Identification of stable housekeeping genes in mouse liver for studying carbon tetrachloride-induced injury and cellular senescence Authors: Keting He, Dongfan Wei, Qiuhong Liu, Xiaoli Liu, Diwenxin Zhou, Shangci Chen, Danhua Zhu, Xiaowei Xu Published: 2024 ABSTRACT AbstractAcute liver injury (ALI) presents a challenging problem worldwide, prompting extensive research efforts. Cellular senescence has been found to be induced following ALI, and targeting cellular senescence has shown therapeutic potential. Therefore, understanding the expression of senescence-related genes in ALI can help to explore pathogenesis and treatment of this common disease. Carbon tetrachloride (CCl4) is commonly used to study ALI. Although polymerase chain reaction (PCR) is a convenient and economical molecular biology technique widely used in basic medicine, research on selecting suitable reference genes to obtain objective and reproducible PCR data is scarce. Moreover, evidence supporting the choice of reference genes for experimental studies of CCl4-induced ALI and hepatic senescence in mice is limited. In this study, we obtained murine livers at four time points (0, 12, 24, and 48 h) following CCl4 treatment. We used five algorithms (geNorm, BestKeeper, NormFinder, delta Ct, and RefFinder) to rank 12 candidate genes in real-time reverse-transcription quantitative PCR (RT-qPCR) experiments. Focusing on cellular senescence in this model, we adopted four senescence-associated secretory phenotype (SASP) genes (Il6, Il1b, Ccl2, and Ccl5) as target genes. Our results confirmed Gapdh and Tbp as suitable reference genes in murine CCl4-induced ALI models. Furthermore, we provide a table of published studies recommending reference genes for various liver disease models. This study provides a valuable reference for enhancing the reliability and reproducibility of ALI molecular findings.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-024-78183-y.Keywords: Polymerase chain reaction, Mouse, Carbon Tetrachloride, Cellular senescence, Reference geneSubject terms: Biological techniques, Diseases, Gastroenterology Introduction Acute liver injury (ALI) and drug-induced liver injury (DILI) are prevalent conditions that represent common causes of acute liver failure globally. DILI is typically divided into two types, intrinsic and idiosyncratic (unpredictable), depending on the mechanism of drug action. Intrinsic DILI is usually predictable, dose-dependent, and inducible by certain drugs, including acetaminophen (APAP), methotrexate, amiodarone, and statins1,2. Various drugs, including APAP, carbon tetrachloride (CCl4), thioacetamide, and furosemide, are commonly used to study intrinsic DILI in animals. Among these, CCl4 is a common reagent for investigating acute or chronic liver disease, and a high dose (≥ 1 mL/kg) causes liver injury that mimics intrinsic DILI3. Cellular senescence, characterized by irreversible cell cycle arrest, plays a vital role in ALI. Senescent cells exhibit distinct features, such as morphological changes, cell cycle arrest, macromolecular damage, and the senescence-associated secretory phenotype (SASP)4. Notably, senescence has been demonstrated in CCl4-induced ALI5–7. However, information regarding the choice of suitable reference genes for cellular senescence in this model is limited. Reverse-transcription quantitative real-time PCR (RT-qPCR) is an important, economical, and convenient molecular method. The Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines outline the essential details necessary for reporting quantitative real-time PCR experiments in scientific publications8. These guidelines highlight the unreliability of normalization using reference genes without validation. Studies typically select reference genes based on previous studies; however, the stability of these genes may not have been experimentally validated. Furthermore, although genes, such as Gapdh and Actb, are commonly used in various models; their expression is affected by the tissue type and experimental conditions9,10. Therefore, the RT-qPCR results are often inaccurate and irreproducible. Here, we aimed to identify and validate reference genes for experimental studies in CCl4-induced ALI. We initially confirmed ALI induction and hepatic senescence in mice following CCL4 treatment. Subsequently, we determined the expression of 12 commonly used genes and ranked the stability of each gene using five algorithms (BestKeeper, geNorm, NormFinder, delta Ct, and RefFinder) to identify suitable reference genes. To validate the results, we selected several SASP genes (IL-6, IL-1β, CCL2, and CCL5), whose expression increases in CCl4-induced ALI11–14. Results CCl4 treatment induces ALI and hepatic senescence All mice received a high dose of CCl4 or olive oil via intraperitoneal injection. Liver injury was assessed using serum transaminase measurement and hematoxylin and eosin (HE) staining at 12, 24, and 48 h post-CCl4 treatment. The mice successfully developed ALI, which peaked at 48 h. Mice in the 48 h treatment group exhibited the highest serum levels of ALT and AST (Fig. 1A), compared with that in the other groups. HE liver staining revealed classic pathological features of CCl4-induced liver injury. Histologically, liver injury was also the most severe at 48 h, with a large area of necrosis (bridging necrosis), damage to the hepatic architecture, and inflammatory cell infiltration (Fig. 1B). Notably, the differences within each group were generally small, indicating the stability of the model. Hepatic senescence was assessed using immunohistochemical (IHC) staining of γH2A.X and p21. Both markers peaked at 48 h, which was consistent with the extent of the liver injury (Fig. 1C and D). (A) Serum ALT and AST of four groups, 48 h was the most severe group. (B) HE Staining (20x) and Pathologic histology score (Ishak Score) of four groups, 12 h group mainly shows that the degeneration of liver cells, no obvious necrosis of liver cells. At 24 h, a large number of liver cells were necrotic, showing bridging necrosis and inflammatory cell infiltration, and reached the peak at 48 h (C) p21 (D) γH2A.X Elevated p21 is one of the markers of cell cycle arrest, and γH2A.X is one of the markers indicating DNA damage, both of which are markers of cellular senescence. With the aggravation of liver injury, they were gradually expressed in liver tissue, mainly in hepatocytes, and reached the peak at 48 h. CV: central vein. Selection and expression profiling of reference genes for RT-qPCR experiments Table 1 provides details about the 12 reference genes including the abbreviation, full name, function, and NCBI ID of each gene. The melt curves confirmed the primer specificity (Fig. 2A). Table 1Information of 12 reference genes.MouseNCBI IDOfficial full nameProduct lengthActb11,461Actin beta154B2m12,010Beta-2 microglobulin104Gapdh14,433Glyceraldehyde-3-phosphate dehydrogenase123Gusb110,006Glucuronidase, beta78Hsp90ab115,516Heat shock protein 90 alpha (cytosolic), class B member 180Ldha16,828Lactate dehydrogenase A114Pgk118,655Phosphoglycerate kinase 1164Hprt115,452Hypoxanthine phosphoribosyltransferase 1142Sdha66,945Succinate dehydrogenase complex, subunit A, flavoprotein (Fp)106Tbp21,374TATA box binding protein119Tfrc22,042Transferrin receptor136Ubc22,190Ubiquitin C106 Information of 12 reference genes. (A) Melt curves of 12 reference genes; (B) Mean ± SD of Cq of each reference gene. We used RT-qPCR to detect the gene expression levels in all 20 samples and collected the quantification cycle (Cq) of each gene in different samples. We divided all samples into five groups: ‘normal control’ (NC, n = 5), ‘12 h’ (12 h, n = 5), ‘24 h’ (24 h, n = 5), ‘48 h’ (48 h, n = 5), and ‘All’ groups (n = 20) and analyzed the results of each group. According to Fig. 1, 48 h seemed to be the most ‘senescent’ group, whether to weight the four time periods is a question to be considered. However, cellular senescence has many markers to be defined, we could not ensure that 48 h was the most senescent or the most important group. Furthermore, CCl4-induced ALI model is commonly used, it can study many mechanisms of ALI. In this article, we used cellular senescence as the entry point to verify the reliability of these algorithms, ‘48 h’ seems to be more significant. But we hope to provide a reference for more researchers, perhaps ‘12’ and ‘24 h’ are more important in their studies. Therefore, when handling the ‘All’ group, we did not consider giving a particular group greater weight. Figure 2B shows the mean ± SD of the Cqs for one specific gene among samples within each group. Stability of candidate reference genes We used five algorithms, namely geNorm, NormFinder, delta Ct, BestKeeper, and RefFinder, to assess housekeeping gene expression stability. RefFinder is a comprehensive tool that combines the aforementioned four algorithms to rerank the candidate genes. BestKeeper BestKeeper evaluates candidate genes and ranks them based on the inspection of calculated variations (SD and CV values) and correlation coefficient15. Higher correlation coefficient values and lower SD and CV values indicate better gene stability. Genes with SD values exceeding 1 should not be considered suitable as reference genes. Apart from assessing the stability of reference genes, it is also possible to evaluate the expression levels of the target genes. B2m and Gusb showed the highest stability among all groups of genes, while Ubc and Hsp90ab1 exhibited the lowest stability (Fig. 3). The ranking of 12 candidate genes using BestKeeper. geNorm geNorm assesses gene stability by examining the consistency of expression ratios of two candidate genes across all sample groups. The algorithm utilizes M, which represents the average pairwise variation of a specific gene with all other reference genes, to characterize gene stability. Gene stability is inversely related to the value of M16. Among the four groups, Gapdh, Tbp, and Pgk1 were identified as the most stable genes, while Hprt1 and Ubc were found to be less stable (Fig. 4). The ranking of 12 candidate genes using geNorm. Normfinder NormFinder can analyze subgroups of samples separately and estimate variation in both the intra- and intergroup expression, calculating the stability of reference genes17. A lower stability value indicates greater gene expression stability, while a higher value indicates the opposite. In our study, Gapdh and Tbp were identified as the most stable genes across all four groups, whereas Ubc and Ldha showed lower levels of stability (Fig. 5). The ranking of 12 candidate genes using NormFinder. Delta CT The Delta CT method is a straightforward approach for identifying suitable reference genes by comparing the relative expression of gene pairs. A high delta Ct value indicates instability in the expression of one or both genes in the pair18. Among the four groups, Gapdh and Tbp were the most stable genes, while Ubc and Ldha were the least stable (Fig. 6). The ranking of 12 candidate genes using DeltaCT. RefFinder The results contained variations owing to the different principles in calculating the four algorithms, complicating determining the best method. RefFinder combines the results of geNorm, NormFinder, BestKeeper, and delta Ct by calculating the geometric mean of each gene weight and re-ranking candidate reference genes based on their stability. A smaller geometric mean value indicates greater gene stability. The website https://blooge.cn/RefFinder/ offers easy access to this tool19. After re-ranking with RefFinder, Gapdh and Tbp were the most stable genes, while Ubc and Hprt1 were the least stable among the four groups (Fig. 7). In Table 2, we listed the results of the 12 candidate genes in the five algorithms, and listed the corresponding rank of each gene in parentheses. The overall ranking of 12 candidate genes using RefFinder. Table 2Comprehensive ranking of 12 candidate genes in five algorithms.BestkeeperSDGeNormM ValueNormFinderStability ValueDelta CTAverage of STDEVRefFinderGeomean of ranking valuesGapdh0.57 (3)0.45 (1)0.27 (1)0.73 (1)1.32 (1)Tbp0.61 (4)0.45 (1)0.29 (2)0.75 (2)2.00 (2)Gusb0.53 (2)0.58 (4)0.42 (3)0.79 (3)2.91 (3)Pgk10.72 (10)0.53 (3)0.48 (4)0.83 (4)4.68 (4)B2m0.53 (1)0.74 (8)0.77 (9)0.99 (9)5.05 (5)Sdha0.62 (5)0.62 (5)0.69 (7)0.93 (6)5.69 (6)Tfrc0.71 (9)0.67 (6)0.65 (5)0.92 (5)6.06 (7)Actb0.69 (8)0.70 (7)0.68 (6)0.94 (7)6.96 (8)Hsp90ab10.91 (11)0.79 (9)0.73 (8)0.98 (8)8.92 (9)Ldha0.63 (6)0.83 (10)0.84 (11)1.05 (11)9.23 (10)Hprt10.68 (7)0.87 (11)0.80 (10)1.03 (10)9.37 (11)Ubc1.16 (12)0.93 (12)1.2 (12)1.26 (12)12.00 (12) Comprehensive ranking of 12 candidate genes in five algorithms. Reference genes stability validation To validate the reference gene ranking obtained from the five algorithms, we selected four target genes: Il6, Il1b, Ccl2, and Ccl5. Senescent cells produce various factors known as the SASP, and these four genes are widely recognized as belonging to the SASP20,21. The most stable (Gapdh and Tbp) and least stable (Ubc and Hprt1) reference genes were selected for the normalization of target gene expression. The sequences of the primers targeting these genes are listed in Table 3. Table 3Primer sequence information for genes.NameForward primerReverse primerProduct length (bp)Ccl2GGTGTCCCAAAGAAGCTGTAGTTAGCTTCAGATTTACGGGTCAACT203Ccl5CTCCAATCTTGCAGTCGTGTTTGACCCTCTATCCTAGCTCATCTCC116GapdhAGGTCGGTGTGAACGGATTTGTGTAGACCATGTAGTTGAGGTCA123IL-1bCAGCACATCAACAAGAGCTTCAGGAGGATGGGCTCTTCTTCAAAGA150IL-6AACCGCTATGAAGTTCCTCTCTGTGGTATCCTCTGTGAAGTCTCCT125TbpCCTTGTACCCTTCACCAATGACACAGCCAAGATTCACGGTAGA119UbcACCACCAAGAAGGTCAAACAGGTAAGACACCTCCCCCATCAC106 Primer sequence information for genes. When Gapdh and Tbp were used as reference genes, the overall patterns of target gene expression was similar, with an initial rise followed by a decline. Nearly all four genes reached their peak at 24 h. When Hprt1 was used as the reference gene, Il1b and Il6 expression peaked at 12 h, Ccl5 expression peaked at 24 h, and Ccl2 expression peaked at 48 h, with Il1b displaying a bimodal pattern. Using Ubc as the reference gene, most genes reached their peak at 48 h, while Il1b showed a decrease followed by an increase in contrast to the findings of other experiments (Fig. 8). Validation of reference genes using four target genes *: p < 0.05, **: p < 0.01. The rest groups have no statistically significant difference. Discussion Cellular senescence plays a notable role in physiological and pathological conditions. Cellular senescence occurs in chronic liver diseases, liver cancers, and cholangiopathies22,23. Recently, additional studies have reported hepatic senescence in ALI models, and attempted to treat ALI via targeting senescence. However, the conflicting results suggest that the role of cellular senescence in ALI is complex5,24. Thus, additional elaborate studies are required. DILI is an important cause of ALI. Although CCl4 is not a drug, it is a valuable model for investigating the mechanisms of DILI. In this study, we explored suitable reference genes in a CCl4-induced mice model using tissues collected at different time points, ranked them using five algorithms, and validated them using four SASP genes. We identified Gapdh and Tbp as the most stable genes, and Hprt1 and Ubc as the least stable genes. In contrast to our findings, a previous study selected stable reference genes in a CCl4-treated human hepatocyte cell line (L02) and found that TUBB2a was the most stable, whereas ACTB, GAPDH, and B2M were the least stable among the candidate genes studied25, despite being commonly used in CCl4-induced DILI models. This discrepancy may be attributed to the diverse cell types within the liver, highlighting that a reference gene suitable for one specific cell type may not be suitable for all the tissues. Furthermore, the previous study used exclusively geNorm to select suitable genes, whereas our study used five different methods (including geNorm) to measure gene stability. A literature survey on RT-qPCR-analyses in human hepatocyte-like cells, such as primary hepatocytes, and studies in human hepatocyte-like cells, such as primary hepatocytes, reported that GAPDH and ACTB were the most commonly used reference genes. However, PSMB6 and CYC1 have been found as the most stable genes in primary human hepatocytes, whereas ACTB and GAPDH were the least stable of the candidate genes studied26. This indicates that the PCR results of some studies are unreliable owing to the selection of unstable reference genes. To assist appropriate reference gene selection, we summarized a list of studies encompassing common liver diseases, such as ALI25–33, chronic liver diseases10,34–39, and liver carcinoma40–45, in Table 4. This compilation underscores the necessity of careful assessment before conducting RT-qPCR experiments, particularly when using the same reagent to treat different experimental subjects or using the same experimental subject with various reagents. For example, a previous study explored the stability of reference genes in an APAP-exposed primary rat hepatocyte and reported that B2m and 18 S were the most stable genes, whereas Hprt and Gapdh were the least stable. However, in HepG2 cells, PPIA and GAPDH were the most stable, and CDKN1A and HPRT1 were the least stable genes32. Another study, using HepG2 and L02 cells to select reference genes in four liver-cell injury models (ethanol, hydrogen peroxide, APAP, and CCl4), yielded disparate conclusions25. Moreover, the expression levels of reference genes may vary in different stages of the same experimental subject, such as the follicular and luteal phases in female subjects46,47. Table 4The most and least stable genes in common liver diseases.ModelSpeciesPCRAlgorithmsThe most satble gene(s)The least stable gene(s)YearAuthorJournalAcute liver injury1,4-Bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP) induced direct hyperplasiaFemale C57BL/6 mice, 11–12 week oldReal time RT-PCRgeNorm and NormFinderPPIA and RPL418 S and TFRC2008Takagi S et al.Biochem Biophys Res Commun.2/3 partial hepatectomy (PH)Female wild-type C57BL/6 mice, 10–12 weeks old,RT-PCRgeNorm and NormFinderPPIA and TBPGUSB and TFRC2008Tatsumi K et al.Biochem Biophys Res Commun.2/3 PHSprague-Dawley rat regenerating hepatocytesRT-PCRgeNorm18SrRNA-UBCHK1 and B2M2010Wang GP et al.Mol Biotechnol.Hepatic stellate cellsACTB-HK118SrRNA and B2MKupffer cellsACTB-GADPHHK1 and UBCBiliary epithelial cellsB2M-ACTBGAPDH and HK1Sinusoidal endothelial cells18SrRNA-UBCGAPDH and HK1Pit cellsB2M-UBCACTB and HK1Dendritic cellsB2M-ACTBUBC and GapdhOval cellsB2M-UBCHK1 and GAPDHAcute alcoholic liver injuryMale Imprinting Control Region (ICR) mice, 8–10 weeks oldRT-qPCRgeNorm, NormFinder, and BestKeeperHPRT118 S2018Wang S et al.Int J Mol Med.concanavalin A (ConA)Male C57BL/6 mice, 8–10 weeks oldReal time RT-PCRgeNorm, NormFinder, and BestKeeperRPLP0ACTB2010Shi G et al.Anal Biochem.Acetaminophen (APAP)HepG2 cellsRT-qPCRgeNormGAPDH and PPIACDKN1A and HPRT12010Fox BC et al.Toxicol In Vitro.primary rat hepatocytesRT-qPCR18 S and B2MHPRT1 and GapdhAPAP, carbon tetrachloride (CCl4), D-galactosamine, and thioacetamideMale Wistar ratsRT-qPCRgeNorm, Normalization, and BestKeeper18 S and B2M/2012Lardizábal MN et al.PLoS One.APAPHepatoma carcinoma cell (HepG2)qPCRgeNorm, NormFinder, and BestKeeperGUSBACTB, GAPDH, YWHAZ, and B2M2020Chen J et al.Biomed Res Int.Human hepatocyte cell line (L02)TBPEthanolHepG2 cellsTBPL02 cellsTBPHydrogen peroxideHepG2 cellsTBPL02 cellsTBPCCl4HepG2 cellsCYC1L02 cellsTUBB2aDifferent cells treated with various drugs and culture conditionsPrimary human hepatocytes, HepG2, and Huh-7.5 cellsRT-qPCRgeNorm, NormFinder, BestKeeper, and an extended delta CT methodPSMB6 MDH1 (overall ranking)HPRT Gapdh (overall ranking)2014Riedel G et al.PLoS One.Hepatitis C virus (HCV) infectionHumanRT-PCRSpearman’s rank correlation, NormFinder, BestKeeper and geNorm.(GUSB and SFRS4)β-glucuronidase and splicing factor arginine/serine-rich 4HPRT1, β-actin, GAPDH and 18 S rRNA2011Congiu M et al.Liver Int.HCV chronic hepatitis with or without liver cirrhosisHumanRT-PCRgeNorm and NormFinderβ-glucuronidase (GUSB)GAPDH, SMARCA5, TBP, B2M, GMFG, and RNA18S52013Yamaguchi H et al.Exp Mol Pathol.Beta-casomorphin peptides (BCM7/BCM9)Male C57BL/6 mice,7–8 weeks oldRT-qPCRgeNorm, NormFinder and BestKeeper.PPIA and SDHATFR2 and TBP2023Kumar A et al.PLoS One.Non-alcoholic fatty liver disease (NAFLD) /high fat diet (HFD)C57/BL6J miceRT-PCRgeNormYWHAZEIF4A2 and CYC12012Bruce KD et al.Liver Int.HFDC57BL/6J miceRT-PCRgeNorm, NormFinder, and BestKeeperHPRT and GAPDHTFRC(Transferrin receptor) and RPLP0 (Ribosomal protein, large, P0)2011Xu L et al.Mol Biotechnol.HFDMale C57BL/6J miceRT-qPCRgeNorm, DeltaCT, BestKeeper, and NormFinderHRPT、YWHAZ and RPLP0UBC Gapdh2020Fan X et al.Front Nutr.Caloric restrictionMale C57BL/6 miceRT-qPCRgeNorm, NormFinder and BestKeeperHmbsPGK12016Gong H et al.Sci Rep.Liver cancerHCCHepaRG cellsRT-qPCRgeNorm, NormFinder and BestKeeperTBP, SDHA, and UBCACTB and S182011Ceelen L et al.Biochem Pharmacol.Hepatitis virus B (HBV) related HCCHumanqRT-PCRgeNorm and NormFinderHPRT and TBPACTB and GAPDH2008Gao Q et al.J Cancer Res Clin Oncol.Tumor and non-tumor tissues from untreated HBV-related HCCHumanRT-qPCRgeNorm and NormFinderC-terminal binding protein 1(CTBP1) and UBC18 S and GAPDH2013Liu S et al.Cancer Biomark.HCV induced HCC, including pre-neoplastic lesions (cirrhosis and dysplasia)HumanqPCRgeNorm and NormFinderribosomal protein large 41 (RPL41) and SFRS4 (arginine/serine-rich splicing factor 4)GAPDH and RPS20 (ribosomal protein small 20)2007Waxman S et al.BMC Genomics.HCCHuman liver cancer cell lines Hep3B, HepG2, SK-HEP-1, SNU-182, and HUH7/human liverRT-PCRgeNorm and NormFinderhydroxymethyl-bilane synthase (HMBS)B2M and UBC2008Cicinnati VR et al.BMC Cancer.HBV-related HCC (with-metastasis and without-metastasis )HumanRT-PCRgeNorm and NormFinderTBP and HPRT118s and ACTB2009Fu LY et al.BMC Cancer.TCPOBOP: 1,4-Bis[2-(3,5-dichloropyridyloxy)]benzene PH: Partial hepatectomy ConA: Concanavalin A CCL4: Carbon tetrachloride TAA: Thioacetamide D-galN: D-galactosamine H2O2: Hydrogen peroxide NAFLD: Non-alcoholic fatty liver disease HFD: High fat diet HCV: Hepatitis C virus HBV: Hepatitis virus B BCM: Beta-Casomorphin ICR: Imprinting Control Region. The most and least stable genes in common liver diseases. TCPOBOP: 1,4-Bis[2-(3,5-dichloropyridyloxy)]benzene PH: Partial hepatectomy ConA: Concanavalin A CCL4: Carbon tetrachloride TAA: Thioacetamide D-galN: D-galactosamine H2O2: Hydrogen peroxide NAFLD: Non-alcoholic fatty liver disease HFD: High fat diet HCV: Hepatitis C virus HBV: Hepatitis virus B BCM: Beta-Casomorphin ICR: Imprinting Control Region. According to Table 2, we observed that, overall, Bestkeeper presents a different gene stability ranking compared to Normfinder, GeNorm, and Delta CT. This discrepancy is primarily due to RefFinder outputting Bestkeeper’s ranking based solely on the SD value, while ignoring CV and correlation coefficient. Lower SD lower CV, and higher correlation coefficient mean better stability. However, we found that the three results are not matching. SD: B2m (most stable) < Gusb < Gapdh < Tbp < Sdha < Ldha < Actb < Hprt1 < Pgk1 < Tfrc < Hsp90ab1 < Ubc.CV: Gusb(most stable) < Tbp < Sdha< Tfrc< Hprt1 < Gapdh < B2m < Pgk1 < Ldha < Actb < Hsp90ab1 < Ubc. Correlation coefficient: Gapdh (most stable) > Tbp > Pgk1 > Hsp90ab1 > Tfrc > Gusb > Actb > Ubc > Hprt1 > Sdha > Ldha > B2m. When combining the three indexes, we found that Tbp, Ubc, and Gapdh were the most stable genes, and this results was similar to geNorm, Normfinder, and delta CT. In addition, our results showed that the four ‘SASP’ genes (RNA) peaked before p21 and γH2A.X (protein). This was intriguing, as the SASP is typically produced by senescent cells, suggesting that these genes might increase before the onset of senescence. However, considering that gene transcription precedes translation, it is possible that p21 and other senescence-related genes are elevated before the 12 or 24-hour mark, but we did not measure their transcript levels at those earlier time points. Even if these genes appear after the onset of ‘cellular senescence’, we cannot definitively conclude that they are produced as SASPs by senescent cells. This is because these genes are not specific to senescence and often appear in various biological processes such as necrosis, cancer, and inflammation48,49. To further address this issue, we could employ methods such as in vitro culturing of senescent cells to detect the expression of relevant genes, as well as single-cell sequencing. These methods would allow us to more precisely elucidate the problem at the cellular level. Nonetheless, we chose these genes because they are frequently referred to as “SASPs” in previous studies. It is possible that these earlier studies encountered similar issues to ours. Identifying senescence-specific SASPs or other specific markers would significantly enhance our understanding of this field. This study had some limitations. First, liver injury will last for several days following CCl4 treatment, but our study just focused on the early stage (before 48 h) of ALI. Second, although the 12 candidate genes used in this study are commonly used, there may be more stable genes, and combining techniques such as RNA sequencing could help us to find suitable reference genes. Finally, these experiments used whole liver tissue; therefore, the results are only applicable to studies using the whole liver. As the liver contains epithelial cells and non-epithelial cells, some genes may exhibit cell-specific changes. For example, cellular senescence occurs mainly in hepatocytes in CCl4-induced liver injury 5,7. Therefore, further studies involving different cell types are required for a comprehensive understanding. Conclusion In this study, we confirmed ALI and hepatic senescence following CCl4 treatment, explored suitable reference genes in a CCl4-induced mice model, ranked them using five algorithms (BestKeeper, geNorm, NormFinder, delta Ct, and RefFinder), and validated them using four SASP genes (IL-6, IL-1β, CCL2, and CCL5). Among the four groups (NC, 12 h, 24 h, 48 h), Gapdh and Tbp were the most stable genes, whereas Hprt1 and Ubc were the least stable genes. Our study is the first to reveal suitable reference genes in a CCl4-induced ALI murine model and will serve as a reference for future research endeavors in this field. Materials and methods Ethical statement All the animal studies were conducted according to the protocol approved by the Tab of Animal Experiment Ethical Inspection of The First Affiliated Hospital, Zhejiang University School of Medicine. The study was conducted in compliance with the ARRIVE guidelines. All experiments were performed in accordance with relevant guidelines and regulations. Animals C57BL/6 male mice aged 6–8 weeks were obtained from the Experimental Animal Center of Zhejiang Academy of Medical Sciences. The mice were allowed to adapt at the Laboratory Animal Center for one week. ALI was induced in 15 mice weighing between 20 and 22 g by administering a high dose of CCl4 (Aladdin, China) (1.5 uL/g weight, dissolved in olive oil at a ratio of 1:3) via intraperitoneal injection. As a control group, five mice received an equivalent amount of olive oil only. Mice were anesthetized using isoflurane and sacrificed. Liver tissues and blood were collected from the CCl4-treated groups after 12, 24, and 48 h. Blood samples were obtained from the retroorbital venous plexus and then centrifuged at 4 °C for 15 min at 2,200 rpm. The livers were rinsed with PBS solution; some sections were used for histopathological analysis while the remaining tissue was cryopreserved. Determination of liver enzyme activities Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities were tested using a Mindray automatic biochemical analyzer BS-220 (Shenzhen, China). Histology and Ishak score The liver tissue samples were treated with 4% paraformaldehyde for one day, then preserved in paraffin and sliced into 5 μm sections. The sections were hematoxylin-eosin stained, and digital images were captured using a NanoZoomer 2.0RS scanner. Hematoxylin-eosin-stained livers were evaluated by a single pathologist based on Ishak Score50. Immunohistochemistry Paraffin sections were deparaffinized with xylene followed by a gradient concentration of ethanol, and then subjected to antigen retrieval in citrate antigen retrieval buffer. Subsequently, tissue sections were incubated in 3% hydrogen peroxide solution to block endogenous peroxidase activity in the dark, followed by washing and blocking with 3% BSA. Sections were then incubated overnight at 4 °C with primary antibodies against p21 (1:1000, Abcam, ab188224) and γ-H2AX (1:2000, Abcam, ab81299). After washing the sections in PBS, species-specific HRP-conjugated secondary antibodies corresponding to the primary antibodies were added and incubated at room temperature for 50 min. DAB staining was subsequently performed, and nuclear counterstaining was conducted using hematoxylin. After staining completion, sections were dehydrated, mounted, and observed under a microscope. qRT-PCR Liver tissue samples were processed using TRIzol (Sigma-Aldrich) to extract total RNA. Subsequently, cDNA was synthesized using PrimeScrip RT Master Mix (TaKaRa). To analyze gene expression levels, a commercially available PCR array comprising 12 reference genes (wc-Reg mRNA004) sourced from WeGene Biotech was used. Primers (Table S1) required for amplification were procured from Qingke Biotechnology Corporation (Beijing, China). The Applied Biosystems QuantStudio 5 Real-Time PCR System facilitated the execution of qRT-PCR experiments. The 2-ΔΔCT method was used to quantify gene expression data. Gapdh, Tbp, Hprt1, and Ubc were used as reference genes for normalization. Statistical analysis The data were expressed as the means ± standard deviation (SD). The ranking of gene stability was completed by WeGene Biotech and an online tool (https://blooge.cn/RefFinder/ ). When comparing the same variable among multiple groups, a One-way ANOVA was used if the data met the assumptions of normal distribution and homogeneity of variance. If the data were normally distributed but exhibited heterogeneity of variance, a Welch ANOVA was employed. When the data did not satisfy a normal distribution, the Kruskal-Wallis test was utilized. When the p value is less than 0.05 we think there are statistically significant. All statistical analyses were performed using GraphPad Prism 9 software. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors would like to thank Dr. Jinlin Chen, the pathologist of State Key Laboratory for Diagnosis and Treatment of Infectious Diseases for her technical support. Thanks WeGene Biotech for their technical support in the analysis of reference genes. This research was funded by National Key R&D Program of China (grant No. 2020YFE0204300) and the Fundamental Research Funds for the Central Universities (grant No. 2022ZFJH003). Author contributions Conceptualization, X.X.; methodology, D.Z.; software, K.H.; validation, K.H. and D.W.; formal analysis, Q.L.; investigation, K.H.; resources, D.W.; data curation, Q.L.; writing—original draft preparation, K.H., D.W., and Q.L.; writing—review and editing, X.X.; visualization, D.Z and S.C.; supervision, X.L. and D.Z.; project administration, X.X.; funding acquisition, D.Z. and X.L. All authors have read and agreed to the published version of the manuscript. Data availability Data are available from the corresponding author upon reasonable request. Declarations Competing interests The authors declare no competing interests. Ethics declarations The animal study protocol was approved by the Institutional Review Board of the First Affiliated Hospital, Zhejiang University School of Medicine (protocol code 2024 − 1207). All experiments were performed in accordance with relevant guidelines and regulations. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Keting He, Dongfan Wei and Qiuhong Liu contributed equally to this work. Contributor Information Danhua Zhu, Email: zhudanhua@zju.edu.cn. Xiaowei Xu, Email: xxw69@zju.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Data are available from the corresponding author upon reasonable request.