Cancer prevention in the era of precision oncology
The cancer research community has struggled to understand the early events leading to cancer development. It is likely that multiple signaling events are operational at various stages along the carcinogenesis continuum, contributing to considerable molecular tumor heterogeneity both between target organs as well as within a given tumor.1 Successful precision cancer prevention will require an understanding of the genetic, epigenetic, signaling, microenvironment, and immune factors that drive the development of cancer from initiation through premalignancy to invasive cancer.
Given the limited access to human premalignant lesions, nonclinical models are important for investigations of pathways and markers of early carcinogenesis. Carcinogen‐driven and genetically engineered mouse models can be useful for mechanistic studies. The main strengths of nonclinical models have been in understanding mechanisms of agent action, providing preliminary evidence of agent efficacy, and agent toxicology and pharmacokinetic profiling. Efforts like the “Collaborative Cross” attempt to model population diversity by outbreeding common inbred mice strains2 and may also prove informative for the identification and evaluation of biomarkers to better predict preventive agent response. Unfortunately, findings in mice do not necessarily hold true with respect to human biology; therefore, the translatability of nonclinical conclusions can be limited.
Observational studies in humans are rich resources of testable hypotheses about cancer etiology and strategies for interventions. In the studies of infectious agents and cancer, recognition of the relationship between specific human papillomavirus (HPV) strains and cervical cancer ultimately drove development of agents that protect against infection and reduce cervical premalignancy incidence.3 Similarly, understanding the link between hepatitis B virus (HBV) infection and hepatocellular carcinoma (HCC) pointed to HBV vaccination as an HCC prevention strategy. Analyses showed that, indeed, HCC incidence was lower among vaccinated cohorts decades after Taiwan launched a universal HBV immunization program.3
Bringing the precision of vaccines against infectious agents to vaccine development for cancers not associated with infection has been more challenging. However, accumulating data suggest that vaccines directed against aberrantly expressed self‐antigens may activate the immune system to detect and clear preinvasive lesions without triggering an autoimmune reaction. One such antigen is mucin 1 (MUC1), which is expressed on the apical surface of normal ductal epithelial cells but abnormally expressed and glycosylated in premalignant lesions and cancers.4 Similarly, proteins like insulin‐like growth factor 1 receptor (IGF1R), insulin‐like growth factor binding protein 2 (IGFBP2), and human epidermal growth factor receptor 2 (HER2) can serve as potential antigens due to their differential expression patterns and roles in the pathogenesis of cancer.4 Multivalent vaccines can help address heterogeneity in cancer initiation. In these examples, molecular profiling of cancers and early‐stage lesions served to identify candidate antigens that are now being tested in clinical trials.
In the absence of a thorough understanding of the early molecular events in the genesis of most common epithelial cancers, precision approaches to target identification have relied on evidence gathered from epidemiologic, animal, and clinical trial sources and have already had some success.