Breast Cancer as an Infectious Disease

“…there must be environmental factors that influence breast cancer induction and progression.”

For over 100 years, scientists (including Brian MacMahon in the USA and Dimitrios Trichopoulos in Greece) have sought to understand the causes of breast cancer, with epidemiological studies demonstrating a diverse and sometimes bizarre range of risk factors [1]. These risk factors have included age, early onset of menarche and late-age menopause, being left-handed, becoming overweight following menopause and having a mother with breast cancer. It was also demonstrated that early-age, full-term pregnancy offered some protection. Interestingly, being overweight prior to menopause reduces the risk of breast cancer [2]. Large birthweight has also been shown to increase the risk of breast cancer [3]. More recently, genome-wide association studies have demonstrated that certain patterns of inheritance of specific DNA loci are associated with an increased risk of breast cancer [4]. This suggests that the accumulation of changes in a number of genes that, individually, only modestly increase the risk is associated with breast cancerigenesis. However, by far the most significant risk factor is female gender. Differing levels of estrogens may well explain some of these risk factors (e.g., early-age menarche, late-age menopause and postmenopausal obesity are associated with increased exposure to estrogens, while fetal exposure to high levels of estrogen and growth hormones are associated with large birthweights) [5].

“While there is no direct evidence for the means of transmission of human papillomavirus in breast tissue, there are a number of intriguing clues.”

However, with the exception of female gender, none of these factors explains the most striking epidemiological risk. This is the four- to sixfold higher risk of breast cancer in Western as compared with most Asian women; a risk which dramatically increases among Asian women and their descendants following migration [6]. Accordingly, there must be environmental factors that influence breast cancer induction and progression. It is reasonable to speculate that migration may lead to changes in food consumption (which in turn alters circulating estrogen levels), the acquisition of oncogenic viruses or a combination of both. Our interest has been the association of oncogenic viruses and breast cancer.

The search for viruses that may cause breast cancer also has a decades' long history dating back to the work of John Bittner at the Jackson Laboratory (ME, USA), in the 1930s. Bittner discovered what he termed a milk transmission factor, which was conveyed in murine mother's milk and led to mammary tumors late in the recipient mouse's life [7]. This was later shown to be the mouse mammary tumor virus. Unfortunately, despite intense investigations during the Nixon ‘war on cancer’ years of the 1970s and more recent work initiated by Beatriz Pogo (NY, USA), definitive evidence for a causal role for mouse mammary tumor virus in human breast cancer has remained elusive.

More recently, a role for viruses in human breast cancer has been met with skepticism, not unlike that leveled at the pioneering work of Harald zur Hausen when he suggested a causal role for human papillomavirus (HPV) in cervical cancer [8]. Despite this disbelief, zur Hausen's work was ultimately very well recognized, and he was awarded the Nobel Prize in 2008.

There are some 200 different HPV serotypes but only a small number of these, the so-called ‘high-risk’ types, are associated with cancer. After many years of study, the mechanism by which the high-risk HPV types transform cells and cause cancer is now well understood. While integration of the HPV genome and a number of HPV gene products are important, the inactivation of the cellular tumor-suppressing proteins p53 and p110^RB by E6 and E7, respectively, are of primary importance.

A role for HPV in other cancers has been speculated, and it is now well accepted that HPV (in conjunction with Epstein-Barr virus) has a causal role in head and neck cancers [9,10]. In 1992, Lonardo et al. identified high-risk HPV type 16 virus in nearly 30% of human breast tumors [11]. However, this work was met with disbelief. How could HPVs even reach the breast, let alone cause breast cancer?

Fortunately, there has been considerable interest in this area, and high-risk HPVs in breast tumors have been reported in breast cancer specimens in some 15 different countries [12]. It must be noted, however, that ten of the studies reviewed by Mendizabal et al. failed to detect HPV in breast cancer [12]. In fact, detection of HPV in breast cancer has proven to be extremely difficult. In our laboratory, we have used standard PCR on DNA extracts from fresh frozen breast tumors to identify HPV sequences. This required repeated PCR analyses using SYBR Green for greater sensitivity; however, we were finally able to confirm the presence of HPV type 18 (type confirmed by sequencing) in 48% of Australian breast cancer specimens tested [13]. The difficulty in detecting HPV in breast cancer appears to be due to the extremely low viral load as indicated by the report from Khan et al., in which there was a 4000-fold lower load in breast cancer than in cervical cancer in Japanese women [14]. The low viral load may account for the lack of detection of HPV in the studies reported by Mendizabal et al. [12].

“The difficulty in detecting human papillomavirus in breast cancer appears to be due to the extremely low viral load.”

The issue with the use of PCR as a technique for detecting viral sequences is the possibility of false-positive detection through contamination [15]. The use of in situ PCR maintains the sensitivity of PCR while eliminating the possibility of contamination. In addition, in situ PCR is able to localize the target sequence within the tissue as well as within the cells. Our laboratory has recently reported the detection of high risk HPV (primarily HPV-18) in the nuclei of cancer cells within breast cancer specimens [16] from a separate cohort of Australian breast cancer specimens from those reported in 2005 [13]. However, in situ PCR remains subject to false-positive and -negative outcomes, and the scientific community has been justifiably skeptical regarding studies of breast cancer based on these PCR techniques. Thus, corroborative data that are not dependent on amplification of DNA are required.

The de Villiers' group has succeeded in confirming and supplementing their findings based on standard PCR with in situ hybridization, and were able to show that a range of HPV types were present in breast tumors of US women [17]. In our laboratory, we have supplemented PCR and in situ PCR data with the identification of the viral oncoprotein HPV E6 by immunohistochemical techniques, and traditional light microscopy in order to identify HPV-associated koilocytes. Koilocytes are large epithelial cells with a vacuolated halo surrounding a dense nucleus. They are specific and indicative of HPV infection being caused by the action of HPV E5 and E6 oncoproteins [18]. Importantly, their occurrence in the cervix is commonly a sign of precancerous change [19].

A causal role for HPV infection in cancer induction requires evidence that correlates the presence of HPV proteins with enhanced cellular proliferation and markers such as low (or undetectable) p53 and p110^RB protein levels and mRNA expression of p21 and p16^INK4A. Hennig et al. have demonstrated significantly reduced levels of p53 in HPV-positive as compared with HPV-negative invasive breast tumors [20]. While we have not been able to consistently demonstrate low p53 protein levels with the presence of HPV E6 in invasive breast tumors, it has been demonstrated that low p53 protein and the presence of HPV do correlate in noninvasive ductal carcinoma in in situ breast tumors (Lawson JS et al., Unpublished Data).

The presence of high-risk HPV in archival Australian breast tumors has been demonstrated by the use of a combination of both standard and in situ PCR. The presence of HPV in these breast tumors was confirmed by the detection of HPV E6 protein and the associated presence of koilocytes. In addition, the presence of HPV E6 and koilocytes was demonstrated in the normal breast skin and breast lobules of some women who had cosmetic breast surgery [19], indicating HPV infection prior to cancerigenesis.

“…low p53 protein and the presence of human papillomavirus do correlate in noninvasive ductal carcinoma in in situ breast tumors.”

The scientific challenge has become a question of causation and possible modes of transmission since it is realized that HPV may not be merely an innocuous passenger.

While there is no direct evidence for the means of transmission of HPV in breast tissue, there are a number of intriguing clues. We have reported that HPV-positive breast tumors are more common in younger than older women [21]. This observation parallels the high incidence of cervical HPV infections in young women — over 50% of women under the age of 30 years as compared with 5% over the age of 50 years [22]. The same HPV types have been identified in women who have suffered both cervical and breast cancer, suggesting a common infection [23,24]. HPV sequences have been identified in white blood cells and the placentas of unaffected women and, accordingly, transmission of HPV by circulating blood is a possibility [25]. Recently, HPV-positive cancers of the mouth, head and neck have been shown to be more common following oral sexual activities [26]. While there will be continued dis-belief regarding this matter, these observations increasingly suggest that some breast cancers are sexually transmitted.

When considered in the context of previously published evidence related to HPV and breast cancer, these recent findings strongly suggest a causal role for HPVs in some breast cancers. However, we must wait for the development of further evidence before this relationship can be stated definitively.

The immediate importance of this work is that it brings with it the possibility, for the first time, that primary preventative measures for some breast cancers are likely to be immediately available. This is because the high-risk HPV types that we and many others have identified in breast tumors (principally HPV types 16 and 18) are the same types for which the new HPV vaccines are most effective [27]. These vaccines are already available and being used on a worldwide basis.