Birnbaum, LS and SE Fenton. 2002. Cancer And Developmental Exposure to Endocrine Disruptors. Environmental Health Perspectives doi:10.1289/ehp.5686.

Birnbaum and Fenton review a wide array of experimental evidence from animals showing that exposure to endocrine-disrupting compounds in early development can cause cancer and/or increase sensitivity to cancer-causing agents later in life.

Their review then highlights how few human studies have been built upon this understanding. Almost all human epidemiological research into cancer risk from contaminant exposures examines chemical levels only at the time of diagnosis or afterward. This approach misses entirely the period of developmental sensitivity to exposures that animal studies have identified [a recent example being the Long Island Breast Cancer study].

Birnbaum and Fenton summarize this with two important questions about the way most human studies have been conducted:

One interesting pattern that emerges from their review of animal experiments is that in utero exposure to several endocrine disrupting compounds (including dioxin, atrazine and bisphenol A) can alter mammary gland development in ways that prolong the period of sensitivity to carcinogens. This suggests a different way of thinking about the contribution of these contaminants to carcinogenesis: even if they don't cause cancer directly, they contribute to cancer risk by increasing vulnerability. No epidemiological study has ever attempted to test for this sort of effect in people.

Birbaum and Fenton begin their review with a very succinct summary of the basic reasons why early developmental stages through puberty are especially vulnerable to chemical exposure:

• The pace of and the nature of change in a developing fetus or child is dramatically enhanced compared with that in an adult. An embryo and fetus is changing quickly, with rapid cycles of cell division and growth, and massive changes in the patterns of gene activation over time. These cycles provide extensive opportunities for mistakes to occur and be incorporated into the organism. Sometimes these mistakes are mutagenic, sometimes they are based on changes outside the genes. Comparable periods of cell division, differentiation and growth are long since over in an adult. Hence the chances for mistakes to be made and incorporated aren't nearly as common in an adult compared to an embryo or fetus.

• Second, physiological barriers such as the blood-brain barrier are not yet complete in the womb.

• Finally, the enzymatic mechanisms that work to detoxify contaminants in adults are not fully developed until after birth.

They then cite two examples offering "unequivocal evidence" from human studies demonstrating that developmental exposures can cause cancers in children and young adults. This comes from studies from ionizing radiation and the synthetic estrogen diethylstilbestrol.

Studies of a range of other human exposures suggest causal relationships between developmental exposure and subsequent cancers, but the evidence, while strong, is not as conclusive as for DES and radiation. These other studies implicate occupational exposures of parents to brain cancers in children, pesticides, paints, paint thinners and solvents in causing leukemia, and cigarette smoke and childhood cancer, among others. In this discussion Birnbaum and Fenton also comment on one of the chief obstacles impeding epidemiological studies of childhood cancers: they are so uncommon in the general population that prospective studies rarely have a sufficient sample size to find positive results.

Turning to animal studies where experimental studies are possible, Birnbaum and Fenton observe that "data from experimental animal studies for developmental exposures and early lifestage or adult cancer is far more extensive and convincing than the current epidemiological data."

They review an extensive literature showing conclusively that prenatal and early postnatal exposure to various types of radiation and to many different chemicals cause cancers later in life in the exposed animals.

Induced tumors span the gamut: skin carcinogenesis, liver ovarian, uterine and pituitary tumors from prenatal x-ray exposure; respiratory tumors from a wide array of mutagens, including ethyl-nitrosourea (ENU); uterine tumors from ENU, dimethylbenz[a]anthracene (DMBA) and urethane; nervous system tumors in a wide range of mammalian species from ENU, etc.

In careful, elegant experiments, scientists have also shown that developmental exposure can heighten sensitivity to carcinogens later in life. For example, studies of the industrial chemical, ethylene thiourea (ETU), found that perinatal exposures alone did not affect cancer risk. Individuals that had been exposed perinatally, however, developed more cancers when exposed in adulthood than did others, also exposed in adulthood, who had not been exposed perinatally (Chhabra et al. 1992).

They conclude this section of their review:

Birnbaum and Fenton then turn their focus specifically to endocrine-disrupting compounds, beginning with a reminder that fetal exposure to one endocrine disruptor, diethystilbestrol, is clearly linked to vaginal cancer in young adult women and possibly linked to testicular cancer in young adult men. Beyond those cases, there have been almost no human epidemiological studies examining links between developmental exposures and subsequent cancer risk.

Experimental work with animals is more extensive, however, and as above with exposures in general, it shows convincingly links between exposures and cancer causation. The classic case here also is diethylstilbestrol, but addresses many other compounds also.

For example, when pregnant mice are injected with genistein (a phytoestrogen abundant in soy), their female daughters develop mammary gland tumors. Neonatal mice injected with very low levels of genistein develop uterine cancer. Despite these results and the popularity of soy formula, according to Birnbaum and Fenton "there is a dramatic lack of epidemiologic studies evaluating the effect of maternal (fetal) or infant soy
consumption and correlation with breast, uterine, or testicular cancer."

Polyhalogenated aromatic hydrocarbons
Birnbaum and Fenton provide an excellent short synopsis of experimental work on the polyhalogenated aromatic hydrocarbons, that family of brominated and/or chlorinated compounds that include the polychlorinated biphenyls (PCBs), polybrominated biphenyls (PBBs) and dioxins. Animal experiments establish definitive causal links between developmental exposure to an array of these compounds and adverse effects later in life, including cancers.

Specifically with respect to dioxin, "there have been at least 18 published animal cancer studies in rats, mice, hamsters, and fish demonstrating cancer positive outcomes in both sexes and at multiple sites. Experimental studies have also demonstrated that dioxins are potent tumor promoters, enhancing both the incidence and multiplicity of tumors at multiple sites following initiation with a direct acting mutagen."

Birnbaum and Fenton acknowledge that evidence from animal experiments about dioxin's link to mammary tumors is contradictory, but they point to very recent human epidemiological work indicating an association between developmental exposure and heightened breast cancer risk. In general, the published animal research on in utero exposure indicates that dioxin induces changes in mammary gland development and structure that prolong the developmental period of sensitivity to carcinogenesis. One interesting set of animal experiments involved prenatal exposure to dioxin followed by exposure to DMBA at sexual maturity. This more than doubled the number of mammary tumors. Another suggests effects on maternal pituitary weight and prolactin levels consistent with elevated estrogen levels.

Atrazine and Bisphenol A
Recent studies of these two compounds show that in utero exposure can prolong the period of sensitivity to carcinogens. Atrazine also clearly alters the pattern of mammary gland development around puberty.

Birnbaum and Fenton's concluding paragraph, below, should be read by all epidemiologists contemplating work on endocrine disruption and carcinogenesis, as well as by policy advocates, reporters and editorial writers. It effectively rebukes any claims (e.g., in the New York Times) that existing studies on the links between EDCs and cancer risk exonerate the contaminants.