Chromosome aberrations induced by the non-mutagenic carcinogen acetamide involve in rat hepatocarcinogenesis through micronucleus formation in hepatocytes

Chromosome aberrations (CAs), i.e. changes in chromosome number or structure, are known to cause chromosome rearrangements and subsequently tumorigenesis. However, the involvement of CAs in chemical-induced carcinogenesis is unclear. In the current study, we aimed to clarify the possible involvement of CAs in chemical carcinogenesis using a rat model with the non-mutagenic hepatocarcinogen acetamide. In an in vivo micronucleus (MN) test, acetamide was revealed to induce CAs specifically in rat liver at carcinogenic doses. Acetamide also induced centromere-positive large MN (LMN) in hepatocytes. Immunohistochemical and electron microscopic analyses of the LMN, which can be histopathologically detected as basophilic cytoplasmic inclusion, revealed abnormal expression of nuclear envelope proteins, increased heterochromatinization, and massive DNA damage. These molecular pathological features in LMN progressed with acetamide exposure in a time-dependent manner, implying that LMN formation can lead to chromosome rearrangements. Overall, these data suggested that CAs induced by acetamide play a pivotal role in acetamide-induced hepatocarcinogenesis in rats and that CAs can cause chemical carcinogenesis in animals via MN formation.

Keywords: Acetamide; Micronucleus test; Chromosome aberrations; Large micronuclei; Hepatocarcinogenesis; Chromosome rearrangements

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s00204-021-03099-9.

Chromosome aberrations (CAs), i.e. structural or numerical aberrations in chromosomes, are a known cause of tumorigenesis and malignant transformation via the propagation of substantial genomic instability (Santaguida and Amon [27]; Tijhuis et al. [34]). CAs are mainly caused by abnormal mitoses and/or chromosomal fragmentations (Hatch and Hetzer [12]; Guo et al. [8]) that result in the formation of micronucleus (MN), which is a small nuclear structure in the cytoplasm. Although MN had been thought to be simply a hallmark CA, multiple studies involving cultured cells or plants suggested that MN is also an active driver of genomic instability, and has different characteristics from those of primary nucleus such as loss of nuclear proteins or disrupted chromatin structure (Hatch et al. [13]; Liu et al. [17]; Guo et al. [8]; Kneissig et al. [16]). These characteristics lead to DNA damage in MN and could be a cause of clustered chromosome rearrangements, termed chromothripsis or chromoplexy, which induce drastic and massive gene mutations (Stephens et al. [30]; Crasta et al. [4]; Zhang et al. [41]; Russo and Degrassi [26]; Pellestor and Gatinois [23]). These chromosome rearrangements were recently recognized in a range of human tumors (Cortés-Ciriano et al. [3]; Voronina et al. [40]), suggesting that CAs play a crucial role in tumorigenesis and malignant transformation.

The frequent occurrence of chromosome rearrangements in human tumors suggests that clastogenic and aneugenic carcinogens, which both cause MN formation in mammalian cells, may contribute to in vivo tumorigenesis via chromosome rearrangements (Guo et al. [8]; Tweats et al. [35]). Indeed, an in vivo MN test, a common tool for detecting the clastogenic or aneugenic potential of chemicals, can successfully detect MN formation in target organs of carcinogenic agents, such as skin, lung, gastrointestinal tracts and liver (Uno et al. [36]; Hamada et al. [10]). Notably, recent studies reported that the MN assay in liver has high sensitivity and specificity for detecting hepatocarcinogenicity of both genotoxic and non-genotoxic carcinogens (Shigano et al. [28]; Hamada et al. [10]), suggesting that the CAs underlie the mechanisms of various chemical-induced hepatocarcinogenesis. However, the involvement of CAs via MN formation in chemical carcinogenesis remains to be fully elucidated, largely because examination of the characteristics of MN in animal models has been more limited.

Acetamide (CAS 60-35-5) is a food contaminant (Vismeh et al. [39]) that is known to be a potent hepatocarcinogen in rats. In chronic studies using F344 rats, oral administration of 2.36% acetamide in the diet for one year induced hepatocarcinoma in 42/47 males and 36/48 females (Fleischman et al. [7]). Several in vitro and in vivo genotoxicity studies including in vivo bone marrow MN assay in rats showed negative results for this compound (Abbott et al. [1]; Moore et al. [20]). Moreover, our in vivo mutation assay using F344 gpt delta transgenic rats demonstrated that acetamide showed no mutagenicity at carcinogenic doses even at the target site for carcinogenesis (Nakamura et al. [21]). Overall, these data do not support the direct involvement of the mutagenic action of acetamide in hepatocarcinogenesis. On the other hand, in histopathological analyses we found that carcinogenic doses of acetamide promoted formation of basophilic cytoplasmic inclusion derived from nuclear components such as MN in hepatocytes (Nakamura et al. [21]). Abnormal hepatocyte mitoses including chromosome misalignment or multipolar spindles, which are generally known to be a cause of CAs, were concurrently observed with the cytoplasmic inclusion (Nakamura et al. [21]). Based on these data, we hypothesized that CAs through the formation of cytoplasmic inclusion may be involved in the potent hepatocarcinogenicity of acetamide.

In the current study, we use the in vivo liver MN assay to clarify whether acetamide induces CAs in rat liver at carcinogenic doses. Focusing on histopathologically-detectable MN-like inclusion, we investigate various molecular pathological features of MN in in vivo models. Time-dependent changes observed in MN-like structure may imply the possible involvement of chromosome rearrangements in carcinogenesis. Taken together, the data suggest that CAs play an essential role in non-mutagenic acetamide-induced hepatocarcinogenesis in rats via MN formation.

Acetamide, mitomycin C (MMC), colchicine (COL) and ethyl methanesulfonate (EMS) were purchased from Tokyo Chemical Industry (Tokyo, Japan), Kyowa Hakko Kirin (Tokyo, Japan), Sigma-Aldrich (St. Louis, USA) and Fujifilm Wako Pure Chemical Corporation (Osaka, Japan), respectively. In the in vivo micronucleus test, acetamide was mixed into powdered basal diet (CRF-1; Oriental Yeast, Tokyo, Japan) every week and kept at room temperature until use. MMC and COL were dissolved in distilled water immediately before use. For the in vivo comet assay, acetamide was dissolved and EMS was diluted in distilled water immediately before use.

The protocols for experiments involving animals described in this study were approved by the Animal Care and Utilization Committee of the National Institute of Health Sciences. Specific pathogen-free, 5-week-old male F344 rats were obtained from Japan SLC (Shizuoka, Japan). The animals were housed in polycarbonate cages (five rats per cage) with hardwood chips for bedding. The animals were maintained in a room with a barrier system, under conditions of controlled temperature (24 ± 1 °C), humidity (55 ± 5%), air exchange (18 times/h) and lighting (12-h light/dark cycle). Animals had free access to tap water. For in vivo comet assay, animals also had free access to an Oriental CRF-1 solid basal diet. After a 1-week acclimatization period, the animals were divided into experimental groups consisting of five rats per group using the weight-stratified randomization method based on body weights that were measured just before starting the test chemical administration. Three rats were randomly selected for the COL-treated group.

In the in vivo micronucleus test, animals had free access to powdered basal diet with 0%, 0.625%, 1.5% or 2.5% (w/w) acetamide for 4 weeks. These same dose levels were also used in our previous study (Nakamura et al. [21]). Animals in one positive control group were orally administered MMC by gavage at 0.25 mg/kg/day for 2 weeks. In the other positive control group, animals were intraperitoneally administered COL at 0.06 mg/kg/day once daily. Since one animal in the COL-treated group was found dead on Day 3, the remaining animals were sacrificed on that day,

During the experimental period, the animals were observed daily for any clinical signs and mortality. In the acetamide-treated groups, body weight and food consumption were measured once weekly. In the MMC-treated group, body weights were measured once weekly. At the end of treatment period for the acetamide-treated groups, the animals were fasted overnight and then anesthetized with isoflurane before collection of blood samples from the abdominal aorta. The animals were sacrificed by exsanguination from the abdominal aorta, and liver and bone marrow were extirpated. Cells were collected from the bone marrow using fetal bovine serum (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). The livers were weighed, and portions of the left and right lobes were fixed in 10% neutral-buffered formalin for histopathological examination and immunofluorescent analysis. The remaining tissue was stored at − 80 °C for western blotting and real-time RT-PCR analyses.

Bone marrow cells collected at necropsy were smeared on a slide glass. After drying, the cells were fixed in methanol, and the slides were stored at room temperature until analysis. Acridine orange solution was dropped on the slide glass and spread with cover glass and then the slide was served with the observation by fluorescence microscope (BX-51, Olympus, Tokyo, Japan). Micronucleated polychromatic erythrocytes (MNPCEs) were recorded based on observation of 4,000 polychromatic erythrocytes (PCEs) and the MNPCE/PCE ratio was calculated.

Liver micronucleus assays were conducted with a slightly modified version of a previously reported method (Hamada et al. [10]). Briefly, a small portion of the left lobe of the liver fixed with 10% neutral-buffered formalin was cut into approximately 3 mm cubes with a razor and washed with water. Then, about ten cubes were incubated in 10 mL 12 M potassium hydroxide solution for 16 h at room temperature. After incubation, the cubes were thoroughly washed and homogenized with water, followed by filtration through a cell strainer (pore size: 100 µm). Hepatocyte (HEP) suspensions were centrifuged at 50 × g for 5 min and washed three times with 10% neutral-buffered formalin. Pelleted HEPs were resuspended in approximately 2 mL 10% neutral-buffered formalin to prepare HEP suspensions that were stored at 4 °C until microscopy analyses.

For microscopic observation, HEP suspensions were mixed and stained with the same volume of a solution of 2 × SYBR Gold (SYGO; 10,000 × concentrate, Life Technologies, Inc., Carlsbad, CA, USA) in TE buffer (pH 7.5–8.0), and then the mixture was dropped onto a glass slide to which coverslips were applied. Each of the specimens on slides was observed with a fluorescent microscope (BX-51, Olympus) with a B-excitation filter (wavelength: 420–490 nm). For each sample, around 2000 parenchymal HEPs were analyzed with a magnification of ≥ 200 ×. The number of micronucleated HEPs (MNHEPs) was recorded as previously reported (Shigano et al. [28]; Hamada et al. [10]). In the current study, large micronucleated HEPs (LMNHEPs) were defined as hepatocytes with MN that showed diameters 1/2 to 1/4 that of primary nuclei. The number of MN with such diameters was also recorded.

To characterize large micronucleus (LMN) in hepatocytes, FISH analysis for HEP specimens was conducted using a modification of a method that was previously reported for analysis of bone marrow erythrocyte specimens (Takeiri et al. [31]). Template DNA derived from hepatocytes from male F344 rats, and primers (5′-TCCCGCTTGGAACGAAGAGA-3′ and 5′-TTCTATATCCCGAACAGTCC-3′) specific for centromeric satellite I (de Stoppelaar et al. [5]; Takeiri et al. [31]) were used for PCR amplification. PCR was conducted using Ex Taq (Takara Bio, Shiga, Japan) and a previously reported amplification protocol (Takeiri et al. [31]). PCR products were purified using NucleoSpin Gel and a PCR Clean-up kit (Macherey–Nagel, Düren, Germany). Around 2 μg of the resulting DNA from the PCR amplification was used for a digoxigenin labeling reaction with a DIG DNA Labeling Kit (Roche Diagnostics, Mannheim, Germany).

HEP slides were prepared by smearing HEP suspensions on a MAS-coated slide glass (Matsunami Glass Ind. Ltd., Osaka, Japan). After drying, the slides were dehydrated with methanol, dried and stored at − 20 °C. After rehydration in distilled water, the slides were incubated with 500 μg/mL proteinase K solution [proteinase K (Sigma-Aldrich) diluted in 10 mM Tris–HCl, 0.1 mM EDTA (pH 8.0)] for 20 min at 37 °C to eliminate the hepatocyte cytoplasm and its non-specific fluorescence. After washing in PBS, the slides were denatured with 2 × SSC/70% formamide at 70 °C for 5 min, and then dehydrated with a series of ethanol. Then, 15 μL of estimated 500 ng/mL DIG-labeled DNA probe in hybridization buffer (Nippon Gene Co, Tokyo, Japan) was dropped on the slides, which were then covered with a cover slip and sealed with rubber cement. After heating on a hot plate at 80 °C for 10 min, the slides were incubated in a humidified container overnight at 37 °C. The cover slip was then removed and the slides were washed with 2 × SSC and 2 × SSC/50% formamide at 37 °C for 5 min. After blocking with 10% bovine serum, anti-mouse anti-digoxigenin antibody (1:200, Roche Diagnostics) was applied to the slides for 2 h at room temperature followed by incubation with Rhodamine Red-X-labeled anti-mouse antibody (1:50, Jackson ImmunoResearch Laboratories, PA, USA) for 1 h. After staining with 1 × SYGO solution for 10 min for nuclear counterstaining, the slides were embedded in VECTASHIELD HardSet Antifade Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA, USA) under a cover slip for mounting. Fluorescent signals from centromeres on LMN among up to 30 large MNHEPs were observed in the 1.25% acetamide-treated group (n = 5) and colchicine-treated group (n = 1; only animal in which LMNs were observed) with a fluorescence microscope (BX-51, Olympus).

In vivo alkaline comet assay was carried out according to the Organization for Economic Co-operation and Development (OECD) test guideline 489 (OECD [22]). Animals were orally administered with 0, 200, 600 or 2000 mg/kg acetamide for three consecutive days. Animals were necropsied three hours after the last dosing, and the livers were extirpated and weighed. In the positive control group, animals received EMS at 200 mg/kg/day for two consecutive days. Three hours after the last dosing the animals were necropsied and the livers were extirpated. A small portion of the left lobe of the liver was cut into small pieces in HBSS medium (without Ca2+/Mg2+, with 1 mM EDTA) containing 10% DMSO to obtain a cell suspension. The cell suspension was then mixed with 0.5% low-melting point agarose (Takara Bio Inc., Shiga, Japan) in PBS and deposited on a MAS-coated slide glass. The slide was placed in a lysis solution (2.5 mM NaCl, 0.1 M EDTA, 10 mM Tris, 10% DMSO and 1% Triton X-100, pH 10) and incubated overnight at 4 °C. The DNA was allowed to unwind in electrophoresis buffer (0.3 M NaOH, 1 mM EDTA, pH 13) for 20 min. The slides were placed into a horizontal electrophoresis tank and exposed to 0.7 V/cm (300 mA) for 30 min. Then, the slides were washed twice in neutralization buffer (0.4 M Tris, pH 7.5) and dehydrated in ethanol for at least 5 min. After the slides were blinded randomly and stained with SYGO, at least 150 randomly selected cells per slide were visualized under a fluorescent microscope and submitted to image analysis using Comet Assay IV software (Perceptive Instruments Ltd., Bury St. Edmunds, England). The percentage of DNA in the comet tail (% tail DNA) was scored to evaluate the extent of DNA damage in individual cells.

Paraffin-embedded sections of livers from all groups were routinely prepared and stained with hematoxylin and eosin for histopathological examination.

The primary antibodies used for immunofluorescence analyses are listed in Supplementary Table S1. Antigens were retrieved by autoclaving with Target Retrieval Solution (pH 9, Dako, Carpinteria, CA) for double immunofluorescent staining of β-tubulin and phosphor-histone H3, or with sodium citrate buffer (pH 6) for others. Next, tissue sections were immersed in 3% H2O2/methanol solution and incubated with 10% normal goat serum. The tissue sections were then simultaneously incubated overnight at 4 °C with sets of two or three primary antibodies. The sections were simultaneously incubated with Alexa Fluor 488-labeled anti-mouse antibody (1:100, Jackson ImmunoResearch, Inc.) and Rhodamine Red-X-labeled anti-rabbit antibody (1:50, Jackson ImmunoResearch, Inc.), or Rhodamine Red-X-labeled anti-mouse antibody (1:50 or 1:100, Jackson ImmunoResearch, Inc.) and Alexa Fluor 488-labeled anti-rabbit antibody (1:100, Jackson ImmunoResearch, Inc.) for 1 h at room temperature. VECTASHIELD HardSet Antifade Mounting Medium with DAPI (Vector Laboratories) was used to embed the slides before observation with an All-in-One Fluorescence Microscope BZ-X710 (Keyence Corp., Osaka, Japan). For quantitative analyses of emerin- or barrier-to-autointegration factor (BAF)-overexpressing hepatocytes in animals at Week 4, at least ten fluorescent images at 40 × magnification were randomly obtained per animal and the number of positive cells among at least 1000 hepatocytes in each image was counted. For quantitative analyses focusing on LMN, slides for all animals in the 2000 mg/kg acetamide group on Day 3 and the 1.25% group at Week 4 were examined. The number of LMNs among 50–100 LMNs from different areas on each slide that were positive for primary antibody-positive LMNs were counted. β-catenin was used to visualize cell membranes. Double immunofluorescent staining images of β-tubulin and phosphor-histone H3 were acquired at 63 × magnification using a Zeiss LSM710NLO confocal laser scanning microscope (Carl Zeiss, Jena, Germany). The fluorescence intensities were analyzed using Zeiss ZEN software (Carl Zeiss).

For electron microscopic examination, small portions of liver tissue fixed in 10% neutral-buffered formalin were additionally fixed with 2.5% glutaraldehyde and postfixed with 1% OsO4. The samples were then dehydrated using ascending grades of alcohol and embedded in epoxy resin. Ultrathin sections were stained with an EM Stainer (Nisshin EM CO., Ltd., Tokyo, Japan) and lead citrate followed by examination using a JEM 1400 transmission electron microscope (JEOL Ltd., Tokyo, Japan).

Small portions of livers from all animals at Week 4 were homogenized with a Teflon homogenizer with ice-cold RIPA lysis buffer (Fujifilm Wako Pure Chemical Co.) containing mammalian protease inhibitor cocktail and phosphatase inhibitor cocktails 2 and 3 (Sigma-Aldrich). Samples were homogenized and centrifuged at 15,000 × g for 30 min, and the resulting supernatants were used. Protein concentrations were determined using the Advanced Protein Assay (Cytoskeleton, Denver, CO, USA) with bovine serum albumin as a standard. Samples were separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to 0.45 µm PVDF membranes (Millipore, Billerica, MA, USA). Membranes were incubated at 4 °C overnight with primary antibodies targeting STING (anti-TMEM173/STING rabbit polyclonal antibody, diluted 1:1000; #19851-1-AP; Proteintech Group, Inc., Rosemont, USA), TBK1 (anti-TBK1/NAK rabbit monoclonal antibody diluted 1:1000; #3504), phospho-TBK (Ser172) (anti-phospho-TBK1/NAK (Ser172), diluted 1:1000; #5483, Cell Signaling Technology) and β-actin (anti-β-Actin rabbit polyclonal antibody, diluted 1:1000; #4967, Cell Signaling Technology). Secondary antibody incubation was performed using horseradish peroxidase-conjugated secondary anti-rabbit (1:2000; #7074) or anti-mouse antibodies (1:2000; #7076, Cell Signaling Technology, Inc.) for 1 h at room temperature. Protein detection was facilitated by chemiluminescence using ECL Prime (GE Healthcare Japan Ltd., Tokyo, Japan). Densitometric analysis of protein or phosphorylated protein signals were normalized to β-actin levels in the same tissue sample. Protein extracted from a F344 rat thymus was used as a positive control.

Microtubule polymerization assays were performed with a fluorescence-based tubulin polymerization assay kit (Cytoskeleton, Denver, CO, USA) according to the manufacturer's instructions. Acetamide was diluted in distilled water at a concentration of 0.01, 0.1, 1, or 10 mM. Colchicine (Tokyo Chemical Industry) was diluted to 10 μM in distilled water.

Small portions of livers taken from animals in the vehicle control group and the group treated with 2000 mg/kg acetamide on Day 3 were soaked overnight in RNAlater-ICE (Life Technologies) at − 20 °C, and total RNA was extracted using an ISOGEN kit (Nippon Gene Co) according to the manufacturer's instructions. cDNA copies of total RNA were obtained using a High Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific, Waltham, MA, USA). All PCR reactions were performed with primers (Assay ID) for rat Aurka (Rn00821847_g1), Aurkb (Rn01460656_m1), Plk1 (Rn00690926_m1) and Bub1 (Rn01484954_m1), and TaqMan Rodent GAPDH Control Reagents (#4308313, Thermo Fischer Scientific) as an endogenous reference. PCR was carried out with an Applied Biosystems 7900HT FAST Real-Time PCR System using TaqMan Fast Universal PCR Master Mix and TaqMan Gene Expression Assays (Thermo Fischer Scientific). Expression levels of the target gene were calculated using the relative standard curve method and were determined as ratios to Gapdh levels. Data are presented as fold-change value of treated samples relative to controls.

Variance in data for body and liver weights, % tail DNA, emerin- or BAF-overexpressing hepatocytes and protein expression levels was checked for homogeneity by Bartlett's test. Dunnett's multiple comparison tests and Steel's test were applied for homogenous and heterogenous data, respectively. The data for MNPCE/NCE ratio, MNHEP/HEP ratio and large MNHEP/HEP ratio was analyzed with a conditional binomial test using Kastenbaun and Bowman's table (Kastenbaum and Bowman [15]). For variance in data on immunofluorescent analyses focusing on LMN, homogeneity was tested by an F test. Student's t test was conducted for homogeneous cases, whereas Welch's t test was applied for heterogeneous cases. The significance of differences from positive control data in the comet assay was analyzed with Student's t test.

To examine the potential of acetamide to induce CAs in rat liver, we performed an in vivo MN assay in both liver tissue and bone marrow after a 4-week administration of acetamide at doses including carcinogenic doses of 1.25% and higher (Jackson and Dessau [14]). During administration, body weight gain was significantly decreased in the 2.5% group from Week 2, whereas food consumption did not obviously differ among the groups (Supplementary Fig. S1A, B). After 4 weeks of administration, the final body and liver weights were significantly decreased for the 1.25% and 2.5% groups (Supplementary Table S2), and histopathological examination of the liver showed various hepatic changes in both groups (Supplementary Table S3). These results corresponded with our previous 13-week acetamide-administration study (Nakamura et al. [21]). In the MN assay, none of the treatment groups showed significant changes in MN erythrocytes, whereas the 1.25% and higher groups had significant increases in MN hepatocytes (Fig. 1A, B). These data indicated that acetamide specifically induced CAs in the target organ at carcinogenic doses. To examine the involvement of direct DNA damage in MN formation, we conducted an in vivo liver comet assay for a 3-day oral administration of acetamide. No changes in % tail DNA were observed even at the maximum dose of 2000 mg/kg (Fig. 1C), suggesting that acetamide can induce CAs without inducing direct DNA damage.

Graph: Fig. 1 Acetamide specifically induces formation of micronuclei in hepatocytes without direct DNA damage. A Representative photos of micronucleated hepatocytes (MNHEPs) in in vivo liver micronucleus assay (left panel, yellow arrow; micronucleus). Dashed line indicates the cell border. The ratio of MNHEP/HEP is quantified on the right. B Quantification of the ratio of micronucleated erythroblasts/erythroblasts in an in vivo bone marrow micronucleus assay. C % tail DNA in an in vivo liver comet assay. *, **, ##: Significantly different from the respective control groups at p < 0.05, 0.01, 0.01 respectively. Scale bar is 10 µm

In an in vivo liver MN assay, hepatocytes with LMN were observed in the groups given 0.625% acetamide and higher (Fig. 2A). LMNs were also observed in the aneugen COL group, but not in the clastogen MMC group (Fig. 2A), and most were positive for centromere probes in a FISH analysis (Fig. 2B). Abnormal mitoses of hepatocytes, such as lagging chromosome and chromosomal misalignment, were also observed in acetamide-treated livers examined in the MN assay and immunofluorescent analysis (Supplementary Fig. S2A, B). These results suggested that acetamide could exhibit aneugenic action. Histopathological examination of the livers from groups given 0.625% and higher doses at Week 4 revealed formation of basophilic cytoplasmic inclusion (CI) (Fig. 2D, Supplementary Table S2). Since the CI, which is composed of nuclear components (Nakamura et al. [21]), corresponded morphologically to LMN, we found that the LMN was detectable histopathologically as CI. The LMN was observed not only in non-dividing hepatocytes but also in dividing cells, where the LMN was often located at the end of the M-phase plate (Fig. 2C). Interestingly, double immunofluorescent staining of histone H3 phosphorylated at Serine 10, an M-phase marker, and β-tubulin, a spindle fiber marker, revealed that the LMN in dividing cells expressed phosphorylated histone H3 at levels similar to those of normal mitotic chromosomes, but the spindle fibers were not attached to the LMN (Fig. 2D). Thus, the LMN seemed to be derived from abnormal mitotic chromosomes. Notably, the LMN was also induced by a 3-day administration of acetamide, which did not induce hepatic injury or increase mitosis (Supplementary Table S4), suggesting that the LMN was induced as a primary effect of acetamide. Taken together, the LMN appeared to be derived from abnormal mitosis caused by direct effects of acetamide.

Graph: Fig. 2 Histopathological detection of centromere-positive LMN as cytoplasmic inclusions derived from mitosis. A Representative images of large MNHEPs in an in vivo liver micronucleus assay (left panel). Dashed line indicates the cell border. The ratio of large MNHEP/HEP is quantified on the right. B Representative images of LMNHEPs induced by acetamide and colchicine (COL) treatment detected in fluorescence in situ hybridization (FISH) with a centromere probe (upper panels). The ratio of centromere-positive LMN in up to 30 LMN is quantified in the lower panel. C The LMN is detected histopathologically as basophilic cytoplasmic inclusions (a). In dividing cells, the LMN is frequently located at the end of the M-phase plate (b). D The LMN expresses phospho-histone H3 (Ser10), but spindle fibers detected by β-tubulin were not attached to it (left panel). The relationship of colocalization of phospho-histone H3 and β-tubulin in the M-phase plate is plotted on the right. All yellow arrowheads represent LMN. Data represent the mean ± SD. All scale bars are 10 µm

To clarify the cause of abnormal mitoses, we first assessed how acetamide affected tubulin polymerization in a microtubule polymerization assay. Acetamide had no effect on tubulin polymerization at the maximum concentration of 10 mM (Supplementary Fig. S2C). Real-time RT-PCR analyses of liver tissue after 3-day administration of acetamide at 2000 mg/kg/day also revealed no significant changes in expression of Aurka, Aurkb, Plk1 and Bub1, which are activated at the spindle assembly checkpoint (SAC) in abnormal mitosis (Maillet et al. [18]) (Supplementary Fig. S2D). These results indicated that acetamide did not directly affect spindle formation in mitoses. On the other hand, acetamide treatment resulted in altered nuclear shape in hepatocytes at Week 4 (Supplementary Fig. S3A, B), which corresponded with our previous report (Nakamura et al. [21]) and suggests that it affects the nuclear envelope (NE). Since changes in the NE could be responsible for chromosomal mis-segregation (Smith et al. [29]; Liu et al. [17]), we next performed immunohistochemical analyses of the expression of NE-associated proteins, such as emerin, an inner NE protein, and BAF, an adaptor protein that anchors chromatin with inner NE proteins in the primary nuclei. Immunofluorescent staining demonstrated that the number of hepatocytes having abnormal expression of emerin or BAF in nuclei was significantly increased in groups given 1.25% and higher acetamide at Week 4 (Supplementary Fig. S3C, D). In addition, a slight, but not significant, increase in hepatocytes having abnormal BAF expression was observed in the 0.625% group at Week 4 (Supplementary Fig. S3D). Hepatocytes containing abnormal NE were also frequently observed on Day 3 (Supplementary Fig. S3E—G). Taken together, these data suggested that acetamide affects NE-associated proteins in hepatocytes to alter nuclear morphology and cause abnormal mitosis.

To clarify the potential involvement of CAs in acetamide-induced tumorigenesis, we focused on the histopathologically detected LMN and examined its biological significance. We first performed immunofluorescent analyses for NE-associated proteins, such as lamin B1 and lamin A/C, as well as the lamin-interacting proteins emerin and BAF. Strikingly, whereas most of the LMNs partially or fully expressed lamin B1 in the NE on Day 3, most of the LMNs had lost lamin B1 expression at Week 4 (Fig. 3A). Similarly, lamin A/C was also depleted in these LMNs (Fig. 3B). LMNs showed emerin overexpression on Day 3 and Week 4, and the number of emerin-expressing LMNs was decreased at Week 4 (Fig. 3C). In contrast, the region of BAF overexpression was expanded to encompass the entire LMN at Week 4 (Fig. 3D). Changes in the expression of these NE-associated proteins, which are essential for NE integrity (Smith et al. [29]; Liu et al. [17]), suggested that NE integrity was lost in the LMN.

Graph: Fig. 3 LMN shows abnormal expression of nuclear envelope-associated proteins. A LMN shows three expression patterns for lamin B1: fully positive (a), partially positive (b), or fully negative (c). The upper panels show respective illustrations and representative images of these expression patterns. The lower panel shows the average ratio of LMN with each lamin B1 expression pattern on Day 3 and Week 4 (n = 5 in each group). At least 50 LMNs were analyzed for each animal. B Representative image of LMN having concurrent loss of lamin B1 and lamin A/C expression at Week 4 (upper panels). The average ratio of LMN showing lamin B1 and lamin A/C double positive, lamin B1 or lamin A/C single positive, and lamin B1 and lamin A/C double negative staining at Week 4 (n = 5) is shown in the lower panel. At least 50 LMNs were analyzed for each animal. C LMN shows overexpression of emerin compared with the primary nucleus (upper panel). The average ratios of emerin-overexpressing LMN on Day 3 and Week 4 are quantified in the lower panel. D The LMN mainly shows two staining patterns of barrier-to-autointegration factor (BAF); partial staining (a) or pan-staining (b). The lower panel shows the average ratio of BAF-partial or pan-staining LMN on Day 3 and Week 4 (n = 5 in each group). At least 50 (A, B, and D) or 100 (C). LMN were analyzed for each animal. Data represent the mean ± SD. *, **: Significantly different from the 0% group at p < 0.05 and 0.01, respectively. All scale bars are 10 µm

As NE-associated proteins are essential to maintain chromatin structure (van Steensel and Belmont [38]; Ranade et al. [25]), we further investigated the chromatin states of LMN. Stronger staining intensity of SYBR Gold was seen in the LMN instead of the primary nucleus (Fig. 2A). Regions of the NE with lamin B1 depletion or BA overexpression in LMN displayed stronger DAPI staining intensity (Fig. 4A, B), which corresponded to high-electron density nuclear structures demonstrated by electron microscopy (Fig. 4C). These results implied that the LMN contains condensed chromatin, termed heterochromatin. Indeed, immunofluorescent analysis showed strong intensity of H3K9me3, a heterochromatin marker, in LMN on Day 3 and the area of intense staining had expanded at Week 4 (Fig. 4D). Correspondingly, LMNs that were negative for H4K8ac, a euchromatin marker, were significantly increased at Week 4 (Fig. 4E). Altogether, the abnormal morphology of NE resulted in altered chromatin structures in LMN, leading to excessive chromatin condensation.

Graph: Fig. 4 LMN shows more heterochromatinization associated with NE abnormality. A Representative images of LMN showing partially positive (upper left) or fully negative lamin B1 (lower left) staining. The relationship of lamin B1 and DAPI staining intensities in the former and latter are shown in the upper and lower right, respectively. PN; primary nucleus. B Representative images of BAF-partial (upper left) or pan-staining LMN (lower left). The relationship of BAF and DAPI staining intensities in the former and latter are shown in the upper and lower right, respectively. C Representative electron microscopic images of LMN on Day 3 (a, 12,000 × magnification) or Week 4 (b, 15,000 × magnification). Insertion represents NE with double membranes (30,000 × magnification). D The LMN shows two staining patterns for H3K9me3: partial staining (a) or pan-staining (b). The average ratios of H3K9me3-partial or pan-staining LMN on Day 3 and Week 4 are quantified in the lower panel (n = 5 in each group). E Representative image of LMN that do not express H4K8ac (upper panel). The lower panel shows the average ratio of H4K8ac-negative LMN on Day 3 and Week 4. At least 50 LMN were analyzed for each animal (D and E). Data represent the mean ± SD. **: Significantly different from the 0% group at p < 0.01. Scale bars in A, B, D and E are 10 µm

To examine the influence of the altered NE and chromatin structure on genomic DNA in LMN, double immunohistochemical staining of histone H2AX phosphorylated at serine 139 (γ-H2AX), a DNA damage marker, and 53BP1, a DNA repair marker, was conducted. Surprisingly, the LMN showed pan-staining of γ-H2AX (Fig. 5A), suggesting massive DNA damage. Although single positive staining of 53BP1 was predominant in the LMN on Day 3, at Week 4 most LMNs showed single positive staining for γ-H2AX or double positive staining for γ-H2AX/53BP1 (Fig. 5A). These results indicated that longer duration of acetamide treatment caused accumulation of DNA damage in the LMN.

Graph: Fig. 5 Acetamide induces massive DNA damage in the LMN and increases expression of cGAS/STING pathway-associated proteins in the liver. A Representative images of LMN co-expressing γ-H2AX and 53MP1 (left panels). On the right the average ratio of γ-H2AX/53BP1 double negative, 53BP1 single positive, γ-H2AX single positive, or γ-H2AX/53BP1 double positive LMN on Day 3 and Week 4 (n = 5 in each group) is shown. At least 50 LMNs were analyzed for each animal. B Western blotting analyses of STING, TBK1, and phospho-TBK1 (Ser172) in the liver at Week 4 (n = 5 in each group). Data represent the mean ± SD. *, **: Significantly different from the 0% group at p < 0.05 and 0.01, respectively. Scale bar is 10 µm

The loss of NE integrity may lead to rupture of LMN. Ruptured MN in cancer cells are known to expose genomic DNA to the cytoplasm, which activates the cGAS/STING pathway, an innate immune inflammatory signal (Bakhoum et al. [2]; Hatch [11]; Dewhurst [6]). As such, we carried out immunoblotting analyses to examine the expression levels of STING and TBK1, a downstream protein activated by STING, in the liver of animals treated with acetamide for 4 weeks. The amounts of STING and TBK1 proteins were significantly increased in animals given 1.25% and higher doses of acetamide (Fig. 5B). Furthermore, expression levels of TBK1 phosphorylated at Serine 172 tended to increase in the 1.25% and higher dose groups, although these changes were not significant and dose dependence was not observed (Fig. 5B). In addition to the overexpression of STING and TBK1, increases in phosphorylated TBK1 protein implied activation of the cGAS/STING pathway in the liver, suggesting that the loss of NE integrity in LMN led to exposure of damaged DNA in the cytoplasm of hepatocytes.

To investigate the mechanisms of CAs underlying chemical carcinogenesis, we first assessed the potential of acetamide, a non-mutagenic hepatocarcinogen in rats, to induce CAs. An in vivo MN assay revealed that acetamide-induced MNs in hepatocytes, but not erythrocytes, showing that acetamide specifically induced CAs in rat liver at carcinogenic doses. In addition, centromere-positive LMNs and abnormal mitoses in hepatocytes that were seen in acetamide-treated rats were similar to those induced by the aneugen colchicine, a disruptor of microtubules. A liver comet assay showed that acetamide did not induce direct DNA damage in hepatocytes. These data indicated that acetamide likely induced CAs by interfering with spindle formation in mitoses. However, acetamide did not affect microtubule polymerization, and did not show SAC activation during the early phase of treatment. We revealed that hepatocytes from acetamide-treated animals had irregularly shaped nuclei even during the early phase of treatment. This nuclear morphology was accompanied by several features including absence of lamin A/C or abnormal expression of BAF, which together could impede proper chromosomal segregation (Qi et al. [24]; Liu et al. [17]). These results indicate that acetamide likely causes abnormal chromosomal segregation through induction of altered nuclear morphology in hepatocytes that in turn leads to CAs. Therefore, it was concluded that acetamide showed clastogenicity in rat liver, which seemed to be resulted from aneugenic action. Considering that acetamide did not show any in vivo mutagenicity (Nakamura et al. [21]) and DNA damages in the liver, acetamide would be classified as a non-mutagenic genotoxic carcinogen. The finding that acetamide-induced MNs in hepatocytes but not bone marrow erythrocytes implies that metabolic activation in the liver plays a crucial role in the induction of CAs (Uno et al. [37]; Hamada et al. [10]). Further investigations on the metabolic profiles and the underlying mechanism of aneugenic activity of acetamide will be needed.

The LMN detected in the liver MN assay correspond to CI that was histopathologically observed in the liver and was accompanied by abnormal mitosis, suggesting that this characteristic structure was derived from CAs. Immunofluorescent analyses focusing on this nuclear structure revealed that the LMN had several molecular pathological features that differed from primary nuclei. For example, the LMN showed abnormal expression of NE-associated proteins, and a corresponding increase in heterochromatinization. The changes of these features showed a time-dependent progression with acetamide treatment. Moreover, damaged genomic DNA contained in the LMN was evident from increased expression of γ-H2AX, which is known to be associated with shifts in 53BP1 expression in the later phase of MN formation (Tang et al. [32]), implying progression of DNA repair in MN. Although in this study most LMNs at 3 days of acetamide treatment only expressed 53BP1, after 4 weeks of acetamide treatment the number of LMNs expressing γ-H2AX and co-expressing γ-H2AX and 53BP1 was markedly increased. These results suggested that longer duration of acetamide treatment could cause DNA damage associated with insufficiency of DNA repair in LMN. Together with the fact that NE-associated proteins were depleted and abnormal chromatin structures are prone to DNA damage due to vulnerabilities to mechanical stress or obstructed recruitment of DNA repair proteins to damaged chromatin sites (Misteli and Soutoglou [19]; Terradas et al. [33]; Guo et al. [8]; Kneissig et al. [16]; Halfmann et al. [9]), the progressive alterations in NE and chromatin structure seen in the LMN were likely to promote accumulation of massive DNA damage.

Recent reports described drastic mutational chromosome rearrangements caused by the formation of MNs in cancer cells in a variety of human tumors (Liu et al. [17]; Cortés-Ciriano et al. [3]; Voronina et al. [40]). Interestingly, chromosome rearrangements such as chromothripsis or chromoplexy can occur in MNs and are associated with various features of MNs that differ from primary nuclei including abnormal NE and chromatin structures (Liu et al. [17]; Guo et al. [8]; Kneissig et al. [16]). The abnormal NE is prone to rupture and allow subsequent leakage of chromosomal DNA from the MNs to the cytoplasm (Guo et al. [8]; Dewhurst [6]). Substantial numbers of gene mutations would occur after reincorporation of rearranged chromosomes from MNs into primary nuclei through mitosis (Terradas et al. [33]; Guo et al. [8]). In the current study, we demonstrated by immunofluorescence and electron microscopy analyses that LMNs had the same molecular pathological features as those reported for MN in multiple in vitro studies (Hatch et al. [13]; Liu et al. [17]; Guo et al. [8]; Kneissig et al. [16]). In addition, several MNs were detected histopathologically in hepatocytes that possessed the same features as LMNs (Supplementary Fig. S4). Immunoblotting analyses indicated that acetamide could activate the cGAS/STING innate immune pathway in the liver that could involve generation of cytoplasmic DNA in hepatocytes. If damaged genomic DNA is incorporated into the primary nucleus of a daughter hepatocyte after mitosis, massive gene mutation via chromosome rearrangements might occur. In our previous study using gpt and Spi− assay, acetamide did not showed any in vivo mutagenicity in the liver (Nakamura et al. [21]). This result implied that these assays could not detect mutations at chromosome levels and/or there were few occurrences of chromosomal rearrangements in non-tumoric liver. Taken together, our results suggested that both LMNs and MNs induced by acetamide could cause hepatocarcinogenesis through chromosome rearrangements.

Some groups hypothesized that chromosome rearrangements that arise through the formation of MNs might underlie chemical carcinogenesis (Guo et al. [8]; Tweats et al. [35]; Pellestor and Gatinois [23]). In addition, a recent report of results for liver MN assays revealed that not only genotoxic hepatocarcinogens, but also some non-genotoxic hepatocarcinogens, could induce MNs in hepatocytes (Hamada et al. [10]). This finding implies that CAs contribute to tumorigenesis by non-genotoxic carcinogens. However, there currently is no direct evidence of the mechanism of chemical carcinogenesis originating from MN formation, since few carcinogens have genotoxic potential that manifests only as CAs. In addition, animals have lower tolerance for long periods of aneugen administration (Shigano et al. [28]) and detection sensitivity of MNs is lower in chemical-treated animal models. Since in vivo and in vitro cancer cells obtained from human cancers are already transformed (Zhang et al. [42]), animal models are essential for rigorous investigation of the involvement of CAs in carcinogenesis including tumor initiation. In the current study, we found that the non-mutagenic hepatocarcinogen acetamide was a pure inducer of CAs. Moreover, the detection of LMNs by histopathology enabled clarification of the biological significance of MNs in animals and the involvement of them in acetamide-induced hepatocarcinogenesis. Together with the fact that long-term administration of acetamide until tumorigenesis occurs is well-tolerated in rats (Jackson and Dessau [14]; Fleischman et al. [7]), a rat model of acetamide-induced hepatocarcinogenesis could be a competent animal model for understanding the mechanisms underlying tumorigenesis originating from CAs. Overall data indirectly but strongly suggested that CAs induced by acetamide play a crucial role to hepatocarcinogenesis. To reveal the direct contribution of CAs to carcinogenesis, we are currently using whole genome sequencing in acetamide-induced rat liver tumors.

In conclusion, the current study revealed that the non-mutagenic hepatocarcinogen acetamide-induced CAs specifically in rat liver, and LMNs derived from CAs could be histopathologically detected as CI in hepatocytes. The LMN showed abnormal expressions of NE proteins, increased heterochromatinization, and massive DNA damage, and these changes were progressed with time-dependency on duration of treatment. Overall, the data from this study implicate the involvement of chromosome rearrangements in acetamide-induced hepatocarcinogenesis, suggesting that CAs play a critical role in this process in rats. This study using a rat model treated with acetamide provides strong evidence that CAs result in chemical carcinogenesis in animals via formation of MNs.

We thank Ms. Ayako Saikawa and Ms. Yoshimi Komatsu for expert technical assistance in processing histological materials.

Kenji Nakamura is an employee of Ono Pharmaceutical Co., Ltd., Osaka, Japan. The authors declare no conflicts of interest associated with this manuscript.

Below is the link to the electronic supplementary material.

Graph: Supplementary file1 (DOCX 55 kb)

Graph: Supplementary file2 (DOCX 8219 kb)

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.