HYPOTHESIS:
Underarm Cosmetics are a Cause of Breast Cancer
by Philippa D Darbre

THE HYPOTHESIS
I propose a novel hypothesis that the chemical constituents of antiperspirant/deodorant cosmetics applied to the underarm area are a cause of breast cancer.

BACKGROUND
Breast cancer is the major cancer of women in the western world, but shows a rising incidence all over the world (Ursin et al, 1994). Epidemiological evidence demonstrates that 90% of breast cancers are environmental in origin (Lipworth, 1995) and linked to a western lifestyle, but the specific environmental causes have never been identified. In recent years, two tumour suppressor genes, BRCA1 (Miki et al, 1994) and BRCA2 (Wooster et al, 1995), have been cloned which confer susceptibility to breast cancer, but loss of function of these genes accounts for a maximum of 5-10% of breast cancer (Easton et al, 1993). The main identified risk factors are hormonal and linked, in particular, to lifetime exposure to oestrogen through variations in menarche, menopause, childbirth and personal choices such as use of the contraceptive pill or hormone replacement therapy (Lipworth, 1995). Diet, smoking and alcohol can also exert influences (Lipworth, 1995). However, these known risk factors confer only a small increased risk and the main causative agents of breast cancer remain unaccounted for.

A novel, unexplored but plausible explanation for the cause of breast cancer is the use of underarm cosmetics. These chemicals are applied repetitively and frequently to an area directly adjacent to the breast area. They are not rinsed off but left always on the skin. They are used by women with ever increasing frequency and by ever younger girls before puberty. They are now also being used by some men in increasing amounts. The progressive increase in use of these cosmetics in the western world over the past 100 years is illustrated by US sales figures (Laden and Felger, 1988). Sales in 1914 reached sufficient levels to support national advertising, and rose from there to US business worth $30million in 1947 to $300million in 1970 to over $1billion in 1983 (Laden and Felger, 1988). Although these cosmetics are regulated as over-the-counter drugs (Laden and Felger, 1988), there are no indications on the containers of the safe level of usage or whether the constituent chemicals are safe for prolonged use by young children before puberty. Contrary to common belief, there is great diversity in use of these cosmetics across the population. My own unpublished survey within the University of Reading shows that women use a variety of different products, each containing different types and amounts of chemicals and that frequency of use varies from never to more than 5 times a day. Such diversity in usage provides ample possibility for cancer to arise through quantity used, through pattern of usage or through individual susceptibility to specific product formulations.

ANATOMICAL SITE OF BREAST CANCER
The strongest supporting evidence for a role for underarm cosmetics in breast cancer comes from published clinical observations showing disproportionately high incidence of breast cancer both in the upper outer quadrant of the breast and in the left breast. Numerous clinical studies, dating back decades, have shown that the upper outer quadrant of the breast is the most frequent site of carcinoma. This basic observation has now become textbook fact (Haagensen, 1971) and for countries as different as India (Hussain et al, 1994), the West Indes (Raju and Naraynsingh, 1989), and Italy (Azzena et al, 1994), and irrespective of race within any one country (Patterson et al, 1998). The upper outer quadrant is also the most frequent site of breast cancer in men as well as in women (Jaiyesimi et al, 1992; Rizk et al, 1994). The accepted explanation is that this region of the breast contains a greater proportion of the epithelial target tissue but evidence for this explanation seems to be largely anecdotal and as written by Haagensen himself (Haagensen, 1971) "This is a purely anecdotal explanation, but I know of no other".

An alternative explanation of these studies could simply be that the upper outer quadrant is the local area adjacent to which the underarm cosmetics are applied. Since they are applied in large amounts, they may simply penetrate through the skin of the local area without even invoking any major physiological carrier such as blood or lymphatics. It is interesting to note that the disproportionate incidence of female breast cancer in the upper outer quadrant rises with year of publication, from 30.9% in 1926 (Lane-Claypon, 1926) to 43-48% in 1947-1967 (Truscott, 1947; Harnett, 1948; Nohrman, 1949; Smithers et al, 1952; Donegan and Spratt,1967) and to 60.7% in 1994 (Azzena et al, 1994). If this observation of published trends is not just a reflection of different study populations and reflects some real increasing incidence in the upper outer quadrant relative to other quadrants with time, then this would question the explanation as being due always to more epithelial tissue in that region. Other workers have also queried explanatory dogma through their studies showing an even distribution of cancer between quadrants in large and small breasts, despite the less marked quadrant distribution of tissue in the smaller breasts (Rimsten, 1976).

Another set of numerous studies, also without adequate explanation, show that the left breast is more prone to development of cancer than the right breast in both female (Busk and Clemmesen, 1947; Harnett, 1948; Smithers et al, 1952; Garfinkel et al, 1959; Haagensen,1971) and male (Busk and Clemmensen, 1947; Jepson and Fentiman, 1998) breast cancer. This has been attributed to more epithelial cells on the left side of the breast due to preferential vascular supply to the left side of the body during intrauterine cardiac development (Jepson and Fentiman, 1998) but a simpler and equally plausible reason could relate simply to the right-handed nature of a majority of the population resulting in a greater application of chemicals to the left underarm area. This could be tested by study of the quadrant incidence in left- and right-handed people, if such data were available.

Lastly, the reasons for enhanced risk of (Chen et al, 1999) or even coexistence of (Fenig et al, 1975) contralateral breast disease have yet to be explained. This type of pathological pattern is more consistent with a general intolerance to chemicals of cosmetics applied under both arms than to the equivalent of several coincidental monoclonal initiation events (Ponten et al, 1990).

CONSTITUTENTS OF UNDERARM COSMETICS AND POSSIBLE ROLE IN CARCINOGENESIS:

i) Chemical constituents
If underarm cosmetics play any role in breast cancer, then the challenge will be to identify specific chemical culprits. All ingredients will have been tested by current safety guidelines (Laden and Felger, 1988), but the effects of long-term use over an entire lifetime by the whole global population of women (and increasingly by men) can only be investigated retrospectively. The main active ingredients of underarm cosmetics are:

1) Antiperspirant agents
2) Deodorant agents
3) Preservatives

Minor components include fragrance and colourings.

The antiperspirant component acts to block the sweat ducts, so preventing escape of sweat onto the body surface (Laden and Felger, 1988). The main active agents are metal salts, principally aluminium chlorhydrate and the aluminium zirconium chlorhydrate glycine complexes. Their mechanism of action is not fully established but is thought to involve the formation of a physical plug at the top of the sweat duct (Laden and Felger, 1988) which is composed of a combination of precipitated salts and damaged cells. This plug then prevents the secretion of sweat. The deodorant components are antimicrobial agents which act to kill bacteria on the body surface (Laden and Felger, 1988). Since it is the bacterial action on sweat which generates the undesirable odour from sweating, the deodorant is designed to eliminate the smell (Laden and Felger, 1988). Finally, since consumers expect long shelf life from these cosmetics, preservatives are added, especially to stick, roll-on and cream formulations, to prevent microbial and fungal growth in the containers during long-term storage.

ii) Possible mechanism of action
Any carcinogenic action by the constituent chemicals of underarm cosmetics would require a minimal combination of initiating and promoting agents. The simple suggestion would be that underarm cosmetics act by delivering to the breast area a regular supply of components capable of both initiation and promotion of breast cancer, and which are simply absorbed through the skin into the local breast area. However, it is also possible that there would be interactions with other physiological, dietary and environmental agents, possibly with the underarm cosmetics acting as the final insult.

iii) Initiation
Initiation would require agents capable of binding to DNA and causing DNA damage and mutation. The active antiperspirant agents include aluminium and zirconium salts (Laden and Felger, 1988), and there is evidence that aluminium can bind to DNA (Karlik et al, 1980; Ahmad et al, 1996). There are also reports of accumulation of aluminium in human breast material (Mulay et al, 1971) and in mammary tumours of rats (Ogoshi et al, 1994). Another alternative source of initiating agents could arise from chemicals secreted in the sweat itself (Stowe and Plaa, 1968) if they could accumulate at the point where sweat ducts are blocked by the antiperspirant agents. This has been suggested in an anonymous e-mail widely circulated across the world over the past two years, but although it has stimulated discussion, most conclusions have been negative due to the fact that the source of the e-mail has, unfortunately, never been identified.

iv) Promotion
Promotion of breast cancer would require agents capable of enabling growth of breast epithelial cells. In view of the well established role of oestrogen in growth of breast cancer cells both in vivo and in vitro (Miller, 1996), the most likely promotional candidate would have oestrogenic properties. It is possible that promotion could result from exposure to physiological oestrogens, since established risk factors for breast cancer include lifetime exposure to oestrogen (Lipworth, 1995). However, the human population is now also exposed to a variety of pollutant chemicals which can mimic the action of oestrogen and which, due to their lipophilic nature, are stored in human breast fat (Darbre, 1998). These include a variety of pesticides such as DDT (Smith, 1999) and polychlorinated biphenyls (PCBs) (Dobson and van Esch, 1993). Some PCB congeners can mimic the action of oestrogen (Nesaretnam et al, 1996; Nesaretnam and Darbre, 1997) and can enhance breast cancer in an animal model (Nesaretnam et al, 1998). Such chemicals accumulate in body fat over a lifetime but can be released during times of fasting or dieting and can be released into breast milk during lactation (Darbre, 1998).

However, it is also possible that underarm cosmetics contain themselves oestrogenic chemicals capable of being absorbed through skin and acting locally. There is now evidence that parabens (alkyl esters of p-hydroxybenzoic acid) can mimic the action of oestrogen (Routledge et al, 1998; Byford et al, 2001) and parabens are used as preservatives in over 13,000 cosmetic formulations including underarm cosmetics in concentrations of up to 1% (Elder, 1984). Parabens have been shown to bind to the oestrogen receptors of rat uterus (Routledge et al, 1998; Blair et al, 2000) and of MCF7 human breast cancer cells (Byford et al, 2001). They show oestrogenic activity in yeast cell assays (Routledge et al, 1998) and in regulating gene expression and growth of oestrogen-responsive human breast cancer cells (Byford et al, 2001). Their oestrogenic activity has also been detected in vivo in fish (Pedersen et al, 2000) and in the immature rodent uterine weight assay (Routledge et al, 1998; Hossani et al, 2000). Interestingly, oestrogenic activity in the latter assays was found only when administration was subcutaneous and not oral (Routledge et al, 1998; Hossani et al, 2000) suggesting that topical application of parabens in cosmetics could provide an oestrogenic stimulus. The issue of penetration of parabens through the human underarm skin needs now to be studied, but parabens have been shown capable of penetrating skin in animal studies (DalPozzo and Pastori, 1996) and our own preliminary unpublished work has shown that parabens can be detected in human breast fat by either simply thin-layer chromatography or by high-pressure liquid chromatography followed by mass spectrometry.

INHERITED SUSCEPTIBILITY TO BREAST CANCER
About 5-10% of female breast cancer is inherited as a genetic susceptibility with an increased risk of developing breast cancer (Easton et al, 1993). In recent years, two of the susceptibility genes BRCA1 (Miki et al, 1994) and BRCA2 (Wooster et al, 1995) have been cloned and sequenced. BRCA1 has been implicated to play a primary role in maintenance of DNA repair; BRCA2 is a DNA binding protein but its function has yet to be fully understood (Hilakivi-Clarke, 2000). Both these genes act as tumour suppressor genes in that loss of their function is associated with increased risk of breast cancer (Hilakivi-Clarke, 2000), presumably through an inability to maintain normal DNA repair systems. Although much has therefore now been learnt about the inheritance of susceptibility, the question remains as to susceptibility to what? It is possible that the explanation lies in susceptibility to use of underarm cosmetics. Loss of BRCA1 or BRCA2 will have compromised the ability of the breast cells to repair DNA damage and it is possible that this could lead to increased sensitivity to any DNA damaging chemicals of underarm cosmetics.

BENIGN BREAST DISEASE
Breast cancer represents only around 5% of clinical abnormalities of the breast (Haagensen, 1971). Other major problems of the breast include fibroadenomas and breast cysts (Haagensen, 1971). Interestingly, the upper outer quadrant is not only the most common site of the tumour in cancer but also of the abnormalities in many benign breast conditions including fibroadenoma and breast cysts (Rimsten, 1976) and phyllodes tumour (Stebbing and Nash, 1995). It is interesting to note that breast cysts are caused by blocked breast ducts (Haagensen, 1971). The reason for their high incidence rate remains unknown but they are of considerable concern since their presence can be an indicator of cancer to come (Dixon et al, 1999). Since antiperspirants act by blocking sweat ducts (Laden and Felger, 1988) and breast cysts result from blocked breast ducts (Haagesen, 1971), it is possible that breast cysts might also arise from underarm cosmetic use. Application of underarm cosmetics is rarely so precise that ingredients fall only on the underarm site and not partly also on adjacent breast areas. There is no reason to believe that antiperspirants should not do the same to adjacent breast ducts as to the sweat ducts, if sufficient cosmetic is applied and if the cosmetic is applied carelessly to areas beyond the immediate underarm site, particularly since breast is essentially a modified sweat gland (Anderson, 1991).

CONCLUSIONS:
Whether or not this hypothesis proves to be correct, it deserves serious testing by both scientists and clinicians. The nature of the chemicals in these cosmetics and the lack of any advice about safe quantity or frequency of application should be of major concern alone. However, their use by ever younger children before puberty is also an extremely worrying trend. There is evidence that the origin of breast cancer can be many years before the symptoms become visible and that there may even be a specific time period before puberty when the breast is particularly susceptible to carcinogenesis (Russo and Russo, 1987). Studies of Japanese survivors of the atomic bomb in 1945 have shown that the highest risk of radiation-induced breast cancer was in girls who were prepubertal at the time of exposure (Tokunaga et al, 1979).

If there proves to be any truth in the hypothesis, then underarm cosmetics can be given up without addictive or life-threatening consequences. Furthermore, since these cosmetics are applied voluntarily, then women would have, for the first time, an opportunity to choose to reduce their own personal risk of breast cancer.
European Journal of Cancer Prevention, 10:389-393, 2001.
Division of Cell and Molecular Biology, School of Animal and Microbial Sciences, The University of Reading, PO Box 228, Whiteknights, Reading, RG6 6AJ, UK. E-mail: p.d.darbre@reading.ac.uk

PHILLIP DAY'S COMMENT: Ms Darbre is to be applauded for producing her hypothesis which I believe deserves to be rapidly distributed around the world. So many chemicals used in personal care products have not been adequately tested for their long-term effects. This is also true for many foods we consume, such as white bread, sugar, coffee and others, that are taken for granted. My own investigations into the antiperspirant/deodorant conundrum have led me actively to encourage the public to seek safe alternatives to the chemical products being sold without a care in every supermarket across the world. Stay natural with safer, gentler products. And if you have chronic body odour, investigate a colon cleanse, a change of diet to raw fruits and vegetables, and increase water intake to 2-3 litres a day.

REFERENCES
Ahmad R, Naoui M, Neault JF, Diamantoglou S, Tajmir-Riahi HA (1996). An FITR specroscopic study of calf thymus DNA complexation with Al (III) and Ga (III) cations. J Biomol Struct Dyn 13: 795-802.

Anderson TJ (1991). Genesis and source of breast cancer. British Medical Bulletin 47: 305-318.

Azzena A, Zen T, Ferrara A, Brunetti V, Vasile C, Marchetti M (1994). Risk factors for breast cancer. Eur J Gynaecol Oncol 15: 386-392.

Blair RM, Fang H, Branham WS et al. (2000) The estrogen receptor relative binding affinities of 188 natural and xenochemicals: structural diversity of ligands. Toxicol Sci 54: 138-153.

Busk T, Clemmesen J (1947). The frequencies of left- and right-sided breast cancer. Brit J Cancer 1: 345.

Byford JR, Shaw LE, Drew MGB, Pope GS, Sauer MJ, Darbre PD. Oestrogenic activity of parabens in MCF7 human breast cancer cells. J Steroid Biochem Molec Biol (2001, in press).

Chen Y, Thompson W, Semenciw R, Mao Y (1999). Epidemiology of contralateral breast cancer. Cancer Epidemiol Biomarkers Prev 8: 855-861.

DalPozzo A, Pastori N (1996). Percutaneous absorption of parabens from cosmetic formulations. Int J Cosmetic Sci 18: 57-66.

Darbre PD (1998). Environmental contaminants in milk: the problem of organochlorine xenobiotics. Biochem Soc Trans 26: 106-112.

Dixon JM, McDonald C, Elton RA, Miller WR (1999). Risk of breast cancer in women with palpable breast cysts: a prospective study. The Lancet 353: 1742-1745.

Dobson S, van Esch GJ (1993). Polychlorinated Biphenyls and Terphenyls, ed.2. Environmental Health Criteria 140. World Health Organisation, Geneva, Switzerland.

Donegan WL, Spratt JS (1967). Cancer of the second breast. In: JS Spratt, Donegan WL Cancer of the Breast,Philadelphia: WB Saunders, p 179.

Easton D, Ford D, Peto J (1993) Inherited susceptibility to breast cancer. Cancer Surveys 18: 95-113.

Elder RL (1984) Final report on the safety assessment of methylparaben, ethylparaben, propylparaben and butylparaben. Journal American College Toxicology 3: 147-209.

Fenig J, Arlen M, Livingston SF, Levowitz BS (1975). The potential for carcinoma existing synchronously on a microscopic level within the second breast. Surg Gynecol Obstet 141: 394-396.

Garfinkel L, Craig L, Seidman H (1959). An appraisal of left and right breast cancer. J Natl Cancer Inst 23: 617.

Haagensen CD. (1971). Diseases of the Breast, 2nd ed. Philadelphia: WB Saunders.

Harnett WL (1948). A statistical report on 2529 cases of cancer of the breast. Brit J Cancer 2: 212.

Hilakivi-Clarke,L (2000). Estrogens, BRCA1, and breast cancer. Cancer Research 69: 4993-5001.

Hossaini A, Larsen JJ, Larsen JC (2000). Lack of oestrogenic effects of food preservatives (parabens) in uterotropic assays. Food and Chemical Toxicology 38: 319-323.

Hussain MA, Sawkat A, Tyagi SP, Reza H (1994). Incidence of cancer breast at Aligarh. J Indian Med Assoc 92: 296-297.

Jaiyesimi IA, Buzdar AU, Sahin AA, Ross MA (1992). Carcinoma of the male breast. Ann Intern Med 117: 771-777.

Jepson AS, Fentiman IS (1998). Male breast cancer IJCP 52: 571-576.

Karlik SJ, Eichhorn GL, Lewis PN, Crapper DR (1980). Interaction of aluminium species with deoxyribonucleic acid. Biochem 19: 5991-5998.

Laden K, Felger CB. (1988). Antiperspirants and deodorants. Cosmetic Science and technology Series volume 7. New York: Marcel Dekker.

Lane-Claypon JE (1926). A further report on cancer of the breast with special reference to its associated antecedent conditions. Reports on Public Health and Medical Subjects no 32,London: Ministry of Health.

Lipworth L (1995). Epidemiology of breast cancer. Eur J Cancer Prevention 4: 7-30.

Miki Y, Swensen J, Shattuck-Eidens D et al .(1994). A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266: 66-71.

Miller WR (1996). Estrogen and Breast Cancer, Chapman and Hall.

Mulay IL, Roy R, Knox BE, Suhr NH, Delaney WE (1971). Trace-metal analysis of cancerous and non-cancerous human tissues. J Natl Cancer Inst 47: 1-13.

Nesaretnam K, Corcoran D, Dils RR, Darbre P (1996). 3,4,3',4'-Tetrachlorobiphenyl acts as an oestrogen in vitro and in vivo. Mol Endocrinol 10: 923-936.

Nesaretnam K, Darbre P (1997). 3,5,3',5'-Tetrachlorobiphenyl is a weak oestrogen agonist in vitro and in vivo. J Steroid Biochem Molec Biol 62: 409-418.

Nesaretnam K, Hales E, Mohammed S, Krausz T, Darbre P (1998). 3,4,3',4'-Tetrachlorobiphenyl (TCB) can enhance DMBA-induced mammary carcinogenesis in rat. Eur J Cancer 34: 389-393.

Nohrman BA (1949). Cancer of the breast. Acta Radiol, Suppl 77.

Ogoshi K, Yanagi S, Moriyama T, Arachi H (1994). Accumulation of aluminium in cancers of the liver, stomach, duodenum and mammary glands of rats. J Trace Elem Electrolytes Health Dis 8: 27-31.

Patterson SK, Helvie MA, Joynt LK, Roubidoux MA, Strawderman M (1998). Mammographic appearance of breast cancer in African-American women. Acad Radiol 5: 2-8.

Pedersen KL, Pedersen SN, Christiansen LB, Korsgaard B, Bjerregaard P (2000) The preservatives ethyl-, propyl-, and butylparaben are oestrogenic in an in vivo fish assay. Pharmacol Toxicol 86: 110-113.

Ponten J, Holmberg L, Trichopoulos D, Kallioniemi OP, Kvale G, Wallgren A, Taylor-Papadimitriou J (1990). Biology and natural history of breast cancer. Int J Cancer Suppl 5, 5-21.

Raju GC, Naraynsingh V (1989). Breast cancer in West Indian women in Trinidad. Trop Geogr Med 41: 257-260.

Rimsten A (1976). Symptoms and signs in benign and malignant tumours of the breast. Upsala J Med Sci 81: 54-60.

Rizk SN, Assimacopoulos CA, Ryan JJ (1994). Male breast cancer: three case reports and review of the literature. SDJ Med 47: 343-346.

Routledge EJ, Parker J, Odum J, Ashby J, Sumpter JP (1998) Some alkyl hydroxy benzoate preservatives (parabens) are estrogenic. Toxicol and Appl Pharmacol 153: 12-19.

Russo J, Russo IH (1987).Biology of disease. Biological and molecular bases of mammary carcinogenesis. Laboratory Investigation 57: 112-137.

Smith D (1999) Worldwide trends in DDT levels in human breast milk. Int J Epidemiol 28: 179-188.

Smithers DW, Rigby-Jones P, Galton DAG, Payne PM (1952). Cancer of the breast. Brit J Radiol, Suppl 4.

Stebbing JF, Nash AG (1995). Diagnosis and management of phyllodes tumour of the breast: experience of 33 cases at a specialist centre. Ann R Coll Surg Engl 77: 181-184.

Stowe CM, Plaa GL (1968). Extrarenal excretion of drugs and chemicals. Annu Rev Pharmacol 8: 337-356.

Tokunaga M, Norman JE, Asano M, Tokuoka S, Ezaki H, Nishimori I, Tsuji Y (1979). Malignant breast tumors among atomic bomb survivors, Hiroshima and Nagasaki, 1950-74. J Natl Cancer Institute 62: 1347-1359.

Truscott BM (1947). Carcinoma of the breast. Brit J Cancer 1: 129.

Ursin G, Bernstein L, Pike MC (1994). Breast Cancer. Cancer Surveys 19/20: 241-264.

Wooster R, Bignell G, Lancaster J et al .(1995). Identification of the breast cancer susceptibility bene BRCA2. Nature 378: 789-792.