Heritable one-hit events defining cancer prevention?

Abstract

Over 100 y ago (1902–1914) Theodor Boveri suggested a role for mutations in cancer.^1,2 Boveri’s ideas were derived from the then “just-emerging” chromosome theory of inheritance. While demonstrating chromosomal aberrations as a cause of genetic imbalance, Boveri suggested that possible causes of malignancy may include events such as aneuploidy that are now defined as gene mutations, asserting all the while that malignancy occurs at the cellular level. Indeed, studies to date essentially uniformly show that cancer is a genetic disease.^3,4

Keywords: :

• cancer prevention
• cancer syndromes
• Knudson
• two-hit
• one-hit
• driver mutations
• passenger mutations

The mechanisms that underline genetic predisposition to cancer have been clarified through a series of elegant observations by Alfred G Knudson,^3,5 who hypothesized that germline mutations occur in one allele of a tumor suppressor gene followed by somatic inactivation, or loss of function, of the remaining normal allele through mutations, deletions, or epigenetic repression (defined as loss of heterozygosity [LOH]). The Knudson “two-hit” hypothesis has been largely validated in essentially all forms of autosomal heritable cancer and, in principle, has been extended to sporadic forms of cancer as well, albeit at a greater level of complexity.⁶ Other studies suggest non-mutational events in the causation of cancer, commonly defined as a field effect, including epithelial–mesenchymal transition (EMT).^7-10 These non-mutational events largely rely on alterations of the cancer phenotype, presumably affecting the course of carcinogenesis. Nonetheless, it is universally accepted that mutations constitute an overriding presence in the etiology of cancer.

Cancer prevention is defined, in general, as the treatment of ostensibly normal individuals, including those who are at identifiable risk, over extended periods with essentially non-toxic agents. Multistep carcinogenesis posits that rate-limiting mutations accumulate in a single cell and its progeny, bringing about histopathologic alterations in the target cells. The protracted time required for these alterations to occur permits the testing of pharmacologic and/or dietary interventions to prevent or delay the transition to malignancy. Thus, early intervention would be optimally performed on persons at high risk of developing a specific cancer, such as those individuals who carry a germline mutation or epigenetic changes in gene(s) known to impart such risk. Targeted individuals also include persons with the potential for metachronous cancers, obese/chronically inflamed individuals, as well as those who might be at increased risk due to dietary and environmental exposures. Importantly, secondary malignancies in children who have been successfully treated for early onset cancers, whether of the hereditary or of the sporadic type, constitute a special case where effective prevention modalities of these secondary cancers may be identified

Heritable cancer syndromes represent a paradigm of the two-hit hypothesis for cancer progression.^3,5 There are about 50 forms of heritable autosomal dominant cancer syndromes recorded in humans.⁶ While comprising only about 5% of all cancers, they appear to recapitulate, in general, most forms of known sporadic human cancers. As such, exploration of these syndromes may bring clarity to our understanding of sporadic cancer forms and the potential role for cancer prevention. For example, several distinct pathways have been implicated in the etiology of heritable breast cancer, including Li-Fraumeni syndrome (LFS), BRCA1/2 (also termed hereditary breast and ovarian cancer syndrome, or HBOCS), and Cowden syndrome (CS), wherein the p53,^11-14 BRCA1/2,^15-18 and PTEN,¹⁹ respectively play a major role. These same pathways are highly relevant in sporadic breast cancer as well.⁶ Whether these distinct pathways merge into a key proximal pathway that would ultimately lead to clinical breast cancer, or whether these pathways remain distinct and apart from one another, each sufficient to ultimately lead to breast cancer, possibly with a different histological outcome, is unclear. Hence, comparing abnormal pathways in LFS, BRCA1/2, or Cowden syndrome with abnormal pathways in the sporadic forms of breast cancer might permit better diagnosis and more effective intervention modalities for the prevention of breast cancer. With that in mind, it should be recognized that mutation types and their occurrence within specific functional domains of tumor suppressor genes among different individuals (molecular heterogeneity) are likely to represent confounding factors in heritable and sporadic cancers.⁶

Experimental support for this approach is provided by on-going studies, using expression arrays, proteomics, or metabolomics. To date, these studies demonstrated that phenotypically normal human cells heterozygous for a cancer-predisposing, loss-of-function mutation show abnormalities that are consistent with “one-hit” effects in a variety of autosomal dominant forms of cancer, including FAP,^20-22 TSC2,²³ VHL,²³ BRCA1/2,²⁴ LFS,²⁵ familial melanoma,²⁶ and several as yet to be published, including newer insights on VHL, FAP, HNPCC, and BCNS (in preparation). In this regard, earlier studies originally based on “the ubiquitous fibroblast”²⁷ that was used as a surrogate for the cells of tumor origin indicated a systemic presence of some of these biomarkers, including their occurrence in heritable cancers and in the corresponding sporadic cancers.²⁸ Based on the early discoveries of these biomarkers, a flowchart of events that define the initiated state at the cellular/phenotypic level and its relationship to cancer predisposition in the autosomal dominant cancer syndrome has been proposed.^20,28

The clinical relevance of alterations at the “one-hit” level to the cancer process is highlighted by the fact that essentially 80–100% of patients with a heritable cancer syndrome will develop cancer,⁶ indicating that observations in phenotypically normal cells/tissues from these individuals are directly associated with actual cancer development. These results have also led us to formulate a distinction between the autosomal and recessive forms of heritable cancer, wherein genetic information in the former at the cellular level is associated with cancer initiation, whereas genetic information at the cellular level in the latter is presumably consistent with the early phases of cancer promotion.²⁸ In this regard, developmental abnormalities that are concordant with increased cancer risk in some of the dominantly heritable cancer syndromes,⁶ may at some level provide further insights about the development of specific cancers.

Importantly, abnormal alterations that occur in normal cells/tissues, derived from patients representing different forms of heritable cancer, appear to capture not only changes that are unique to a specific pathway(s) for a specific heritable cancer form, but they also include a subset of “global changes”, whether systemic or at the target site, that seem to be shared by different forms of heritable cancers at the tumor level, including sporadic forms of cancer.⁶ Collectively, these global changes can be defined as the cancer trait.^20,28 In general, they included changes in cytoskeletal organization and biogenesis, cell adhesion, apoptosis inhibition, oxidative stress responses, and mitochondrial damage, as well as an altered metabolism.^6,20-26

Most recently, experimental considerations have led to the suggestion that only a limited number of genetic alterations may be absolutely required for tumor formation, referred to as driver mutations. If this number is reasonably accurate, it suggests that most of the genetic changes in tumor cells represent collateral damage or passenger mutations.⁴ Thus, characterization of phenotypically normal cells/tissues from the heritable cancer syndromes (one-hit), which do not include confounding secondary effects that occur at the two-hit tumor stage, should facilitate discovery of the earliest molecular genetic changes/targets that represent driver mutations in sporadic cancer as well, enabling prioritization and validation of relevant biomarkers and optimal targets at the “two-hit” tumor stage. A more precise identification of early biomarkers and valuable drug targets would guide development of personalized prevention strategies for sporadic cases of cancer.

The salient features which provide the experimental rationale for the use of dominantly heritable cancer forms to identify relevant molecular pathways and targets during cancer progression can be summarized as follows:

• Genetic changes/pathways seen in phenotypically normal cells/tissues from autosomal dominant forms of cancer are present in the tumors from which they arise and in the corresponding sporadic cancers without confounding secondary tumor effects.

• Genetic changes/pathways seen in phenotypically normal cells/tissues from autosomal dominant forms of cancer are relatively well understood in the corresponding sporadic cancers.

• Genetic changes/pathways in one-hit phenotypically normal cells/tissues that are found in heritable forms of cancer represent cancer initiation/predisposition at the cellular level.

• Cancer initiation in one-hit phenotypically normal cells/tissues from patients with autosomal forms of cancer, and to varying degree in the sporadic forms of cancer, also includes changes that are shared by different forms of cancer, which, collectively, can be defined as “global expression of the cancer trait”.

Two recent examples that capture the process from recognition of relevant biomarker/pathways during cancer development to the conceptual development of potential interventional approaches based on “one-hit” cells are as follows:^25,26

Li-Fraumeni syndrome (LFS) is a rare autosomal dominant disorder characterized by germline mutations in TP53 and the early onset of multiple forms of cancer, including cancers of the breast, soft tissues, brain, and adrenal gland, which constitute the LFS “core cancers”, with breast cancer being the most dominant.^11-13 While clinical protocols exist to provide recommendations for LFS screening, diagnosis and follow-up is complicated by factors such as the wide variety of phenotypes associated with LFS as well as the difficulty in identifying mutations in TP53. Because TP53 is one of the most frequently mutated genes in sporadic cancer, using LFS as an example can also benefit the treatment or prevention of sporadic cancers. Although mutations affecting p53 are present in virtually all human cancers, “stress-induced” non-mutational activation of p53 occur very early in cancer progression and may precede and perhaps facilitate mutational activation associated with p53 and possibly other cancer causing genes.^29,30

Herbert et al.,²⁵ have observed that phenotypically normal breast epithelial cells with the heterozygous germline M133T mutation in TP53 exhibited significant alterations that were largely centered around the p53 pathway. Of particular interest was the observation that the anti-apoptotic gene BIRC3 was significantly upregulated in LFS-derived normal epithelial cells compared with controls, as well as increased pro-inflammatory IL-1β gene expression in LFS-derived stromal cells. A similar increase in and dependency on BIRC3 gene expression was demonstrated using a heterozygous TP53 mouse model of osteosarcoma,³¹ frequently diagnosed in LFS patients.^11-13

Importantly, treatment of LFS cells with p53-rescue drugs such as CP-31398 together with PRIMA-1 (now in phase I trials), synergistically reduced BIRC3 and pro-inflammatory IL-1β mRNA expression, resulting in a pronounced reduction in cell growth compared with either drug alone or no treatment. This may suggest that intervention of these pathways by targeting TP53 directly may be beneficial in preventing cancer progression in LFS patients. In this regard, CP-31398 has been shown to be highly effective in the prevention of preclinical models of ovarian cancer, non-melanoma skin cancer, both squamous cell carcinoma (SCC), and basal cell carcinoma (BCC), colon cancer and rhabdomyosarcoma, where mutations in TP53 play a major role in the etiology of these cancers.³² Therefore, using this example an early interventional approach can be proposed that may selectively inhibit tumors in LFS patients expressing the mutated TP53 phenotype.

Familial melanoma (FM) is a dominantly heritable cancer that is associated with mutations in the tumor suppressor gene, CDKN2A/p16.^33,34 In FM, a single inherited “hit” occurs in every somatic cell, enabling interrogation of cultured normal skin fibroblasts (SFs) from FM gene carriers as surrogates for the cell of tumor origin, namely the melanocyte. Recently, considerable insights from p16 familial melanoma (FM) patients suggest that genomic profiling of phenotypically normal skin fibroblasts can help identify novel molecular targets for chemoprevention, including early specific biomarkers of melanoma risk among individuals who are heterozygous for CDKN2A mutations (which encodes for p16) within FM families.²⁶ These potential biomarkers may enable mechanism-based early detection of melanoma and personalized prevention strategies to target sporadic melanoma. For example, this study shows that occurrence of CDKN2A mutations of the more “benign” V126D-p16 or the more “disruptive” R87P-p16 mutation within a given FM family is mutually exclusive, supporting the notion that personalized, cohort-based, intervention modalities, including clinical outcome, should be recognized. In this instance, we posit that, direct or up/downstream “pharmacologic normalization” of the R87P-p16 mutation,²⁵ individually and/or together with BRAF intervention³⁵ may enable more effective strategies to delay or even prevent FM, including the occurrence of melanoma in individuals who are otherwise at risk. In this regard, proclivity toward pancreatic cancer that is likely to be caused by CDKN2A,³⁶ especially among affected carriers of the R87P-p16 mutation, should be explored for early diagnosis and intervention of pancreatic cancer.

In conclusion, analyses of primary, phenotypically normal cell cultures from persons with heritable cancer syndromes indicate that loss-of-function mutations of tumor suppressor genes are associated with detectable changes, including gene expression, proteomics, and metabolomics. In many cases, changes detected in phenotypically normal, one-hit cells, are consistent with the known biology of these genes in homozygous mutant tumor cells that are found in various cancer types. These alterations may represent early molecular changes in the process of tumorigenesis that could serve as risk biomarkers and targets for preventive agents. Significantly, the majority of non-hereditary cancers are mutant for a selected group of driver genes.⁴ These same genes appear to underwrite the development of many dominantly heritable forms of cancer, and, as such, may be relevant to the management of a much larger group of cases of sporadic cancers.

Epilogue

In the mid-1980s, insights into mechanisms of differentiation and cancer arose unexpectedly from earlier studies on “single-hit” human adult skin fibroblasts (HASF) obtained from familial adenomatous polyposis (FAP) patients, a dominantly heritable cancer syndrome that predisposes to colorectal cancer (vide supra), wherein FAP-derived HASF showed increased sensitivity to transformation by the Kirsten Murine sarcoma virus-Ras bearing oncogene (KiMSV).³⁷ Significantly, this has been recently attributed, at least in part, to a substantial decrease in the Ras suppressor protein (RSU-1) from FAP-derived HASF.^21,22 Subsequent studies demonstrated that FAP-derived HASF undergo cellular reprogramming upon infection with viral oncogenes. Specifically, HASF-transduced with KiMSV or the Snyder–Thelien Feline Sarcoma Virus (ST-FeSV) differentiated into preadipocytes, then to adipocytes (in the presence of glucocorticosteroids) or macrophages, respectively (reviewed in ref. 38). More recently, ectopic expression of a set of transcription factors implicated in tumorigenesis and defined as oncogenes, such as Oct4, Sox2, Myc, Klf4, has been shown to reset the epigenetic state of differentiated HASF to pluripotency.^39,40 Other pluripotency factors, namely nanog and Lin28, which can substitute for Myc and Klf4 to reprogram human somatic cells, have been described.^39,40 Notably, in the latter studies, cellular reprogramming caused pluripotency of HASF, whereas transduction with viral oncogenes resulted in the terminal differentiation of HASF.³⁸ Together, these studies raise the possibility that cell transformation and normal physiologic reprogramming, including pluripotency, share common pathways and may represent variations of similar biological motifs.

Oncogenes have been characterized by their role in the development of cancer.^41-43 Here we posit that, in principle, activated c-proto-oncogenes/oncogenes can reprogram a normal cell phenotype, bringing about the de novo expression of a different, yet non-cancerous cell phenotype. For example, conversion of human fibroblasts into adipocytes by an activated c-ras proto-oncogene, the cellular homolog of v-Ras gene, in the presence of glucocorticosteroids may have profound implications on processes of adipocyte recruitment in vivo, including obesity and inflammation. Similarly, the cellular homolog of the v-Fes oncogene, upon activation, may give rise to macrophages as a means of increasing the number of such cells during acute/chronic inflammatory disease, immune-deficient states, and cancer (vide supra). The recently identified axis of molecular changes linking obesity and inflammation to increased risk of breast cancer in ER-positive post-menopausal women⁴⁴ may serve as a paradigm for this supposition.

Thus, physiological processes that are influenced by constitutively activated cellular proto-oncogenes to “oncogene,” often defined as “gain-of-function mutations,” may range from highly reversible reactions to essentially irreversible ones, in much the same way as do enzymatic reactions. In the extreme case of mutated cellular oncogenes, reversibility during “normal” physiological processes may be due to increased instability of the mutated oncogenes and/or their products. Oncogenesis, on the other hand, may be caused by increased stability of the activated cellular proto-oncogenes, leading to a long-term, sustained, presence of the corresponding oncogenes and/or their products. It seems plausible that such contrapuntal forces, which are presumably cell context/environment-dependent, may not be mutually exclusive. Oncogenes, therefore, may be instrumental during cellular reprogramming as descendant of constitutive genes whose role is to facilitate normal growth and differentiation, but when stably dysregulated, they may cause cancer.