2018 Volume 43 Issue 6 Pages 359-367
Screening prostatic carcinogens is time-consuming due to the time needed to induce preneoplastic and neoplastic lesions. To overcome this, we investigated alternative molecular markers for detection of prostatic carcinogens in a short period in rats. After treatment with 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), expression of high-mobility group protein B2 (HMGB2) was up-regulated in rat ventral prostate. To evaluate the applicability of HMGB2 in the early detection of carcinogenicity of chemicals using animal models, we examined HMGB2 expression in prostate of rats. Six-week-old male F344 rats were gavaged for four weeks with a total of eight individual chemicals, divided into two categories based on prostate carcinogenicity. Animals were sacrificed at the end of the study and HMGB2 immunohistochemistry was performed. HMGB2 expression in least one prostate lobe was significantly increased by all four prostate carcinogens compared with the controls. In contrast, the four chemicals that were not carcinogenic in the prostate did not cause HMGB2 up-regulation. Additionally, high HMGB2 expression in neoplastic lesions in both rat and human was detected. Therefore HMGB2 expression may be a good screening tool for the identification of potential of prostate carcinogens.
The standard method used to assess the carcinogenic potential of chemicals is a two-year carcinogenicity study with rodents. However, this model bioassay is time-consuming and extremely expensive to conduct. Additionally, the previous models of prostate carcinogenesis with carcinogenic chemicals require long periods to induce neoplastic lesions in the rodent prostate (Shirai et al., 1985, 1997). Consequently it takes a long time to determine that a chemical is a prostate carcinogen with animal models. The development of more appropriate analytical methods and procedures for the identification of novel biomarkers, that can be performed simultaneously with general short-term toxicity studies, is needed. To this end, we investigated alternative molecular markers for detection of prostatic carcinogens in a short period in rats.
To investigate the new marker for detection of prostate carcinogens, we analyzed mRNA expression in the ventral prostate of rats treated by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), which is known to be a prostate carcinogen (Shirai et al., 1997) and administrated to humans via food (Wakabayashi et al., 1993), and detected the high mobility group box 2 (HMGB2) as an up-regulated gene. HMGB2 is one of the HMG protein families, acting as chromatin-binding factor that binds DNA and promotes access to transcriptional protein assemblies on specific targets (Bianchi and Agresti, 2005; Agresti and Bianchi, 2003). We previously reported that HMGB2 is highly expressed in hepatocellular carcinoma compared to adenoma and related to cell proliferation (Suzuki et al., 2009). We also focused on Ki-67 as cell proliferation marker because chemicals are first evaluated for proliferative activity in various tissues (Cohen and Arnold, 2016).
In the present study, we performed immunohistochemical analysis of HMGB2 expression in chemically treated rats to evaluate the potential of HMGB2 expression as a screening tool for identifying potential prostate carcinogens. In experiments 1 and 2, DNA microarray analysis was used to screen for molecular markers for the early detection of prostate carcinogens in samples from rats treated with a positive control chemical (PhIP). In experiment 3, a candidate marker (HMGB2) was immunohistochemically evaluated in samples from rats treated with two other positive control chemicals (DMAB and MNU) via different administration routes. In experiments 4 and 5, samples from rats treated with known prostate carcinogens (PhIP, BOP, DMAB and MNU) and carcinogens that target non-prostate tissues (MNNG, DMBA, DMN and DMH) were analyzed by using a toxicological program (Repeated Dose 28 Day Oral Toxicity Study in Rodents) in order to consider the 3Rs (Replacement, Reduction and Refinement).
PhIP was synthesized in the NARD Institute (Osaka, Japan), with a purity of 99.9%. N-nitrosobis(2-oxopropyl)amine (BOP) was obtained from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). 7,12-dimethylbenz[a]anthracene (DMBA) and N-methyl-N-nitrosourea (MNU) were obtained from Sigma-Aldrich Co. LLC (St. Louis, MO, USA). N-methyl-N’-nitro-N-nitrosoguanidine (MNNG), dimethylnitrosamine (DMN) and 1,2-dimethylhydrazine (DMH) were obtained from Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). 3,2’-dimethyl-4-aminobiphenyl (DMAB) was obtained from Matsugaki Pharmaceutical Co. (Osaka, Japan).
The animal experiment was performed in compliance with protocols approved by the Institutional Animal Care and Use Committee of Nagoya City University School of Medical Sciences. Five-week-old male F344 rats were obtained from Charles River Japan (Kanagawa, Japan). They were housed in plastic cages with hardwood chip bedding in an air-conditioned room at 23 ± 2°C and 55 ± 5% humidity with a 12-hr light/dark cycle and maintained on a basal certified diet (Oriental MF; Oriental Yeast Co., Tokyo, Japan) and tap water ad libitum.
Six-week-old rats received an intragastric injection of PhIP at a dose of 300 mg/kg body weight as an initiation procedure. The dose of PhIP was sufficient to induce rat prostate carcinoma (Shirai et al., 1997; Hikosaka et al., 2004). Two rats were sacrificed under anesthesia at days 1, 2 and 7 for experiment 1, and four rats at day 2 for experiment 2. Two and four non-treated rats were also sacrificed by exsanguination under anesthesia for experiments 1 and 2, respectively. The urogenital complex of each rat was removed as a whole together with the seminal vesicles, then the ventral prostate was weighed. A part of the prostate gland was immediately frozen in liquid nitrogen and stored at -80°C until it was processed for experiment 1. The remainder of the prostate in experiment 1 and all prostate gland for experiment 2 were fixed in 10% phosphate-buffered formalin. The liver and kidneys were removed and fixed in 10% phosphate-buffered formalin for experiment 2. The tissues were routinely processed to paraffin embedded sections and stained with hematoxylin and eosin (H&E).
Six-week-old rats received an intragastric or subcutaneous injection of DMAB at a dose of 200 mg/kg body weight or an intragastric or intraperitoneal injection of MNU at a dose of 50 mg/kg body weight as an initiation procedure. The dose of each chemical is approximately related to the carcinogenic dose for rat prostate carcinogenesis (Bosland and Prinsen, 1990; Gonçalves et al., 2013). At experimental day 2, rats were sacrificed by exsanguination under deep isoflurane anesthesia. The urogenital complex of each rat was removed as a whole together with the seminal vesicles. The tissues were routinely processed to paraffin embedded sections and stained with H&E.
The present study was performed as two separate experiments using the same protocols (experiments 4 and 5). Animals were randomly divided into five groups of four to six rats each. Five times per week for 4 weeks, the animals were given the following chemicals by gavage (corn oil as the vehicle): PhIP (15 mg/kg), BOP (5 mg/kg), MNNG (10 mg/kg) or DMBA (5 mg/kg) in experiment 4 and DMAB (5 mg/kg), MNU (5 mg/kg), DMN (2 mg/kg) or DMH (5 mg/kg) in experiment 5. PhIP, BOP, DMAB and MNU were reported as prostate carcinogens (Shirai et al., 1997, 2000). MNNG (stomach, esophagus and small intestine), DMN (liver and kidney), DMH (colon and liver) and DMBA (mammary and skin) were reported as carcinogens, but not as prostate carcinogens when administrated alone (Zusterzeel et al., 1999; TOXNET: https://toxnet.nlm.nih.gov/; The Carcinogenic Potency Database: https://toxnet.nlm.nih.gov/cpdb/index.html). The dose of each chemical was defined with approximately 1/20 of LD50 of each chemical (Huggins and Sugiyama, 1966; Habs et al., 1978; TOXNET: https://toxnet.nlm.nih.gov/; Concise International Chemical Assessment Documents: http://www.who.int/ipcs/publications/cicad/en/). The method was developed to investigate chemicals for their carcinogenic potential on the rat prostate. Body weight was measured weekly. At experimental week 4, rats were sacrificed by exsanguination under deep isoflurane anesthesia, and the urogenital complex of each rat was removed as a whole together with the seminal vesicles, then the ventral prostate was weighed. The tissues were routinely processed to paraffin embedded sections and stained with H&E.
Neoplastic lesions of ventral prostate were utilized from previous studies as follows: 6-week-old animals were gavaged with PhIP (200 mg/kg) twice per week for 10 weeks, and then sacrificed at week 60 (Hikosaka et al., 2004). The 6-week-old animals were subcutaneously injected with DMAB (50 mg/kg) once per fortnight for 20 weeks, and then sacrificed at week 60 (Shirai et al., 1995).
Forty-eight formalin-fixed paraffin-embedded samples of organ-confined prostate cancers were obtained from patients undergoing total prostatectomy. None of the cancer patients had previously undergone chemotherapy, radiotherapy, or hormonal therapy. The use of study specimens for analyses was approved by the research ethics committee of Nagoya City University School of Medical Sciences.
Prostate sections were treated with rabbit monoclonal HMGB2 antibody (Abcam plc, Cambridge, UK), rabbit monoclonal Ki-67 for rats or mouse monoclonal Ki-67 for human (Agilent Technologies, Santa Clara, CA, USA), followed by staining with BOND-MAX (Leica Biosystems, Wetzlar, Germany) according to the manufacturer’s instructions. The number of HMGB2- or Ki-67-labeled cells in at least 1000 luminal cells in each prostate lobe was counted to determine labeling indices.
Total RNA of two samples each from the PhIP and control group was isolated from ventral prostate tissues of rats two days after PhIP treatment en bloc by phenol-chloroform extraction (ISOGEN, Nippon Gene Co. Ltd., Toyama, Japan). Gene expression analysis was performed using a Rat Genome 230 2.0 Array (Affymetrix, Santa Clara, CA, USA) according to the manufacturer’s instructions. The RNA expression of ventral prostate from PhIP-treated rats was compared with that from control rats. After global median normalization, data cleansing was performed to remove the values for which fluorescence intensity was less than 100. The genes for which expression were more than two-fold increased, or reduced to less than half, in PhIP-treated rats as compared to control rats were selected.
Total RNAs from ventral prostate tissues were reverse-transcribed with the PrimeScript® RT reagent Kit (Takara Bio Inc., Shiga, Japan), and real-time RT-PCR was performed using a LightCycler (Roche Diagnostics GmbH, Mannheim, Germany). The quantitative value of clusterin was normalized to endogenous cyclophilin. HMGB2 RT-PCR primers were 5’- CGTCTGCCTTCTTCCTGTTT-3’ and 5’- TCTTCTTTGAGCCTGTTGGC-3’. Cyclophilin RT-PCR primers were 5’-TGCTGGACCAAACACAAATG-3’ and 5’-GAAGGGGAATGAGGAAAATA-3’.
Statistical analyses were performed with mean ± standard deviation (S.D.) values using 1-way ANOVA and Dunnett’s test by Prism ver. 6 (GraphPad Software, Inc., La Jolla, CA, USA). P < 0.05 was considered statistically significant. The Pearson correlation coefficient was used to estimate the relationship between HMGB2 and Ki-67 correlation in rat prostate tissue and neoplasm, and human prostate cancer.
In experiment 1, there was less histological difference between the time points. In ventral prostate, the highest positivity of Ki-67 labeling index was detected at day 2 after PhIP injection (Fig. 1). Meanwhile, there was less difference of Ki-67 labeling indices among the time points in both lateral and dorsal prostates (Fig. 1). Therefore, we analyzed mRNA expressions of ventral prostate in rats treated with PhIP at day 2.
Ki-67 labeling indices in prostate lobes with PhIP-treatment. After the PhIP treatment, rats were sacrificed at days 1, 2 and 7. Ki-67 labeling index each lobe with time course was plotted.
In experiment 2, to investigate alternative molecular markers for detection of prostatic carcinogens, microarray analysis was performed (GEO: GSE95214). Genes which were up- or down-regulated by PhIP are listed in Table 1. After selection, 27 up-regulated and 11 down-regulated genes were detected, some related to cell cycle (cyclin-dependent kinase inhibitor 1A (p21), cell division cycle 2 homolog A (S.pombe), cyclin-dependent kinase regulatory subunit 2, cell division cycle associated 3, cyclin B2 and A2). Among these genes, we focused on high mobility group box 2 (HMGB2), because we previously detected HMGB2 as prognostic marker in hepatocarcinogenesis (Suzuki et al., 2009). The expression of HMGB2 was confirmed with Real-Time RT-PCR. The expressions of HMGB2 in ventral prostate of PhIP-treated rats were 2.4 times higher than that of control rats, but statistical analysis could not be performed due to the number of samples.
With the prostate tissue in experiment 2, the HMGB2 labeling index in ventral prostate of PhIP-treated rats (25.8 ± 3.7%) was significantly higher than that of controls (16.1 ± 1.6%; P < 0.01). Meanwhile, there was less difference of HMGB2 labeling indices in lateral (28.6 ± 2.6 and 27.6 ± 2.2%) and dorsal (28.8 ± 2.1 and 27.1 ± 0.9%) prostate between PhIP-treated and control rats, respectively. HMGB2 labeling indices in liver (6.0 ± 0.9 and 5.8 ± 1.0%) and renal tubules of kidneys (10.5 ± 0.9 and 10.5 ± 1.0%) were not different between PhIP-treated and control rats, respectively. The Ki-67 labeling index in ventral prostate of PhIP-treated rats (8.6 ± 2.5%) was also significantly higher than that of controls (2.57 ± 0.9%; P < 0.01). There was less difference of Ki-67 labeling indices in lateral (3.9 ± 1.2 and 2.8 ± 1.8%) and dorsal (3.2 ± 1.3 and 2.8 ± 1.1%) prostate between PhIP-treated and control rats, respectively.
To clarify the difference in treatment route, DMAB and MNU were administered by gavage or prostate carcinogenesis route (subcutaneous injection for DMAB and intraperitoneal injection for MNU). In the results, HMGB labeling indices were significantly up-regulated by both chemicals with both treatment routes (Table 2). Meanwhile, Ki-67 labeling indices were significantly up-regulated by only DMAB with both treatment routes (Table 2). Interestingly, despite the treatment method used, HMGB2 and Ki-67 indices were similar in each prostate lobe (Table 2).
*, **, ***: P < 0.05, 0.01, 0.001, compared to Control.
To consider the 3Rs (Replacement, Reduction and Refinement), we investigated the detection method of prostatic carcinogens with a toxicological program, and focused on the Repeated Dose 28-Day Oral Toxicity Study in Rodents (OECD Test Guideline 407). We utilized prostate carcinogens (PhIP, BOP, DMAB and MNU) and carcinogens which target non-prostate tissues (MNNG, DMBA, DMN and DMH) for this study. In experiments 4 and 5, body weight gain was significantly reduced in MNU-treated rats compared with that of the control at week 4 (Table 3). Relative ventral prostate weight was significantly reduced in DMBA-treated rats compared with that of the control at week 4 (Table 3). There were no histological differences in prostate tissue among the groups (data not shown). Prostate carcinogens significantly increased HMGB2 positivity in at least two prostate lobes compared to the control in rats (Table 4). PhIP induced high HMGB2 expression in ventral lobe compared to other lobes. BOP significantly induced HMGB2 expression in lateral and dorsal lobes. Significant expression of HMGB2 was detected in all lobes of DMAB- and MNU-treated rats. Prostate carcinogens also increased Ki-67 positivity in prostate lobes compared to the control in rats (Table 4). HMGB2 positivity has a strong correlation with Ki-67 positivity (Pearson r of ventral, lateral and dorsal lobes; 0.63, 0.74 and 0.84, respectively, P < 0.001). Meanwhile, there was no statistical significant difference in HMGB2 and Ki-67 positivity after exposure to non-prostatic carcinogens in all prostate lobes compared with respective controls (Table 4).
*: Significantly different from Control group, p < 0.05.
*, **, ***: Significantly different from Control group, p < 0.05, 0.01, 0.001.
For further analysis, we examined expression level of prostate neoplasms, with chemically induced rat prostate neoplasia and human prostate cancer samples. The HMGB2 labeling index in neoplasia of rat ventral prostate induced by PhIP and DMAB was significantly higher (P < 0.001 and 0.01, respectively) than that in non-neoplastic area (Figs. 2A and B). A correlation between HMGB2 and Ki-67 positivities in rat prostate neoplasia was not detected (P > 0.05), possibly due to the small sample size. In human prostate samples, the HMGB2 labeling index was also significantly higher (P < 0.001) than in non-neoplastic epithelial cells (Figs. 2A and B). Additionally, a correlation between HMGB2 and Ki-67 positivities in human prostate cancer was detected (Pearson r = 0.44, P < 0.01).
HMGB2 labeling indices in prostate neoplastic lesions of rat and human. Photographs (A) and data (B) for immunohistostaining of HMGB2. Percentage values were expressed as the mean ± S.D.: P < 0.01 (**) and 0.001 (***) compared to the each normal area, respectively.
In the present study, we investigated HMGB2 expression profiles induced by eight chemicals exhibiting carcinogenic/noncarcinogenic characteristics to the prostate using the repeated dose 28-day oral toxicity study. Our data clearly demonstrated that all four prostate carcinogens (PhIP, BOP, DMAB and MNU) significantly increased the number of HMGB2-positive cells after four weeks of administration. In contrast, the four chemicals that are not carcinogenic to prostate (MNNG, DMBA, DMN and DMH) did not induce this change.
HMGB2 has been reported to regulate cellular proliferation in other cancers (Suzuki et al., 2009; Wu et al., 2013; Syed et al., 2015; Shin et al., 2013). To consider the relationship between HMGB2 and Ki-67 expression in the prostate, high correlations between HMGB2 and Ki-67 indices were detected in all prostate lobes of experiments 4 and 5. Meanwhile, PhIP did not increase HMGB2 positivity in the liver or kidney, which are not carcinogenic target organs for PhIP (Ito et al., 1997; Shirai et al., 1997). We investigated alternative molecular markers including HMGB2 from the ventral lobe with highest Ki-67 positive point in PhIP-treated rats. The correlation ratio between HMGB2 and Ki-67 positivities in human prostate cancer was lower than that observed in animal experiments. This indicates that HMGB2 may have other mechanisms in tumor development such as chemoresistance and/or radioresistance (Suzuki et al., 2009; Wu et al., 2013; Syed et al., 2015; Shin et al., 2013). In this study, Ki-67 also has similar potential to HMGB2 to detect the prostate carcinogens in experiments 4 and 5. However, higher sensitivity of HMGB2 positivity was presented compared to Ki-67 positivity in experiments 3.
To consider HMGB2 expression in each prostate lobe for prostate carcinogenesis, significant higher positivity of HMGB2 was detected in the ventral prostate compared to other lobes in PhIP-treated rats. In the previous study, PhIP only induced prostate cancer in the ventral prostate lobe (Shirai et al., 1997). In MNU-treated rats, HMGB2 posivitity was high in all prostate lobes. MNU was reported to induce prostate cancer in the ventral and dorsolateral lobes (Gonçalves et al., 2013). In DMAB-treated rats, HMGB2 posivitity was high in all prostate lobes. DMAB is only known to induce prostate cancer in the ventral lobe (Shirai et al., 1985). However, DMAB has been found to induce androgen-independent prostate cancer in dorsolateral lobe with testosterone propionate treatment (Shirai et al., 1991). Meanwhile, in BOP-treated rats, HMGB2 posivitity was high in dorsolateral lobes. BOP was reported to induce prostate cancer in the mainly dorsolateral lobes, but also in the ventral lobe (Pour and Stepan, 1987; Pour, 1983). As gene expression analysis of the rodent prostate indicated that gene expression pattern differed among the lobes (Berquin et al., 2005), our results suggest that HMGB2 positivity may be correlated with carcinogenic potential in each lobe of prostate, but not completely.
To administer the chemicals, we utilized gavage in experiments 4 and 5. The rats received PhIP and BOP by gavage, and for DMAB and MNU subcutaneous and intravenous/intraperitoneal injections were used, respectively (Shirai et al., 2000; Pour and Stepan, 1987; Gonçalves et al., 2013). In the current study, irrespective of the administration route used, HMGB2 positivities in DMAB- and MNU-treated rats were approximately same in each prostate lobe in experiment 3. In experiments 4 and 5, independent of the administration route used, high HMGB2 positivities were detected in the prostate of DMAB- and MNU-treated rats. The carcinogenic potential of DMAB or MNU via gavage in prostate was not reported, but MNU was reported to induce prostate carcinoma via both intravenous (50 mg/kg body weight) and intraprostatically (15 mg/kg body weight) injections (Schleicher et al., 1996). The problems of dose and route for carcinogenic potential of chemicals were still uncertain, but, these data indicate that the prostate samples from the standard repeated dose 28-day oral toxicity study may be useful for the screening of prostate carcinogens with our model.
High expression of HMGB2 was not only detected in the rat prostate treated with PhIP and DMAB after 2 days, and 4 weeks, but also detected in neoplastic lesions of prostate induced by PhIP- and DMAB-treatment. Therefore, high expression of HMGB2 may be an early event in rat prostate carcinogenesis and may also continuously remain within tumors. In addition, high HMGB2 positivity was detected in neoplastic lesions of both rat and human prostate samples. HMGB2 is reported as a prognostic marker in various malignant tumors including human hepatocellular and bladder carcinomas and glioblastoma (Kwon et al., 2010; Wu et al., 2013; Wang et al., 2013), but not prostate cancer. In our results, HMGB2 also has the possibility to be a prognostic marker in human prostate cancer.
HMGB2 is a member of the HMG proteins which are essential, and highly dynamic, constituents of mammalian chromosomes that participate in all aspects of chromatin structure and function, including DNA repair processes. HMGB proteins are reported to inhibit DNA repair in vivo, which may be intimately involved with the accumulation of genetic mutations and chromosome instabilities frequently observed in cancers (Reeves and Adair, 2005). As all prostate carcinogens in this study are mutagens, which induce DNA damage, high expression of HMGB2 may be related with progressive effects for prostate carcinogenesis. To focus on the non-mutagens in prostate carcinogens, androgen receptor agonists testosterone and testosterone propionate, are reported to induce hyperplasia and tumors in rat prostate (Cha et al., 2015; Shirai et al., 2000). Further studies are needed to determine the functions of HMGB2 in prostate carcinogenesis when hormone related substances are involved.
In conclusion, we demonstrated that increased expression of HMGB2 may be a good screening tool to identify potential prostate carcinogens. Further work is needed to investigate the role of HMGB2 expression specifically as a biomarker for prostate cancer. Immunohistochemistry of HMGB2 can be performed simultaneously with the standard repeated dose 28-day oral toxicity study and may be readily applied for retrospective analyses using formalin-fixed and paraffin-embedded tissues from previous studies. Although further analysis of the specificity, dose-dependency, and applications of this method to other organs is required, HMGB2 expression may be a good screening tool for identification of potential of prostate carcinogens.
This work was supported by the Grant-in-Aid for Cancer Research from the Ministry of Health, Labour and Welfare of Japan, the Health and Labour Sciences Research Grant for the Research on Risk of Chemical Substances from the Ministry of Health, Labour and Welfare, Japan and the Japan Society for the Promotion of Science, Tokyo, Japan, for Scientific Research (No. 15K08352). We gratefully acknowledge the expert technical assistance of Koji Kato, Junko Takekawa and Yuko Nagayasu.
The authors declare that there is no conflict of interest.