Environmental chemical exposures and breast cancer

As a hormone-sensitive condition with no single identifiable cause, breast cancer is a major health problem. It is characterized by a wide range of contributing factors and exposures occurring in different combinations and strengths across a lifetime that may be amplified during periods of enhanced developmental susceptibility and impacted by reproductive patterns and behaviours. The vast majority of cases are oestrogen-receptor positive and occur in women with no family history of the disease suggesting that modifiable risk factors are involved. A substantial body of evidence now links oestrogen-positive breast cancer with environmental exposures. Synthetic chemicals capable of oestrogen mimicry are characteristic of industrial development and have been individually and extensively assessed as risk factors for oestrogen-sensitive cancers. Existing breast cancer risk assessment tools do not take such factors into account. In the absence of consensus on causation and in order to better understand the problem of escalating incidence globally, an expanded, integrated approach broadening the inquiry into individual susceptibility breast cancer is proposed. Applying systems thinking to existing data on oestrogen-modulating environmental exposures and other oestrogenic factors characteristic of Westernisation and their interactions in the exposure, encompassing social, behavioural, environmental, hormonal and genetic factors, can assist in understanding cancer risks and the pursuit of prevention strategies. A new conceptual framework based on a broader understanding of the “system” that underlies the development of breast cancer over a period of many years, incorporating the factors known to contribute to breast cancer risk, could provide a new platform from which government and regulators can promulgate enhanced and more effective prevention strategies.

cases, it reflects the growing number of environmental exposures that impact on the sensitive workings of oestrogen metabolism [39,40]. Evidence from experimental, ecological and epidemiological research collectively examined by expert committees have led to recommendations that the complexities of breast cancer causation should now be explored using a systems approach in which trans-disciplinary research should also be encouraged [23, 24,41]. Sometimes referred to as "team science", trans-disciplinary research is conducted by investigators from a number of disciplines, ranging from the genetic and molecular to the societal and regulatory, working together to devise new conceptual, theoretical and methodological innovations that move beyond discipline-specific approaches towards an integrated solution to a problem [42].
The observation; "We are not students of some subject matter, but students of problems. And problems may cut across the borders of any subject matter or discipline." [43] is exemplified in a recurring hypothesis for breast cancer causation. This proposes that particular characteristics of Westernisation in rapidly developing and newly industrialised countries are contributing to increased incidence in their female populations [44]. "Westernisation" is an umbrella term that attracts mixed and, in many parts of the world, unfavourable reactions. It leads in major changes in cultural behaviours that are protective of health whilst introducing the environmental agents and exposures instrumental to modernization and industrialization. Such agents include large numbers of endocrine disrupting chemicals (EDCs) which have been shown in multiple studies and confirmed in authoritative reviews, as being capable of modulating highly sensitive internal biochemical and hormonal processes [23, 24,45].
On-going research contributes greatly to our growing understanding of breast cancer at a molecular and cellular level and has led to the development of drugs including selective oestrogenreceptor modulators as therapeutic and preventative agents. This approach which is reductionist in nature has given rise to a proliferation of specialisations, funding streams, research projects and reporting but which are, by their nature, geared neither to explaining the problem of increasing incidence worldwide nor promoting prevention.
Molecular approaches towards understanding breast cancer whilst they have the capacity to identify preclinical biomarkers and therapeutic interventions to halt progression of the disease are not aimed specifically at prevention. The objective is, essentially, to cure a condition, not to prevent it from arising in the first place. Prevention in its most basic sense requires a different perspective. It has been proposed that developments in computer technology now provide a means of analysing large, disparate bodies of data using a non-linear, systems approach as a means of overcoming the limitations of the reductionist paradigm when it comes to human health and disease [46][47][48][49]. These developments provide the opportunity to consider increased incidence of breast cancer in a wider, environmental context.
Whilst it is not the role of reductionism to address the multi-factorial origins of cancer, its findings can validate and inform systems biology and thereby the approach proposed here which would help establish whether cancer is what emerges when human interaction with the environment goes wrong.

Breast cancer trends
Global trends since 1973 show increased incidence rates of breast cancer in virtually all regions of the world [38]. The average annual rate of increase in 187 countries between 1980 and 2010 (30 years) has been calculated as 3.1% [50] while incidence worldwide has increased by 20% since 2008. It is now the second most common cancer globally accounting for 12% of all new cancers and 25% of call cancers in women [51].
In the UK, 33% of all new cancers are breast cancers and lifetime risk is currently one in eight [35]. In the US from 2001 to 2010 and notwithstanding a fall of 0.4% in AS and time-adjusted incidence rates in all cancers, breast cancer incidence neither increased nor decreased significantly [52,53]. However, data from the North American Association of Central Cancer Registries (NACCR), show that absolute numbers of breast cancers are currently on the increase [54].
This increase has been argued as attributable to the combination of a growing and an ageing population coupled with more widespread screening [55]. However, accumulating evidence also points to environmentally driven epigenetic changes impacting on endocrine function as a further plausible explanation for the escalation in those breast cancers (the majority) exhibiting hormonal disruption [56]. Rising trends in obesity have been forecast to add significantly to incidence both in the UK and the US especially in post-menopausal women where obesity has been identified as an independent risk factor for ER+ breast cancer [57].
Leaving aside "hereditary" cancers that constitute a distinct, but comparatively small, sub-group of breast cancers the vast majority of cases (>80%) do not display the inherited genetic component exemplified by mutations in the BRACA1, BRACA2 and TP50 genes [58]. It has been proposed that this large majority may reflect instead an accumulation of epigenetic mutations driven by exposures to EDCs [27,45,56,59,60] which amplify the risk of breast cancer via their modulating effects on oestrogen metabolism [61][62][63]. They may also reflect higher use in certain countries of HRT and oral contraception [40,58,64] illustrated by a decline of 8.8% in breast cancer incidence in the US in women aged 40-79 following widespread abandonment of HRT use between 2000 and 2005 [65].
Increases in incidence of breast cancer are not consistent across the globe [38]. The women of East Asia have, traditionally, experienced the lowest rates of breast cancer incidence in the world [62]. However, although rates of incidence are still lower in Asia than in the West, a significant rate of increase has been seen in this region over the past several decades. In Japan and Singapore rates have doubled, and in South Korea tripled, in the past 40 years while China's urban registries have recorded increases in incidence of up to 30% in the decade to 2007 [66][67][68]. These rapid escalations in incidence require explanation and particularly those in South Korea where, in 2012, breast cancer incidence was the highest in Asia (overtaking Hong Kong), with an AS rate of 52.1 per 100,000 of population [69]. Also noteworthy is that the median age of breast cancer diagnosis in China has been reported as being nearly 10 years younger than in the either the US or Europe [70].

Environmental chemical exposures
It has been estimated that there are over 140,000 synthetic chemicals on the market worldwide today with another 1,000 or more being added every year [71]. Over 90% have not been tested for their effects on human health [72]. An estimated 85,000 synthetic chemicals are registered in the United States but with toxicological screening data available for only about 7% [73]. Various bodies, including the President's Cancer Panel (PCP) [74], the IOM [75] and IBCERCC [24] have advocated further study of these environmental chemicals to better understand their potential to cause disease.
The challenges for researchers in doing so are substantial. In some cases, while the impact of an environmental chemical on an adult may be known, its developmental effects on the foetus and in childhood may not be well documented (if at all), and measurable effects in adults may be a reflection of long-term cumulative exposure starting in utero [76,77]. This proposition was articulated by Heindle [78] as the "foetal basis of adult disease" which describes the interaction between the external environment and an individual's genes as a determinant of health/disease in later life beginning in the womb. As well as age at exposure, because some chemicals may be particularly harmful at specific stages of breast tissue development [43], other difficulties arise from duration of exposure and concentrations that can be hard to measure [79], possible effects of "mixing" where a chemical under review may, in combination with another, become toxic [80] and many EDCs exhibit J-type dose response curves [61].
Immigration to an industrialised country, increases exposure to environmental pollutants [80][81][82] and breast cancer risk in later life may be higher in women whose mothers immigrate from undeveloped or developing countries to developed ones during pregnancy or when their daughters are children or adolescent [83]. Migration studies show that women who move from countries with low breast cancer incidence rates soon acquire any higher risk prevailing in the adopted country [84]. Women from Asian countries moving to the United States have been found to experience an 80% increase in risk after just one decade [85,86], a finding replicated in other studies [84,87], while age at immigration is a factor determining whether a woman carries the risk of her country of origin with her or acquires the risk of the country of destination [83].
The proposition that environmental factors may be influencing breast cancer rates has been endorsed in studies of twins. One of the largest of these found that where one twin developed breast cancer the most significant contribution to her risk were the environmental exposures unique to her as opposed to her twin [88]. Inherited genes were found to contribute 27%, shared environmental factors 6% and non-shared environmental factors 67%, of the risk. The study was subsequently re-analysed by others who concluded that "genetic susceptibility makes only a small to moderate contribution" to the incidence of breast cancer [89]. Further studies have since identified more genes that may increase breast cancer risk [90][91][92] but the most recent research suggests that it is epigenetic alterations that are the primary initiators of cancer development [93][94][95][96][97][98] and that these "epimutations" may be caused by environmental chemicals with endocrine disrupting potential [60,99].

Endocrine disrupting chemicals
A large and growing number of studies focusing on endocrine disruption, and specifically oestrogen modulation, point to commonly encountered environmental chemicals as contributing to the increase of oestrogen-positive breast cancer incidence in post-menopausal women [100]. Such studies have been the subject of authoratative and detailed reviews [22,45,101].
An endocrine disrupting compound or chemical (EDC) is defined by the US Environmental Protection Agency (EPA) as "an exogenous agent that interferes with synthesis, secretion, transport, metabolism, binding action, or elimination of natural blood-borne hormones that are present in the body and are responsible for homeostasis, reproduction, and developmental process." EDCs include oestrogen, rated by IARC as a Group One Carcinogen [102]. The 2009 Endocrine Society Scientific Statement cites a large body of evidence supporting the proposition that EDCs contribute to breast cancer risk via their effects on oestrogen metabolism [25].
Both the timing and duration of exposure to EDCs have been proposed as critical in terms of breast cancer incidence [24,101]. It has been observed that in the decades since World War II rates of breast cancer incidence have risen in parallel with the proliferation of synthetic chemicals some now identified as EDCs. A woman's lifetime risk of breast cancer in the US in the 1940's was 1 in 22. Today, the risk is 1 in 8 (12%) [1]. To what extent this increase is attributable to EDCs is the subject of intense debate. However, a woman's risk aged 40 and over has been assessed as 2.5 fold higher in those exposed in utero to the synthetic oestrogen diethylstilbestrol (DES) than in women of the same age who were not exposed, demonstrating the potential of intrauterine events to contribute to the development of cancer in later life [104][105][106].
A survey in 2007 indicated that 216 chemicals had, by that date, been registered by international and national regulatory agencies as having been shown in at least one study to be implicated in mammary gland carcinogenesis [39]. Of those 216, 25 were from occupational exposures (in more than 5000 women), 10 had been registered with the US Food and Drug Administration (FDA) as food additives, and 73 either had been or were present in consumer products or as contaminants in food [45]. Although only six were directly identified as being EDCs, this class of chemicals have been shown not only to persist in the environment but, being fat-soluble, to accumulate in body fat including breast tissue (which is fatty) where they can remain for long periods [107,108].
Breast tissue cells are considered to be at their most susceptible to the carcinogenic effects of EDCs starting in utero and continuing through puberty and adolescence up to the first full-term pregnancy [109]. A study that examined the effects of exposure shortly before birth to polycylic aromatic hydrocarbons (PAHs) from air pollution, vehicle exhaust, tobacco smoke and grilled foods showed altered oestrogen metabolism and an increased risk of breast cancer in the daughters of mothers exposed to higher levels in late pregnancy [110]. The effects of childhood and adolescent exposures to EDCs are difficult to assess but one study in 2007 showed that exposure in childhood or early adolescence to the then widely used but since banned pesticide DDT led to a five-fold increase in breast cancer before the age of fifty [111].
Early, or "precocious" puberty has also been proposed as being a consequence of exposure in preceding years to EDCs [61,112] and which has been demonstrated in animal studies [76,113,114]. Biro and co-workers [115] noted significant increases in early puberty in girls in the US between the early 1990s and 2004-2006 (when their study was conducted) and which the authors propose may be linked to a general increase in BMI over the past 20 years, itself a possible consequence of exposures to environmental agents.
Undisclosed chemicals represent another major potential source of human exposure to EDCs [116][117][118]. Fragranced consumer goods such as air fresheners, cleaning materials and personal care products (PCPs) contain ingredients that emit EDCs often as volatile organic compounds (VOCs). Despite the coming into force in the European Union in June 2007 of REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) neither this, nor legislation in the US require full disclosure of chemical ingredients in PCPs or other chemically-based consumer products where these are described and classified as a "fragrance" notwithstanding that VOCs emitted from such "fragrances" may be individually regulated as being toxic or hazardous [59,119].
Assessing the risk posed by such EDCs even individually is complex and controversial. Largescale epidemiological studies of women with breast cancer exploring the effects of environmental chemicals in combination with other factors have produced contradictory results [120][121][122] serving to illustrate how complex some of the interactions may be and how difficult it is to assign or attribute risk between exposures and in proportion to one another.

Impacts of EDCs on oestrogen metabolism
Oestrogen is lipophilic and accumulates in fatty tissue [123]. Breast tissue is fatty and therefore a potential storage site both for endogenous oestrogen and for fat-soluble EDCs from the environment. Oestrogen metabolism is central to breast cancer, and elevated levels of oestrogen have been consistently associated with increased risk with numerous animal studies demonstrating that exposure to environmentally encountered chemicals in general use in developed countries, such as bisphenol-A (BPA), produce measurable effects on oestrogen metabolism [124][125][126]. Oestrogen receptors (ERs) located in the nucleus of a cell bind with their designated endogenous ligands but can also bind with a range of exogenous synthetic compounds such as BPA which can then impact on cell signaling thorough a number of different pathways [127,128].
The potential for a biological effect is dependent upon the cell type, the concentration of the EDC in the target cell, and the binding affinity of that compound for an ER receptor [129]. Some EDCs exhibit relatively low binding affinities for the ER, suggesting that relatively high concentrations of the compounds are required to induce a response [130]. Studies suggest that EDCs may exert more potent effects during cell growth and particularly in utero when the foetus may be more sensitive to low concentrations of oestrogens [131]. Munoz-de-Toro and co-workers [113] argued that the primary target of bisphenol-A (BPA) is the foetal stroma, the only mammary tissue that expresses oestrogen receptors in utero and that BPA therefore alters stroma-epithelial interactions that mediate mammogenesis. In addition to this direct effect on mammary gland development, it is further postulated that BPA affects the hypothalamus and therefore the regulation of mammotrophic hormones at puberty, another critical window of susceptibility to endocrine disruption [132].
Studies in the 1960s and 1970s started to indentify the oestrogenicity of a number of industrial compounds and the pesticides o,p-DDT, kepone, methoxychlor, phenolic derivatives and polychlorinated biphenyls (PCBs). In the late 1990s, other environmental chemicals were added to the list including the pesticides toxaphene, dieldrin and endosulphan, several different compounds used in the food industry, antioxidants such a t-butylhydroxyanisole; plasticizers such as benzylbutylphthalate and 4-OH-alkylphenols along with substances used in dental restorations, such as BPA [133]. Later, complex mixtures of pollutants occurring in the environment such as diesel exhaust particles were also shown to have oestrogenic activity and were classified as EDCs [134].
Multiple studies demonstrate the potential of synthetic oestrogens (such as BPA) from environmental exposures to lead to long-term health problems in wildlife where they have been shown to impact on fecundity and fertilisation rates [103,123,[135][136][137][138]. Oestrogens are important regulators of the metabolic requirements for reproduction and have thereby been implicated in the aetiology inter alia of the feminization of fish and testicular dysgenesis including hypospadias, cryptorchidism [139,140] and testicular cancer [141]. In humans, exposure to animal oestrogens is primarily through diet with meat and dairy foods a principal source [142] and it has recently been reported that dietary oestrogens may increase ER+ breast cancer risk in post-menopausal women whose diets are higher in red meat and lower in whole grains and coffee [143].
Exogenous EDCs also include naturally occurring animal, phyto-and mycoestrogens that have the capacity to mimic endogenous oestrogen action and thereby affect hormone levels and/or bind to the ER [144]. Some have been shown to be hormonally active at extremely low concentrations, similar to endogenous hormones [145,146] and as being able to cross the placental barrier and enter foetal circulation [147][148][149][150]. It has also been demonstrated that those with smaller molecular structure are able to cross the blood-brain barrier which is still underdeveloped at birth [151].

Discussion
Up until now, breast cancer has been approached on the basis that understanding it fully can be achieved by reducing it to manageable components and examining each of them in turn and in depth. Such methods have done much over the past decades to elucidate the processes involved in the aetiology of the disease at both molecular and cellular level leading to improved treatment and, in consequence, falling mortality rates in most developed countries [38].
Such progress notwithstanding, it cannot be said with certainty that exposure to any one of the established risk factors (or any combination thereof) "causes" breast cancer. This makes an effective strategy for primary prevention difficult both to design and to deliver. In the meantime, incidence continues to rise across the world. A different approach is required, one that is not in competition with the reductionist paradigm but supplemental to it. Systems thinking, which tackles issues that are embedded in complexity and especially those that involve human activity would be appropriate to the task because it addresses the interconnectedness that is usually absent in the reductionist approach.
Many of the problems facing mankind comprise complex, dynamic systems consisting of multiple contributory factors that are both interdependent and self-organising and involve human interactions. A self-organising system has been defined as one "in which elements interact in order to achieve a global function or behaviour. This function or behaviour is not imposed by one or few elements, nor is it determined hierarchically. It is achieved dynamically as the elements interact with one another. These interactions produce feedbacks that regulate the system" [152]. In seeking to understand such problems, science has focused, historically, on understanding and testing individual components of the system rather than developing a more holistic and integrated understanding of "the problem" in its totality. As a consequence, one or more crucial aspects of the system (and particularly their interactions) may be excluded from the formulation of the "solution". This means in turn that management of "the problem" is likely to be less effective as well as encouraging adoption and pursuit of a solution that may, in the absence of a full understanding of the underlying cause, create as many (or more) problems than it solves. A systems approach to problem solving focuses on "the problem" instead of "a solution" and looks at the interactions between its various components [153]. It involves a holistic, step-by-step process the aim of which is to understand the problem in its totality, proposing that the solution will emerge from a fuller understanding of the problem [154] but even where it does not, focus will usefully have been shifted to goals that can be achieved by regulation and policy making [155].
The aetiology of cancer (as distinct from its biological hallmarks once developed) fits this description of a complex system well. Increased incidence of breast cancer (the problem) emerges from the self-organisation of external and internal factors interacting with one another (the system) giving rise to the behaviour (cancer). It has therefore been proposed that by looking for patterns within the exposome that reflect human homeostasis (i.e. the non-cancerous state) and identifying when and what changes or perturbations arise from environmental exposures, some consensus on causation of cancer may be achieved [28]. The exposome, a concept first proposed by Wild [156], represents "the totality of exposures received by a person during life, encompasses all sources of toxicants and, therefore, offers scientists an agnostic approach for investigating the environmental causes of chronic diseases." [157]. This concept of the exposome is consistent with developments in epigenetic research which have shown that cancer is preceded by abnormalities arising from the epigenetic reprogramming of gene expression and specifically those "guardian" genes that have been found to be protective against cancer [158].
Bearing in mind the probable number of environmental EDCs with potential specifically for oestrogenic activity, the biochemical complexities of oestrogen metabolism (including interactions with other agents) and oestrogen's pivotal role in the development of ER+ breast cancer, research into the effects of individual EDCs on breast cancer risk needs to be complemented by a more multiconnected, web-like conceptual framework in which to examine escalating incidence. Focused on prevention, such an integrated, holistic approach would aim to synthesise knowledge of the molecular and biological processes involved in breast cancer with a growing parallel body of literature on environmental exposures that have been shown to modulate oestrogen metabolism.
A number of models designed to calculate breast cancer risk have been developed over the past three decades. The Gail (1989) model (now incorporated in The Breast Cancer Risk Assessment Tool (BCRAT)) calculates, by reference to five largely non-genetic variables, the probability of developing breast cancer over a given lifespan. Later models include BRCAPR0 (after the BRCA1 and BRACA2 genes) designed to assess probability of germline mutations in high-penetrance genes; the Claus risk tables [159] aimed at estimating risk of breast cancer in women with a first degree family history of ovarian cancer; the Tyrer-Cuzick model and the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADCEA). Between them they incorporate some, but not all, established risk factors and the Claus model does not take into account any non-hereditary factors. Validation studies have produced mixed results [160,161].
What none of these models does, is to take into account potentially modifiable environmental exposures across a lifetime nor assess the potential for their interactions or cumulative effects in the development of breast cancer. A multi-level mathematical model of post-menopausal breast cancer causation addressing these and other factors, including social and behavioural, has been constructed by Hiatt and co-workers for the California Breast Cancer Research Programme (CBCRP) [162]. The model incorporates the strength of evidence for, and the relative strengths of associations arising out of the interactions between, risk factors in four designated and hierarchically constructed domains. What Hiatt's model clearly demonstrates is the complex origins of cancer lending strength to the proposition that breast cancer is the result of interactions between numerous factors over a protracted period that in particular, but as yet unclear, combinations and levels of exposure lead from health to disease.
Whilst aimed specifically at identifying risk of post-menopausal breast cancer in California in response to marked increases in incidence in the preceding years, the CBCRP model (or parts of it) could usefully be applied in the analysis of cancer scenarios in developing countries, particularly in East Asia, which have seen rapid increases in incidence over the past several decades.
This study suggests that prevention of breast cancer depends not only on avoiding and reducing exposure to traditional risk factors but also on recognition and elimination of exposures to carcinogens especially oestrogens and other chemicals in the environment capable of upsetting finely tuned internal hormonal balance. This is made more difficult in developing countries with high rates of incidence where there are conflicting priorities between aspirations for economic development and prosperity on the one hand and concerns over environmental degradation and loss of cultural traditions beneficial to health on the other. Difficult though it may be to achieve, understanding and explaining the dramatic rise in incidence of breast cancer particularly in East Asia, requires assessment of the likely interactions between the established risk factors and the emerging risks posed by EDCs and other environmental exposures characteristic of inustrialsiation.
Reducing exposure to such compounds will require a political and cultural shift so that protection of public health is regarded as paramount. This in turn will demand a comprehensive re-thinking and re-ordering of priorities by science, industry and government together with co-operation between independent organisations and institutions. All share a joint responsibility for delivering primary prevention which is first about identifying and then eliminating the causes of a disease before it affects people. Historically this has been based on a common sense approach together with recognition of either proved or suspected causative agents. A new conceptual framework incorporating the factors known to contribute to breast cancer risk together with those known to confer protection from it, could provide a new platform from which government and regulators can promulgate enhanced and more effective prevention strategies based on a broader understanding of the "system" that is breast cancer. Systems thinking has evolved to understand the reality of how such interactions affect the whole (in this case increasing incidence) and is therefore consistent with the holistic approach to primary breast cancer prevention proposed here.

Conclusion
Current strategies to halt the rise in incidence of breast cancer worldwide have not been effective. In the European Union, where there are more than twice as many new breast cancers annually than in any other site; a woman is diagnosed with breast cancer every 2.5 minutes [38]. Major advances in cancer treatment have not been accompanied by similar achievements in reducing cancer incidence through increased prevention. Globally, rates rose by 20% between 1980 and 2010 and emerging evidence suggests that age of onset is falling. Incidence is projected to continue rising. At the same time, a growing number of compounds in the environment are being shown to have the capacity, whether individually or in combination, to interfere adversely with normal oestrogen metabolism. What is the connection?
We may not be certain what causes breast cancer but there is a good body of evidence to show how risk can be mitigated. What is required is a way of linking that evidence to risk factors so as to identify the best protection "package" for women, deliverable via upstream changes in policy-making, aimed at creating a safer environment and better health culture. In setting about this daunting task it should be borne in mind that risk alone is not a cause of illness. Risk is the result of exposure to a hazard. If either hazard or exposure can be removed from the equation there is no risk. Avoiding hazard and reducing exposure are therefore key to decreasing risk and in this regulators and policy-makers will be key.