Hereditary cancer syndromes as model systems for chemopreventive agent development

Seminars in Oncology 43 (2016) 134–145

Contents lists available at ScienceDirect

Seminars in Oncology journal homepage: www.elsevier.com/locate/ysonc

Hereditary cancer syndromes as model systems for chemopreventive agent development Farzana L. Walcotta,n, Jigar Patela, Ronald Lubetb, Luz Rodriguezc, Kathleen A. Calzoned,n a

National Institutes of Health, National Cancer Institute, Division of Cancer Prevention, Bethesda, MD, USA Consultant to National Institutes of Health, National Cancer Institute, Division of Cancer Prevention, Chemopreventive Agent Development Research Group, Bethesda, MD, USA c National Institutes of Health, National Cancer Institute, Division of Cancer Prevention, Gastrointestinal & Other Cancers Research, Bethesda, MD, USA d National Institutes of Health, National Cancer Institute, Center for Cancer Research, Genetics Branch, Bethesda, MD, USA b

a r t i c l e i n f o

abstract

Keywords: Chemoprevention Hereditary cancer Lynch syndrome Li-Fraumeni syndrome HBOC

Research in chemoprevention has undergone a shift in emphasis for pragmatic reasons from large, phase III randomized studies to earlier phase studies focused on safety, mechanisms, and utilization of surrogate endpoints such as biomarkers instead of cancer incidence. This transition permits trials to be conducted in smaller populations and at substantially reduced costs while still yielding valuable information. This article will summarize some of the current chemoprevention challenges and the justification for the use of animal models to facilitate identification and testing of chemopreventive agents as illustrated though four inherited cancer syndromes. Preclinical models of inherited cancer syndromes serve as prototypical systems in which chemopreventive agents can be developed for ultimate application to both the sporadic and inherited cancer settings. Published by Elsevier Inc.

1. Introduction The search for the ideal natural, synthetic, or biologic agents to reverse, suppress, or prevent cancer has been the aim of cancer chemoprevention research, beginning in 1976 with Dr Michael Sporn’s [1] creation of the term "chemoprevention". The approvals of tamoxifen and raloxifene by the FDA (1999 and 2007, respectively) for breast cancer chemoprevention, or more precisely risk reduction, were successes that have yet to be achieved by other agents such as aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs) for colorectal cancer and finasteride for prostate cancer prevention. Disappointing results from large trials and budgetary constraints have led to a shift in focus from funding large, phase III randomized trials to smaller earlier phase studies focused on safety, mechanistic elucidation, and biomarker development.

n Corresponding authors. Farzana L. Walcott, MD, MPH (For questions regarding LFS and LFS clinical trial NCT01981525), National Institutes of Health, National Cancer Institute, Division of Cancer Prevention, Cancer Prevention Fellowship Program, 9609 Medical Center Dr, #5E522, Rockville, MD 20850. Kathleen A. Calzone, PhD, RN, APNG, FAAN, National Institutes of Health, National Cancer Institute, Center for Cancer Research, Genetics Branch, 37 Convent Dr, Building 37, Room 3039, MSC 4256, Bethesda, MD. E-mail addresses: [email protected]; fl[email protected] (F.L. Walcott); [email protected] (K.A. Calzone)

http://dx.doi.org/10.1053/j.seminoncol.2015.09.015 0093-7754/Published by Elsevier Inc.

Biomarkers of drug effect may serve as surrogate endpoints for cancer incidence and drug toxicity. Another approach to overcoming the burden of large trials is to study very high-risk populations with germline mutations. Among individuals with such inherited genetic changes the rate of cancer development is much higher, allowing the use of smaller sample sizes than do trials involving moderately increased risk populations. Another advantage of the inherited syndrome approach is the well-defined genetic cancer predisposition of the cohort, which contrasts with the use of populations at moderately increased risk that may be considerably more heterogeneous at the molecular level. The use of genetic syndromes also facilitates the development of agents that target the relevant mutations. The value of testing chemopreventive agents in these high penetrance syndromes extends beyond the syndromes themselves to possible relevance for prevention of equivalent cancers in the sporadic setting. In addition, representative animal models of human hereditary cancers driven by germline mutations in a single gene or family of genes can be very useful. Inserting the relevant mutation into the genome of an animal, results in an imperfect but valuable model of human disease. Despite routine use, animal models of cancer have certain limitations, especially the fact that the physiology of a rodent is in many ways dissimilar to that of a human. Rodents have a much shorter life span, and in general tumors that arise in rodents are not as genetically complex at the chromosomal level as

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

135

Table 1 Mouse models of hereditary cancer syndromes. Animal model

Description

Outcome

APC mutant mice (Min, APC codon 1638 mutation)

Have mutation in APC gene relevant for Gardner syndrome (FAP) as well as most sporadic colon cancers.


Mice develop multiple adenomas but most are in small intestine
Mice respond to multiple agents that are effective in humans

MSH or MLH knockout mice Lynch/HNPCC

Have knockout of MSH or MLH while humans (Lynch/HNPCC) have mutations or methylation (sporadic colon cancer).

p53 knockout mice (LFS)

Typically knockout p53 mutations. Partially overlapping tumor spectra between human and mouse (osteosarcomas and lymphomas but not carcinomas and no metastasis). Tumor spectrum is dependent on strain of mouse.

(NSAIDs, DFMO)


Results in mice supported celecoxib trial in FAP that gave a positive result


Mouse tumors like the humans exhibit MSI phenotype
Mouse tumors do not show mutations in TGFβRII or BAX, which are


mutated in humans, due to lack of nucleotide repeats in coding region of mice. Mice appear to be responsive to NSAIDs as are humans.


Mice with p53 mutations or knockouts respond to agents which are effective in the organ in which they occur.

p53 R270H or R172H LFS mouse model with specific p53 arginine to histidine missense mutation at codon 172 or 270 corresponding to human hotspot knockin mice mutation at codon 175 or 273, respectively. (Li-Fraumeni)


Mice develop osteosarcomas, lymphomas, and carcinomas (similar

BRCA1/2 mice


In BRCA1-deficient mice, a p53 alteration must also be present in

Often knockout or partial deletion of BRCA1/2 in mice versus typically nonsense mutations in humans

to human LFS), with metastasis occurring to lymph nodes, lung, liver, and brain.





order to get a reasonable number of tumors. Most human BRCA1 tumors have P53 mutations Resulting mammary tumors have multiple genomic changes in mice and humans. PARP inhibitors are relatively effective in both mice and humans.

Abbreviations: APC, adenomatous polyposis coli; DFMO, difluoromethylornithine; FAP, familial adenomatous polyposis; HNPCC, hereditary nonpolyposis colon cancer; LFS, Li- Fraumeni syndrome; MSI, microsatellite instability; NSAIDs, nonsteroidal anti-inflammatory drugs.

human tumors. Cancers in animals typically exhibit less systematic amplification or deletion of specific chromosomal regions that are affected in human cancers. A key difference between the species is evident in the nature of the mutation. In man most hereditary syndromes are driven by one mutated allele in the germline and a second mutation in or loss of the intact allele in the tumor tissue. Interestingly, in animals, mutation or loss of a single allele (eg, BRCA1/2, MSH or MLH) in the germline does not routinely yield an animal with a tumor phenotype. Yet, mutations or knockouts of both copies of the gene in the germline can result in developmental changes or even embryonic toxicity, as in the case of BRCA1/2 [2]. This has prompted the development of models (Table 1) where the mutation or knockout of these genes is accomplished selectively in the target tissue. Herein we will address the current challenges of chemoprevention and the rationale for using inherited cancer syndromes as model systems for identifying and testing chemopreventive agents. While more than 50 hereditary cancer predisposition syndromes have been identified [3], four major inherited cancer syndromes that have accepted clinical genetic testing, established animal models, and ongoing chemoprevention efforts will be discussed: hereditary breast and ovarian cancer syndrome, LiFraumeni syndrome, familial adenomatous polyposis, and Lynch syndrome.

2. Hereditary breast and ovarian cancer syndrome, BRCA1 and BRCA2 Having a family history of breast and/or ovarian cancer has long been recognized as a risk factor for these malignancies [4–5]. Of the dominant, high-penetrance susceptibility alleles identified to date, mutations in BRCA1 and BRCA2 associated with hereditary breast and ovarian cancer syndrome (HBOC) are the most prevalent affecting approximately 1/400–800 in the general population

[6]. The prevalence is higher in populations such as Ashkenazi Jewish and Icelandic populations due to founder mutations [7–9]. 2.1. Mutation spectrum Both BRCA1 and BRCA2 are large genes: 24 exons encoding 1,863 amino acids for BRCA1 [10]; and 27 exons encoding 3,418 amino acids for BRCA2 [11]. Hundreds of different mutations span each gene, with more than 1,700 different BRCA1 and 2,000 different BRCA2 mutations, polymorphisms, and variants reported in the Breast Cancer Information Core, an online BRCA1/2 mutation database [12]. 2.2. Penetrance data Since the identification of BRCA1 in 1994 [10] and BRCA2 in 1995 [13], several studies have described the cancer penetrance of these mutations. The breast and ovarian cancer risks vastly exceed the risks found in the general population. Two separate metaanalyses have been conducted that help clarify the risks for breast and ovarian cancer by age 70, which are summarized in Table 2. These risks are lower in these population-based studies than in the original Breast Cancer Linkage Consortium data where risks for BRCA1or BRCA2 mutation-associated breast cancers approached 85%. Of note, the latter data were subject to family ascertainment bias [14,15]. Individual risks do vary based on personal, environmental, and genetic modifiers. Additionally, mutations in the 3´ and 5´ end of the BRCA1 and BRCA2 genes confer higher risks for breast cancer while mutations in the central portion of the BRCA1 and BRCA2 confer higher risks for ovarian cancer but lower overall breast cancer risks [15]. 2.3. Mechanistic data Both BRCA1 and BRCA2 are tumor suppressor genes and critical to chromosome structure preservation and numeric control during

136

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

Table 2 Hereditary breast/ovarian cancer and Lynch syndrome gene mutations and cancer-specific risks. Mutated gene

Breast cancer (female)

Ovarian cancer

Colorectal cancer

Endometrial cancer

Cancer risk (95% CI)-Risk by age 70 (95%CI)

Cancer risk (95% CI)-Risk by age 70 (95% CI)

Cancer risk (95% CI)

Cancer risk (95% CI)

49% (29%–85%) [122] 52% (31%–90%) [122] 18% (13%–30%) [122] 75% (65%–85%) [123]

57% (22%–82%) [122] 21% (9%–82%) [122] 17% (8%–47%) [122] 12% (0%–27%) [123]

Hereditary breast/ovarian cancer syndrome BRCA1 57% (47%–66%) [139] - 65% (44%–78%) [140] BRCA2

45% (31%–56%) [140] - 49% (40%–57%) [139]

Lynch syndrome MLH1 MSH2 MSH6 EPCAM

39% (18%–54%) [140] - 40% (35%–46%) [139] 11% (2.4%–19%) [140] - 18% (13%–23%) [139] 20% (1%–66%) [122] 38% (3%–81%) [122] 1% (0%–3%) [122]

Abbreviation: CI, confidence interval.

the cell cycle [16]. Chromosomal instability resulting from loss of BRCA1 or BRCA2 function is what leads to carcinogenesis. Both genes are involved in DNA double strand break repair but through different mechanisms. BRCA1 signals the presence of DNA damage, while BRCA2 participates in DNA repair by controlling RAD51 recombinase [16]. Other functions are also being investigated including protein ubiquitination and chromatin remodeling [17].

2.4. Chemoprevention for BRCA1/2 Current chemoprevention trials in BRCA1/2 mutation carriers listed on clinicaltrials.gov are few and mainly limited to short-term studies assessing tolerability and tissue-specific changes (ie, breast), after a brief period of intervention. There are 10 active chemoprevention trials currently listed which include BRCA1/2 mutation carriers (Table 3). While agents such as selective estrogen receptor modulators (SERMs) have been shown to be effective at reducing the risk of ER þ breast cancer, in BRCA1/2 mutation carriers 75% of breast cancers are ER- [18]. Further, triple-negative tumors (ER/progesterone receptor [PR]/Her2-negative) are found in approximately 70% of breast tumors in BRCA1 and 20% in BRCA2 mutation carriers [19]. The utility of these agents are questioned in BRCA mutation carriers for this reason, and because of inadequate efficacy data in mutation carriers. However, a recent pooled analysis of data collected from multiple cohorts of mutation carriers found a decreased risk of contralateral breast cancer with adjuvant tamoxifen use in BRCA1 and BRCA2 mutation carriers [20]. Currently, of major interest is a class of agents targeting the DNA repair pathway: poly (adenosine diphosphate [ADP]-ribose) polymerases (PARPs) inhibitors. Currently there are no chemoprevention trials involving PARP inhibitors and their utility in chemoprevention for BRCA1/2 mutation carriers remains an area of debate due to concerns about long-term toxicity, secondary tumors, and drug resistance. PARPs are a large family of enzymes that play a key role in the repair of DNA single-strand breaks [21] via base excision repair mechanisms (BER). The BRCA1 protein is normally required for homologous recombination repair (HRR) to maintain genomic integrity of the cell. BRCA1 mutation carriers normally retain one wild-type copy of the BRCA1 gene that allows for efficient DNA repair; however, loss of heterozygosity (LOH), involving loss of the wild-type copy of the gene, forces cells to rely on BER, which requires PARP. The inhibition of PARP in BRCA1deficient cells inhibits BER and induces these cells to undergo apoptosis. This process appears to be selective for tumor cells, as these cells have lost their wild type BRCA1 allele. The application of PARP inhibitors in this context exemplifies how drugs target tumor

cells via synthetic lethality, ie, the ability of a drug to be selectively toxic in a system with a genetic mutation in a critical gene. Clinical trials using PARP inhibitors for the treatment of cancer in BRCA1 and BRCA2 mutation carriers have exhibited efficacy (reviewed in [22]) with tolerable side effects. The latter consist of fatigue, nausea, vomiting, thrombocytopenia, and mild gastrointestinal symptoms [22,23]. Unfortunately, resistance to PARP inhibitors has also been noted [22]. PARP inhibitors are making their way into the adjuvant setting, but their role in chemoprevention is less clear. Thus far most of the evidence of PARP inhibitors for chemoprevention has been generated in mouse models reviewed in detail below. As appears in Table 3, a number of chemoprevention trials for BRCA1/2 mutation carriers are listed in clinicaltrials.gov. 2.5. Animal models of HBOC, BRCA1/2 The BRCA1 gene has multiple functions, including DNA repair and centrosome duplication. These attributes probably contribute to the high incidence of breast cancer in BRCA1 mutation carriers as well as contributing to the high levels of chromosome alterations observed in BRCA1-related tumors. As is true of some of the other models, an early finding in the case of BRCA1 was that knockout of a single copy of the complete gene resulted in virtually no disease phenotype in the absence of irradiation whereas homozygous knockout of BRCA1 proved to be embryotoxic. The solution to the practical challenge this poses for using BRCA1 models has been the development of conditional knockouts of the BRCA1 gene [24]. Such models involve knockout of portions of the BRCA1 gene in the target tissue, ie, “conditional” knockout. A specific DNA sequence (lox site) is placed in the BRCA1 gene in the germline of the mice and the lox sites are cut by a bacterial recombinase, leading to deletion of the target sequence within BRCA1. The activation of the recombinase is under the control of an organ specific promoter. In the model of breast cancer, the specificity in mice is for mammary tissue such that the conditional knockout of the target BRCA1 sequence is under the control of organ-specific, ie, mammary-specific, promoters such as MMTV (mouse mammary tumor virus), whey acidic protein or βlactoglobulin. This results in mammary-specific removal of the lox-defined site in cells of the mammary gland. A variety of breast cancer transgenic models have been derived. These fall into two main categories: those that depend on an altered genome alone and those that use mice with a modified BRCA1 in conjunction with irradiation to generate tumors [2]. The pure genomic lesions have typically entailed both an alteration in the BRCA1 gene as well as a knockout or alteration in TP53. In fact, mutations in TP53 are observed in the preponderance of BRCA1-driven breast cancer in

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

137

Table 3 Chemoprevention trials that are being conducted in designated hereditary cancer syndromes and are listed on clinicaltrials.gov (trials with unknown status not included). Agent BRCA1 and/or BRCA2 Aromatase inhibitor (letrozole) Recombinant human chorionic gonadotropin GnRH agonist (deslorelin) Mifepristone SERMs (arzoxifene/tamoxifen) Curcumin Cholecalciferol Fenretinide Soy isoflavones Metformin Li-Fraumeni syndrome Metformin FAP Celecoxib/ursodeoxycholic acid Eicosapentanoic acid (EPA) Calcumin (curcumin) Erythromycin Erlotinib (EGFR inhibitor)/sulindac Eflornithine/sulindac Lyophilized black raspberries Curcumin Eflornithine/sulindac Celecoxib/eflornithine Celecoxib Metformin Exisulind (withdrawn) Celecoxib Celecoxib Lynch syndrome Levonorgestrel-releasing intrauterine system Medroxyprogesterone; ethinyl estradiol; norgestrel; Naproxen Celecoxib Dendritic cell vaccination

Clinical trial identifier/phase

Outcome

NCT00673335/III NCT00700778/NL NCT00080756/II NCT01898312/II NCT00253539/II NCT01975363/NL NCT01097278/II NCT00098800/NL NCT01219075/NL NCT01905046/III

Survival post 1st breast cancer diagnosis Gene expression in breast epithelial cells Biomarkers Epithelial cell proliferation Ki-67 in breast tissue Safety and adherence/plasma and breast adipose tissue biomarker analysis Change in mammographic density and Ki-67 in breast biopsies Induction of apoptosis in ovarian stromal and epithelial cells Reduction in MRI volume Testing for atypia in fine needle aspirates after 12 months

NCT01981525/I

Safety and tolerability and effect on circulating IGF-1, insulin, and IGFBP3 and mitochondrial function

NCT00808743/II-III NCT00510692/II-III NCT00927485/NL NCT02175914/IV NCT01187901/II NCT01483144/III NCT00770991/I NCT00641147/NL NCT01245816/III NCT00033371/II NCT00685568/I NCT01725490/II NCT00026468/II-III NCT00503035/II NCT00005094/III

Number and size of duodenal adenomas Change in the number of polyps in the rectum Regression of intestinal adenomas: duodenal, colorectal/ileal polyp number, size Number and size of intestinal adenomas Change in total duodenal and colorectal polyp burden at 6 months Delaying time to the 1st occurrence of any FAP-related event Rectal polyp burden and biomarker modulation Number of polyps Qualitative change in overall colon/rectum/pouch polyp burden Change in number and size of duodenal polyp burden Aberrant crypt foci and colorectal adenoma burden Number and size of polyps in colon/duodenum Polyp prevention Biomarkers and apoptosis in colorectal polyps Colorectal polyp burden

NCT00566644/III NCT00033358/II

Atypical endometrial hyperplasia Biomarkers in endometrial biopsies

NCT02052908/I NCT00001693/I-II NCT01885702/I-II

Polyps in rectosigmoid area Biomarkers Safety and feasibility

Abbreviations: GnRH, gonadotropin releasing hormone; IGF-1, insulin-like growth factor 1; IGFBP3, insulin-like growth factor binding protein 3; SERM, selective estrogen receptor modulator.

humans. Nevertheless, the relatively involved breeding protocol (Cre/lox for BRCA1 combined with altered TP53) makes generation of appropriate animals difficult and time-consuming, limiting their routine use in screening for potential effective agents. The compounds of greatest mechanistic interest for prevention of BRCA1-related mammary cancers have been the PARP inhibitors. ADP ribosylation is involved in repair of single-strand breaks in DNA via BER. The accumulation of multiple single strand breaks caused by PARP inhibitors can proceed to the formation of doublestrand breaks during DNA synthesis. As previously discussed, it is these double-strand breaks that are normally repaired by a complex that includes the BRCA1/2 proteins, making cells deficient in BRCA1/2 particularly sensitive to the toxic and potentially mutagenic effects of such agents. The selectivity of cellular toxicity for the cancer cells, which are uniquely BRCA1/2-deficient, represents the desirable outcome. Jonkers and colleagues showed in tumors in BRCA1-deficient mice that these agents were effective therapeutic agents with or without cisplatinum [25]. Regrettably, tumors tended to develop resistance after continuous treatment. Nevertheless, these results have supported the implementation of recent clinical treatment trials for BRCA1-related cancer employing both a PARP inhibitor and cisplatinum. Recently To et al found that administration of PARP inhibitors early during tumor development in transgenic BRCA1 deficient mice significantly delayed tumor development [26]. These results confirm the expected efficacy of the PARP inhibitors but still leave open two concerns. First, resistance may be more likely to develop following long-term

PARP inhibitor treatment; this may even apply in the prevention setting where the absence of invasive cancer makes the development of resistance less likely. Second, the mechanism by which PARP inhibitors induce tumor cell cytotoxicity in patients with germline BRCA1 mutations is identical to the mechanism by which they induce non-therapeutic deleterious mutations. Two other classes of targeted therapies that have shown efficacy in BRCA1 transgenic models are the AKT inhibitors and the epidermal growth factor receptor (EGFR) inhibitors. The EGFR inhibitors have been proposed to exert their effects early in tumor development, potentially due to expression of EGFR1 in tumor stem cells [27]. Three hormonal regimens have been examined in BRCA1 models [2]. Oophorectomy, much as in humans, was relatively effective in preventing mammary cancer, presumably due to alteration of the development of mammary tissue in the mouse. An anti-progestin agent which inhibits progesterone function was similarly highly effective in mammary cancer prevention, again presumably due to alteration of mammary development. An unexpected result was obtained with tamoxifen, which proved ineffective despite the fact that it reduces risk of breast cancer in humans. This was due to the fact that tamoxifen stimulates breast epithelial cell proliferation in BRCA1-deficient mice [2]. Another chemopreventive agent that has been tested in BRCA1deficient mice is the triterpenoid bardoxolone. This agent, which induces the antioxidant response element and inhibits NF-Kβ, is relatively effective in this model and warrants further examination [28].

138

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

3. Li-Fraumeni syndrome Li-Fraumeni syndrome (LFS), a highly penetrant, autosomaldominant, inherited cancer predisposition syndrome, was first described in 1969 by Li and Fraumeni based on a retrospective analysis of four families with childhood rhabdomyosarcoma [29]. The second most common cancer identified in these families was early age-onset breast cancer (age o30 years). In 1990, using a candidate gene approach, the underlying genetic abnormality in LFS was identified as a germline mutation of TP53 [30]. Though LFS can present with a broad range of cancers, “core” LFS cancers are considered to be soft tissue and bone sarcomas, premenopausal breast cancers, brain tumors (particularly choroid plexus carcinomas), and adrenal cortical carcinomas (ACCs) [31]. These tumors comprise approximately 70% of LFS associated malignancies and tend to occur at significantly younger ages than their sporadic counterparts [31]. The exact prevalence of LFS is unknown. More than 500 LFS families worldwide have been reported thus far [32]. The founder mutation, cDNA: 1010G-A, which encodes the amino acid change p.R337H, has been described in southeast and southern Brazil at an allelic frequency 0.3%, which is higher than any other single TP53 mutation associated with LFS [33–35]. 3.1. Mutation spectrum The TP53 gene is located on the short arm of chromosome 17 (17p13.1) [36]. More than 250 germline mutations have been described throughout the TP53 gene (reviewed in [35]). The majority of germline TP53 mutations known to date are missense mutations found within coding region exons 5 to 8 [32,35,37], which correspond to the DNA binding domain of the p53 protein. Mutations outside the DNA binding domain, as well as rearrangements and deletions have also been described [38]. The previously discussed Brazilian R337H mutation is within the oligomerization domain of the p53 protein.

3.2. Penetrance data Penetrance of inherited TP53 mutations is high, with early estimates being 50% risk of carriers developing cancer by age 40 and 90% by age 60 years [35,39,40]. These numbers have evolved over time, with female carriers having significantly higher cancer risks than male carriers, even after controlling for gender-specific cancers such as breast, ovarian, and prostate cancers (reviewed in [35] and [32]). Subsequent studies have estimated lifetime risks of cancer for male carriers of up to 70% and nearly 100% in female carriers [40]. In the Brazilian population the R337H mutation was initially suggested to be associated exclusively with adrenocortical carcinoma (ACC). However, it has subsequently been shown that the age-related tumor pattern in R337H founder mutation carriers includes cancers compatible with the LFS spectrum, the most common of which is breast cancer [34]. However, cancers in R337H mutation carriers may occur at later ages than in carriers of other TP53 mutations, with approximately 15%–20% penetrance by age 30, and approaching 80%–90% penetrance by age 70 [32,34]. Thus, different germline TP53 mutations may exhibit differences in penetrance. The phenotypic manifestations of breast cancer in LFS patients are also of particular interest in that they differ from those in the sporadic setting. Among sporadic breast cancers up to 25% express high levels of Her2, whereas over two thirds of LFS cancers are positive for Her2 expression, both by fluorescence in situ hybridization (FISH) and immunohistochemistry [41]. In addition, in LFS

a high percentage is positive for ER and PR, with approximately half being positive for both ER and Her2. LFS breast cancers also present a stark contrast to other hereditary breast cancers, such as those associated with germline BRCA1 mutations, where at least one third are triple-negative (ER/PR/Her2-negative) [41].

3.3. Mechanistic data The p53 protein is known for its function as a tumor suppressor protein, the loss of which is strongly associated with increased susceptibility to cancer [42]. Interestingly, recent studies have shown that mutant p53 proteins not only lose tumor-suppressor function, but may also gain novel oncogenic functions with different TP53 hotspot mutations exhibiting unequal gain-of-function (GOF) effects [37,43]. These GOF effects may in turn affect age of onset of certain cancers in LFS [37]. In addition to regulating activities directly related to tumor suppression such as the cell cycle and DNA repair, wild-type p53 regulates mitochondrial respiration and promotes oxidative phosphorylation while decreasing the flux through the glycolytic pathway that is characteristic of many cancers [44]. p53 also promotes the nuclear transcription of genes involved in mitochondrial biogenesis and function [45]. Furthermore, under normal conditions p53 translocates to the mitochondria to maintain mitochondrial genomic integrity [46,47]. Although cancer cells have been thought to rely primarily on glycolysis for energy, increasing evidence indicates that cancer cells also depend on mitochondria for cellular bioenergetics and growth [48,49]. Current thinking is that in order to accommodate their proliferative activities, cancer cells undergo a “reprogramming” of their cellular metabolism that relies on functional mitochondria [49].

3.4. Chemoprevention for LFS Currently, there is no evidence-based approach to successful cancer screening or prevention in patients with LFS. The high penetrance, early age onset of cancer, and the fact that LFS patients are at risk for developing multiple cancers over their lifetime, underscores the need for chemoprevention in this population. Metformin is an oral biguanide approved by the US Food and Drug Administration (FDA) for the treatment of type II diabetes [50,51]. Epidemiologic studies have suggested that metformin is associated with reduced incidence and mortality of diverse cancers in diabetic patients [52], though these results have lately been questioned by recent meta-analyses presented in this issue. Nevertheless, metformin remains of interest as a possible chemopreventive agent in select populations, including those with germline mutations in tumor suppressor genes. The possibility of “synthetic lethality” has been suggested whereby metformin may be cytotoxic only in the context of a genetic defect, such as in tumor suppressor gene TP53 [53,54]. Another mechanism of the potential chemopreventive effect of metformin is its effect on reducing circulating insulin and glucose levels, the direct and indirect activation of AMP-activated protein kinase (AMPK), and the inhibition of mammalian-target-of-rapamycin (mTOR) [55]. However, there is a growing body of evidence pointing to the mitochondria as the primary site of metformin’s action. Specifically, metformin has inhibitory effects on respiratory complex I which may result in decreased oxidative phosphorylation, lower ATP production, and cause an “energy saving” phenotype, all of which may be deleterious for cancer cell proliferation [54]. As stated previously, wild-type p53 promotes oxidative phosphorylation and mitochondrial respiration which is contrary to the bioenergetics of the cancer cell which relies on aerobic glycolysis

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

(Warburg effect) [48] but also requires functional mitochondria for its growth needs. This raises the question of how mutant p53 protein might affect cellular bioenergetics and mitochondrial function and what implications this might have for patients with germline mutations in TP53 who have a high risk of cancer. To help answer this, investigators at the National Heart, Lung, and Blood Institute conducted a pilot study of 20 members of families with LFS with germline TP53 mutations [56]. These mutation carriers served as cases, while the 20 controls consisted of 9 healthy volunteers and 11 members of LFS families without TP53 mutations. Participants engaged in submaximal leg exercise using a foot-pedal apparatus. During use participants experienced a depletion of phosphocreatine levels in the tibialis anterior muscle. Phosphocreatine normally shuttles high-energy phosphate from the mitochondria to the cytosol in order to maintain ATP levels in skeletal muscle during physical activity. The measurement of phosphocreatine regeneration after exercise-induced depletion was done via phosphorus-31 magnetic resonance spectroscopy (31P-MRS), which provides a sensitive gauge of in vivo oxidative phosphorylation capacity of skeletal leg muscle. In patients with germline TP53 mutations, the recovery time for phosphocreatine levels in skeletal muscle after exercise was significantly shorter than in controls (28.7 seconds v 36.7 seconds, respectively, P ¼ .01), suggesting an increased capacity for oxidative phosphorylation in mutation carriers. Further, these investigators also conducted additional studies of mitochondrial function in cultured lymphocytes and myoblasts from members of families with LFS. As compared with noncarriers (wild-type TP53), carriers of TP53 mutations showed increased oxygen consumption. Two key regulators of mitochondrial biogenesis, transcription factor A (TFAM) and cytochrome c oxidase (SCO2), were also observed to be present at higher levels in myoblasts of mutation carriers compared to noncarriers from LFS families. Of note, these in vivo results were supported by animal studies conducted by these investigators in an established mouse model of LFS (see next section: Animal models of LFS). The role that this metabolic “gain-of-function” plays in tumorigenesis in LFS patients remains unclear. This study was small, but the suggestion of increased oxidative phosphorylation in LFS patients, with their high-risk of cancer suggests a role for enhanced mitochondrial function in LFS tumorigenesis. Thus, if metformin decreases oxidative phosphorylation via inhibition of complex I of the mitochondrial respiratory chain, as is suspected, testing of this agent for chemoprevention in LFS patients offers a reasonable, mechanistically rational intervention. These mechanisms are being explored in an ongoing trial, NCT01981525 (Table 3), led by Walcott and Fojo [57], which is being conducted by a National Cancer Institute team in collaboration with National Heart, Lung, and Blood Institute investigators. In this single-arm study of patients with deleterious germline mutations in TP53, the primary endpoints are safety and tolerability of up to 2000 mg/day of metformin. Secondary endpoints include the evaluation of the effect of metformin on circulating levels of insulin, glucose, and insulin-like growth factor (IGF)-1. In addition, evaluation of the effects of metformin on mitochondrial function, utilizing 31P-MRS technology to measure phosphocreatine recovery time as a biomarker of oxidative phosphorylation in the skeletal muscle of these patients, is being performed. The goal is to determine if there is a quantifiable effect of metformin that can serve as a marker of metformin efficacy in larger phase II studies. These types of translational studies may shed light on the mechanism of metformin’s activity in vivo, potentially supporting its use as a chemopreventive agent and establishing biomarkers of metformin efficacy in patients with germline TP53 mutations. At this time, this is the only chemoprevention trial being conducted in LFS patients that can be found on clinicaltrials.gov (Table 3).

139

3.5. Animal models of LFS LFS, with its germline mutation of TP53, is a human syndrome that has been modeled in the mouse. The fact that such germline mutation results in a wide variety of tumors is consistent with TP53 being the most commonly mutated gene in a variety of sporadic human cancers including lung, bladder, esophagus, head and neck. As discussed previously, commonly observed cancers in humans with LFS include breast cancer, sarcomas, blood related cancers, lung cancer, and others. The majority of animal models of TP53 were engineered by knocking out the function of the TP53 gene [58]. The tumor spectra observed in TP53 knockout mice reflects the background of the mice, with lymphomas and sarcomas predominating on a C57Bl6 background. Prevention studies in these animals have shown that caloric restriction increases the latency of lymphomas but has minimal effects on sarcoma formation. As stated above, Neu overexpression/amplification is observed in the preponderance of LFS individuals with breast cancer. Mice overexpressing Neu (MMTV-Neu) provide a logical model for preclinical testing of chemopreventive agents in general. Furthermore, MMTV-Neu mice with an altered TP53 offer a close parallel to humans with LFS. Brown and co-workers showed that both RXR agonists and EGFR inhibitors were highly effective in MMTV-Neu transgenic mice. In combination with anti-estrogens such as tamoxifen, RXR agonists and EGFR inhibitors have been shown to prevent both ER-positive and ER-negative breast cancers in animal models [59]. These results have been confirmed in MMTVNeu mice with heterozygous knockout of TP53 (Lubet, Grubbs unpublished data), further suggesting that the RXR agonists and EGFR inhibitors may be useful in breast cancer prevention. Finally, a multivalent multipeptide (Neu, IGFBP2, IGF1R) vaccine was recently reported to be effective in cancer prevention in an MMTV model [60]. The potential use of such vaccines in various familial syndromes is promising. In studies with transgenic mice expressing a dominantnegative TP53 mutation on an A/j background, You, Lubet and colleagues treated animals with organ-specific carcinogens to induce specific types of tumors [61]. The most important findings included: (1) Mice with a dominant-negative TP53 mutation tended to develop more and generally more aggressive tumors. (2) The targeted organs in these mice responded to organ-specific agents as predicted. Thus, glucocorticoids and RXR agonists were effective in inhibiting lung tumor formation in wild type mice and mice with a dominant negative TP53. In another example, NSAIDs were effective in preventing colon cancer, but not lung cancer or sarcomas, in TP53 mutant mice. These observations lend support to the notion that organ-specific tumors in the context of germline TP53 mutations, mimicking an inherited syndrome, can be prevented by agents known to be effective in the same organ in the sporadic setting. This comparison of agent efficacy in germline versus sporadic tumors of a given tissue type does not address a different question, namely whether tumors in the various organs driven by a single germline mutation (as in LFS) can be prevented by the same agent. Attempts have been made to establish a mouse model of LFS that recapitulates the types of mutations and the tumor spectrum seen in human LFS [62,63]. As stated above, TP53 knockout mice are highly cancer prone but the tumor spectrum of these mice tends to be restricted to sarcomas and lymphomas. Carcinomas in these mice are rare, particularly in TP53-null mice. Mouse models that carry gain-of-function mutations in TP53 (LFS models) more closely parallel human LFS which is characterized by gain-offunction mutations. For example, the mouse TP53 R172H and TP53 R270H knock-in mutations are homologous to the human hotspot R175H and R273H mutations, respectively. These are

140

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

missense mutations with an arginine (R) to histidine (H) change at codons 172 and 270 (human codons 175 and 273, respectively), and have been shown to result in a tumor spectrum similar to that observed in these patients [62,63]. Unlike TP53 knockout models, the LFS mouse models also develop carcinomas, and the osteosarcomas and carcinomas in these models also show a high frequency of metastasis. Along with the previously mentioned pilot study of LFS patients [56], the investigators supplemented their human studies with studies in the TP53 R172H mouse model to evaluate mitochondrial function. They found that mitochondrial oxygen consumption and levels of respiratory-complex proteins in skeletal muscle were increased in the mice with TP53 R172H mutation as compared with wild-type mice. Further, mRNA levels of the mitochondrial biogenesis genes, TFAM and SCO2, were also increased in cells from mice with LFS. In sum, these findings are consistent with in vivo results in humans, supporting the conclusion that increased oxidative phosphorylation capacity exists in LFS patients.

Furthermore, the severity of polyposis and age of onset are related to different mutations in the APC gene [76,77]. Approximately 10%–15% of patients with FAP present with a de novo mutation, although the majority of de novo cases confer no risk of inheritance to siblings. Approximately 11%–20% will exhibit somatic mosaicism of their APC mutation, which is associated with risk of a less severe form of FAP [78]. Biallelic germline mutations in the MYH gene, a base-excision-repair gene, have also been found to be associated with FAP, particularly in FAP patients who test negative for APC mutations [79–81]. Extracolonic manifestations have been attributed to different genotypic mutations within APC; however, such mutations do not always predict the phenotype [82,83]. Extracolonic manifestations include desmoid tumors, periampullary neoplasms, osteomas, odontomas, supernumary teeth, hepatoblastomas, thyroid tumors, and congenital hypertrophy of the retinal pigmented epithelium (CHRPE). 4.2. Mechanistic data

4. Familial adenomatous polyposis Familial adenomatous polyposis (FAP) is the second most common inherited colon cancer syndrome. FAP is diagnosed in individuals with more than 100 adenomatous colorectal polyps, which typically develop around a median age of 16 years [64]. Symptoms may arise in the third decade and the median age for the development of colonic cancer is 35–40 years. Progression to colorectal cancer occurs with nearly complete penetrance by age 40–50 years; however, malignancy in childhood does occur [65]. Patients with FAP may also develop upper gastrointestinal polyps, although these less commonly progress to malignancy as compared to colon and rectal adenomas [66–68]. Sites other than the gastrointestinal tract are at increased risk of tumors and malignancy including: thyroid, liver, brain, bone, adrenal gland, retinal pigment, and subcutaneous tissues. FAP most commonly results from germline mutations in the APC tumor suppressor gene on chromosome 5q21. APC mutations are transmitted in an autosomal dominant fashion. Despite the nearly 100% penetrance seen in individuals with APC mutations, FAP accounts for only approximately 1% of all colon cancers [69]. The FAP incidence ranges between 1 in 8,000 and 30,000 live births, affecting both sexes equally and with a relatively equal worldwide distribution [70]. 4.1. Penetrance data Because the probability of developing cancer is almost 100%, complete colectomy is recommended in cases of classic FAP, if acceptable to the patient, after the onset of puberty; polyp formation and cancer in childhood is very unusual. In one example involving a specific mutation at codon 1309 in exon 15 of APC, onset of symptoms and colon cancer occur about 10 years earlier, and a more severe phenotype is sometimes observed [71]. Classic FAP is associated with mutations extending throughout the APC gene to the 5´ end of exon 15. On the other hand, an attenuated form of the disease, characterized by later onset and a smaller number of polyps, is caused by mutations on the extreme 5´ end or the 3´ end of the APC gene [72] or in the alternatively spliced region of exon 9 [73–75]. APC codes for a protein that inhibits cell proliferation in part by controlling the activity of the oncoprotein, B-catenin. FAP can present as a number of different syndrome variants all of which have a germline mutation in APC. These phenotypic variants include attenuated familial adenomatous polyposis (AFAP), mild or severe polyposis, Gardner’s syndrome, and Turcot’s syndrome.

Somatic mutations in the APC gene are an early event in colorectal tumorigenesis, and can be detected in the majority of sporadic colorectal tumors. APC consists of 8,535 base pairs spanning 21 exons (16 translated exons) and encoding a 2,861– amino acid protein that is expressed in specific epithelial and mesenchymal cells of several fetal and adult human tissues. APC is a large multidomain protein with a molecular mass of 300 kD. The best known function of the APC protein is as part of the Wnt-1 signaling pathway. Binding of Wnt to its receptor, frizzled, leads to the inactivation of glycogen synthase kinase 3β in a cytoplasmic complex with APC, β-catenin, axin, and components of the ubiquitin ligation machinery. This leads to a decrease in βcatenin phosphorylation and inhibits its proteasomal degradation. As a consequence, increased β-catenin is available to bind transcription factors leading to the activation of proliferative genes. In addition, APC has recently been shown to be a multifunctional protein that can affect a variety of fundamental cellular processes, in particular cytoskeletal regulation and chromosomal stability. Improved understanding of both the genetics and biology of APC may in time culminate in preventive or therapeutic strategies specifically targeted at reducing the burden of colorectal cancer. 4.3. Chemoprevention of FAP FAP is an ideal high-risk group in which to test chemopreventive interventions because patients develop innumerable adenomatous polyps. NSAIDs, such as sulindac and COX-2 inhibitors, have historically been shown to be effective at reducing polyp burden in these patients [84]. In a double-blind, placebo-controlled trial conducted by Steinbach et al, 77 FAP patients (25 with intact colons) were randomized to celecoxib at 100 mg or 400 mg twice daily or placebo for 6 months [85]. Endoscopic evaluation performed at baseline and at 6 months showed a 28% reduction in the mean number of polyps in the high-dose celecoxib group compared to placebo. These results were so dramatic that in 1999 celecoxib was given preliminary FDA approval for the indication of cancer prevention in patients with FAP. A larger trial, the Adenoma Prevention with Celecoxib trial, was initiated in 2000 to study the effects of celecoxib in patients with sporadic adenomatous polyps. This study was halted early because of increases in adverse cardiovascular events in patients who were taking celecoxib [84]. Celecoxib still remains on the United States (US) market, though its manufacturer, Pfizer, asked the FDA to withdraw the FAP indication from its label in 2011. Rofecoxib, another COX-2 inhibitor, was withdrawn from the US market in 2004 due to increased reports of

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

cardiovascular toxicity, and trials listed on clinicaltrials.gov using rofecoxib, have been terminated (Table 3). Mechanisms that have been proposed for the effects of COX-2 inhibitors on polyp regression include increased apoptosis, decreased cellular proliferation, and modulation of prostaglandin E2 levels (PGE2). NSAIDs may reverse suppression of apoptosis resulting from loss of APC function [86]. Using the study population from the Steinbach group, another study was conducted to identify the underlying mechanism of COX-2 inhibitors on polyp regression in these patients. Biopsy samples of normal appearing mucosa and adenomatous polyps were analyzed for PGE2, apoptotic and proliferation indices (AI and Ki-67 respectively), and correlation between these modulation of these indices and polyp regression from baseline to 6 months was assessed. No differences were found in the PGE2 levels between normal and adenomatous polyp tissues. However, decreases in Ki-67 and increases in AI in adenomas were significant from baseline to 6 months. Further, these changes corresponded with clinical outcomes, ie, polyp regression. A non-significant difference between the higher dose and placebo groups (P ¼ .077) was found [87]. Other chemoprevention studies in the FAP population have tested combinations of NSAIDs with other agents, in order to allow the use of lower NSAID doses with a goal of decreasing toxicity (reviewed in [84]). Such studies (either in animals or humans) have used combinations of NSAIDs or COX-2 inhibitors with difluoromethylornithine (DFMO), an enzyme-activated irreversible inhibitor of ornithine decarboxylase (ODC), ursodeoxycholic acid (ursodiol), and statins. Curcumin and omega-3-fatty acids also hold potential as chemopreventive agents because of their anti-inflammatory properties. Several randomized, controlled trials of these agents are ongoing [88]. 4.4. Animal models of FAP Mutation of the APC gene, as seen in the Min (multiple intestinal neoplasia) mouse/APCmin was induced by random mutagenesis and is associated with hundreds of adenomatous polyps. APC1638N, with a truncating mutation at codon 1638, has a lower polyp burden. These mouse models recapitulate the range of phenotypes seen in human FAP, eg, the development of scores to hundreds of adenomas, which is associated with loss of the second allele in lesions [89]. These models also hold appeal because mutations of the APC gene are associated with most sporadic colorectal adenomas and cancers in humans. Despite the fact that mutation/loss of a single gene causes multiple adenomas, the model only imperfectly resembles the human phenotype in that adenomas/cancers arise overwhelmingly in the small intestine and not in the colon as in humans. A possible explanation is that in contrast to humans, mice have substantial bacterial content in their small intestines. Nevertheless, adenoma formation in the small intestine has proven to be susceptible to inhibition by the same agents/classes of agents that are effective in humans in both the FAP and sporadic settings [90]. Thus, it has long been known that Min lesions can be inhibited by exposure to nonspecific NSAIDs, eg, sulindac and piroxicam. In fact, sulindac is routinely administered to persons with FAP. The Min mice have proven sensitive to celecoxib [91], which supported the testing of celecoxib in FAP and the subsequent approval of this agent by the FDA for this purpose. As discussed previously, use of celcoxib in FAP has been limited due in part to potential cardiovascular risks of COX-2 inhibitors, which led to removal of the FAP adenoma prevention indication by the manufacturer of celecoxib. The combination of an NSAID plus DFMO in Min mice proved to be highly effective. This supported the successful clinical trial that showed sulindac in combination with DFMO to confer a significant reduction in adenoma recurrence in the sporadic setting [92]. In addition, the

141

EGFR inhibitor gefitinib is relatively effective in the Min model (R.A. Lubet, M. Clapper, unpublished). This gains importance when viewed in the context of a recently completed trial demonstrating efficacy of combined erlotinib and sulindac in individuals with FAP [93]. A wide variety of agents have proven ineffective in the model as well. The supportive results bring up the question of whether a model in which tumors arise in a different site (small versus large intestine) is applicable to human disease. Finally, we have recently found that vaccines against a number of antigens, including COX-2, inhibit adenoma formation in this model (W.F. Broussard, R.A. Lubet, M.L. Disis, unpublished). One of the major morbidities associated with FAP is the development of desmoid lesions, benign but somewhat aggressive fibroblast growths that can require repeated surgeries and can be life threatening. Although these lesions do not routinely develop in mice, insertion of a knockout of TP53 on the Min background enables mice to develop multiple desmoids [94]. Thus, the Min mouse model can partially recapitulate the human model, which it was meant to mimic.

5. Lynch syndrome Lynch syndrome (LS), previously referred to as hereditary nonpolyposis colon cancer (HNPCC) [95], is an autosomal dominantly inherited disorder of cancer susceptibility caused by germline mutations in one of the DNA mismatch repair (MMR) genes. LS was first described by Aldred Warthin in 1913 [96]. In 1966, Henry Lynch reported two large families with hereditary colorectal cancer from the midwest [97]. In the 1990s the underlying gene defects were discovered, mutation of one of the MMR genes: MLH1 (chromosome 3p21.3) [98,99], MSH2 (chromosome 2p22-21) [100], MSH6 (chromosome 2p16) [101,102], and PMS2 (chromosome 7p22.2) [103,104]. Recently two groups reported that a constitutional 3´ end deletion of the gene EPCAM (epithelial cell adhesion molecule) (chromosome 2p21), which is immediately upstream of the MSH2 gene, may cause LS through epigenetic silencing of the nearby MSH2 gene [105,106]. These mutations cause 2%–5% of all colorectal cancers and 10%– 15% of all colorectal cancers diagnosed before age 50 [107,108]. The lifetime risk for developing a colorectal cancer in mutation carriers ranges between 10% and 69%, depending on the gene mutation [109,110]. MSH2/MLH1 mutation-associated colorectal cancers account for about 50% of the excess colorectal cancers observed in first-degree relatives of a familial colorectal cancer case [111]. Estimated carrier frequency of germline mutations in these genes in the population range from about 1 in 300 to 1 in 3,000 [110,112–116]. 5.1. Penetrance data Nearly 90% of the mutations in MMR genes in LS are located in MLH1 and MSH2 and approximately 10% in MSH6 and PMS2 [117]. In addition to cancers of the colorectum, mutation carriers are at substantially increased risk of cancers of the endometrium (27%– 71% lifetime risk [LTR]), stomach (2%–13% LTR), ovary (3%–13% LTR), urothelial (1%–12% LTR), brain (1%–4% LTR), small bowel (4%– 7% LTR), and hepatobiliary tract (2% LTR). Sebaceous gland adenomas and keratoacanthomas are increased in the Muir-Torre variant of LS, which has a phenotype involving multiple cutaneous neoplasms. Individuals with Muir-Torre syndrome have mutations in MSH2 more frequently than MHL1. The cancer risks vary with the specific MMR gene that is mutated within the LS family [118–120]. The diagnosis of these cancers occurs at younger ages than that of the general population [121].

142

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

In 2011 Bonadona et al [122] were able to estimate the cumulative cancer risk in individuals with LS when studying the MLH1, MSH2 and MSH6 genes (see Table 2). The estimated cumulative risks by age 40 years did not exceed 2% (95% confidence interval [CI], 0%–7%) for endometrial cancer nor 1% (95% CI, 0%–3%) for ovarian cancer, irrespective of the gene that was mutated. The estimated lifetime risks for other tumor types did not exceed 3% with any of the gene mutations. Based on these population’s studies, MSH6 mutations are associated with markedly lower cancer risks than MLH1 or MSH2 mutations. Lifetime ovarian and endometrial cancer risks associated with MLH1 or MSH2 mutations were high but did not increase appreciably until after the age of 40 years [122]. In addition, Kempers et al [123] compared the cancer risk between 194 carriers of an EPCAM deletion and 473 carriers of a mutation in MLH1, MSH2, MSH6, or a combined EPCAM-MSH2 deletion. The risk of developing colorectal cancer for EPCAM deletion carriers was similar (70% by age 70) to the risks in carriers of an MLH1 or MSH2 mutation or a combined EPCAM-MSH2 deletion. But it was higher than the risk in MSH6 mutation carriers. In contrast, the risk of endometrial cancer was 12% by age 70 for EPCAM mutation carriers, lower than the risk in carriers of an MSH1 or MSH6 mutation or combined EPCAM-MSH2 deletion. The endometrial cancer risk in EPCAM deletions was lower than the risk in MSH1 carriers. 5.2. Mechanistic data MMR genes encode proteins whose function is to recognize and correct errors during DNA replication. Thus when mutated, they cause errors in the number of repetitive sequences replicated, known as microsatellite instability (MSI). MSI occurs in about 90%– 95% of cancers that develop in LS, and about 15% of sporadic colorectal cancer cases, in which there is somatic methylation of some of the MMR genes [69,124]. BRAF gene mutations are associated with MLH1 promoter methylation in patients who have sporadic MSI tumors with reduced MLH1 expression. Loss of MSH2 protein can occur via at least two mechanisms. In some cases MSH2 loss is associated with an MSH2 gene mutation, while in other cases MSH2 loss results from an EPCAM gene mutation, which causes MSH2 silencing by eliminating EPCAM transcription termination, leading to MSH2 promoter methylation [106]. 5.3. Chemoprevention of LS While a fair number of chemoprevention studies have been conducted for FAP, substantially fewer chemoprevention studies exist for LS. Part of the reason for this may be that FAP is distinguished by polyps which can be used as biomarkers for cancer development, while LS is not typically characterized by polyposis and, at this time, lacks a comparable definitive biomarker. Current chemoprevention trials for LS are listed in Table 3. The largest LS chemoprevention trial to date is the CAPP2 trial. The CAPP program (Colorectal Adenoma Carcinoma Prevention Programme) was launched in 1990. CAPP2 was pivotal because it was the first large-scale double-blind, placebo-controlled aspirin chemoprevention trial in this cohort which had colorectal cancer as its primary endpoint. This trial of 1,000 LS participants, most of whom were carriers of MMR gene variants, was designed with a 10 year follow-up period [125]. Initial analyses of the data at 4 years showed no effect of aspirin on the development of largebowel neoplasia (adenomas and colorectal cancers were pooled for analysis) [126]. However, when data were analyzed 9 years after randomization, there were 50% fewer cancers in the aspirin treated group versus the placebo group. This effect was most pronounced in individuals who took aspirin for over 2 years. This “delayed

cancer chemopreventive effect” only became apparent 3–4 years after beginning aspirin intervention. While the incidence of colon cancer was reduced, there was no apparent effect of aspirin on the development of colorectal polyps. Individuals taking aspirin still developed polyps, but did not go on to develop cancer from their polyps. There was also a trend (P ¼ .07) of aspirin reducing risk of LS associated cancers other than colorectal cancer. Despite this evidence, no standard recommendation exists for aspirin chemoprevention in LS. More long-term data are needed on adverse events in aspirin treatment groups as well as the best effective dose of aspirin associated with the least adverse events. The ongoing CAPP3 study aims to recruit 1,000 LS mutation carriers and test three doses of aspirin (100, 300, or 600 mg daily), with a focus on new tumors and adverse event rate. The mechanism underlying the effect of aspirin on colorectal cancer development in LS is unclear. Aspirin may be pro-apoptotic at early stages of colorectal cancer development, perhaps even preceding adenoma formation [125]. Others have reported reduced microsatellite instability and increased apoptosis in MMR-deficient cells exposed to aspirin, suggesting that aspirin might induce genetic selection for microsatellite stability in a subset of MMR-deficient cells [127,128]. The fact that aspirin had modest effects on individuals with FAP compared to those with LS, and the apparent lack of effect of aspirin on polyp formation in LS patients underscores critical differences in the pathogenesis of these syndromes. Further, endometrial cancer is frequently the first cancer to appear in women with LS [129] and less progress has been made in screening, detection, and prevention in this area. Women with LS may develop type I (sensitive to estrogen stimulation) or type II (not hormonally driven) endometrial cancer. Nevertheless, there is some interest in progestins as a possible chemopreventive agent for women with LS because of the antagonistic effects of progestins on the proliferative effects of estrogen on the endometrium [130]. Whether aspirin could prevent endometrial cancer in patients with LS is unknown as these cancers have not been analyzed separately. 5.4. Animal models of LS Tumors derived from individuals with LS as well as roughly 15% of sporadic colon cancers demonstrate MSI, although in the latter case MSI is driven by inactivation of the MSH or MLH genes by methylation rather than mutation. Among both LS and sporadic MSI tumors, roughly 15% have APC mutations and another 15%– 20% have mutations in beta-catenin. In contrast, roughly 90% of MSI tumors have mutations in TGFβRII and BAX, whereas tumors from FAP or sporadic non-MSI tumors infrequently (o 5%) have mutations in either of these genes. This disparity in the types of mutations in the different tumor settings reflects a comparable difference in the underlying mechanisms driving these colon cancers. Transgenic mice with heterozygous knockout of the MLH or MSH genes failed to show a cancer phenotype [89]. When the knockout was made homozygous, lymphomas developed in addition to intestinal tumors (primarily small intestine). The intestinal tumors showed MSI as expected. Disappointingly, they did not show mutations in TGFβRII or BAX but routinely exhibited mutations in APC. The lack of mutations in TGFβRII or BAX is presumably because the mouse homologues of these genes do not contain tandem nucleotide repeats in their coding regions, as is found in humans. Thus, this alteration in the germline yields mouse lesions that are very different from the human tumors. Interestingly, the intestinal lesions in MLH knockout mice have proved sensitive to the preventive NSAIDs, including aspirin, NO-aspirin and naproxen [131]. This result parallels the observation that extended dosing with aspirin at doses Z 325 mg/day reduced the incidence of colon

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

cancer by 4 50% in individuals with LS [132]. These results are encouraging from the point of view that NSAIDs may be effective chemopreventive agents in typical APC-mediated tumors/adenomas as well as lesions with an MSI phenotype. A recent study with a tissue specific MLH mouse model showed that these tumors are sensitive to treatment with a PI3K inhibitor, a finding that was predicted based on human data showing that MSI tumors have an activated PI3K pathway [133]. Whether tyrosine kinase (TK) inhibitors will prove applicable in a prevention setting is a key clinical question related to anticipated toxicity from these agents.

143

chemoprevention agents should provide insight into preventive interventions for similar cancers in the sporadic setting. With concerted efforts between animal studies and clinical human studies, the ability of chemoprevention to realize its potential is becoming more of a reality.

Conflicts of interest None. References

6. Discussion The concept of designing chemoprevention studies for use in genetically predisposed populations is not new. Inherited mutations have the unique potential to be used as either molecular biomarkers of cancer risk or targets for chemopreventive agents. The penetrance of these genes and the young age-of-onset of cancer in these groups reflect a shorter duration of malignant transformation than in the sporadic setting. This allows for chemoprevention trials that have smaller cohorts followed for shorter periods of time than would be possible for sporadic tumors of the same tissue type [134]. Trials focused on high-risk, genetically predisposed populations are also more cost-effective [135,136]. Another advantage to using genetically based high risk cohorts with a high penetrance of cancer in chemoprevention trials is the increased likelihood of achieving a clinical endpoint and detecting biomarker modulation in response to the intervention. In chemoprevention studies in lower risk populations, responders are much more difficult to tease apart from those who are unresponsive. Responders in this setting may or may not have developed cancer whether they had the intervention of interest or not [137]. To address this limitation, the idea of identifying “molecular risk signatures” as targets for intervention has been proposed (reviewed in [137]). These signatures ideally correspond with subsequent occurrence of invasive cancer and may help to identify individuals who would benefit from intervention versus those who would not. An example is the “p53 signature” which may serve as a precursor to tubal epithelial carcinoma in women with BRCA mutations, identifying abnormal clonal expansion even in tissue that appears to be benign [138]. The challenge of using precursor lesions is the inability to easily monitor these lesions for progression to cancer or response to intervention. Therefore, it is important to develop animal models in which these lesions are recapitulated and to demonstrate the response of these lesions to the drug of interest. Although we have examined transgenic models for only 4 inherited cancer syndromes representing different disease sites, these are illustrative of the promise and limitations of transgenic models for other cancer syndromes. A broader view is presented in a relatively recent review by Jahid and Lipkin [89]. In general, the primary focus of animal studies has been to examine the biology of the model as well as to examine its potential for testing therapeutic agents. However, the same models should prove applicable for testing potential preventive strategies.

7. Conclusion Syndromes with highly penetrant germline mutations have contributed information about agents and their mechanisms that may be applicable to the general population. With animal models that can inform human applications and a focus on inherited predisposition syndromes, these model systems for testing

[1] Sporn MB. Approaches to prevention of epithelial cancer during the preneoplastic period. Cancer Res 1976;36:2699–702. [2] Diaz-Cruz ES, et al. BRCA1 deficient mouse models to study pathogenesis and therapy of triple negative breast cancer. Breast Dis 2010;32(1-2):85–97. [3] Lindor NM, McMaster ML, Lindor CJ, Greene MH. Concise handbook of familial cancer susceptibility syndromes—second edition. J Natl Cancer Inst Mongr 2008;38:1–93. [4] Claus EB, Risch N, Thompson WD. The calculation of breast cancer risk for women with a first-degree family history of ovarian cancer. Breast Cancer Res Treat 1993;28:115–20. [5] Newman B, et al. Inheritance of human breast cancer: evidence for autosomal dominant transmission in high-risk families. Proc Natl Acad Sci U S A 1988;85(9):3044–8. [6] Petrucelli N, Daly MB, Feldman GL. Hereditary breast and ovarian cancer due to mutations in BRCA1 and BRCA2. Genet Med 2010;12(5):245–59. [7] Struewing JP, Abeliovich D, Peretz T, et al. The carrier frequency of the BRCA1 185delAG mutation is approximately 1 percent in Ashkenazi Jewish individuals. Nat Genet 1995;11(2):198–200. [8] Struewing JP, et al. The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J Med 1997;336 (20):1401–8. [9] Thorlacius S, et al. A single BRCA2 mutation in male and female breast cancer families from Iceland with varied cancer phenotypes. Nat Genet 1996;13:117–9. [10] Miki Y, et al. A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 1994;266(5182):66–71. [11] Tavtigian SV, et al. The complete BRCA2 gene and mutations in chromosome 13q-linked kindreds. Nat Genet 1996;12(3):333–7. [12] National Human Genome Research Institute. Breast Cancer Information Core. 2014 5/16/2014 [cited 2014 6/27/2014]; Available from: http://research.nhgri. nih.gov/bic/. [13] Wooster R, et al. Identification of the breast cancer susceptibility gene BRCA2. Nature 1995;378:789–92. [14] Easton DF. Breast Cancer Linkage Consortium. Cancer risks in BRCA2 mutation carriers. J Natl Cancer Inst 1999;91(15):1310–6. [15] Couch FJ, Nathanson KL, Offit K. Two decades after BRCA: setting paradigms in personalized cancer care and prevention. Science 2014;343(6178):1466–70. [16] Venkitaraman AR. Cancer suppression by the chromosome custodians, BRCA1 and BRCA2. Science 2014;343(6178):1470–5. [17] Venkitaraman AR. Cancer susceptibility and the functions of BRCA1 and BRCA2. Cell 2002;108(2):171–82. [18] Atchley DP, et al. Clinical and pathologic characteristics of patients with BRCApositive and BRCA-negative breast cancer. J Clin Oncol 2008;26(26):4282–8. [19] Mavaddat N, et al. Pathology of breast and ovarian cancers among BRCA1 and BRCA2 mutation carriers: results from the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA). Cancer Epidemiol Biomarkers Prev 2012;21 (1):134–47. [20] Phillips KA, et al. Tamoxifen and risk of contralateral breast cancer for BRCA1 and BRCA2 mutation carriers. J Clin Oncol 2013;31(25):3091–9. [21] Weil MK, Chen AP. PARP inhibitor treatment in ovarian and breast cancer. Curr Probl Cancer 2011;35(1):7–50. [22] Sonnenblick A, et al. An update on PARP inhibitors-moving to the adjuvant setting. Nat Rev Clin Oncol 2015;12:27–41. [23] Fong PC, et al. Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA mutation carriers. N Engl J Med 2009;361(2):123–34. [24] Xu X, et al. Conditional mutation of Brca1 in mammary epithelial cells results in blunted ductal morphogenesis and tumour formation. Nat Genet 1999;22 (1):37–43. [25] Rottenberg S, et al. High sensitivity of BRCA1-deficient mammary tumors to the PARP inhibitor AZD2281 alone and in combination with platinum drugs. Proc Natl Acad Sci U S A 2008;105(44):17079–84. [26] To C, et al. The PARP inhibitors, veliparib and olaparib, are effective chemopreventive agents for delaying mammary tumor development in BRCA1deficient mice. Cancer Prev Res (Phila) 2014;7(7):698–707. [27] Burga LN, et al. Loss of BRCA1 leads to an increase in epidermal growth factor receptor expression in mammary epithelial cells, and epidermal growth factor receptor inhibition prevents estrogen receptor-negative cancers in BRCA1-mutant mice. Breast Cancer Res 2011;13(2):R30. [28] Kim EH, et al. CDDO-methyl ester delays breast cancer development in BRCA1-mutated mice. Cancer Prev Res (Phila) 2012;5(1):89–97.

144

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

[29] Li FP, Fraumeni JF Jr. Soft-tissue sarcomas, breast cancer, and other neoplasms. A familial syndrome? Ann Intern Med 1969;71(4):747–52. [30] Malkin D, et al. Germline mutations of the p53 tumor-suppressor gene in children and young adults with second malignant neoplasms. N Engl J Med 1992;326(20):1309–15. [31] Gonzalez KD, et al. Beyond Li Fraumeni syndrome: clinical characteristics of families with p53 germline mutations. J Clin Oncol 2009;27(8):1250–6. [32] Malkin D. Li-fraumeni syndrome. Genes Cancer 2011;2(4):475–84. [33] Achatz MI, Hainaut P, Ashton-Prolla P. Highly prevalent TP53 mutation predisposing to many cancers in the Brazilian population: a case for newborn screening? Lancet Oncol 2009;10(9):920–5. [34] Garritano S, et al. Detailed haplotype analysis at the TP53 locus in p.R337H mutation carriers in the population of Southern Brazil: evidence for a founder effect. Hum Mutat 2010;31(2):143–50. [35] Kamihara J, Rana HQ, Garber JE. Germline TP53 mutations and the changing landscape of Li-Fraumeni syndrome. Hum Mutat 2014;35(6):654–62. [36] Gardner CR, et al. Synthesis of retinoid enhancers based on 2aminobenzothiazoles for anti-cancer therapy. Bioorg Med Chem 2012;20 (23): 6877–84. [37] Xu J, et al. Heterogeneity of Li-Fraumeni syndrome links to unequal gain-offunction effects of p53 mutations. Sci Rep 2014;4:4223. [38] Bougeard G, et al. Molecular basis of the Li-Fraumeni syndrome: an update from the French LFS families. J Med Genet 2008;45(8):535–8. [39] Kumar N. Author’s response. J Cytol 2012;29(3):220. [40] McBride KA, et al. Li-Fraumeni syndrome: cancer risk assessment and clinical management. Nat Rev Clin Oncol 2014;11(5):260–71. [41] Masciari S, et al. Breast cancer phenotype in women with TP53 germline mutations: a Li-Fraumeni syndrome consortium effort. Breast Cancer Res Treat 2012;133(3):1125–30. [42] Vousden KH, Lane DP. p53 in health and disease. Nat Rev Mol Cell Biol 2007;8(4):275–83. [43] Leroy B, et al. The TP53 website: an integrative resource centre for the TP53 mutation database and TP53 mutant analysis. Nucleic Acids Res 2013;41: D962–D969. [44] Matoba S, et al. p53 regulates mitochondrial respiration. Science 2006;312 (5780):1650–3. [45] Gottlieb E, Vousden KH. p53 regulation of metabolic pathways. Cold Spring Harb Perspect Biol 2010;2(4):a001040. [46] Lago CU, et al. p53, aerobic metabolism, and cancer. Antioxid Redox Signal 2011;15(6):1739–48. [47] Zhuang J, et al. Mitochondrial disulfide relay mediates translocation of p53 and partitions its subcellular activity. Proc Natl Acad Sci U S A 2013;110(43): 17356–61. [48] Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 2009;324 (5930):1029–33. [49] Ward PS, Thompson CB. Metabolic reprogramming: a cancer hallmark even Warburg did not anticipate. Cancer Cell 2012;21(3):297–308. [50] Sahin I, et al. Metformin versus flutamide in the treatment of metabolic consequences of non-obese young women with polycystic ovary syndrome: a randomized prospective study. Gynecol Endocrinol 2004;19(3):115–24. [51] Kumbak B, Kahraman S. Efficacy of metformin supplementation during ovarian stimulation of lean PCOS patients undergoing in vitro fertilization. Acta Obstet Gynecol Scand 2009;88(5):563–8. [52] Decensi A, et al. Metformin and cancer risk in diabetic patients: a systematic review and meta-analysis. Cancer Prev Res (Phila) 2010;3(11):1451–61. [53] Buzzai M, et al. Systemic treatment with the antidiabetic drug metformin selectively impairs p53-deficient tumor cell growth. Cancer Res 2007;67 (14):6745–52. [54] Pollak MN. Investigating metformin for cancer prevention and treatment: the end of the beginning. Cancer Discov 2012;2(9):778–90. [55] Viollet B, et al. Cellular and molecular mechanisms of metformin: an overview. Clin Sci (Lond) 2012;122(6):253–70. [56] Wang PY, et al. Increased oxidative metabolism in the Li-Fraumeni syndrome. N Engl J Med 2013;368(11):1027–32. [57] A pilot study of metformin in patients with a diagnosis of Li-Fraumeni syndrome. Cited 2015; Available from: https://clinicaltrials.gov/ct2/show/ NCT01981525?term=Li þFraumeni&recr=Open&rank=4. [58] Donehower LALG. 20 years of studying P53 function in genetically engineered mice. Nat Rev Cancer 2009;9:831–41. [59] Li Y, Brown PH. Prevention of ER-negative breast cancer. Recent Results Cancer Res 2009;181:121–34. [60] Disis ML, et al. A multiantigen vaccine targeting neu, IGFBP-2, and IGF-IR prevents tumor progression in mice with preinvasive breast disease. Cancer Prev Res (Phila) 2013;6(12):1273–82. [61] Lubet R, Wang Y, Zhang Z, You M. Mouse models incorporating alterationsin the major tumor suppressor genes P53 and P16: their use in screening for potential carcinogens, developing further mouse models and screening for chemopreventive and chemotherapeutic agents. Exp Lung Res 2005;31:117–33. [62] Lang GA, et al. Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome. Cell 2004;119(6):861–72. [63] Olive KP, et al. Mutant p53 gain of function in two mouse models of LiFraumeni syndrome. Cell 2004;119(6):847–60. [64] Merg A, et al. Hereditary colorectal cancer—part II. Curr Probl Surg 2005;42 (5):267–333.

[65] Half E, Bercovich D, Rozen P. Familial adenomatous polyposis. Orphanet J Rare Dis 2009;4:22. [66] Attard TM, et al. Multicenter experience with upper gastrointestinal polyps in pediatric patients with familial adenomatous polyposis. Am J Gastroenterol 2004;99(4):681–6. [67] Attard TM, Yardley JH, Cuffari C. Gastric polyps in pediatrics: an 18-year hospital-based analysis. Am J Gastroenterol 2002;97(2):298–301. [68] Debinski HS, et al. Upper intestinal surveillance in familial adenomatous polyposis. Eur J Cancer 1995;31A(7-8):1149–53. [69] Merg A, et al. Hereditary colon cancer—part I. Curr Probl Surg 2005;42(4): 195–256. [70] Gryfe R. Clinical implications of our advancing knowledge of colorectal cancer genetics: inherited syndromes, prognosis, prevention, screening and therapeutics. Surg Clin North Am 2006;86(4):787–817. [71] Aretz S, et al. Frequency and parental origin of de novo APC mutations in familial adenomatous polyposis. Eur J Hum Genet 2004;12(1):52–8. [72] Friedl W, et al. Attenuated familial adenomatous polyposis due to a mutation in the 3’ part of the APC gene. A clue for understanding the function of the APC protein. Hum Genet 1996;97(5):579–84. [73] Knudsen AL, Bisgaard ML, Bulow S. Attenuated familial adenomatous polyposis (AFAP). A review of the literature. Fam Cancer 2003;2(1):43–55. [74] Young J, et al. A family with attenuated familial adenomatous polyposis due to a mutation in the alternatively spliced region of APC exon 9. Hum Mutat 1998;11(6):450–5. [75] van der Luijt RB, et al. APC mutation in the alternatively spliced region of exon 9 associated with late onset familial adenomatous polyposis. Hum Genet 1995;96:705–10. [76] Gayther SA, et al. Regionally clustered APC mutations are associated with a severe phenotype and occur at a high frequency in new mutation cases of adenomatous polyposis coli. Hum Mol Genet 1994;3(1):53–6. [77] Caspari R, et al. Familial adenomatous polyposis: mutation at codon 1309 and early onset of colon cancer. Lancet 1994;343:629–32. [78] Aretz S, et al. High proportion of large genomic deletions and a genotype phenotype update in 80 unrelated families with juvenile polyposis syndrome. J Med Genet 2007;44(11):702–9. [79] Sieber O, et al. Colorectal tumourigenesis in carriers of the APC I1307K variant: lone gunman or conspiracy? J Pathol 2003;199(2):137–9. [80] Lipton L, et al. Carcinogenesis in MYH-associated polyposis follows a distinct genetic pathway. Cancer Res 2003;63(22):7595–9. [81] Al-Tassan N, et al. Inherited variants of MYH associated with somatic G:C-T: A mutations in colorectal tumors. Nat Genet 2002;30(2):227–32. [82] Caspari R, et al. Familial adenomatous polyposis: desmoid tumours and lack of ophthalmic lesions (CHRPE) associated with APC mutations beyond codon 1444. Hum Mol Genet 1995;4(3):337–40. [83] Nieuwenhuis MH, et al. Genotype-phenotype correlations as a guide in the management of familial adenomatous polyposis. Clin Gastroenterol Hepatol 2007;5(3):374–8. [84] Kim B, Giardiello FM. Chemoprevention in familial adenomatous polyposis. Best Pract Res Clin Gastroenterol 2011;25(4-5):607–22. [85] Steinbach G, et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000;342(26):1946–52. [86] Morin PJ, Vogelstein B, Kinzler KW. Apoptosis and APC in colorectal tumorigenesis. Proc Natl Acad Sci U S A 1996;93(15):7950–4. [87] Sinicrope FA, et al. Cell proliferation and apoptotic indices predict adenoma regression in a placebo-controlled trial of celecoxib in familial adenomatous polyposis patients. Cancer Epidemiol Biomarkers Prev 2004;13(6):920–7. [88] Hull MA. Nutritional agents with anti-inflammatory properties in chemoprevention of colorectal neoplasia. Recent Results Cancer Res 2013;191: 143–56. [89] Jahid S, Lipkin S. Mouse models of inherited cancer syndromes. Hematol Oncol Clin North Am 2010;24(6):1205–28. [90] Fischer SM, Hawk ET, Lubet RA. Coxibs and other nonsteroidal antiinflammatory drugs in animal models of cancer chemoprevention. Cancer Prev Res (Phila) 2011;4(11):1728–35. [91] Jacoby RF, et al. The cyclooxygenase-2 inhibitor celecoxib is a potent preventive and therapeutic agent in the min mouse model of adenomatous polyposis. Cancer Res 2000;60(18):5040–4. [92] Meyskens FL Jr, et al. Difluoromethylornithine plus sulindac for the prevention of sporadic colorectal adenomas: a randomized placebo-controlled, double-blind trial. Cancer Prev Res (Phila Pa) 2008;1(1):32–8. [93] A clinical trial of COX and EGFR inhibition in Familial Adenomatous Polyposis patients (FAPEST). 2014 [cited 2014 11/4/2014]; Available from: https:// clinicaltrials.gov/ct2/show/NCT01187901?term=A þclinical þtrial þof þCOX þand þEGFR þinhibition þin þFamilial þAdenomatous þPolyposis þpatien ts&rank=1. [94] Halberg RB, et al. The pleiotropic phenotype of Apc mutations in the mouse: allele specificity and effects of the genetic background. Genetics 2008;180 (1):601–9. [95] Jass JR. Hereditary non-polyposis colorectal cancer: the rise and fall of a confusing term. World J Gastroenterol 2006;12(31):4943–50. [96] Warthin AS. Heredity with reference to carcinoma: as shown by the study of the cases examined in the pathological laboratory of the University of Michigan, 1895-1913. Arch Intern Med 1913;12(5):546–55. [97] Lynch HT, et al. Hereditary factors in cancer: study of two large midwestern kindreds 1966;117(2):206–12.

F.L. Walcott et al. / Seminars in Oncology 43 (2016) 134–145

[98] Bronner CE, et al. Mutation in the DNA mismatch repair gene homologue hMLH1 is associated with hereditary non-polyposis colon cancer. Nature 1994;368(6468):258–61. [99] Papadopoulos N, et al. Mutation of a mutL homolog in hereditary colon cancer. Science 1994;263:1625–9. [100] Leach FS, et al. Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 1993;75(6):1215–25. [101] Miyaki M, et al. Germline mutation of MSH6 as the cause of hereditary nonpolyposis colorectal cancer. Nat Genet 1997;17:271–2. [102] Akiyama Y, et al. Germ-line mutation of the hMSH6/GTBP gene in an atypical hereditary nonpolyposis colorectal cancer kindred. Cancer Res 1997;57: 3920–3. [103] Nicolaides NC, et al. Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 1994;371:75–80. [104] Hendriks YM, et al. Heterozygous mutations in PMS2 cause hereditary nonpolyposis colorectal carcinoma (Lynch syndrome). Gastroenterology 2006;130(2):312–22. [105] Ligtenberg MJ, et al. Heritable somatic methylation and inactivation of MSH2 in families with Lynch syndrome due to deletion of the 3’ exons of TACSTD1. Nat Genet 2009;41(1):112–7. [106] Kovacs ME, et al. Deletions removing the last exon of TACSTD1 constitute a distinct class of mutations predisposing to Lynch syndrome. Hum Mutat 2009;30(2):197–203. [107] Vasen HF, et al. Revised guidelines for the clinical management of Lynch syndrome (HNPCC): recommendations by a group of European experts. Gut 2013;62(6):812–23. [108] Hooper JL. Application of genetics to the prevention of colorectal cancer. Recent Results Cancer Res 2005;166:17–33. [109] Hampel H, et al. Cancer risk in hereditary nonpolyposis colorectal cancer syndrome: later age of onset. Gastroenterology 2005;129(2):415–21. [110] de la Chapelle A. The incidence of Lynch syndrome. Fam Cancer 2005;4 (3):233–7. [111] Aaltonen L, et al. Explaining the familial colorectal cancer risk associated with mismatch repair (MMR)-deficient and MMR-stable tumors. Clin Cancer Res 2007;13(1):356–61. [112] Salovaara R, et al. Population-based molecular detection of hereditary nonpolyposis colorectal cancer. J Clin Oncol 2000;18(11):2193–200. [113] Hampel H, de la Chapelle A. The search for unaffected individuals with Lynch syndrome: do the ends justify the means? Cancer Prev Res (Phila) 2011;4 (1):1–5. [114] Dunlop MG, et al. Population carrier frequency of hMSH2 and hMLH1 mutations. Br J Cancer 2000;83(12):1643–5. [115] Terdiman JP. HNPCC: an uncommon but important diagnosis. Gastroenterology 2001;121(4):1005–8. [116] Boland CR, Shike M. Report from the Jerusalem workshop on Lynch syndrome-hereditary nonpolyposis colorectal cancer. Gastroenterology 2010;138(7):2197. [117] Lynch HT, de la Chapelle A. Hereditary colorectal cancer. N Engl J Med 2003;348(10):919–32. [118] Geary J, et al. Gene-related cancer spectrum in families with hereditary nonpolyposis colorectal cancer (HNPCC). Fam Cancer 2008;7(2):163–72. [119] Vasen HF. Review article: the Lynch syndrome (hereditary nonpolyposis colorectal cancer). Aliment Pharmacol Ther 2007;26(Suppl 2):113–26. [120] Kastrinos F. et al. Risk of pancreatic cancer in families with Lynch syndrome. JAMA. 2009;302(16):1790-5.

145

[121] Umar A, Boland CR, Terdiman JP, et al. Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst 2004;96(4):261–8. [122] Bonadona V, et al. Cancer risks associated with germline mutations in MLH1, MSH2, and MSH6 genes in Lynch syndrome. JAMA 2011;305(22):2304–10. [123] Kempers MJ, et al. Risk of colorectal and endometrial cancers in EPCAM deletion-positive Lynch syndrome: a cohort study. Lancet Oncol 2011;12 (1):49–55. [124] Burgart LJ. Testing for defective DNA mismatch repair in colorectal carcinoma: a practical guide. Arch Pathol Lab Med 2005;129(11):1385–9. [125] Burn J, et al. Long-term effect of aspirin on cancer risk in carriers of hereditary colorectal cancer: an analysis from the CAPP2 randomised controlled trial. Lancet 2011;378(9809):2081–7. [126] Burn J, et al. Effect of aspirin or resistant starch on colorectal neoplasia in the Lynch syndrome. N Engl J Med 2008;359(24):2567–78. [127] Ruschoff J, et al. Aspirin suppresses the mutator phenotype associated with hereditary nonpolyposis colorectal cancer by genetic selection. Proc Natl Acad Sci U S A 1998;95(19):11301–6. [128] McIlhatton MA, et al. Nitric oxide-donating aspirin derivatives suppress microsatellite instability in mismatch repair-deficient and hereditary nonpolyposis colorectal cancer cells. Cancer Res 2007;67(22):10966–75. [129] Lu KH, et al. Gynecologic cancer as a "sentinel cancer" for women with hereditary nonpolyposis colorectal cancer syndrome. Obstet Gynecol 2005;105(3):569–74. [130] Lu KH, et al. Prospective multicenter randomized intermediate biomarker study of oral contraceptive versus depo-provera for prevention of endometrial cancer in women with Lynch syndrome. Cancer Prev Res (Phila) 2013;6 (8):774–81. [131] McIlhatton MA, et al. Aspirin and low-dose nitric oxide-donating aspirin increase life span in a Lynch syndrome mouse model. Cancer Prev Res (Phila) 2011;4(5):684–93. [132] Langley RE, Rothwell PM. Aspirin in gastrointestinal oncology: new data on an old friend. Curr Opin Oncol 2014;26(4):441–7. [133] Kucherlapati MH, et al. Genotype directed therapy in murine mismatch repair deficient tumors. PLoS One 2013;8(7):e68817. [134] Wu X, Patterson S, Hawk E. Chemoprevention—history and general principles. Best Pract Res Clin Gastroenterol 2011;25(4-5):445–59. [135] Reed SD, et al. Effects of family history and genetic polymorphism on the cost-effectiveness of chemoprevention with finasteride for prostate cancer. J Urol 2011;185(3):841–7. [136] Svatek RS, Lotan Y. Cost utility of prostate cancer chemoprevention with dutasteride in men with an elevated prostate specific antigen. Cancer Prev Res (Phila) 2011;4(2):277–83. [137] Johnson KA, Brown PH. Drug development for cancer chemoprevention: focus on molecular targets. Semin Oncol 2010;37(4):345–58. [138] Crum CP, et al. The distal fallopian tube: a new model for pelvic serous carcinogenesis. Curr Opin Obstet Gynecol 2007;19(1):3–9. [139] Chen S, Parmigiani G. Meta-analysis of BRCA1 and BRCA2 penetrance. J Clin Oncol 2007;25(11):1329–33. [140] Antoniou A, et al. Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case Series unselected for family history: a combined analysis of 22 studies. Am J Hum Genet 2003;72 (5):1117–30.