Increased reactive oxygen and nitrogen species (RONS) leading to increased risk of breast cancer

• Jessica Helm
• Evgeniia Kazymova

• Tanja Burgdorf
• Shihori Tanabe

Revision dates for related pages

While the key events described here are likely relevant to all tissues after exposure to IR, it is particularly relevant to the female mammary gland. While ionizing radiation causes many kinds of cancers including leukemia, lung, bladder, and thyroid cancers (BEIR 2006; Preston, Ron et al. 2007), breast cancers are among the cancers most increased by exposure to ionizing radiation (Preston, Ron et al. 2007).

The lengthy and hormone-dependent developmental trajectory of the mammary gland is likely to be a major factor in its susceptibility to breast cancer. Numerous epidemiological and laboratory studies support the requirement for ovarian hormones in the risk of breast cancer from ionizing radiation (Grant, Cologne et al. 2018). Although at first examination breast cancer from ionizing radiation and hormones involve very different processes, in fact the hormone-dependent and ionizing radiation pathways of carcinogenesis intersect at multiple points that are part of breast development leaving the hormone-exposed breast more vulnerable to radiation. Two studies in humans and rats also suggest that IR can increase long term concentrations of circulating estrogen which would further amplify any additive effects, although additional evidence is needed (Suman, Johnson et al. 2012; Grant, Cologne et al. 2018).

One major mechanism promoting breast cancer from ionizing radiation is the proliferation of breast stem cells. Stem cells are considered to be important to initiation because of their long life and capacity to pass on mutations to many progeny. Breast tissue is responsive to estrogen and progesterone, reproductive hormones that rise at puberty and stimulate cellular proliferation with each reproductive cycle and in pregnancy. These hormonal proliferative cycles increase the risk of cancer in breast tissue (Brisken, Hess et al. 2015). IR increases the long term proliferation of stem cells in pubertal but not adult mammary gland (Nguyen, Oketch-Rabah et al. 2011; Datta, Hyduke et al. 2012; Snijders, Marchetti et al. 2012; Suman, Johnson et al. 2012; Tang, Fernandez-Garcia et al. 2014). Replication of stem cells in the IR-exposed breast is therefore particularly elevated during puberty, likely contributing to the increased susceptibility to breast cancer from IR at this age.

Another vulnerability of the breast to IR is a byproduct of proliferation: mutations. Replication itself increases the likelihood of mutations, which add to mutations arising from IR and increase the likelihood of oncogenic transformation (Atashgaran, Wrin et al. 2016). Furthermore, the high replication rate of mammary gland epithelial cells during puberty and pregnancy increases reliance on homologous recombination pathways (Kass, Lim et al. 2016). Disruption of these HR processes by IR-induced mutation or increased demand for repair can increase mutation rates and increase tumorigenesis (Mahdi, Huo et al. 2018). This disruption is particularly relevant for mammary stem cells which are highly replicating and dependent on HR but shift to NHEJ to respond to DNA damage from IR (Chang, Zhang et al. 2015). The consequence of mutations in stem cells is significant, since these cells can clonally expand to generate many mutated progeny. However, errors in stem cell division may not be the sole or primary factor driving cancer from radiation, since excess cancer risk for solid cancers at different sites from the atomic bomb are not clearly related to the number of stem cell divisions at that site (Tomasetti, Li et al. 2017).

The elevated estrogen associated with development and the estrous cycle may also have direct effects that further complement the carcinogenic effects of IR. Estrogen directly increases oxidative stress in virgin (but not parous) mice (Yuan, Dietrich et al. 2016), interferes with DNA repair (Pedram, Razandi et al. 2009; Li, Chen et al. 2014) increases mutations (Mailander, Meza et al. 2006), and increases TGF-b (Jerry, Dunphy et al. 2010). Each of these effects would increase the impact of the same events arising from IR alone.

Inflammation from the estrous cycle may also contribute to tumorigenesis following IR. Cytokines and macrophages play an integral role in mammary gland development and ductal elaboration, with alternating inflammatory, immune surveillance, and phagocytic activity occurring over each estrous cycle (Hodson, Chua et al. 2013; Atashgaran, Wrin et al. 2016; Brady, Chuntova et al. 2016). This inflammation could potentially increase IR-induced DNA damage and mutations and promote tumorigenic and invasive characteristics.

The enhancement of IR induced tumorigenesis by the estrous cycle may be replicated or further enhanced by exogenous endocrine disrupting chemicals. Indeed, evidence suggests that BPA (and presumably other estrogenic chemicals) exposure in utero can increase the mammary gland’s response to progesterone during puberty (Brisken, Hess et al. 2015). This enhancement would presumably also increase the risk of breast cancer from ionizing radiation, since that risk increases with estrogen exposure and the number of menstrual cycles.

Uncertainty arising from extrapolating from rodent and human in vitro studies to human biology

Uncertainty in this pathway arises from inconsistencies in carcinogenesis between rodent and mouse species and strains and from incomplete information about the same mechanisms operating in humans. This raises questions about whether all evidence should be weighted equally.

Almost half of the data included here is from in vitro experiments on human primary or cultured cells, which should have a high degree of relevance for this pathway in humans.  However, most of the human cells are not from mammary gland, and most of the mammary gland derived cells are cancer or immortalized cells that will not respond in exactly the same way as primary cells. Even this human data should therefore be interpreted with some caution.

Most of the remaining data in this AOP is from mice, with a relatively small number of rat studies. As a breast cancer model, mice share important characteristics with humans (Medina 2007; Imaoka, Nishimura et al. 2009). Mice and humans share similar epithelial cell types (Lim, Wu et al. 2010) and a similar developmental regime with the bulk of epithelial development occurring postnatally and accelerating during puberty, with differentiation during pregnancy (Medina 2007). Tumors in humans originate in the terminal ductal lobular unit, a structure that includes the lobule with secretory alveoli and the start of the collecting duct. The developmental terminal end bud structure is thought to be particularly vulnerable to carcinogens because of the presence of stem cells and proliferation, although it is not the only possible site of initiation. Similarly, tumors in mice originate in predominantly in alveoli as well as terminal end buds and small ducts (Medina 2007). Humans are more susceptible to carcinogens around puberty, and pregnancy is protective. Evidence on the role of development and reproduction in mammary carcinogenesis in mice is limited compared with rats but is consistent with sensitivity to radiation around puberty (Imaoka, Nishimura et al. 2009), and parity is protective for chemical carcinogens (Medina 2007). In addition, proliferation contributes to carcinogenesis in both mice and humans (Medina 2007).

However, mice differ from humans in some notable ways (Medina 2007). Mammary tumors are not common in mice, so susceptible strains or tumor-promoting viruses are used to increase spontaneous incidence and response to carcinogenic stimuli. This difference may be partially attributable to hormone responsiveness of tumors. Although tumors in mice depend on hormones for development, breast cancers in rats and humans are frequently hormone receptor positive, while mammary tumors in mice are not (Nandi, Guzman et al. 1995; Medina 2007; Imaoka, Nishimura et al. 2009).

Atashgaran, V., J. Wrin, et al. (2016). "Dissecting the Biology of Menstrual Cycle-Associated Breast Cancer Risk." Front Oncol 6: 267.

Brady, N. J., P. Chuntova, et al. (2016). "Macrophages: Regulators of the Inflammatory Microenvironment during Mammary Gland Development and Breast Cancer." Mediators Inflamm 2016: 4549676.

Brisken, C., K. Hess, et al. (2015). "Progesterone and Overlooked Endocrine Pathways in Breast Cancer Pathogenesis." Endocrinology 156(10): 3442-3450.

Chang, C. H., M. Zhang, et al. (2015). "Mammary Stem Cells and Tumor-Initiating Cells Are More Resistant to Apoptosis and Exhibit Increased DNA Repair Activity in Response to DNA Damage." Stem Cell Reports 5(3): 378-391.

Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation (2006). Health risks from exposure to low levels of ionizing radiation : BEIR VII, Phase 2, National Research Council of the National Academies.

Datta, K., D. R. Hyduke, et al. (2012). "Exposure to ionizing radiation induced persistent gene expression changes in mouse mammary gland." Radiat Oncol 7: 205.

Grant, E. J., J. B. Cologne, et al. (2018). "Bioavailable serum estradiol may alter radiation risk of postmenopausal breast cancer: a nested case-control study." International journal of radiation biology 94(2): 97-105.

Hodson, L. J., A. C. Chua, et al. (2013). "Macrophage phenotype in the mammary gland fluctuates over the course of the estrous cycle and is regulated by ovarian steroid hormones." Biol Reprod 89(3): 65.

Imaoka, T., M. Nishimura, et al. (2009). "Radiation-induced mammary carcinogenesis in rodent models: what's different from chemical carcinogenesis?" J Radiat Res 50(4): 281-293.

Jerry, D. J., K. A. Dunphy, et al. (2010). "Estrogens, regulation of p53 and breast cancer risk: a balancing act." Cellular and molecular life sciences : CMLS 67(7): 1017-1023.

Kass, E. M., P. X. Lim, et al. (2016). "Robust homology-directed repair within mouse mammary tissue is not specifically affected by Brca2 mutation." Nat Commun 7: 13241.

Li, Z., K. Chen, et al. (2014). "Cyclin D1 integrates estrogen-mediated DNA damage repair signaling." Cancer Res 74(14): 3959-3970.

Lim, E., D. Wu, et al. (2010). "Transcriptome analyses of mouse and human mammary cell subpopulations reveal multiple conserved genes and pathways." Breast cancer research : BCR 12(2): R21.

Mahdi, A. H., Y. Huo, et al. (2018). "Evidence of Intertissue Differences in the DNA Damage Response and the Pro-oncogenic Role of NF-kappaB in Mice with Disengaged BRCA1-PALB2 Interaction." Cancer Res 78(14): 3969-3981.

Mailander, P. C., J. L. Meza, et al. (2006). "Induction of A.T to G.C mutations by erroneous repair of depurinated DNA following estrogen treatment of the mammary gland of ACI rats." The Journal of steroid biochemistry and molecular biology 101(4-5): 204-215.

Medina, D. (2007). "Chemical carcinogenesis of rat and mouse mammary glands." Breast Dis 28: 63-68.

Nandi, S., R. C. Guzman, et al. (1995). "Hormones and mammary carcinogenesis in mice, rats, and humans: a unifying hypothesis." Proceedings of the National Academy of Sciences of the United States of America 92(9): 3650-3657.

Nguyen, D. H., H. A. Oketch-Rabah, et al. (2011). "Radiation acts on the microenvironment to affect breast carcinogenesis by distinct mechanisms that decrease cancer latency and affect tumor type." Cancer Cell 19(5): 640-651.

Pedram, A., M. Razandi, et al. (2009). "Estrogen inhibits ATR signaling to cell cycle checkpoints and DNA repair." Mol Biol Cell 20(14): 3374-3389.

Preston, D. L., E. Ron, et al. (2007). "Solid cancer incidence in atomic bomb survivors: 1958-1998." Radiation research 168(1): 1-64.

Snijders, A. M., F. Marchetti, et al. (2012). "Genetic differences in transcript responses to low-dose ionizing radiation identify tissue functions associated with breast cancer susceptibility." PLoS One 7(10): e45394.

Suman, S., M. D. Johnson, et al. (2012). "Exposure to ionizing radiation causes long-term increase in serum estradiol and activation of PI3K-Akt signaling pathway in mouse mammary gland." International journal of radiation oncology, biology, physics 84(2): 500-507.

Tang, J., I. Fernandez-Garcia, et al. (2014). "Irradiation of juvenile, but not adult, mammary gland increases stem cell self-renewal and estrogen receptor negative tumors." Stem Cells 32(3): 649-661.

Tomasetti, C., L. Li, et al. (2017). "Stem cell divisions, somatic mutations, cancer etiology, and cancer prevention." Science 355(6331): 1330-1334.

Yuan, L., A. K. Dietrich, et al. (2016). "17beta-Estradiol alters oxidative damage and oxidative stress response protein expression in the mouse mammary gland." Mol Cell Endocrinol 426: 11-21.

IR appears to be a “complete” carcinogen in the mammary gland in that the toxin acts as an initiator through the formation of oxidative stress and pro-mutagenic DNA damage and (the MIEs) and as a promoter through increasing inflammation and proliferation, similar to many chemical carcinogens (Russo and Russo 1996). We have high confidence in the evidence linking stressor (IR) with adverse outcome (breast cancer).  The weight of evidence for the first pathway from RONS and DNA damage to Mutation and Proliferation is High while the weight of evidence for the second pathway from RONS to Inflammation to Proliferation and Breast Cancer is Moderate. These evaluations are based on the supporting evidence for all KEs and the considerations in Annex 1, and based on the need for additional evidence in the essentiality of Inflammation for the genesis of breast cancer.

This AOP could not address the large number of related topics that interact with the key events described here. These topics include events following IR that may interact with these key events such as immune surveillance (which may change with the inflammatory environment after IR (Schreiber, Old et al. 2011; Barcellos-Hoff 2013; Lumniczky and Safrany 2015); IR effect on survival/apoptosis and interactions of apoptosis with inflammation, mutation, compensatory proliferation, and selection process; changes to DNA repair; and the role of epigenetics in carcinogenesis from IR (Daino, Nishimura et al. 2018). This AOP also does not address other influences on these key events beyond reproductive hormones and typical breast development. Subsequent contributions to this AOP should elaborate on these points.

MIE1: Increase in RONS

Essentiality is High. The most significant support comes from the relatively large number of studies using antioxidants or other interventions to reduce RONS, which show a reduction in DNA damage and mutations. Additional support comes from experiments increasing external oxidants like H2O2, which show that RONS are independently capable of causing DNA damage and mutations. Uncertainties arise from the smaller effects of RONS on DNA damage compared with ionizing radiation. Mammary gland relevance is less certain due to the relatively few experiments in breast tissue.

Multiple studies support the hypothesis that elevated RONS is a key part of the adverse outcome pathway for breast cancer from ionizing radiation. The strongest evidence comes from studies showing that reducing RONS also reduces DNA damage in irradiated cells and bystander cells, including genomic instability observed at later time points after IR. Free radical and NADPH oxidase inhibitors reduce the effect of IR on DNA nucleotide damage, double strand breaks, chromosomal damage, and mutations in isolated DNA and cultured cells (Winyard, Faux et al. 1992; Douki, Ravanat et al. 2006; Choi, Kang et al. 2007; Jones, Riggs et al. 2007; Ameziane-El-Hassani, Boufraqech et al. 2010; Ameziane-El-Hassani, Talbot et al. 2015; Manna, Das et al. 2015) and on nucleotide damage and double strand breaks in vivo (Pazhanisamy, Li et al. 2011; Ozyurt, Cevik et al. 2014). RONS reduction after ionizing radiation also reduces genomic instability in animals and in cloned cell lines (Dayal, Martin et al. 2008; Dayal, Martin et al. 2009; Pazhanisamy, Li et al. 2011; Bensimon, Biard et al. 2016). RONS are similarly implicated in IR effects in bystander cells. Antioxidants (including a nitric oxide scavenger) and oxidase inhibitors added before or after radiation reduce micronuclei and gamma-H2AX formation in bystander cells (Azzam, De Toledo et al. 2002; Yang, Asaad et al. 2005; Yang, Anzenberg et al. 2007). Antioxidant activity also reduces the inflammatory response to IR in animals and cultured skin cells (Berruyer, Martin et al. 2004; Das, Manna et al. 2014; Ozyurt, Cevik et al. 2014; Haddadi, Rezaeyan et al. 2017; Zhang, Zhu et al. 2017).

RONS are sufficient to trigger subsequent key events in this AOP. Extracellularly applied or intracellularly generated ROS (which also facilitates the formation of RNS) are capable of creating DNA damage in vitro including base damage, single and double strand breaks, and chromosomal damage (Oya, Yamamoto et al. 1986; Dahm-Daphi, Sass et al. 2000; Nakamura, Purvis et al. 2003; Gradzka and Iwanenko 2005; Ismail, Nystrom et al. 2005; Driessens, Versteyhe et al. 2009; Berdelle, Nikolova et al. 2011; Lorat, Brunner et al. 2015; Stanicka, Russell et al. 2015) and mutations (Sandhu and Birnboim 1997; Ameziane-El-Hassani, Boufraqech et al. 2010; Seager, Shah et al. 2012; Sharma, Collins et al. 2016). Similarly, decreased antioxidant activity and higher RONS is observed in cells with genomic instability (Dayal, Martin et al. 2008; Buonanno, de Toledo et al. 2011). To our knowledge, no experiments have tested whether elevating intracellular RONS alone in one group of cells can cause bystander effects in another.

Evidence in Mammary Gland

The increase of RONS following IR has been shown in a wide range of cells, in vivo and in vitro, including epithelial cells, and in two studies in mammary epithelial cells (Jones, Riggs et al. 2007; Bensimon, Biard et al. 2016). Both mammary cell studies also show increased RONS and DNA damage over a day after IR in vitro and link DNA damage with elevated RONS.

Uncertainties or Inconsistencies

The mitigating effects of antioxidants on IR-generated DNA damage support the essentiality of RONS in producing DNA damage and mutations. However, externally applied RONS is less effective than IR at generating double strand breaks and mutations (Sandhu and Birnboim 1997; Dahm-Daphi, Sass et al. 2000; Gradzka and Iwanenko 2005; Ismail, Nystrom et al. 2005). One possible explanation for this discrepancy is that IR may elicit a higher concentration of localized RONS than can be achieved with external application of H2O2. IR deposits energy and oxidizes molecules within a relatively small area over a rapid timescale potentially permitting a very high local concentration which could precede or overwhelm local buffering capacity. In contrast, extracellularly applied H2O2 would interact with many antioxidants and other molecules on its way to the nucleus, where the concentration would slowly reach a lower steady state.

As expected for RONS as a key event for DNA damage from IR, DNA damage from IR and H2O2 are additive in cells (Dahm-Daphi, Sass et al. 2000; Driessens, Versteyhe et al. 2009). Unexpectedly however, inhibiting glutathione (which should increase or sustain the effects of RONS), increases DNA damage from H2O2 but not IR. This lack of effect of glutathione inhibition on IR conflicts with multiple studies showing decreased DNA damage from IR with anti-oxidants. One possible explanation is that the concentration or reaction rate of glutathione is already inadequate to buffer the elevated RONS from IR, so further inhibition has no measurable effect.

KE/AO: Increase in DNA damage

Essentiality is High. The essentiality of this MIE to cancer is generally accepted. Supporting evidence comes from application of mutagenic agents: the increase in DNA damage precedes mutations, proliferation, and tumorigenesis. Further indirect evidence comes from evidence for MIE1, in which antioxidants that reduce DNA damage also reduce mutations and chromosomal damage. Finally, mutations in DNA repair genes increase the risk of tumors.

Increases or decreases in DNA damage are associated with corresponding increases or decreases in downstream key events in the pathway to breast cancer. An external agent (ionizing radiation) that increases DNA damage (Padula, Ponzinibbio et al. 2016) also causes chromosomal damage and increased mutations (Sandhu and Birnboim 1997; Jones, Riggs et al. 2007; Denissova, Nasello et al. 2012; Fibach and Rachmilewitz 2015), transforms cells (Yang, Craise et al. 1992; Yang, Georgy et al. 1997; Unger, Wienberg et al. 2010), and causes tumors (Poirier and Beland 1994; Little 2009). Polymorphisms or mutations in DNA repair genes affect tumor formation after ionizing radiation in animals (Yu, Okayasu et al. 2001; Umesako, Fujisawa et al. 2005) and in people (Millikan, Player et al. 2005; Andrieu, Easton et al. 2006; Broeks, Braaf et al. 2007; Bernstein, Haile et al. 2010; Brooks, Teraoka et al. 2012; Pijpe, Andrieu et al. 2012; Bernstein, Thomas et al. 2013). Consistent with these findings, antioxidants that reduce DNA damage from stressors like IR also reduce chromosomal aberrations and micronuclei arising from those stressors (Azzam, De Toledo et al. 2002; Choi, Kang et al. 2007; Jones, Riggs et al. 2007).

Evidence in mammary gland

The majority of research on the effects of IR on DNA damage has been performed in tissues other than mammary gland, but several studies suggest that effects in the mammary gland (and its consequences) would be consistent with other tissues. Oxidative DNA damage in mammary cells increases immediately after exposure to IR (Haegele, Wolfe et al. 1998), and double stranded breaks, micronuclei, and (later) chromosomal aberrations appear two hours to six days after IR exposure in vivo and in vitro (Soler, Pampalona et al. 2009; Snijders, Marchetti et al. 2012; Hernandez, Terradas et al. 2013). Genomic instability was reported in genetically susceptible cells after a month of higher doses of IR (4 doses of 1.8 Gy but not 0.75 Gy) (Snijders, Marchetti et al. 2012).

KE/AO: Increase in mutation

Essentiality is High. The contribution of this MIE to cancer is generally accepted. Evidence comes from knock-out and knock-in experiments, which find that mutations in certain key genes increase tumorigenesis. However, an ongoing debate pits the singular importance of mutations against a significant role for the tissue microenvironment. This debate is fueled by transplant studies that show the importance of tissue environment for tumorigenesis and suggesting that mutations may not be sufficient for tumorigenesis.

Mutations increase transformation in culture (Wang, Su et al. 2011) and proliferation and tumors in mice (Radice, Ferreira-Cornwell et al. 1997; Umesako, Fujisawa et al. 2005; de Ostrovich, Lambertz et al. 2008; Podsypanina, Politi et al. 2008; Francis, Bergsied et al. 2009; Gustin, Karakas et al. 2009; Francis, Chakrabarti et al. 2011; Tao, Xiang et al. 2017). Restoring function in mutated genes regresses tumors in animals (Martins, Brown-Swigart et al. 2006; Podsypanina, Politi et al. 2008). Mutations are common in tumors (Haag, Hsu et al. 1996; Greenman, Stephens et al. 2007; Stratton, Campbell et al. 2009; CGAN (Cancer Genome Atlas Network) 2012; Vandin, Upfal et al. 2012; Garraway and Lander 2013; Vogelstein, Papadopoulos et al. 2013; Yang, Killian et al. 2015) and tumors are largely clonal, suggesting that individual mutations offer the tumor evolutionary advantages (Wang, Waters et al. 2014; Yates, Gerstung et al. 2015; Begg, Ostrovnaya et al. 2016).

Evidence in mammary gland

Many of the studies in support of the proliferative and tumorigenic role of mutations are in mammary gland or breast cancers. Further support for including DNA damage and mutation in the mechanistic pathway linking ionizing radiation with breast cancer comes from the observation that variants in DNA repair genes increase the risk of mammary tumors in animals after IR (Yu, Okayasu et al. 2001; Umesako, Fujisawa et al. 2005) and increase breast cancer after IR (Millikan, Player et al. 2005; Andrieu, Easton et al. 2006; Broeks, Braaf et al. 2007; Bernstein, Haile et al. 2010; Brooks, Teraoka et al. 2012; Pijpe, Andrieu et al. 2012; Bernstein, Thomas et al. 2013). BRCA is perhaps the best known DNA repair gene linked with breast cancer risk, and several studies of these studies have suggested a link between BRCA mutation status and increased susceptibility to breast cancer following ionizing radiation, particularly in women exposed at younger ages (Pijpe, Andrieu et al. 2012).

Uncertainties or Inconsistencies

Mutations alone are not sufficient or even essential for tumor growth in mammary glands. Mammary tumor incidence following ionizing radiation varies significantly by sex and depends on the presence of ovarian hormones (Cronkite, Shellabarger et al. 1960; Segaloff and Maxfield 1971; Shellabarger, Stone et al. 1976; Holtzman, Stone et al. 1979; Holtzman, Stone et al. 1981; Welsch, Goodrich-Smith et al. 1981; Clifton, Yasukawa-Barnes et al. 1985; Solleveld, van Zwieten et al. 1986; Broerse, Hennen et al. 1987; Lemon, Kumar et al. 1989; Inano, Suzuki et al. 1991; Inano, Suzuki et al. 1996; Peterson, Servinsky et al. 2005). Tumor growth from transplanted tumor cells varies with age, parity, and lactational status (Maffini, Calabro et al. 2005; McDaniel, Rumer et al. 2006), and stroma treated with carcinogens or IR supports tumors from pre-malignant epithelial cells (Barcellos-Hoff and Ravani 2000; Maffini, Soto et al. 2004; Nguyen, Oketch-Rabah et al. 2011). While the mechanisms underlying these contextual factors have not been clearly identified, the proliferative effect of hormones on the mammary gland may serve to amplify damaged and mutated cells and modify the stromal environment to increase the likelihood of cellular transformation. Inflammatory responses including the release of cytokines and the activation of inflammatory and anti-inflammatory signaling pathways likely also amplify the effects of DNA damage and mutations through many of the same mechanisms.

KE: Increase in proliferation

Essentiality is High. Cellular proliferation is a key characteristic of cancer cells (Hanahan and Weinberg 2011) and can lead to hyperplasia, an intermediate phase in the development of tumorigenesis. Proliferation also increases the number of cells with mutations, which can further promote proliferation and/or changes to the local microenvironment.

Evidence in mammary gland

Multiple studies show that mammary gland proliferates after IR or chemical carcinogen treatment prior to the appearance of mammary tumors. Epithelial cells proliferate following IR in vitro (Mukhopadhyay, Costes et al. 2010) and in vivo (Nguyen, Oketch-Rabah et al. 2011; Snijders, Marchetti et al. 2012; Suman, Johnson et al. 2012; Tang, Fernandez-Garcia et al. 2014). Increasing proliferation leads to hyperplasia (Korkaya, Paulson et al. 2009). Proliferative nodules and hyperplasia appear in mammary terminal end bud, alveolae, and ducts of rats and mice after exposure to chemical carcinogens (Beuving, Bern et al. 1967; Beuving, Faulkin et al. 1967; Russo, Saby et al. 1977; Purnell 1980) and ionizing radiation (Faulkin, Shellabarger et al. 1967; Ullrich and Preston 1991; Imaoka, Nishimura et al. 2006). Proliferating foci precede the development of tumors (Haslam and Bern 1977; Purnell 1980) and form tumors more effectively than non-proliferating tissue (Deome, Faulkin et al. 1959; Beuving 1968; Rivera, Hill et al. 1981).

Supporting the essentiality of these proliferative processes to tumorigenesis, ACI rats that exhibit no mammary proliferation or hyperplasia following IR are resistant to tumors following IR (Kutanzi, Koturbash et al. 2010). Interventions reducing proliferation in susceptible PyVT and BALB/c mice also reduce mammary tumors (Luo, Fan et al. 2009; Connelly, Barham et al. 2011; Tang, Fernandez-Garcia et al. 2014).

Uncertainties or Inconsistencies

Some studies report carcinogenesis in the absence of hyperplasia (Sinha and Dao 1974) and others do not find increased tumorigenesis from transplanted hyperplasia (Beuving, Bern et al. 1967; Haslam and Bern 1977; Sinha and Dao 1977). The failure of some proliferative foci to form tumors and the regression of some tumors when formed (Haslam and Bern 1977; Purnell 1980; Korkola and Archer 1999) suggests that proliferation may not be sufficient for sustained tumorigenesis in mammary gland.

KE/AO: Increase, ductal hyperplasia

Essentiality is High. Evidence comes from transplant experiments showing that non-proliferating tissue is less tumorigenic than proliferating lesions, and from interventions that reduce both proliferation and tumors. Further evidence comes from animals that are resistant to both mammary gland proliferation and tumors from ionizing radiation. Uncertainty arises from conflicting evidence on the tumorigenicity of hyperplasia, the absence of hyperplasia observed before some tumors, and spontaneous regression of tumors.

Hyperplasia signals the presence of excess proliferation (a key characteristic of cancer cells (Hanahan and Weinberg 2011)) and represents an intermediate phase in the development of tumorigenesis.

Evidence in mammary gland

Multiple studies show that mammary gland proliferates after IR or chemical carcinogen treatment prior to the appearance of mammary tumors. Proliferative nodules and hyperplasia appear in mammary terminal end bud, alveolae, and ducts of rats and mice after exposure to chemical carcinogens (Beuving, Bern et al. 1967; Beuving, Faulkin et al. 1967; Russo, Saby et al. 1977; Purnell 1980) and ionizing radiation (Faulkin, Shellabarger et al. 1967; Ullrich and Preston 1991; Imaoka, Nishimura et al. 2006). Proliferating foci precede the development of tumors (Haslam and Bern 1977; Purnell 1980) and form tumors more effectively than non-proliferating tissue (Deome, Faulkin et al. 1959; Beuving 1968; Rivera, Hill et al. 1981). Adenocarcinomas in rats appear to preferentially form from terminal end bud hyperplasia (Haslam and Bern 1977; Russo, Saby et al. 1977; Purnell 1980), similar to the origin of many breast cancers for humans and for some mice after IR (Medina and Thompson 2000).

Supporting the essentiality of these proliferative processes to tumorigenesis, ACI rats that exhibit no mammary proliferation or hyperplasia following IR are resistant to tumors following IR (Kutanzi, Koturbash et al. 2010). Interventions reducing proliferation in susceptible PyVT and BALB/c mice also reduce mammary tumors (Luo, Fan et al. 2009; Connelly, Barham et al. 2011).

Uncertainties or Inconsistencies

Some studies report carcinogenesis in the absence of hyperplasia (Sinha and Dao 1974) and others do not find increased tumorigenesis from transplanted hyperplasia (Beuving, Bern et al. 1967; Haslam and Bern 1977; Sinha and Dao 1977). The failure of some lesions to form tumors and the regression of some tumors when formed (Haslam and Bern 1977; Purnell 1980; Korkola and Archer 1999) suggests that hyperplasia alone may not be sufficient for sustained tumorigenesis in mammary gland.

KEs: Tissue resident cell activation, Increase, Pro-inflammatory mediators, Leukocyte Recruitment/Activation

Essentiality is Moderate. These key events were reviewed as a group. Evidence comes from using genetic modifications, antibodies, and antioxidants to reduce inflammatory and anti-inflammatory factors. These interventions reduce DNA damage, mutations, and mechanisms contributing to tumorigenesis and invasion. Uncertainty arises from conflicting effects in different genetic backgrounds and in different organs.

Tumors and tumor cells exhibit features of inflammation, and inflammation is generally understood to promote transformation and tumor progression by supporting multiple hallmarks of cancer including oxidative activity and DNA damage, survival and proliferation, angiogenesis, and invasion and metastasis (Iliopoulos, Hirsch et al. 2009; Hanahan and Weinberg 2011; Esquivel-Velazquez, Ostoa-Saloma et al. 2015).

Many of these cancer promoting effects of inflammation can be seen following exposure to ionizing radiation (Bisht, Bradbury et al. 2003; Elahi, Suraweera et al. 2009; Nguyen, Oketch-Rabah et al. 2011; Bouchard, Bouvette et al. 2013; Nguyen, Fredlund et al. 2013; Illa-Bochaca, Ouyang et al. 2014). Inflammatory pathways are commonly activated in breast and mammary cancers following IR (Nguyen, Oketch-Rabah et al. 2011; Nguyen, Fredlund et al. 2013; Illa-Bochaca, Ouyang et al. 2014). Polymorphisms in inflammation genes are associated with breast cancer risk from IR in radiation technologists (Schonfeld, Bhatti et al. 2010) and with susceptibility to intestinal adenoma following IR in mice (Elahi, Suraweera et al. 2009). Cytokines TGF-β and IL6 transform  primary human mammospheres and pre-malignant mammary epithelial cell lines in vitro and make them tumorigenic in vivo (Sansone, Storci et al. 2007; Iliopoulos, Hirsch et al. 2009; Nguyen, Oketch-Rabah et al. 2011), and inflammation related factors COX2 and TGF-β are required for the full effect of IR on DNA damage and transformation in vitro and mammary tumor growth and invasion in vivo (Bisht, Bradbury et al. 2003; Nguyen, Oketch-Rabah et al. 2011).

One mechanism of cancer promotion involves oxidative activity and DNA damage: inflammation in response to IR increases oxidative activity in a positive feedback loop leading to increased DNA lesions and mutations. Oxidative activity mediates the increase in inflammatory markers (TNF-a and neutrophil markers) in bladder and kidney (Ozyurt, Cevik et al. 2014), and TNF-a and neutrophils increase oxidative activity (Jackson, Gajewski et al. 1989; Stevens, Bucurenci et al. 1992; Zhang, Zhu et al. 2017). Inflammatory activity from neutrophils and TNF-a and NF-kB-dependent COX2 and NO damage DNA and increase mutations by increasing oxidative activity (Jackson, Gajewski et al. 1989; Zhou, Ivanov et al. 2005). The mutations can be reduced by blocking the inflammatory factors NF-kB, COX2, TNF-a, or nitric oxide, or with antioxidants (Jackson, Gajewski et al. 1989; Zhou, Ivanov et al. 2005; Zhou, Ivanov et al. 2008; Zhang, Zhu et al. 2017). Antibodies to TNF-a or TGF-β reduce DNA damage in bone marrow (Burr, Robinson et al. 2010; Rastogi, Coates et al. 2012) and CHO cells (Han, Chen et al. 2010). Inhibiting TNF-a also reduces genomic instability in directly irradiated (but not bystander) lymphocytes (Moore, Marsden et al. 2005) and in bone marrow of CBA/Ca mice susceptible to IR-induced leukemia but not resistant C57BL/6 mice (Lorimore, Mukherjee et al. 2011). Inhibiting inflammatory factors NF-kB or iNOS reduces IR-induced bystander mutations in lung fibroblasts (Zhou, Ivanov et al. 2008).

Inflammatory pathways activated by IR are also capable of promoting tumor growth and metastasis. Exposure to IR or RONS sensitizes mammary epithelial cells to respond to TGF-β - which is widely activated by IR (Ehrhart, Segarini et al. 1997). IR and TGF-β signaling leads to an epithelial to mesenchymal (EMT)-like transition, which disrupts the expression and distribution of cell adhesion molecules and multicellular organization and promotes invasion (Park, Henshall-Powell et al. 2003; Andarawewa, Erickson et al. 2007; Andarawewa, Costes et al. 2011; Iizuka, Sasatani et al. 2017). This mechanism resembles wound healing (Koh and DiPietro 2011; Perez, Vago et al. 2014; Landen, Li et al. 2016), but also resembles malignancy - invasive breast cancer cell lines overexpress TGF-β and respond to TGF-β with increased invasion (Kim, Kim et al. 2004; Gomes, Terra et al. 2012). 

The response to TGF-β likely involves an increase in senescence in fibroblasts. IR-induced senescence releases a suite of signaling molecules including pro-inflammatory IL6 and proteases (MMPs) (Tsai, Chuang et al. 2005; Liakou, Mavrogonatou et al. 2016; Perrott, Wiley et al. 2017). The signaling molecules released by IR-senescent fibroblasts promote the disorganized tissue structure of mammary epithelial cells and the growth, EMT, and invasion of breast cancer epithelial cells or mutant epithelial cells (Tsai, Chuang et al. 2005; Liakou, Mavrogonatou et al. 2016; Perrott, Wiley et al. 2017) and 3D mammary tumor models (Sourisseau, Harrington et al. 2011). The induction of senescence in fibroblasts by IR requires TGF-β (Liakou, Mavrogonatou et al. 2016), and the release of the pro-invasive signaling molecules involves an IL-1 dependent activation of NF-kB (Perrott, Wiley et al. 2017). Senescence following IR also selects for a post-senescent variant of epithelial cell that is more conducive to tumorigenesis (Mukhopadhyay, Costes et al. 2010).

Il6 may play an important function in the carcinogenic response to IR. IL6 is expressed in mouse mammary gland after IR (Bouchard, Bouvette et al. 2013). IL6 is produced by IR-senescent fibroblasts, but may also be expressed by epithelial cells after IR since primary human mammospheres and pre-malignant mammary epithelial cell lines respond to IL6 with increased IL6 expression (Sansone, Storci et al. 2007; Iliopoulos, Hirsch et al. 2009). IL6 promotes the mobility and tumorigenesis of normal and breast cancer epithelial cells (Sansone, Storci et al. 2007; Sasser, Sullivan et al. 2007; Studebaker, Storci et al. 2008; Iliopoulos, Hirsch et al. 2009; Iliopoulos, Jaeger et al. 2010). This activity depends on transcription factor NOTCH3, which supports the renewal of stem-like cell populations (Sansone, Storci et al. 2007), and NOTCH has been implicated in multiple other studies in the proliferative response to IR in mammary epithelia (Nguyen, Oketch-Rabah et al. 2011; Marusyk, Tabassum et al. 2014; Tang, Fernandez-Garcia et al. 2014). The NF-kB/IL6/STAT3 signaling pathway generates cancer stem cells in multiple types of breast cancer cells (Iliopoulos, Hirsch et al. 2009; Iliopoulos, Jaeger et al. 2010; Iliopoulos, Hirsch et al. 2011) and is also implicated in colon and other cancers (Iliopoulos, Jaeger et al. 2010). The inflammation related transcription factor NF-kB also contributes to mammary tumorigenesis and metastasis in PyVt mice, in which mammary tumors are induced by expression of an MMTV-driven oncogene (Connelly, Barham et al. 2011). Interestingly, breast cancer fibroblasts and fibroblasts from common sites of breast cancer metastasis (bone, lung) express IL6. IL6 is required for the growth and tumor promoting effects of these fibroblasts on ER-positive cancer cells in vitro and in vivo. ER-negative breast epithelial cells release autocrine IL6 and may therefore be less dependent on IL6 from fibroblasts, although IL6 also transforms these cells (Sasser, Sullivan et al. 2007; Studebaker, Storci et al. 2008; Iliopoulos, Hirsch et al. 2009).

Inflammation is suspected to play a role in the indirect effects of radiation, in which cells not directly targeted by radiation exhibit effects including DNA damage and RONS (Lorimore and Wright 2003; Mukherjee, Coates et al. 2014; Sprung, Ivashkevich et al. 2015). In addition to the IR-induced release of inflammatory signals that are diffusible and can trigger systemic immune responses, inflammatory factors COX2 and TGF-β are produced in bystander cells that are not directly irradiated but are exposed to irradiated cells or media (Zhou, Ivanov et al. 2005; Zhou, Ivanov et al. 2008; Chai, Calaf et al. 2013; Chai, Lam et al. 2013; Wang, Wu et al. 2015).

Inflammatory factors TGF-β, TNF-a, COX2, and NO are implicated in the RONS (Shao, Folkard et al. 2008; Zhou, Ivanov et al. 2008; Wang, Wu et al. 2015), DNA damage (Dickey, Baird et al. 2009; Han, Chen et al. 2010; Dickey, Baird et al. 2012; Chai, Calaf et al. 2013; Chai, Lam et al. 2013; Wang, Wu et al. 2015) and mutations (Zhou, Ivanov et al. 2005; Zhou, Ivanov et al. 2008) observed in bystander cells and in the appearance of genomic instability (Moore, Marsden et al. 2005; Natarajan, Gibbons et al. 2007; Lorimore, Chrystal et al. 2008; Lorimore, Mukherjee et al. 2011) after IR. Further evidence for inflammation in indirect effects of IR come from tumors arising from mammary epithelial cells transplanted into IR exposed cleared fat pads: inflammation-related genes and pathways are upregulated or enriched in the gene expression patters of these indirectly IR-induced tumors (Nguyen, Oketch-Rabah et al. 2011; Nguyen, Fredlund et al. 2013; Illa-Bochaca, Ouyang et al. 2014).

Evidence in mammary gland

Many of the studies above that link inflammatory signals with increased oxidative activity, senescence, EMT, bystander effects, genomic instability, and tumorigenesis, and metastasis use mammary tissue. Since inflammation-related signals are reported after IR in mammary gland (Barcellos-Hoff, Derynck et al. 1994; Dickey, Baird et al. 2009; Datta, Hyduke et al. 2012; Snijders, Marchetti et al. 2012; Bouchard, Bouvette et al. 2013; Wang, Wu et al. 2015) inflammation likely contributes to many of the effects of IR in this tissue.

Uncertainties or inconsistencies

The effects of inflammation can be both pro and anti-tumorigenic. For example, in addition to TGF-β’s role in EMT, in mammary epithelial cells TGF-β is essential to apoptosis of DNA damaged cells including damage following ionizing radiation (Ewan, Henshall-Powell et al. 2002), thus limiting genomic instability (Maxwell, Fleisch et al. 2008). Inflammatory factors TNF-a and COX2 play a similar role in bone marrow of C57BL/6 mice (Lorimore, Rastogi et al. 2013). By eliminating cells with severe DNA damage and curtailing genomic instability, apoptosis (and therefore TGF-β or TNF-a) limits the appearance of major (possibly carcinogenic) mutations following ionizing radiation. However, apoptosis (and thus TGF-β or TNF-a) can indirectly promote tumorigenesis through compensatory proliferation (Loree, Koturbash et al. 2006; Fogarty and Bergmann 2017).

Genetic background also influences the interaction between inflammation and tumorigenesis. Polymorphisms in inflammatory genes influence susceptibility to intestinal cancer following IR (Elahi, Suraweera et al. 2009). In the SPRET outbred mouse higher baseline TGF-β during development decreases tumor incidence following lower doses of IR (0.1 Gy), possibly by reducing ductal branching and susceptibility (Zhang, Lo et al. 2015). Conversely, the BALB/c mouse susceptible to mammary tumors after IR has a lower baseline TGF-β (and a polymorphism in a DNA damage repair-related gene). Early (4 hours) after low dose (0.075 Gy) IR BALB/c mice have suppressed immune pathways and macrophage response but increased IL6, COX2, and TGF-β pathway activation in mammary gland compared to the tumor-resistant C57BL/6 mouse (Snijders, Marchetti et al. 2012; Bouchard, Bouvette et al. 2013).  By 1 week after IR, the BALB/c mice show TGF-β -dependent inflammation in the mammary gland, and by 1 month after IR, their mammary glands show proliferation (Nguyen, Martinez-Ruiz et al. 2011; Snijders, Marchetti et al. 2012), suggesting that TGF-β is associated with inflammation, proliferation, and mammary tumorigenesis in these mice. Consistent with this pattern, BALB/c mice that are heterozygous for TGF-β are more resistant to mammary tumorigenesis following IR (Nguyen, Oketch-Rabah et al. 2011). However, the BALB/c mouse also has a polymorphism in a DNA repair gene associated with IR-induced genomic instability (Yu, Okayasu et al. 2001), making it difficult to distinguish potentially overlapping mechanisms.

While inflammatory signals are associated with bystander effects including DNA damage, genomic instability, and mutation, these effects vary between organs in vivo (Chai, Calaf et al. 2013; Chai, Lam et al. 2013), by genotype (Coates, Rundle et al. 2008; Lorimore, Chrystal et al. 2008; Lorimore, Mukherjee et al. 2011), and by cell type (Chai, Calaf et al. 2013). Further research will be required to identify all the underlying factors determining differences in bystander effects, but one variable is the appearance of a protective apoptotic response to cytokines under some conditions (Lorimore, Mukherjee et al. 2011; Lorimore, Rastogi et al. 2013).

One major piece of conflicting evidence comes from a direct test of the essentiality of inflammation to IR-induced carcinogenesis. In a mouse model of lymphoma, a mutation preventing the PIDD/NEMO dependent activation of NF-kB blocks early IR-induced activation of NF-kB (4-24 h) and production of TNF-a (5-48 h) but not lymphoma, suggesting that activation of these inflammatory factors is not essential in this time period (Bock, Krumschnabel et al. 2013). However, this study examined only day one post-IR time points for NF-kB activity, and did not block production of IL6. Later activation of NF-kB or activation of other inflammation-related factors including IL6 and TGF-β could therefore potentially have contributed to lymphoma.

Support for biological plausibility of KERs 

Empirical support for KERs