Carcinogen-specific mutations in preferred Ras-Raf pathway oncogenes directed by strand bias

Different Oncogenes and Reproductive Histories Shape the Progression of Distinct Premalignant Clones in Multistage Mouse Breast Cancer Models

A remote carcinogen exposure can predispose to breast cancer onset decades later, suggesting that carcinogen-induced mutations generate long-lived premalignant clones. How subsequent events influence the progression of specific premalignant clones remains poorly understood. Herein, multistage mouse models of mammary carcinogenesis were generated by combining chemical carcinogen exposure [using 7,12-dimethylbenzanthracene (DMBA)] with transgenes that enable inducible expression of one of two clinically relevant mammary oncogenes: c-MYC (MYC) or PIK3CAH1047R (PIK). In prior work, DMBA exposure generated mammary clones bearing signature HrasQ61L mutations, which only progressed to mammary cancer after inducible Wnt1 oncogene expression. Here, after an identical DMBA exposure, MYC versus PIK drove cancer progression from mammary clones bearing mutations in distinct Ras family paralogs. For example, MYC drove cancer progression from either Kras- or Nras-mutant clones, whereas PIK transformed Kras-mutant clones only. These Ras mutation patterns were maintained whether oncogenic transgenes were induced within days of DMBA exposure or months later. Completing a full-term pregnancy (parity) failed to protect against either MYC- or PIK-driven tumor progression. Instead, a postpartum increase in mammary tumor predisposition was observed in the context of PIK-driven progression. However, parity decreased the overall prevalence of tumors bearing Krasmut, and the magnitude of this decrease depended on both the number and timing of pregnancies. These multistage models may be useful for elucidating biological features of premalignant mammary neoplasia.

Prolactin synergizes with canonical Wnt signals to drive development of ER+ mammary tumors via activation of the Notch pathway

Prolactin (PRL) cooperates with other factors to orchestrate mammary development and lactation, and is epidemiologically linked to higher risk for breast cancer. However, how PRL collaborates with oncogenes to foster tumorigenesis and influence breast cancer phenotype is not well understood. To understand its interactions with canonical Wnt signals, which elevate mammary stem cell activity, we crossed heterozygous NRL-PRL mice with Apc^Min/+ mice and treated pubertal females with a single dose of mutagen. PRL in the context of Apc^Min/+ fueled a dramatic increase in tumor incidence in nulliparous mice, compared to Apc^Min/+ alone. Although carcinomas in both NRL-PRL/Apc^Min/+ and Apc^Min/+ females acquired a mutation in the remaining wildtype Apc allele and expressed abundant β-catenin, PRL-promoted tumors displayed higher levels of Notch-driven target genes and Notch-dependent cancer stem cell activity, compared to β-catenin-driven activity in Apc^Min/+ tumors. This PRL-induced shift to dominant Notch signals was evident in preneoplastic epithelial hyperplasias at 120 days of age. In NRL-PRL/Apc^Min/+ females, rapidly proliferating hyperplasias, characterized by β-catenin at cell junctions and high NOTCH1 expression, contrasted with slower growing lesions with nuclear β-catenin in Apc^Min/+ females. These studies demonstrate that PRL can powerfully modulate the incidence and phenotype of mammary tumors, shedding light on mechanisms whereby PRL elevates risk of breast cancer.

Dimethylbenz(a)anthracene-induced mammary tumorigenesis in mice

Every year, over 2 million women are diagnosed with breast cancer. Although considerable progress was made within the last years in cancer prevention, diagnosis and treatment, breast cancer is still responsible for over 600,000 of deaths per year. Over the years, numerous mouse models have been developed to understand breast cancer etiology and progression. Among those, mammary carcinomas induced by carcinogen, such as 7,12-dimethylbenz[a]anthracene (DMBA), has been widely used. Generally, 30-70% of mice exposed to 4-6 weekly doses of 1mg of DMBA during the peripubertal period (4-10 weeks of age) will develop mammary tumors within 150-200 days after the first exposure, that sometime metastasize to the lungs. As a result, DMBA-induced tumorigenesis is thought to be an accurate and relevant model to study breast cancer as it closely mimics this multistep process. This chapter presents the typical protocol used in mice to induce mammary gland tumors using DMBA. The influence of the number of doses and the total burden of DMBA given, as well as of the age and strain of the mice on mammary gland incident and on tumor onset are discussed. The current knowledge regarding mechanisms involved in DMBA-induced tumorigenesis is also presented.

A Multistage Murine Breast Cancer Model Reveals Long-Lived Premalignant Clones Refractory to Parity-Induced Protection

Breast cancers evolve in a multistage process that can span decades after a carcinogenic exposure. It follows that long-lived precursor breast lesions persist in a subclinical state prior to completing malignant transformation, yet widely used breast cancer models lack an experimental framework for targeting premalignant disease. Inspired by classic multistage skin carcinogenesis protocols, we combined chemical carcinogenesis with transgenic mouse modeling to resolve mouse mammary carcinogenesis into discrete initiation and progression stages. At the initiation stage, exposure to the carcinogen 7,12-dimethylbenzanthracene (DMBA) generated "initiated mammary epithelial cells" (iMEC) by introducing a stereotyped Hras^Q61L driver mutation. Whether DMBA exposure occurred during puberty or adulthood, mice efficiently acquired iMEC clones that eluded detection by conventional histology, yet were long lived, persisting in a clinically silent state for months in the absence of a cooperating event. At the progression stage, inducible activation of oncogenic Wnt signaling drove rapid and synchronous transformation of latent iMECs into overt mammary carcinomas, while Wnt activation in neighboring normal mammary epithelium yielded only benign hyperplasia over this same time period. Although early parity (completion of a full-term pregnancy) reduces breast cancer risk in some contexts, standard parity-induced protection schemes failed to eliminate iMECs in our multistage model, suggesting Wnt-responsive iMECs are maintained by hormone-independent mechanisms. Variations on our multistage modeling strategy may help to identify and validate cellular and molecular targets for breast cancer chemoprevention.

Evolution of Relapse-Proficient Subclones Constrained by Collateral Sensitivity to Oncogene Overdose in Wnt-Driven Mammary Cancer

Targeted cancer therapeutics select for drug-resistant rescue subclones (RSCs), which typically carry rescue mutations that restore oncogenic signaling. Whereas mutations underlying antibiotic resistance frequently burden drug-naive microbes with a fitness cost, it remains unknown whether and how rescue mutations underlying cancer relapse encounter negative selection prior to targeted therapy. Here, using mouse models of reversible, Wnt-driven mam-mary cancer, we uncovered stringent counter-selection against Wnt signaling overdose during the clonal evolution of RSCs. Analyzing recurrent tumors emerging during simulated targeted therapy (Wnt withdrawal) by multi-region DNA sequencing revealed polyclonal relapses comprised of multiple RSCs, which bear distinct but functionally equivalent rescue mutations that converge on sub-maximal Wnt pathway activation. When superimposed on native (i.e., undrugged) signaling, these rescue mutations faced negative selection, indicating that they burden RSCs with a fitness cost before Wnt withdrawal unmasks their selective advantage. Exploiting collateral sensitivity to oncogene overdose may help eliminate RSCs and prevent cancer relapse.

Keller and Gunther show that Wnt-driven mammary cancers challenged with simulated targeted therapy (Wnt withdrawal) undergo clonal evolution, which stringently selects for mutations that restore a “just right” level of oncogenic signaling. Therefore, cancer relapses emerge from rare subclones that are encumbered by an untapped vulnerability to oncogene overdose.

Pubertal and adult windows of susceptibility to a high animal fat diet in Trp53-null mammary tumorigenesis

Premenopausal breast cancer is associated with increased animal fat consumption among normal weight, but not overweight women (Farvid et al., 2014). Our previous findings in obesity-resistant BALB/c mice similarly showed promotion of carcinogen-induced mammary tumorigenesis by a diet high in saturated animal fat (HFD). This effect was specific to pubertal versus adult HFD. This study identifies the effects of HFD during puberty versus adulthood in Trp53-null transplant BALB/c mice and investigates its mechanism of enhancing tumorigenesis. Either pubertal or adult HFD is sufficient to increase incidence of Trp53-null mammary tumors. Puberty-restricted HFD exposure promoted tumor cell proliferation, increased angiogenesis, and increased recruitment of total and M2 macrophages in epithelial tumors. Adult-restricted exposure to HFD similarly increased proliferation, angiogenesis, recruitment of total and M2 macrophages, and additionally reduced apoptosis. Adult HFD also increased incidence of spindle cell carcinomas resembling claudin-low breast cancer, and thus adult HFD in the Trp53-null transplantation system may be a useful model for human claudin low breast cancer. Importantly, these results on Trp53-null and our prior studies on DMBA-induced mammary tumorigenesis demonstrate a pubertal window of susceptibility to the promotional effects of HFD, indicating the potential of early life dietary intervention to reduce breast cancer risk.

Chemical Carcinogenesis Models of Cancer: Back to the Future

Over a century has elapsed since the first demonstration that exposure to chemicals in coal tar can cause cancer in animals. These observations provided an essential causal mechanistic link between environmental chemicals and increased risk of cancer in human populations. Mouse models of chemical carcinogenesis have since led to the concept of multistage tumor development through distinct stages of initiation, promotion, and progression and identified many of the genetic and biological events involved in these processes. Recent breakthroughs in DNA sequencing have now given us tools to dissect complete tumor genome architectures and revealed that chemically induced cancers in the mouse carry a high point mutation load and mutation signatures that reflect the causative agent used for tumor induction. Chemical carcinogenesis models may therefore provide a route to identify the causes of mutation signatures found in human cancers and further inform studies of therapeutic drug resistance and responses to immunotherapy, which are dependent on mutation load and genetic heterogeneity.