Early-onset colorectal cancer research: gaps and opportunities

Abstract

The incidence rates of sporadic early-onset colorectal cancer (EO-CRC) are increasing rapidly in the USA and globally. Birth cohort analyses strongly suggest that changes in early life exposures to known or unknown risk factors for CRC may be driving EO disease, but the etiology of EO-CRC remains poorly understood. To address the alarming rise in sporadic EO-CRC, the National Cancer Institute and National Institute of Environmental Health Sciences convened a virtual meeting that featured presentations and critical discussions from EO-CRC experts that examined emerging evidence on potential EO-CRC risk factors, mechanisms and translational opportunities in screening and treatment.

Keywords::

• colorectal cancer
• early-onset disease
• meeting report
• molecular oncology

The National Cancer Institute and National Institute of Environmental Health Sciences held the ‘Early-Onset Colorectal Cancer Think Tank’ virtual meeting on 24–25 September 2020. More than 400 people participated from more than 100 organizations inside and outside the USA, including representatives from academia, industry, government and patient advocacy organizations. The agenda is on the meeting website [1].

The incidence rates of sporadic early-onset colorectal cancer (EO-CRC) are rapidly increasing in the USA [2] and globally [3]. Emerging evidence suggests that EO-CRC may have distinct etiology from hereditary CRC and late-onset CRC (LO-CRC). The relatively recent increase in incidence suggests a role for environmental exposures in EO-CRC, however, causes remains poorly understood [4]. An NIH Think Tank was held to identify research priorities that will improve our understanding of the connections between environmental exposures and biological mechanisms relevant to EO-CRC and translate that knowledge for prevention, screening, diagnosis and treatment.

Risk factors & causes

Although incidence rates are increasing, EO-CRC (diagnosis <50 years old) remains relatively rare (∼6–8/100,000 in the USA [4]), creating challenges to establish causation. Germline mutations that predispose to EO-CRC are found in 15–20% of cases. However, the incidences of familial syndromes are stable and are not likely contributing to the overall rise in EO-CRC cases. Most EO-CRC is found in the rectum and distal colon, and the tumors are microsatellite stable, chromosomal instable and have significantly fewer BRAF and APC mutations than LO-CRC, suggesting tumor extrinsic risk factors influence mutational and other molecular signatures in the majority of cases.

Determining changes in exposures to established and novel risk factors in early life and young adulthood requires large, well-annotated datasets. Data-rich studies and surveys exist (e.g., National Health and Nutrition Examination Survey, Nurses’ Health Study), but many are hampered by lack of race/ethnic diversity, lack of stool and blood samples (with most collected at a single time point) and small numbers of cases that are often not optimally annotated. Therefore, questions remain regarding which exposures are driving EO disease, and how they mechanistically contribute to its etiology.

Exposures related to energy balance (diet and physical activity) may modulate risk for EO-CRC, and models that examine how these exposures accumulate and interact over the lifespan are needed. For example, sedentary behavior and obesity associated with enhanced risk, but whether they behave independently or interact is unknown. Diet has a clear role in risk for EO-CRC, with typical western diets associated with higher risk, and diets rich in fruits and vegetables associated with lower risk. The benefits of prudent diets may be mediated via effects on the gut microbiome, thus animal models in which diets and microbiota can be specifically manipulated will be useful for clarifying the individual and combined effects of dietary components.

A key question in understanding the factors responsible for increasing incidence of EO-CRC is whether EO- and LO-CRC are the same disease, or if EO-CRC is caused by unique underlying biological mechanisms that are impacted by different risk factors. A recent analysis of polygenic risk scores developed using genome-wide association study data found that a different set of risk variants associated with EO- versus LO-CRC [5]. Mutation frequencies of key tumor genes (e.g., APC, CTNNB1 and BRAF^V600E ) differ between EO- and LO-CRC [6]. Studies designed to correlate known or novel environmental exposures (or combinations of exposures, i.e., the exposome), with EO-CRC-specific molecular characteristics will help to understand their contribution to risk, as will whole exome sequence analysis comparisons of EO- and LO-CRC [6,7].

Potential mechanisms

Specific molecular characteristics of sporadic EO-CRC may provide clues about the critical risk factors and gene/environment interactions involved in tumor development. Several types of molecular signatures (genomic, epigenetic, proteomic, metabolomic and microbiome profiling) have been used to characterize EO-CRC tumors and to suggest biologic mechanisms that drive the disease. Future studies should consider the effects of putative risk factors on molecular signatures in both transformed cells and in stromal cells and the associated microbiome. For example, antibiotic use, Vitamin D deficiencies or comorbid conditions like obesity and irritable bowel disease (IBD) can directly affect GI inflammation, oxidative stress and DNA damage in developing CRC lesions [8–10].

Comparative analyses of integrated multi-omic datasets from EO- and LO-CRCs found that a key regulator of glutathione production during oxidative stress responses, NRF2, is increased in EO-CRC but not in LO-CRC, suggesting that NRF2 mediated responses may play a key role in EO-CRC [11]. Epigenetic defects leading to hypomethylation of retrotransposons may be important in a subset of EO-CRC [12]. Conversely, APC mutation-negative CRCs have hypermethylated regions, including the promoter of the E3 ligase, RNF43 [13]. Thus, post-translational modifications that activate oncogenes may be another important signaling mechanism in EO-CRC. Mutations or epigenetic changes that inhibit RNF43 can result in loss of Wnt signaling regulation and aberrant stem cell activation [14].

The effects of many EO-CRC risk factors are likely mediated by their interactions with and changes to the GI microbiome. The human microbiota has the capacity to produce a diet-dependent metabolome that impacts host tissue and microbial community homeostasis and function [15]. Microbial metabolites derived from bioactive dietary components are important regulators of immune and metabolic functions during carcinogenesis and may provide future opportunities as targets for novel EO-CRC prevention and treatment strategies.

Screening, prevention & treatment

In the near term, identifying a cost-effective and feasible screening approach for those under age 50 may have the greatest impact on EO-CRC burden. Modeling work and screening modality trials that assess the risks and benefits of lowering the screening age could help determine which US FDA-approved CRC screening test is most effective, both clinically and economically. Research to develop and validate emerging approaches for early detection of EO-CRC, including liquid biopsy, is also needed [16].

Improved risk stratification for younger populations using risk scores and prediction models will help design precision cancer prevention and screening approaches [17]. Several high priority questions remain unanswered: what is the adenoma to carcinoma sequence in EO-CRC; do epigenetic and genetic alterations parallel LO-CRC and do familial cases skew the mutation spectrum; what is the true prevalence of adenomas and asymptomatic individuals in the average risk population; is there a difference in the prevalence of EO-colorectal adenoma among different racial/ethnic groups; and what is the risk for subsequent CRC in EO-adenoma patients?

Current CRC clinical guidelines do not use age as a determinant of EO-CRC therapy. EO-CRC patients tend to be diagnosed at later stages and receive more aggressive treatment, which may improve outcomes relative to older patients, but also can have a significant impact on quality of life. Data comparing survival in younger and older patients is inconsistent, although worse survival is observed for patients younger than 35 years old. This could be attributable to delay in diagnosis or because of distinct biologic features [18]. A recent study showed that consensus molecular subtype 1 tumors, which are more prevalent in EO cases, are less responsive to cetuximab [19]. To develop more targeted treatments, a better understanding of the molecular mechanisms, immune response (particularly given the relatively high incidence of mismatch repair deficiency in the EO population), and role of the microbiome underlying EO-CRC is needed. Questions also remain concerning the most effective use of radiation and/or surgery to improve quality of life without reducing survival, and whether aggressive surgery for metastases improves outcomes.

Conclusion

More effective detection, diagnostic, prevention and treatment strategies are needed to improve EO-CRC outcomes. It will be critical to identify environmental and lifestyle factors contributing to EO-CRC and to develop strategies to ameliorate or manage these exposures, and approaches to identify and target relevant biologic mechanisms. National Cancer Institute and National Institute of Environmental Health Sciences support these studies via existing funding opportunities [20] and research programs (DNTP’s Carcinogenesis-Health Effect Innovations).

Research gaps and opportunities include:

• Need for larger, more diverse cohorts with environmental exposures across the lifespan that include stool and serum samples and integration of consensus molecular subtype information.

• Better molecular markers of exposure are needed to measure the relative contribution of putative environmental factors of risk and to better understand how these exposures affect biological mechanisms in EO-CRC.

• Integrated multi-omic approaches in carefully stratified cohorts to agnostically determine how risk factors (e.g., obesity, diet, infection, drugs and metabolic syndromes) may effect molecular changes in EO-CRC.

• Determine the nonAPC driven Wnt/beta catenin activating mechanisms that promote intestinal epithelial cell (IEC) stem cell activation in EO-CRC (e.g., Line 1 insertions and RNF43 suppression).

• Use of artificial intelligence algorithms and novel models like organoids that link risk exposures to molecular signatures of EO-CRC and identify potential preventive and therapeutic targets.

• Additional collaborative resources such as biorepositories that include digitized tissue images to develop consensus histopathology, electronic health records (EHRs) and annotated family history and exposure data.

Acknowledgments

The authors are thankful to the staff from NCI and NIEHS who helped plan and facilitate this meeting (C Allegra, P Daschner, R Divi, Z Farhat, R Flores, K Helzlsouer, TK Lam, X Li, H Loomans-Kropp, S Mahabir, S Nelson, S Nothwehr, L O’Connor, A Pandiri, E Ramirez-Pena, L Reinlib, G Riscuta, S Ross, H Seifried, R Sinha, JV Tricoli, A Umar, A Wali and M Young).

Authors are thankful to the Think Tank participants, especially the speakers and patient advocates, who contributed their critical expertise and perspective.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.