Issue 28 – Breast Cancer

Rethinking the origins of breast cancer

Naval medical officers examining mammogram images

BPA has been implicated in breast cancer but is not a mutagen - so how might non-mutagenic agents initiate cancer? (Image: Wikimedia)

Many forms of breast cancer are hormone-dependent, with tumour growth accelerated by exposure to oestrogens. There is concern that some man-made chemicals with oestrogen-like properties, such as bisphenol-A (BPA), may cause breast cancer. BPA is not, however, thought to directly damage cell DNA – as an endocrine disrupting chemical, it only interferes with hormone signalling.

Understanding how endocrine disruptors might cause cancer could, in fact, have profound implications for our understanding of what causes tumours to appear in the body. This month, we ask: If endocrine-disrupting chemicals don’t cause genetic mutations, how can they cause breast cancer – and what does this mean for understanding carcinogenesis?

Claims of carcinogenicity of EDCs

There has been growing concern about the potential for endocrine-disrupting chemicals (EDCs), substances which disrupt “the synthesis, release, transport, metabolism, binding, action or elimination of natural hormones in the body”, to cause cancer. Campaign groups have long been sounding the alarm, while in June the US President’s Cancer Panel, in its bi-annual review of cancer (pdf), highlighted these concerns as an emerging priority issue in understanding and preventing cancer.

Concerns about the carcinogenicity of EDCs should be puzzling. Very few EDCs are formally listed by the International Agency for Research on Cancer (IARC) as carcinogens. The common understanding of how cancers are initiated, known as the Somatic Theory of Carcinogenesis, is that cellular DNA becomes corrupted by introduction of an agent which directly damages the cell’s DNA. The mutation causes the cell to start multiplying uncontrollably, leading to formation and growth of a tumour.

Although EDCs interfere with hormone signalling, most are not mutagens. It is therefore not obvious how, within the bounds of somatic theory, they should cause cancer. So why are there concerns about EDCs and cancer?

The hormone-dependency of breast cancers

It is already known that oestrogens affect breast cancer. The hormonal responsiveness of breast cancer has been a subject of published research in the journal Cancer since the 1950s and there have been trials of hormonal therapies for breast cancer since the 1960s. A review study published in Cancer in 1960 found a pronounced increase in oestrogen-responsive breast tumours in women after menopause, providing early confirmation that changes in hormone levels have a significant effect on breast cancer. It is now acknowledged that a number of forms of breast cancer are hormone-dependent and drugs such as tamoxifen have been developed to address exactly these types of cancer.

The basic hypothesis is that risk of developing breast cancer increases with longer lifetime exposure to ovarian hormones, which is why age at menarche, first pregnancy, menopause, lactation and giving birth are all considered risk factors for the disease. Most research is focused on how hormones change the rate of cancer progression or remission by stimulating breast cell division and supporting the growth of oestrogen-dependent tumours, rather than how hormones might actually cause cancer.

How EDCs might cause breast cancer

Hormones do have a clear role in stimulating cell division. In the breast, ductal cells (the cells associated with milk delivery) proliferate most rapidly when oestrogen levels are highest during the menstrual cycle. Because cell multiplication requires copies of DNA to be made each time a cell divides, the longer that proliferation lasts then the more likely it is an error in copying the DNA will result in a cell dividing uncontrollably, resulting in a tumour.

Researchers speculate that small, permanent increases in oestrogen levels, as might be caused by continuous exposure to EDCs through diet, would lengthen the period of ductal cell proliferation in each menstrual cycle. Over a lifetime the sum total period of proliferation is longer, increasing the number of cell divisions and therefore the risk of copying errors in the DNA – and in turn the development of breast tumours.

Excess oestrogen exposure during developmental periods has also been shown to limit the ability of cells to deal with carcinogenic challenges: experiments have shown that rats which are exposed prenatally to dioxin then again to a chemical carcinogen at puberty are much more likely to develop mammary tumours than rats exposed to dioxin alone. Like dioxin, BPA has been observed to make rats more sensitive to carcinogens: BPA plus a sub-carcinogenic dose of nitrosomethylurea makes rats develop mammary cancers, where exposure to nitrosomethylurea on its own does not.

EDCs as a challenge to somatic theory

Both of these mechanisms are consistent with the premise that cancer is caused by genetic mutations, with EDCs either increasing the chances of mutation by encouraging cell division, or by increasing the susceptibility of damage to DNA by mutagens.

There are, however, aspects of hormone-related cancers which are difficult to explain for a theory of carcinogenesis based on mutation, which is why in a paper recently published in Nature Reviews, two researchers, Professors Ana Soto and Carlos Sonnenschein of Tufts University School of Medicine (USA), argue that EDCs lend support to a different way of understanding carcinogenesis: that cancer is not a result of mutation but of breakdown in communication between tissue types – the so-called Tissue-Field Theory of Carcinogenesis.

Soto and Sonnenschein believe it is hard for somatic theory to explain why hormone related cancers regress after hormone withdrawal, as has been demonstrated in both breast and prostate cancers. Also, if cancer is a disease of uncontrolled cell proliferation, they ask why it does not progress so much faster than it actually does – although cells can multiply relatively quickly, some cancers have latency periods of ten or more years. Furthermore, if exposure to EDCs occurs throughout the body, and cancer is a matter of mutation of DNA, why is it that cancers tend to occur in a particular place?

The somatic theory also struggles to explain some very interesting experimental phenomena which show that changing a cell’s environment can cause it to start – and stop – multiplying uncontrollably. The body’s organs are made up of parenchymal cells (the functional elements of an organ, such as epithelial cells) and stromal cells which provide the framework or scaffold.

In breast cancer, animal experiments have shown that when cancerous epithelial cells (the cells which line hollow organs and glands and those that make up the outer surface of the body) are recombined with stroma cells of a living organism, then the cancer disappears, suggesting that breast cancer has something to do with abnormal structure of the breast, as well as (or even instead of) mutations in individual cells.

The experiment with epithelia and stroma indicates that the existence of at least some breast tumours is dependent on its cellular environment, meaning uncontrolled multiplication is not entirely determined by the inner workings of the cell. This strongly suggests that cancer is not, at least in these cases, caused by mutation, but by interference in communication between tissue types – in essence, the epithelial cells in question become disoriented and, instead of sitting quietly, start multiplying.

EDCs and a less bottom-up theory of carcinogenesis

Soto and Sonnenschein believe the idea that DNA mutations are the cause of cancer is a very bottom-up, overly reductionist view of carcinogenesis, implying that organism development is simply the unfolding of a genetic program fully contained within the DNA of the cell. In this picture, cancer is solely the result of the program data, the DNA, becoming corrupted and telling the cell to do things it should not. Breast structure would not, on this model, determine how individual breast cells behave.

Soto and Sonnenschein believe the relation is much more two-way; that the cellular and organism environment are also able to determine the organisms appearance and its propensity to developing cancer – essentially, that higher-level structures can determine the behaviour of its component parts. Soto and Sonnenschein think it more likely that cancers are caused by altered reciprocal interactions between cells, and this is why exposure to hormones at the wrong levels at the wrong times is able to cause cancer.

This is particularly important pre- and perinatally and during puberty, when organs are developing. At these times hormones mainly affect tissue structure and organisation. Any alterations at these times are irreversible, unlike in adults where hormone effects are mainly activational and reversible. During developmental periods, hormones can potentially pre-program an increased risk of cancer for the organism’s entire life, which is why women who were temporarily exposed to DES in the womb see more reproductive cancers in later life.

There is also evidence that hormone exposures can have these effects at much lower doses prenatally than needed for adult health effects: doses of BPA five times lower than needed to induce proliferative effects in the uterus of prepubescent mice can cause morphological and functional changes in the uterus of the mouse foetus.

What does this mean for cancer prevention?

Whether or not cancer is caused by DNA mutation or breakdown in communication between cell types is mainly of intellectual significance, as the whys and wherefores of carcinogenesis do not change the need for identification and control of carcinogens.

However, there is an implication for risk assessment, the process by which it is decided whether or not a chemical poses enough potential for harm to warrant it being banned. Explanatory deficiencies in the somatic theory and growing prominence of the tissue-field theory show we have to be prepared to accept that the role which chemicals such as endocrine disruptors play in carcinogenesis may be more subtle than simple mutation of the DNA – and allow for these possibilities when assessing the risks these chemicals pose to health. Risk assessments need to be protective in spite of an incomplete understanding of the model and the evidence, so people are protected from harm while scientific understanding of how the environment affects health shifts and changes – hence calls for precautionary approaches to chemicals regulation.

Further reading

The Endocrine Society (2009) Scientific Statement on Endocrine Disruptors [pdf]. esp. pp13-16

Soto A and C Sonnenschein (1999) The Society of Cells. (The New England Journal of Medicine’s review of the book is can be read on

Soto A and C Sonnenschein (2010) Environmental causes of cancer: Endocrine disruptors as carcinogens.

New paper calls for changes in research practices to help regulators

Table showing PCB-Neurotox Research Diversity

A table, extracted from the study, which shows the diversity in approaches to assessing neurotoxicity of PCBs, which the researchers say undermines efforts to determine the harm they do. Click image to enlarge.

Faced with challenges of systematically reviewing evidence from life sciences, in a new paper researchers make recommendations for how future research approaches could meet requirements of the weight-of-evidence approach favoured by regulators.

Researchers examining whether or not existing studies allow certainty about the neurotoxicity of PCBs have concluded that existing research practices are not adequately supportive of current weight-of-evidence approaches to evaluating potential harm.

The reviewers attempted systematic reviews with meta-analysis of research into the effects on human neurodevelopment of PCBs, a banned class of chemical used in electrical equipment, in order to test whether or not there is enough comparable evidence to come to a conclusion.

Regulators generally use a weight-of-evidence approach to draw conclusions about whether or not there is a causal relationship between exposure to an environmental chemical and a health outcome. Weight-of-evidence evaluation is only possible, however, if the studies under review are adequately comparable.

The researchers chose PCBs because it is a class of chemicals for which there is a substantial body of peer-reviewed literature, but for which there is also a substantial amount of disagreement about the presence of a causal link between pre- and neonatal PCB exposure and neurodevelopmental outcomes.

Although 60 studies met the researchers’ criteria after a literature review, diversity in approach to identifying PCB exposure and measuring its effects meant the researchers were unable to confirm whether or not any particular study replicated the findings of another.

If replicability of findings can’t be tested it is not possible to use a weight-of-evidence approach to determine, even among 60 studies, whether or not PCBs cause neurodevelopmental problems.

Systematic review is commonplace in evidence-based medicine and critical for drug approval and determining which clinical interventions are most likely to succeed in a given circumstance.

The researchers argue that, in order to support regulation based on a weight-of-evidence approach, fundamental changes need to be made to how research in the life sciences is carried out. Their recommendations include:

  1. Developing a baseline metric of exposure which can be used by all future studies
  2. Ensuring any data collected are available for WOE assessment rather than kept secret
  3. Convening of a multi-stakeholder group should be convened to discuss best practices
  4. Journals making it a requirement that summary statistics are archived

There have been recent calls for reform along similar lines in medical trials, with recommendations that data be made generally available and that medical journals should focus on studies which review data, rather than on publishing the results of clinical trials.

5&5: News and science highlights from June

5 Studies

Parabens act like oestrogens in pre-pubescent female rats: Study concluding that long-term exposure to parabens can suppress hormonal responsiveness and alter the structure of reproductive target tissues in female rats. The study also suggests that parabens are thyrotoxic during a critical stage of development in female rats.

Higher urine BPA levels associated with impotence: New research finding that men with higher exposure to BPA, determined by measuring levels in urine, are more likely to experience reduced sexual function. Higher levels increased the likelihood.

Transfer of bisphenol A across the human placenta: Human study finding “bisphenol A at environmentally relevant levels can transfer across the human placenta, mainly in active unconjugated form.” This is significant because risk assessments generally assume BPA to be in inactive, conjugated form, and therefore not a problem.

Exposure to PFCs and ADHD in U.S. Children Aged 12-15 Years: Higher exposure to PFCs, often used in stain-resistant and non-stick surfaces, is found to be consistent with increased odds of ADHD in children. Given that exposure to PFCs is widespread, the researchers recommend follow-up studies.

Animal study suggests BPA alters breast hormone expression via non-hormonal mechanism: Study on mice finding that BPA and DES alter breast hormone expression via an epigenetic mechanism and that BPA has equal strength to DES via this mechanism. BPA is generally thought to affect the breast because of its hormone-mimicking qualities; this study indicates it can do this in other ways as well.

5 Stories

Scientists Devise New “Benign by Design” Drugs, Paints, Pesticides and More: Scientific American looks at green chemistry and whether it will take new regulation to ensure its principles of low-energy and low-toxicity chemicals manufacture will be implemented more broadly.

‘Dirty dozen’ foods carry most pesticide residue: CNN reports on pesticide levels in US foods, looking at findings that soft-skinned fruits, lettuces and celery carry far more pesticide contamination (up to 67 types for celery) than onions, cabbage or grapefruit, and that substituting the 12 worst can cut your dietary intake of pesticides by up to 80%.

Triclosan Comes under Scrutiny: Environmental Health Perspectives reports on how the use of triclosan, added to products as an antibacterial agent, is coming under increased scrutiny – new research suggests hazards to health, increasing prevalence in the environment and unclear antibacterial benefits in consumer products.

Do Green Building Standards Minimize Human Health Concerns? LEED certification has played an important role in catalysing green building in the US. Its focus on energy, however, may be encouraging building practices which, although energy-efficient, contribute to sick building syndrome.

French lawmakers ban baby bottle chemical: French lawmakers passed legislation in June banning baby bottles containing the chemical bisphenol-A, suspected of harming human development.


RSS feed for comments on this post.

  1. It is in my view unfortunate and counter-productive to breakthrough development in cancer research, to continue to present theories of carcinogenesis as mutually exclusive. They clearly are not and more likely part of a more comprehensive whole. If the tissue field theory gains more support from critical experimental test then certainly more human and financial resources should be devoted to discovery based upon the new paradigm. However,I do not subscribe to the notion that the somatic mutation theory should be abandoned entirely in its’ favor. The theory that tumor suppressor genes and oncogenes collaborate in many different tumor types has much phenomenological support. In addition, empirical observation also supports three other competing models for carcinogenesis, the epigenetic model, the mitochondrial model and the aneuploid model. All of which clearly have experimental support in the literature. What is greatly needed as we move forward towards a better understanding for the initiation of cancer and its early progression is a unifying and simplifying hypothesis that accommodates the essential elements of these five competing models, all of which have significant empirical support.

    • Thanks for the helpful comment. Re-reading this article, I think we could have made it clearer that this shouldn’t be seen as an either/or issue. It was interesting to us how EDCs might challenge a purely somatic model of carcinogenesis. It’s all quite hypothetical – epigenetic considerations are also clearly very important and interesting. The challenge with writing about environmental causes of cancer, as such a complex disease, is knowing where even to begin.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Blog at
Entries and comments feeds.

%d bloggers like this: