa and Takeshi Toyota, Division of Pathology, National Institute of Health Sciences, 3-25-26, Tonomachi, Kawasaki-ku 210-9501, in collaborative studies on RNA-Seq. We thank Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript before the first submission.
Original works were partly supported by KAKENHI (18310047) (C. Furihata, T. Watanabe, and T. Suzuki), a High-Tech Research Center Project of Private Universities with a matching fund subsidy from the Japanese Ministry of Education, Culture, Sports, Science and Technology (C Furihata) and the Project on Regulatory Harmonization and Evaluation of Pharmaceuticals, Medical Devices, Regenerative and Cellular Therapy Products, Gene Therapy Productsm and Cosmetics from the Japan Agency for Medical Research and Devlopment, AMED (16mk0102010j003, T. Suzuki).
Ethics approval and consent to participate
For studies involving animals, ethical approvals were obtained from the institutions where the original studies were conducted.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Furihata, C., Suzuki, T. Short-term in vivo testing to discriminate genotoxic carcinogens from non-genotoxic carcinogens and non-carcinogens using next-generation RNA sequencing, DNA microarray, and qPCR. Genes and Environ 45, 7 (2023). https://doi.org/10.1186/s41021-023-00262-9
- DNA microarray
- Rodent short-term in vivo test
- Genotoxic carcinogen
- Non-genotoxic carcinogen