Abstract
DES carcinogenicity has been investigated in 2 mouse knockout models, the Xpa homozygous knockout, and the combined Xpa homozygous and p53 heterozygous knockout. Wild-type (WT) mice were also included. Xpa mice received diets containing DES at concentrations of 0, 100, 300, and 1500 ppb for 39 weeks; Xpa/p53 and WT mice received diets containing 0 or 1500 ppb. There were 15 of each sex per group. Both Xpa and WT mice had a similar incidence of tumors at the high dosage of 1500 ppb, including pituitary adenomas in 4 WT mice and 7 Xpa mice, and single incidences of osteosarcoma (Xpa), T-cell lymphoma (WT and Xpa), and testicular interstitial cell adenoma (WT and Xpa). The incidence of tumors was higher in the Xpa/p53 mice at 1500 ppb, mainly attributable to 5 osteosarcomas in males and 2 in females, but also 4 pituitary adenomas, testicular interstitial cell adenomas in 4 males, and single incidences of cerebral glioma, phaeochromocytoma, and cervical fibrosarcoma. The incidence of osteosarcomas was related to the severity of fibro-osseous lesions in the bone marrow. It was concluded that for carcinogenicity screening, Xpa mice were no more sensitive than wild-type mice for compounds like DES, but the Xpa/p53 model showed an increased sensitivity.
Introduction
DES was used extensively for the treatment of possible miscarriage during the 1950s and 1960s. However in 1971 it was identified as a transplacental carcinogen, causing unusual clear-cell carcinomas of the cervix of adolescent girls who had been exposed prenatally (Herbst and Scully, 1970; Herbst et al., 1971).
A large amount of research has been performed to define the nature of the carcinogenic potential of DES, and this has been summarized in various reviews (IARC, 1979, 1987; Herbst and Bern, 1981). Pharmacology and toxicology have also been performed and reviewed (IARC, 1979; Marselos and Tomatis, 1993). The effects that have been revealed can largely be ascribed to DES acting through estrogen receptors, despite the fact that DES is nonsteroidal.
DES and its metabolites are nonmutagenic in Ames tests (Glatt et al., 1979; Marselos and Tomatis, 1993), but it does cause clastogenicity in human lymphocytes and Chinese hamster fibroblasts (Bishun et al., 1977; Ishidate and Odashima, 1977). DES has also been demonstrated to cause aneuploidy in the bone marrow of mice and in primary rat fibroblasts, and it is also capable of disrupting microtubules resulting in abnormal or arrested mitotic spindles (Chrisman, 1974; Chrisman and Hinkle, 1974; Tucker and Barrett, 1986; Sakakibara et al., 1991; de Stoppelaar et al., 1997).
A large number of carcinogenicity studies have been performed in rats, mice, and hamsters. In mice and hamsters, similar transplacental carcinogenicity occurs to that in man (Rustia, 1979; McLachlan et al., 1980; Newbold and McLachlan, 1982), but in rats, prenatal exposure to DES results in mammary and pituitary tumors (Boylan and Calhoon, 1979). Postnatal dietary exposure results in mammary carcinomas, ovarian tubular adenomas, and pituitary adenomas in mice, and mammary fibroadenomas and pituitary adenomas in rats (Gass et al., 1964; Gibson et al., 1967; Highman et al., 1977, 1980; Phelps and Hymer, 1983; Greenman et al., 1984).
The recent availability of various transgenic and knockout models has provided an opportunity to investigate further the carcinogenicity of known human carcinogens such as DES. For this reason, DES has been included in the International Life Sciences Institute (ILSI) program for the evaluation of these alternative models, as an example of a nonmutagenic carcinogen acting through a “hormonal” mechanism. The study described below was part of the ILSI program, and evaluated the effects of DES in 2 models, the Xpa −/−knockout mouse (Xpa), and a double knockout model, the Xpa−/−/p53+/− mouse (Xpa/p53).
The characteristics of the Xpa knockout mouse have been described in detail (de Vries et al., 1995; van Kreijl et al., 2001; van Steeg et al., 2001). Briefly, the model has a knockout of both alleles of the Xpa gene, and the protein coded by this gene is currently thought to be responsible for verification of NER-related DNA damage. Mice with this knockout have no residual NER-activity.
Similarly, much has been published about the p53 knockout mouse (Storer et al., 2001), and the use of this model as part of a double Xpa/p53 knockout model has been described by van Kreijl et al. (2001). The assumption is that by combining a model with no NER activity with one that has deficient tumor suppressor function, a very sensitive model to carcinogens has been created.
A study duration of 9 months was used. This duration was decided by the Xpa Assay Working Group after initial studies with positive controls indicated that 6 months was probably insufficient to obtain a robust tumor response (van Kreijl et al., 2001).
A preliminary version of these data was presented as a poster at the ILSI Alternatives to Carcinogenicity Testing (ACT) Workshop in Leesburg, VA, USA, November 1–3, 2000 (McAnulty and Skydsgaard, 2001).
Materials and Methods
This study was run in 2 parts. The first part was a 4-week dose–range-finding study with Xpa mice to determine suitable dietary concentrations for the second part of the investigation, the main 9-month study.
Animals
For the dose–range-finding study, 25 male and 25 female C57BL/6 Xpa mice were obtained from RIVM, Bilthoven, The Netherlands. For the main study, 75 male and 75 female C57BL/6 Xpa mice, and 30 male and 30 female C57BL/6 Xpa/p53 mice were obtained from RIVM, Bilthoven, The Netherlands. A further 30 male and 30 female C57BL/6 wild-type mice were obtained from IFFA Credo, l’Arbresle, France. The animals were at least 6 weeks old at the start of the acclimatization period, and all animals were allowed at least 1 week’s acclimation before the start of treatment.
Husbandry
During the studies the animals were housed singly in Type II Macrolone cages (22 × 16 × 14 cm) in a room maintained at 21 ± 3°C, relative humidity of 55 ± 15%, with 10 changes of filtered air per hour, and 12 hours light and 12 hours dark per day. The animals in the dose–range-finding study were fed Altromin 1314 powdered rodent diet, while those in the main study were fed Altromin 1321N powdered rodent diet (both diets supplied by Chr. Petersen A/S, DK-4100 Ringsted, Denmark). The latter specially formulated diet for carcinogenicity studies was used following agreement to use this diet by all participants in the Xpa program. Drinking water from the public supply, and acidified with hydrochloric acid, was available ad libitum.
Test Articles
DES was obtained from Sigma, the Lot No. 106H0643, and this was the lot used in all DES studies in the ILSI program. The positive control substance used was benzo[a]pyrene (B[a]P), also obtained from Sigma, and food-quality corn oil used as vehicle for the positive controls was obtained from a local supermarket.
Methods (Dose-Range-Finding Study)
Groups of 5 male and 5 female Xpa mice were allocated to the following treatments:
DES was dissolved in acetone and used to prepare a series of premixes in lactose by Scantox, followed by evaporation of the acetone. The DES premixes in lactose were mixed in Altromin 1314 diet at the concentrations shown above by Altromin, Germany, and stored at 4°C until used. The animals were fed continuously with the various diets for 28 days; dietary concentrations were not adjusted with respect to body weight change.
The animals were examined daily for any clinical signs or deaths. Body weights were recorded at the start of treatment, and then twice-weekly up until termination, when a final weight before necropsy was recorded. Food intake was recorded daily during the first week of treatment, and twice-weekly thereafter. After 28 days’ treatment, blood samples were taken for analysis of various hematology and clinical chemistry parameters. All animals underwent a full necropsy. The weights of the brain, heart, kidneys, liver, spleen, testes, and thymus were recorded where possible. A full list of organs and tissues were retained in fixative, as well as any abnormal tissues. These were histologically processed and slides prepared using conventional hematoxylin and eosin staining; all slides from groups 1 and 5 were examined microscopically, as well as gross lesions from all animals.
Methods (Main Study)
Groups of 15 male and 15 female mice were allocated to the following treatments:
The selection of the high dose of DES was based on the histopathological results in the dose–range-finding study. The rationale for dose selection is presented in the Results section.
DES premixes were prepared in the same way as in the dose–range-finding study, and these were mixed in Altromin 1321N diet at the concentrations shown above; all these procedures were performed by Scantox. The DES diets were fed continuously to the DES-treated groups for 39 weeks. The 3 control groups (group Nos. 1, 6, and 8) received untreated Altromin 1321N diet. The positive control group received B[a]P in corn oil by oral gavage at a dose volume of 10 ml/kg on 3 days each week (Monday, Wednesday, Friday) for up to 39 weeks.
The animals were examined daily for any clinical signs or deaths. Body weights were recorded at the start of treatment, weekly up until week 13, and then every 2 weeks until termination, when a final weight before necropsy was recorded. Food intake was recorded weekly, and this information was used to calculate achieved dosage levels in the DES-treated groups. In the week before scheduled termination, blood samples from surviving animals were taken for analysis of various hematology and clinical chemistry parameters (numbers of erythrocytes, leukocytes, lymphocytes, and platelets, and the activities in plasma of alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and lactate dehydrogenase (LDH)). For ALP, an isoenzyme analysis was performed to determine the origin of the enzyme in plasma. There were no survivors in the DES-treated wild-type group and only 1 female survivor in the B[a]P-treated Xpa group, and therefore no clinical pathology data were available for these groups. All animals that died, were sacrificed in extremis, or reached scheduled termination, underwent a full necropsy. The weights of the brain, heart, kidneys, liver, pituitary, spleen, thymus, and testes were recorded if possible. A full list of organs and tissues were retained in fixative, as well as any abnormal tissues and tissue masses. Tissue trimming was performed according to RITA guidelines (Bahnemann et al., 1995). The tissues were histologically processed and slides prepared using conventional hematoxylin and eosin staining, and all slides were examined microscopically.
Statistical Analysis
Data were processed to give group mean values and standard deviations, where appropriate. Thereafter each continuous variable was tested for homogeneity of variance with Bartlett’s test. If the variance was homogenous, possible intergroup differences were assessed with Dunnett’s test. If the variance was heterogeneous, each variable was tested for normality by the Shapiro–Wilk method. If the distribution was normal, possible intergroup differences were identified with Student’s t-test. Otherwise the possible intergroup differences were assessed by the Kruskal–Wallis test. If any significant intergroup differences were detected, the subsequent identification of the groups was carried out with the Wilcoxon Rank–Sum test. In the main study, tumor incidence was compared using Fisher’s exact test.
The positive control animal data was not included in the statistical analysis of the DES-treated groups. In the statistical analysis in the main study, groups 3, 4, and 5 were compared with group 1, group 2 was compared with group 1, group 6 was compared with groups 1 and 8, group 7 was compared with groups 5 and 6, and group 9 was compared with groups 5 and 8.
Results
Dose–Range-Finding Study
There were no clinical signs in the Xpa mice receiving up to 2000 ppb DES in the diet for 28 days, and no deaths occurred. There were body weight losses in both males and females at 1000 and 2000 ppb, and compared with controls, the deficit in weight after 28 days’ treatment was approximately 5–6% at 1000 ppb, and 7.5–9% at 2000 ppb. Food consumption was unaffected in the males, but was lower than controls in females (ca. 25%) at 500 ppb and above.
At all concentrations of DES there was a reduction in the number of circulating leukocytes. All classes of leukocytes were affected to some extent, but at 500 ppb and above the reduction was attributable primarily to lower numbers of lymphocytes. There was also a tendency to reduced numbers of platelets, but this was only statistically significant in males at 500 ppb and above, and females at 2000 ppb. Erythrocyte parameters were unaffected.
The most pronounced clinical chemistry change was an increase in the activity of plasma alkaline phosphatase. This occurred in both sexes in all DES-treated groups, and was dietary concentration-related. At the highest concentration of 2000 ppb, the increase compared with controls was 3- to 5-fold. Plasma bilirubin concentration tended to increase with increasing dietary concentration, while urea concentration tended to decrease.
At necropsy there were very few macroscopic findings. The seminal vesicles of 4 of the 5 males that received DES at 2000 ppb were diminished in size. Of the organs that were weighed, there were statistically significant reductions in the absolute and body weight-relative organ weights of the kidneys at 1000 and 2000 ppb in males, and also in the absolute weight in females at 2000 ppb. The absolute and relative weights of the thymus were reduced in both males and females at 500 ppb and above.
Dose–Range-Finding Study Histopathology
Male Reproductive System
Minimal to marked oligospermia was apparent in the epididymides in all DES-treated groups. However, no disturbance of spermatogenesis was observed in the testes. There was a decrease in the alveolar contents of the prostate gland in all treated groups, and in the seminal vesicles at 500 ppb and higher. Slight to marked squamous metaplasia of the glandular epithelium of the prostate gland (coagulating gland) was apparent at 2000 ppb.
Female Reproductive System
Corpora lutea were absent from the ovaries in all treated groups. Minimal to marked DES concentration-related endometrial glandular hyperplasia was also present in all treated groups, as was minimal to slightly increased keratinization of the vagina.
Thymus
Minimal to marked cortical atrophy of the thymus was found in both sexes at 500 ppb and above.
Spleen
Minimal to moderate extramedullary erythropoiesis was present in all treated groups.
Adrenals
A variable incidence of minimal to slight vacuolation and increased degeneration of the x-zone was seen in males at 1000 ppb and higher, and females at 500 ppb and higher.
Bone
In all treated males and in females at 500 ppb and higher there was a minimal-to-moderate increase of trabecular bone in the medullary cavity of the proximal metaphysis of the femur. This was also seen in the sternum of all treated males at 500 ppb and higher, but not in treated females.
Liver
At 1000 and 2000 ppb there was a decrease in glycogen content of hepatocytes in some animals of both sexes. This was also seen in 2 control males, but in the controls the effect was diffuse, whilst in the treated animals it occurred specifically in the periportal area.
Dose Selection for the Main 9-Month Study
Dietary concentrations for the main study were selected on the basis of the results of the dose range-finding study described above, and by comparison with the results of studies performed in p53 mice by Searle (Storer et al., 2001), and in rasH2 mice by Mitsubishi-Tokyo (Mutai et al., 2001). The preliminary study described above failed to establish a MTD at dietary concentrations up to 2000 ppb for 4 weeks. However, there were microscopic changes in the male and female reproductive organs, thymus, spleen, adrenals, and bones; in particular, minimal to marked dose-related endometrial glandular hyperplasia was apparent in all treated groups. In the 6-month Searle study in p53 mice, there was a high incidence of ruptured endometrial cysts with inflammation at 500 and 1000 ppb, but in the 6-month Mitsubishi-Tokyo study in rasH2 mice there was no evidence of endometrial cysts up to 1000 ppb. Because of the histopathological changes seen in the 4-week Xpa study, it was decided to use 1500 ppb as the highest dietary concentration in order to avoid the possibility of endometrial cysts developing.
Main 9-Month Study
Clinical Signs and Mortality
The Xpa mice treated with DES for 9 months exhibited only a small number of clinical signs. One male receiving 300 ppb (group 4) had unilateral testicular enlargement, while in 6 receiving 1500 ppb (group 5) this condition was bilateral. One Xpa male and 6 females receiving 1500 ppb (group 5) had distended abdomens, and a further female in this group exhibited rectal bleeding. Viability of the DES-treated Xpa mice, particularly the females, was high. One male at 100 ppb (group 3) and 2 at 300 ppb (group 4) were found dead without showing previous signs, while no Xpa females died at these dietary concentrations. At 1500 ppb (group 5), 3 males died without showing previous signs, whilst another 5 were sacrified in extremis because of testicular enlargement. The 3 females that died at 1500 ppb (group 5) were sacrificed in extremis, 2 because of distended abdomens, and 1 because of the rectal bleeding mentioned earlier.
The Xpa/p53 mice that received 1500 ppb DES (group 9) showed a similar profile of clinical signs to the Xpa mice at the same dietary concentration, with 5 males showing testicular enlargement, and 11 females with distended abdomens. In addition, 2 males and 2 females showed hindlimb gait abnormalities. The viability of the male double knockouts treated with DES was similar to that of the Xpa males, but viability of the female double knockouts was considerably lower, with only 2 animals surviving until the end of the study. Two male and 2 female Xpa/p53 mice died without showing any previous signs, and the rest were sacrificed in extremis because of testicular enlargement, distended abdomens, or gait abnormalities.
The wild-type animals treated with 1500 ppb DES (group 7) showed the same spectrum of clinical signs, and at a similar incidence, as the 2 knockout strains. However, viability of the treated animals was lower than in the knockout animals, and all were dead before the end of the study. The majority of animals were sacrificed in extremis for the same reasons as the other genotypes.
The Xpa mice treated with the positive control agent B[a]P (group 2) showed no clinical signs, other than becoming subdued with a rapid decline in condition immediately before death. There was an isolated death of a female in week 3 of treatment, but otherwise deaths did not start until week 20 onwards. Deaths occurred rapidly from about week 31 onwards, and all but 1 female died before the end of the study.
Body Weight and Food Consumption
In agreement with the dose–range-finding study, during the first month of the study there was no effect on weight gain of Xpa mice at the dietary concentrations of 100 and 300 ppb (groups 3 and 4; Table 1). At 1500 ppb (group 5), the Xpa mice either had a lower rate of weight gain than controls, or lost weight. Male Xpa/p53 mice treated with 1500 ppb DES (group 9) also lost weight during the first month of the study, but females were unaffected. The wild-type mice that received 1500 ppb (group 7) were similar to the Xpa mice, either losing weight or gaining weight more slowly than wild-type controls.
During the remainder of the study, there was a clear sex difference in body weight response to DES exposure. In all 3 genotypes, males treated with 1500 ppb had lower weight gains than controls, while females had higher weight gains than controls (Table 1). Weight gain of Xpa females that received 300 ppb (group 4) was also slightly greater than that of controls. There were no apparent effects of DES treatment on food consumption.
Body weight change of the Xpa mice that were administered the positive control B[a]P (group 2) was similar to that of controls up until deaths started to occur. Thereafter, the animals lost weight in the period immediately preceding death, and this was accompanied by a reduction in food consumption.
Hematology
As in the 28-day study, leukocyte numbers were reduced in the Xpa mice that received 1500 ppb DES (group 5), and numbers were also reduced in the Xpa/p53 mice (group 9; Table 2). Also, as before, the reduction was primarily attributable to a loss of lymphocytes. There was also a reduction in platelet counts at all DES concentrations in both genotypes. There were no wild-type-treated animals and only 1 B[a]P-treated Xpa animal that survived to termination to provide blood samples.
Unlike the 28-day study, there was also a reduction in erythrocyte numbers in both Xpa and Xpa/p53 males at 1500 ppb DES (groups 5 and 9), and at all dietary concentrations of DES in the females (Table 2). This was accompanied by reduced hemoglobin concentrations, and an increase in mean corpuscular volume.
Clinical Chemistry
The most marked clinical chemistry change was an increase in the plasma activity of alkaline phosphatase in all DES-treated groups of Xpa males (groups 3, 4, and 5; Table 3). This was also seen in the 28-day study, but in that study females were also affected, whereas in the 39-week study there was no effect in the Xpa females. However, in the Xpa/p53 animals treated at 1500 ppb (group 9), both males and females showed the increase. Isoenzyme analysis revealed that in the majority of animals the origin of the alkaline phosphatase was the liver.
There were also sex differences in many of the other clinical chemistry parameters (Table 3). The activities of alanine and aspartate aminotransferases, and lactate dehydrogenase tended to be increased in Xpa and Xpa/p53 males (groups 5 and 9), but were reduced in females. The concentration of bilirubin was increased in males, but lower in females. The only consistent change between the sexes was a tendency to lower urea concentrations. These differences were at variance with the results obtained in the 28-day study, where the changes in these parameters showed no sex difference.
Necropsy and Organ Weights
Various tissue masses were found at necropsy. These were mainly in the Xpa/p53 group that had received 1500 ppb DES (group 9), with soft, firm and hard masses being found in the musculoskeletal system, spleen, and liver in 6 males and 5 females, and also masses in the cervix of the females. One wild-type male that had received 1500 ppb (group 7) had a soft mass in the chest cavity, and a female from the same group had a pale, soft mass attached to the ventral part of the spinal column.
In one Xpa male and 2 Xpa/p53 females there was a marked increase in the size of the thymus, and this was accompanied by enlargement of lymph nodes in the male and 1 of the females. However, in the majority of DES-treated animals thymus size was reduced (Table 4), and at 1500 ppb in all genotypes (groups 5, 7, and 9) several animals had atrophy of the thymus. Conversely, in all treated groups, the size of the spleen was increased (Table 4).
Enlargement of the pars distalis of the pituitary was apparent macroscopically in a large number of animals of both sexes and all genotypes at 1500 ppb DES (Groups 5, 7, and 9), and also in 1 Xpa female that received 300 ppb (group 4). This was confirmed in these groups by marked increases in pituitary weight (Table 4).
Both enlarged and small testes were observed in males that had received 1500 ppb DES (groups 5, 7, and 9). In the Xpa mice the overall effect was a lower testicular weight (Table 4), but in the Xpa/p53 mice the mean weight tended to be slightly greater than that of controls. Several of the females from all genotypes that received 1500 ppb (groups 5, 7, and 9) had an enlarged cervix, and this was also seen in a few Xpa females that received 300 ppb (group 4).
Unilateral or bilateral cystic dilation of the pelvis of the kidneys, associated with marked distension of the urinary bladder, was observed in both sexes and all genotypes receiving 1500 ppb DES (groups 5, 7, and 9). This was the cause of the distended abdomens mentioned earlier in the Clinical Signs section.
The livers of both Xpa and Xpa/p53 mice that received 1500 ppb DES (groups 5 and 9) were increased in weight compared with their respective controls, and this also occurred in Xpa females that received 100 and 300 ppb (groups 3 and 4).
There were relatively few macroscopic signs at necropsy of the B[a]P positive control group of Xpa mice (group 2). Two males and 4 females had enlarged spleens, and a few animals had enlarged lymph nodes. One female also had an enlarged nodular thymus. Also in a few animals there were black/brown foci in the nonglandular stomach, and raised areas in the small intestine.
Main 9-Month Study Histopathology
Malignant Neoplastic Lesions
The only malignancies that occurred in the Xpa mice were in 2 males of the 1500 ppb DES group (group 5; Table 5). There was an osteosarcoma in the sternum of 1 animal, with invasion into the surrounding tissues (Figure 1A), but no evidence of metastases. The second male had a lymphoma of T-cell origin, with metastasis to several organs.
The incidence of tumors was highest in the Xpa/p53 group that had received 1500 ppb DES (group 9; Table 5). Osteosarcomas were found in 5 males and 2 females. In 4 of the males the osteosarcoma was found in either the femur or tibia with invasion to the surrounding tissues; in 3 of the males metastases were found in the liver (Figure 1B), and in 1 of these there were also metastases in the spleen. In 1 male no evidence of metastasis was found. In the fifth male, metastases from an osteosarcoma were found in the liver, but the primary tumor was not found. In 1 of the affected females there were osteosarcomas in the sternum, nasal cavity, and femur, with invasion to the surrounding tissues. In the second female, an osteosarcoma was found on the inner side of the cranium, with metastases in the spleen. Two of the double knockout females that had received 1500 ppb had T-cell lymphomas that had metastasized to several organs. Another male from this group had a rare glioma in the cerebrum, and 1 female had a cervical fibrosarcoma with invasion of the rectum and vagina. There was also 1 Xpa/p53 female from the control group (Group 8) that had a mammary adenocarcinoma.
In the wild-type animals exposed to 1500 ppb DES (group 7), there was only 1 malignant tumor found, a T-cell lymphoma in a male (Table 5).
Benign Neoplastic Lesions
Pituitary adenomas were found in both sexes of all genotypes at 1500 ppb DES (groups 5, 7, and 9; Table 6). Five male and 2 female Xpa mice were affected, and in both the Xpa/p53 and wild-type groups 2 males and 2 females had adenomas. Testicular interstitial cell adenomas were observed in all 3 genotypes at 1500 ppb (groups 5, 7, and 9), although the incidence was greatest in the Xpa/p53 males (4 affected), with only single incidences in the other 2 groups. However, 3 Xpa males that had received 300 ppb (group 4) also had interstitial cell adenomas. There was also a single incidence of an adrenal phaeochromocytoma in a male Xpa/p53 mouse at 1500 ppb DES (group 9).
Proliferative Lesions
Minimal to marked hyperplasia of the pars distalis was observed in the majority of animals of all genotypes that received 1500 ppb DES (groups 5, 7, and 9; Table 7). Minimal to slight hyperplasia of the pars distalis was also seen in the Xpa females that had received 300 ppb (group 4). Minimal hyperplasia of the pars intermedia was found in about half of the animals that received 1500 ppb DES (groups 5, 7, and 9), and about half the Xpa females that received 300 ppb (group 4).
Hyperostosis occurred in nearly all animals that received 1500 ppb DES (groups 5, 7, and 9; Table 7). There was also a high incidence in both sexes of the Xpa mice at 300 ppb (group 4), and also a low incidence in female Xpa mice at 100 ppb (group 3). In all mice treated at 1500 ppb (groups 5, 7, and 9), minimal to moderate fibro-osseous lesions (FOLs) were present in both of the bones examined (sternum and femur) (Figure 1C). Although hyperostosis can be included in the description of FOLs, it was possible in this study to diagnose hyperostosis as an isolated lesion, and FOLs could be distinguished by the presence of a more dominant fibrovascular component (Long and Leininger, 1999). FOLs were more severe in the Xpa and Xpa/p53 mice (groups 5 and 9) than in the wild-type mice (group 7).
Minimal focal glandular hyperplasia was seen in the mammary glands of about two-thirds of the females that received DES at 1500 ppb, regardless of genotype (groups 5, 7, and 9).
In approximately two-thirds of the males of all genotypes receiving 1500 ppb DES (groups 5, 7 and 9), minimal-to-moderate interstitial cell hyperplasia was observed in the testes. Moderate hyperplasia was also found in 2 males that had received 300 ppb (group 4).
Slight subcapsular hyperplasia of Type A (spindle) cells and Type B (polygonal) cells was observed in the adrenals as a treatment-related change in males of all genotypes (groups 1, 3, 4, 5, 7, 8 and 9), whereas in females this change was present to a minimal to moderate degree in both control and treated animals, with no variation attributable to treatment.
Other Lesions
Thymic atrophy was observed in the majority of animals of all 3 genotypes at 1500 ppb DES (groups 5, 7, and 9; Table 8), but only in a single case at the lower dietary concentrations in the Xpa mice (group 4). In the spleen, there was a moderate to marked increase in extramedullary hematopoiesis in the majority of animals at 1500 ppb (groups 5, 7, and 9); in the Xpa mice, this observation showed a minimal to moderate increase at 300 ppb (group 4), while at 100 ppb (group 3) there was a minimal increase in females only. A minimal-to-slight increase in extramedullary hematopoiesis was also observed in the liver in the majority of animals that received 1500 ppb (groups 5, 7 and 9). Cystic dilation of the pelvis of the kidneys occurred at 1500 ppb in all 3 genotypes (groups 5, 7, and 9; Table 8), although the incidence was low in the Xpa/p53 mice (group 9).
Minimal-to-moderate hypertrophy and vacuolation of the cortical cells of the adrenals was observed in the majority of animals at 1500 ppb (groups 5, 7, and 9; Table 8), although the severity was slightly more pronounced in the knockout strains than the wild-type mice There also tended to be an increase in the incidence of ceroid pigment accumulation at the corticomedullary junction in all DES-treated groups (groups 3, 4, 5, 7, and 9).
Minimal-to-moderate aortic arteritis and periarteritis were observed in several animals at 1500 ppb (groups 5, 7, and 9; Table 8), particularly in wild-type females (group 7). This was considered to be the cause of death in a number of the animals. However, other than in a single male, this observation was not recorded in the treated Xpa/p53 mice (group 9).
DES in all treated groups had anticipated effects on the reproductive system (groups 3, 4, 5, 7, and 9; Table 8). In males there was testicular tubular atrophy, oligospermia in the epididymides, squamous metaplasia in the coagulating gland, and diminished contents of the seminal vesicles. In females there was atrophy of the ovaries, endometrial gland hyperplasia, deposition of hyaline material in the endometrium, and stromal mucoid changes in the cervix.
B[a]P-Treated Positive Control Xpa Mice (group 2)
Malignant lymphomas of B-cell origin were found in 2 males and 3 females, and in a further female a malignant lymphoma of T-cell origin was found. In all affected animals metastases were observed in the liver and other organs. Adenocarcinomas were found in the jejunum of 3 males and 3 females; 1 of these males also had an adenocarcinoma in the duodenum, and in 1 of the females this tumor occurred in the ileum along with a duodenal adenoma. In 1 female there was a squamous cell carcinoma in the nonglandular stomach.
Adenomas were observed in the caecum of 1 male, in the jejunum of 3 females, and also in the ileum of 1 of these females. Papillomas were present in the nonglandular stomach of 2 males and 2 females.
Nonproliferative changes in the B[a]P group were limited to marked atrophy of the thymus, extramedullary hematopoiesis in the spleen, focal hepatic necrosis, accumulation of ceroid pigment in the cells of the adrenal corticomedullary junction, epididymal oligospermia, and atrophy of the ovaries and uterus.
Discussion
This study has demonstrated that the Xpa and Xpa/p53 mouse knockout models are sensitive to the carcinogenic potential of DES. However, the degree of sensitivity of the Xpa model was no greater than that of wild-type mice, treated over the same period of 39 weeks. The Xpa/p53 model, on the other hand, showed a greater sensitivity than wild-type mice in terms of the number and range of tumors that occurred; the Xpa/p53 model also had the advantage over wild-type animals of a better survival rate, and a lower incidence of nonproliferative lesions. While not so sensitive, the Xpa mice had the best survival of the 3 genotypes.
As the Xpa and Xpa/p53 models have a knockout of both alleles of the Xpa gene (de Vries et al., 1995; van Kreijl et al., 2001), it could be speculated that the difference in sensitivity of the Xpa/p53 model is attributable to the loss of 1 allele of the p53 gene. The most significant tumor response in this study was the incidence of osteosarcomas in the Xpa/p53 mice, and it may be significant that in p53 hemizygous mice, osteosarcomas are among the most common spontaneous tumor types (Mahler et al., 1998). However, in all of the p53 studies in the ILSI database, the incidence of osteosarcomas in controls was only 0.5% in males and 0.7% in females (Storer et al., 2001). In the current study there were no osteosarcomas in the control Xpa/p53 mice, despite the longer study period of 39 weeks compared with 26 weeks in the p53 studies, and the incidence of osteosarcomas in the DES-treated group was 33% in males and 13% in females. Further, in the 26-week p53 ILSI study with DES, there was no occurrence of osteosarcomas (Storer et al., 2001). It therefore seems unlikely that the high incidence of osteosarcomas in DES-treated Xpa/p53 mice can be attributed to the deficiency of the p53 gene alone. It may be associated with a synergism between the 2 genotypes, or some other factor such as the genotoxic potential of DES, reduced immunosurveillance, or effects on steroid or bone metabolism.
DES is sometimes referred to as a “nongenotoxic carcinogen,” but this is inaccurate. Using Ames tests, DES and 11 of its metabolites have been shown to be nonmutagenic (Glatt et al., 1979; Marselos and Tomatis, 1993); however, in a wide range of tests, DES has been shown to be clastogenic, including induction of chromosome aberrations in vitro in Chinese hamster fibroblasts and human lymphocytes (Bishun et al., 1977; Ishidate and Odashima, 1977), sister chromatid exchanges in human fibroblasts and mouse cervical epithelium cells (Rudiger et al., 1979; Hillbertz-Nilsson and Forsberg, 1989), formation of abnormal or arrested mitotic spindles in embryonic hamster cells (Tucker and Barrett, 1986; Sakakibara et al., 1991), and the latter accounting for increased aneuploidy in vivo in mouse bone marrow and embryonic cells, and in rat primary fibroblasts (Chrisman, 1974; Chrisman and Hinkle, 1974; de Stoppelar, 1997). In addition, DES has been demonstrated to cause unscheduled DNA synthesis in vitro (Martin et al., 1978), and has also been shown to form DNA adducts in various tissues in vivo (Gladek and Liehr, 1989, 1991). The clastogenic and other genotoxic properties of DES may be responsible for some of the cellular effects, both proliferative and non-proliferative, seen in the current study. However, clastogen tests in vitro, are associated with many false positives in predicting carcinogenicity, although it has been stressed elsewhere that the effects of DES cannot be attributed to its estrogenic properties alone (IARC, 1987), an observation applying to both reproductive and nonreproductive tissues. The loss of the Xpa gene alone does not increase the carcinogenic potential of any genotoxic influence of DES, but it may be that a synergistic relationship with a p53 deficiency does increase susceptibility, as seen in the Xpa/p53 mice.
Cell populations affected by DES include those involved in production of circulating blood cells, i.e., the hematopoietic and lymphoblastic systems. In the current study a reduction in leukocytes was observed in both the 4- and 39-week studies, with circulating lymphocytes particularly affected, although in the 4-week study there were also significant effects in the other cell types. Further, in the 39-week study, there was a reduction in circulating erythrocytes. The effect on hematopoiesis and lower peripheral blood cell counts has been known for many years (Lavenda and Wong, 1954), and in the current study we have demonstrated there are compensatory increases of extramedullary hematopoiesis in the spleen and liver. Also in the current study there was a marked thymic atrophy, affecting the cortex in particular, and contributing to the reduced numbers of circulating lymphocytes. This has been reported previously, and in particular T-cell numbers and function are affected, with a wide range of immunological sequelae in both animals and humans (Marselos and Tomatis, 1993; Burke et al., 2001; Karpuzoglu-Sahin et al., 2001). The effects on T-cell function might have an influence on immunosurveillance for precancerous or cancerous cells, but it is considered unlikely that immunodeficiency is directly related to the carcinogenic potential of DES.
In the preliminary 4-week study, an increase in trabecular bone was observed in the DES-treated animals. In the main study, hyperostosis was observed in the sternum and femur of all animals that received the high dose of DES, and also occurred in the majority of Xpa mice that received 300 ppb DES, and some of the Xpa females at 100 ppb. This response is well-documented for compounds with estrogenic activity (Woodard et al., 2002). However, it was also possible to clearly identify fibro-osseous lesions (FOLs) that were distinct from the hyperostosis. The term FOL was introduced in 1980 (Sass and Montali, 1980) to describe fibro-osseous changes seen in the bone marrow of B6C3F1 female mice. The lesion is characterized by accelerated osteoclastic bone resorption with concurrent fibroplasia (Long and Leininger, 1999). The overall appearance is one of accelerated bone turnover, with resorption encroaching on the center of the medullary cavity, followed by bone formation.
The occurrence of FOLs in B6C3F1 female mice is thought to be attributable to excessive estrogen production by ovarian cysts (Sass and Montali, 1980; Albassam et al., 1991), but the same lesion is not seen in CF1 mice, despite the presence of the same cysts (Albassam et al., 1991). However, administration of estrogens to C57BL and C3H female mice also results in FOLs (Silberberg and Silberberg, 1970; Highman et al., 1981), but administration of DES to CF1 female mice again fails to cause FOLs (Johnson, 1987). The mice used in the current study were all derived from the C57BL/6 strain, and the occurrence of FOLs in all 3 genotypes, including the wild-type animals, indicates that both females and males of this strain are susceptible to this lesion.
It has been suggested that osteosarcomas in mice arise from these areas of osteofibrosis and bony trabecular proliferation in the medullary cavity (Highman et al., 1981). The results from the current study are consistent with this suggestion, and it may be that FOLs could be considered a preneoplastic lesion in mice. In this context it would be of interest to investigate if DES can induce sarcomas in CF1 mice, the strain that did not develop FOLs. Why the Xpa/p53 mice should be more sensitive than the other 2 genotypes is unclear, but maybe the combined deficiency of the Xpa and p53 genes makes the cells of the bone marrow more sensitive to the genotoxic effects of DES, and also inhibits the possibility for nucleotide excision repair and cell-cycle control in damaged cells.
Another aspect of the development of hyperostosis and FOLs is that the 2 lesions may have also affected hematopoiesis in the bone marrow, and contributed to the effects on circulating blood cell numbers discussed earlier. A further aspect of the development of FOLs is the possible link with an increase in plasma alkaline phosphatase activity that has been conjectured to originate from the skeleton (Albassam et al., 1991). In the current study a marked increase in alkaline phosphatase activity was observed, but surprisingly isoenzyme analysis revealed that this, in fact, originated from the liver.
In conclusion, this investigation of the effects of DES in Xpa mice over a 39-week treatment period has demonstrated that the model is sensitive to the proliferative and carcinogenic potential of DES, but that the sensitivity is no greater than the same treatment in wild-type mice. The Xpa/p53 double-knockout model was found to have a greater sensitivity to DES, and gave a very robust carcinogenic response. A similar result, although less robust, has been described for another estrogenic compound, estradiol-17β (Steenhof et al., 2001). Several mutagenic compounds have also been demonstrated to be positive in Xpa/p53 mice, including B[a]P, p-cresidine, 2-acetamidofluorene, and cyclosporin A (Beems et al., 2001; van Kreijl et al., 2001; van Steeg et al., 2001). The only mutagenic carcinogen not to have given a positive response in the Xpa/p53 model is phenacetin (van Kreijl et al., 2001), although there was extensive karyomegaly in the kidneys. It may be that the double knockout is a good potential candidate for general carcinogenicity screening of both mutagenic and clastogenic compounds.
However, an important caution in the interpretation of these results is that the carcinogenicity induced in the mice used in this study does not predict the human carcinogenic response. Human carcinogenicity of DES is associated with preterm exposure, and the target tissues are within the reproductive tract. In this study in genetically modified mice, treatment was postnatal and the tumor response was in the skeleton. It may therefore be the case that the coincidence of carcinogenicity occurring in both mice and humans is fortuitous.
Footnotes
Acknowledgments
We would like to acknowledge the considerable contribution of Scantox in providing resources and financial support for conducting this study. We are also indebted to Dr. Peter Brinck (now of Novo Nordisk, Målev, Denmark) for managing the early phases of the study. We received considerable support in the assessment and interpretation of the results of this study from the members of the Xpa Assay Working Group, in particular the Working Group Chairman, Dr. Coen van Kreijl (RIVM, Bilthoven, The Netherlands), the Pathologist Contact, Dr. Rudolf Beems (also RIVM), and for data interpretation, Dr. Ronny Fransson-Steen (AstraZeneca, Södertälje, Sweden). We are also indebted to Dr. Maurice Cary (Cary Consulting Services, Basel, Switzerland) for his valuable comments on the draft manuscript.