Perinatal exposures may lead to increased risk of childhood cancers, as well as those later in life. Preconceptional parental, transplacental, and/or neonatal exposures may be involved. Studies with animal models are utilized to increase understanding of underlying cellular and molecular mechanisms. We have been focusing on two aspects of perinatal carcinogenesis: (1) exposure of mice during early life to a bacterium, resulting in development of liver tumors; and (2) transgenerational effects of exposures and experiences of male mice on tumorigenesis in offspring, including (a) the occupational metal chromium(III), and (b) food deprivation. (1) Perinatal exposure required for marked increase in liver tumors associated with infection by Helicobacter hepaticus. The bacterium Helicobacter hepaticus, a relative of the human stomach carcinogenic bacterium Helicobacter pylori, occurs as a natural liver infection in mouse colonies and causes liver tumors, associated with a long period of hepatic inflammation. However, Helicobacter hepaticus injected into adult mice causes few liver tumors. Early hepatic changes occurred, including an increase in oxidative DNA damage in livers, but later resolved, even though bacteria were still found in feces. These results suggested that the mice successfully defended against the liver attack by the bacteria. We pursued the hypothesis that perinatal exposure was required for tumors to develop. We found that liver tumors could be caused to develop by establishing the infection by intraperitoneal injection of female mice, then breeding these. Multiple liver tumors, including carcinoma, occurred in 40% of their male offspring. There was typical Helicobacter-associated hepatitis in these livers. Interestingly, other protocols that resulted in weaker infection, including intragastric infection of the mothers, and infection via feces in dirty cages, resulted in comparable levels of hepatitis, but few or no tumors. Thus hepatitis was separable from tumorigenesis, just as chronic viral hepatitis in humans does not always lead to liver cancer. Perinatal acquisition of the infection from strongly infected mothers, probably just after birth, established conditions that resulted months later in appearance of liver tumors. DNA damage by a bacterial toxin in the vulnerable neonatal liver, and/or induction of immune tolerance, may be among the important factors. We have established a model that can be used to explore such questions. Our results also point to the perinatal period as a particularly critical time in the acquisition of chronic human infections that are associated with cancer, including hepatic viruses, stomach bacteria, and bladder parasites. In addition, viruses are postulated to contribute to causation of childhood hematopoietic and neurogenic neoplasms. (2) Male mediated transgenerational carcinogenesis (a) Effects of the metal chromium(III). A possible mechanism of paternal effects on childhood cancer is male-mediated transgenerational carcinogenesis. Exposure of male mice to chromium(III), an environmental/occupational metal, results in increased neoplasms and other lesions in the offspring. The nature of these changes suggested hormonal involvement, and we have discovered alterations in serum glucose, corticosterone, IGF1, and the thyroid hormone T3. It is also important to discover the molecular mechanism in the sperm, by which changes-in-gene-expression signals are passed to offspring. Use of representational difference analysis, bisulfite sequencing, and pyrosequencing has revealed hypomethylation in the spacer-promoter region of the ribosomal RNA (rRNA) gene in sperm from Cr(III)-treated males. Ribosomal RNA is implicated in both growth and cancer. To test whether paternally-mediated effects are inherited by offspring, we have examined the gene in embryos and in tissues of young adult offspring, including liver and lung. Extensive statistical analyses of large numbers of litters and of offspring have now been completed, and some surprising and interesing results are clear. Single nucleotide polymorphisms, including one in the spacer-promoter and four alleles in the main promoter of the rRNA gene, are involved. The ratio of the alleles for the main promoter was actually changed by exposure of the fathers, selectively in the offspring lungs. This suggests tissue-specific genomic effects, either selective allelic expansion or gene conversion. If confirmed it will be a qualitatively new finding. The nature of the allele in the main promoter correlated with degree of methylation at five CpG sites in the spacer-promoter, in a tissue specific way. Importantly, these relationships between gene polymorphisms and gene methylation at a distant regulatory site were significantly modulated by treatment of the fathers. Both the chromium(III), and the acid saline vehicle administered as a control, had such effects. Both chemical treatments were found to be stressors of the fathers, causing changes in serum corticosterone and glucose. Types and degrees of stress in the fathers could be part of the underlying mechanism. Potential implications of this for humans are obvious and worth detailed study. (b) Dietary manipulation of fathers. The results with chromium(III) above suggested that stress could underlie the male-mediated transgenerational effects, with physiological changes in hormones and metabolic parameters as components. In humans, nutritional experiences of men, either famine or food excess, have been linked to disease states of their children and grandchildren. Stress could be part of this phenomenon, particularly for famine. Therefore we subjected male mice to short fasting periods (24 hrs) several weeks before mating. Remarkably, this treatment resulted in a highly significant reduction in average serum glucose in both male and female offspring. Alterations in offspring corticosterone, IGF1, and body weight were also noted though more variably. We have now repeated and confirmed this finding. We find that standardization of the time of weaning of the fathers improves the reliability of the model. Consistent results have been obtained for IGF1 as well as glucose. There are interesting implications of these results, beyond cancer risk, as blood glucose levels obviously relate to obesity and diabetes. What if the current epidemic in childhood obesity relates not only to what the children eat, but also to what, and how much, the fathers ate before the children were conceived? Currently we are extending work with the model, to include treatment with corticosterone as a controlled way of mimicking stress, and adding insulin and leptin to the panel of assays.
Showing the most recent 10 out of 22 publications
