Mice heterozygous for the ptch1 gene (ptch1 mice) are known as a valuable model of medulloblastoma, a common brain tumor in children. To increase the incidence and reduce the time required for tumor development, allowing for evaluation of modifier effects on medulloblastoma in a short time, we attempted to develop an early induction model of medulloblastoma in ptch1 mice initiated with N-ethyl-N-nitrosourea (ENU). Ptch1 mice and their wild-type littermates received a single intraperitoneal injection of ENU (10, 50 or 100 mg/kg) on postnatal day 1 (d1) or 4 (d4), and histopathological assessment of brains was conducted at 12 weeks of age. The width of the external granular layer (EGL), a possible origin of medulloblastoma, after injection of 100 mg ENU on d1 or d4 was measured in up to 21-day-old mice. Cerebellar size was apparently reduced at the 50 mg dose and higher regardless of genotype. Microscopically, early lesions of medulloblastomas occurred with a high incidence only in ptch1 mice receiving 10 mg on d1 or d4, but a significant increase was not observed in other groups. Persistent EGL cells and misalignment of Purkinje cells were increased dose-dependently. Although EGL was strikingly decreased after ENU injection, strong recovery was observed in mice of the d1-treated group. In summary, neonatal treatment with ENU is available for the induction of medulloblastoma in ptch1 mice, and 10 mg of ENU administered on d1 appeared to be an appropriate dose to induce medulloblastoma.
Medulloblastoma, a primitive neuroectodermal tumor that develops in the cerebellum, is the most common brain tumor of childhood. The etiology of childhood brain tumors remains largely unknown, but previous studies have suggested associations between childhood brain tumors and chemicals such as pesticides and nitrates.[2-4] These studies are mainly epidemiological, and there are few investigations into the effects of chemical exposure during development on childhood tumor using animal models.
Recent genomic approaches have demonstrated the existence of four distinct subtypes with demographic, transcriptional profiles and clinical outcome.[5, 6] In these subtypes, the tumors with activation of the Sonic hedgehog (Shh)-Ptch signaling pathway belong to the SHH group, and are considered to arise from granule cell precursors (GCPs) in the external granular layer (EGL) of the developing cerebellum. Since Shh signaling is known to drive proliferation in the GCPs, it has been suggested that the pathway dysregulation resulting from genomic alterations of its components presumably drives medulloblastoma formation.
Mice heterozygous for the ptch1 gene (ptch1 mice) are an important model for medulloblastoma. Homozygous knockout mice die during embryonic development with defects in the nervous system. Heterozygous mice survive to adulthood, and 14–20% develop medulloblastoma several months after birth.[8, 9] Histological and marker expression analysis of the brain tumor has revealed that they closely resemble human medulloblastoma, and similar to human cases, the origin of medulloblastoma in ptch1 mice is thought to be residual EGL cells that failed to exit proliferation. Activation of the Shh pathway has also been confirmed in the tumors of ptch1 mice, and this mouse model is equivalent to the SHH group in human cases. While the study of the molecular mechanism underlying medulloblastoma formation has progressed, few studies on the modifying effects of chemicals on tumor development have been conducted.
Although ptch1 mice are a valuable model for studying medulloblastoma, the low frequency and long latency for tumor development are disadvantages for detection of the modulatory effects on medulloblastomas, especially in the case of tumor suppressive compounds. Previous reports showed that neonatal irradiation dramatically increased the incidence of medulloblastoma, and it has been suggested that tumorigenesis in ptch1 mice follows a multi-step process.[10-12] So far, medium-term carcinogenicity bioassays based on the multi-step cancer development (initiation promotion model) have been established in many organs to detect modifying effects on tumor development in a short term.[13, 14] N-ethyl-N-nitrosourea (ENU) is a very common initiator and is known to induce nervous system tumors including medulloblastoma in newborn mice.[15, 16] Therefore, to increase the incidence and decrease the time required for tumor development, we attempted to induce medulloblastoma using ENU in ptch1 mice.
Materials and Methods
Ptch1 heterozygous knockout mice, generated by replacing exon 1 and 2 of the ptch1 gene with a LacZ/neomycin cassette, were obtained from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained in our laboratory. They were housed in polycarbonate cages with wood chip bedding and kept in an air-conditioned animal room with basal diet (CRF-1; Oriental Yeast, Tokyo, Japan) and tap water available ad libitum.
Twenty-five dams were used in experiment 1. Ptch1 mice and their wild-type (WT) littermates received a single intraperitoneal injection of ENU (Nacalai Tesque, Kyoto, Japan) dissolved in saline. The highest dose selected was 100 mg/kg based on a previous study,[15, 16] and 50 and 10 mg/kg were set as medium and low doses, respectively. Postnatal day 1 (d1) or 4 (d4) was chosen as administration day because they are periods of active proliferation in the EGL. Intact mice were used as the controls. Each group was composed of at least 10 ptch1 mice and five WT mice from three to five dams (Table 1). Daily observation for clinical signs and mortality was conducted throughout the study. We set the duration as 12 weeks, because mortality caused by malignant lymphoma increased after 13 weeks of age in a preliminary study. At 12 weeks of age, all animals were subjected to autopsy under deep isoflurane anesthesia. The brains, thymus, spleen and macroscopic lesions were removed and fixed in neutrally-buffered 10% formalin. The tissues were routinely processed for paraffin embedding, sectioned and stained with hematoxylin and eosin.
Table 1. The number and survival rate (%) of ptch1 and wild-type mice examined at 12 weeks of age
No. of litters
d1, postnatal day 1; d4, postnatal day 4; ENU, N-ethyl-N-nitrosourea.
In experiment 2, ptch1 and WT mice from 16 dams received 100 mg/kg of ENU intraperitoneally on d1 or d4, and the brains of each of the 9–13 mice were removed on d4, 7, 14 or 21. For comparison, the brains from 12 intact mice including both genotypes were also collected at the same time point, respectively, and the tissues were processed routinely for histological examination.
The animal protocol was reviewed and approved by the Animal Care and Use Committee of the National Institute of Health Sciences, Japan.
All cerebella were examined in midsagittal section. Based on a previous report, the stage of neoplastic lesions in the cerebellum was classified according to the size as follows: hyperproliferation of EGL, micronodule, nodule, microtumor and full-blown tumor (Fig. 1).
Persistent EGL cells were classified into the focal lesion and the diffuse or zonal lesions, according to type of distribution. The degrees of focal lesions were divided as follows: grade 1, only one to two very small clusters consisting of about 10 cells; grade 2, a few clusters consisting of about 10–30 cells; and grade 3, several clusters consisting of 30 cells or more in the midsagittal section of the cerebellum. Diffuse lesions were divided as follows: grade 1, persistent EGL cells distributed diffusely in a part of the cerebellum; and grade 2, persistent EGL cells distributed diffusely in most areas of the cerebellum. In addition, when the cell clusters were observed parallel to the granular layer, we classified the lesions into zonal type.
Photomicrographs of midsagittal sections of the cerebellum were taken with a digital camera attached to a microscope (DP71; Olympus, Tokyo, Japan), and measurement was performed using image analysis software (WinROOF, Version 5.7.1; Mitani, Tokyo, Japan). The areas of the cerebellum and granular layer were measured for each genotype, and the animals with a large tumor were eliminated. The average width of the EGL of each mouse was determined by tenth measurements selected at random from the entire cerebellum. Because the width did not differ according to genotype, we counted the values of both mice together.
As for data of the areas of the cerebellum and granular layer, values of the d1-treated and d4-treated groups were compared with the corresponding controls by one-way anova or the Kruskal–Wallis test. When statistically significant differences were detected, Dunnett's multiple comparisons test was used for comparison between the control and treatment groups. Incidence of histopathological findings was compared using Fisher's exact probability test. Width of EGL was analyzed by the Student's or Welch's t-test following a test for equal variance.
Most mice were asymptomatic throughout the study. Eight ptch1 mice and two WT mice were found dead or moribund. Hydrocephalus was found in three ptch1 mice, and the cause of death in the other cases could not be determined. At 12 weeks of age, there were no intergroup differences in survival rate in both groups (Table 1). A significant difference was not detected in final body weight (data not shown).
Reduction of cerebellar size was apparent at 50 mg regardless of genotype (Fig. 2), and morphometric analysis revealed a significant decrease in the areas of the cerebellum and granular layer in the groups treated with 50 or 100 mg of ENU (Fig. 3).
The incidence of medulloblastoma in ptch1 mice of the control group was 19% (3/16) (Fig. 4). In contrast, 11 of 15 (73%) and six of 10 (60%) mice developed medulloblastoma in the groups receiving 10 mg on d1 or d4, respectively. At 50 mg, medulloblastoma occurred in seven of 15 (47%) mice treated on d1 and in two of 10 (20%) mice treated on d4. At 100 mg, the tumor incidence was 27% (6/22) in d1-treated mice and 29% (4/14) in d4-treated mice. Most were regarded as an early stage of medulloblastoma, and a significant increase in the incidence was detected only in the groups receiving 10 mg. In WT mice, there was no medulloblastoma occurrence in either group.
As previously reported, focal lesions of persistent EGL in subpial position were common in ptch1 mice and occasionally found in WT mice (Fig. 5A). The incidence and degree of subpial EGL foci were not greatly influenced by ENU treatment (Fig. 5E). In contrast, the persistent EGL cells were distributed diffusely in the molecular layer in mice receiving ENU at 10 and 50 mg (Fig. 5B), and the persistent EGL cells showed a zonal distribution parallel to the granular layer in the molecular layer at 100 mg in both ptch1 and WT mice (Fig. 5C,F). The persistent EGL cells of focal and diffuse types were well-differentiated, and cellular atypia and mitosis were rarely seen, unlike in medulloblastoma cells. In accordance with the morphometric analysis, reduction of the number of granule cells was histologically apparent starting at 50 mg. Also, misalignment of Purkinje cells was noted from 50 mg regardless of genotype (Fig. 5D,G).
Other than medulloblastoma, there was no occurrence of neural tumors in ptch1 and WT mice. Thymic lymphoma, which is induced by ENU in mice, was increased dose-dependently. The incidence in ptch1 mice was 14% in the group receiving 100 mg on d1 and 10% and 50% in the groups receiving 50 and 100 mg on d4, respectively. Also in WT mice, thymic lymphoma occurred in 25% and 40% of the d1-treated mice at 50 and 100 mg, respectively, and in 11%, 18% and 67% of the d4-treated mice at 10, 50 and 100 mg, respectively. Hydrocephalus and rhabdomyosarcomas, which are known to occur spontaneously in ptch1 mice, were found in three ptch1 mice each, and ENU did not affect their incidence.
In experiment 2, there were no intergroup differences in body weight throughout the study (data not shown). The width of EGL in the control group peaked at postnatal day 7 and then decreased gradually and disappeared at postnatal day 21 (Fig. 6). In mice receiving 100 mg on d1, the width of EGL was strikingly reduced at d4. Then, EGL considerably recovered at d7 and was wider than the control group at d14. Similarly, the width of EGL in mice of the d4-treated group was significantly decreased at d7 and then became wider than that in the control at d14. But, the degree of recovery after damage by ENU was milder than that in the d1-treated group. Misalignment of Purkinje cells was apparent, especially in the d1-treated group. A difference in EGL development between ptch1 and WT mice was not observed in the present study.
Medulloblastomas were observed frequently in 12-week-old ptch1 mice receiving 10 mg of ENU. In ptch1 mice, preneoplastic lesions of medulloblastoma were seen at a high incidence at 2–3 weeks of age, and these lesions spontaneously regressed or matured by 10 weeks of age.[17, 18] Therefore, in the present study, the lesions observed at 12 weeks of age were regarded as selected lesions that directly led to tumor, although most lesions were regarded as an early stage of medulloblastoma. It is believed that ptch1 haploinsufficiency alone is not sufficient to induce tumor formation and that additional genetic lesions are required for tumorigenesis.[9, 11] Because ENU is a multipotent genotoxic carcinogen, ENU possibly enhanced tumorigenesis by increasing DNA damage similar to radiation and p53 inactivation.[11, 19]
As for the dose of ENU, 10 mg appeared to be appropriate for medulloblastoma induction based on the high incidence of medulloblastoma and low impact on cerebellum structure and lymphoma. In contrast, 50 and 100 mg of ENU reduced the size of the cerebellum, and significant induction of medulloblastoma was not observed. In experiment 2, EGL cells, a possible origin of medulloblastoma, were strikingly decreased after 100 mg of ENU. Consequently, the damage of ENU to EGL cells was so severe at a high dose that stockpiles of the damaged EGL cells to develop medulloblastoma could be limited.
Timing of administration is also an important factor for medulloblastoma induction, and in the present study, d1 was more effective than d4. In experiment 2, the response to ENU damage was different between d1 and d4-treatment. The strong recovery of EGL after ENU injection that was observed in the d1-treated group may be a reason for enhanced tumor induction. A previous study demonstrated that EGL cells at d1 showed resistance to radiation-induced cell death and p53 induction compared with EGL cells at d10, despite evident ongoing proliferation at both d1 and d10. Such differences in cellular response to ENU damage might also contribute to the increased tumor induction in the d1-treated group.
Focal lesions of persistent EGL in the subpial position were common in ptch1 mice and occasionally found in WT mice. In addition, the lesions of persistent EGL cells distributed diffusely or zonally in the molecular layer were induced by ENU in both mice. But, the incidence of persistent EGL cells did not correlate with medulloblastoma induction. Persistent EGL has been observed in other mutant mice, and it is believed that persistence of the EGL alone is not sufficient for tumorigenesis because none of the mutant mice showing persistent EGL other than ptch1 mice develop medulloblastoma. Furthermore, morphological alterations like ectopic EGL cells were induced by chemicals such as methyl-azoxy-methanol and cisplatin during cerebellar histogenesis in rats and mice, along with disappearance of the basket cells or damage of glial fibers.[21, 22] Because persistence of the EGL was found in both ptch1 and WT mice, this phenomenon may occur independently of the Shh-Ptch pathway, although damage of glia and basket cells was not apparent in our study. Misalignment of Purkinje cells was reported in rat cerebellum following prenatal exposure to X-irradiation, and the decrease in Reelin, which is expressed in the granule cell at a high level and helps regulating processes of neuronal migration and positioning in the developing brain, was thought to be a possible cause. Therefore, misalignment of Purkinje cells observed in our study also may result from a marked decrease in EGL cells damaged by ENU.
Currently, human medulloblastomas are classified as four distinct subtypes by genomic approaches,[5, 6] and future development of therapeutics and new drugs for medulloblastoma should take account of these subtypes. Tumors of ptch1 mice are thought to be equivalent for those of the SHH group in human, therefore, our model will be a useful tool for testing new drugs targeting the Shh pathway. Additionally, cerebellar development occurring after birth in mice is of good benefit for investigation of the effects of chemical exposure during development on medulloblastoma. Because medulloblastoma of ptch1 mice resembles human cases in histology, marker expression and cell origin, our model will be informative for human risk assessment of environmental chemicals and new drugs during development. In contrast, it should be recognized that ptch1 mouse is unsuitable for the research of medulloblastoma other than the SHH group. Also, this model might not be applicable for elucidation of onset of human medulloblastoma in some cases, because our model is developed for rapid screening of environmental chemicals and new drugs and there might be some differences between induced and spontaneous tumors at the molecular level.
In summary, ENU is available for the induction of medulloblastoma in ptch1 mice, and 10 mg of ENU administered on d1 was an appropriate dosing condition. Induction of medulloblastoma by ENU may be a good model for studying modifier effects on medulloblastoma tumorigenesis because this method allows for high tumor frequency and shorter latency and there was no occurrence of neural tumors other than medulloblastoma in the brain. This model will be helpful for human risk assessment of environmental chemicals and new drugs targeting the Shh pathway.
We thank Ms Tomomi Morikawa, Ayako Kaneko and Yoshimi Komatsu for technical assistance in conducting the animal study. This study was supported by Health and Labour Sciences Research Grants, Research on Risk of Chemical Substances, Ministry of Health, Labour and Welfare, Japan (H22-Toxicol-003).