Obesity is associated with increased risk of ≥13 cancers. Of all cancers attributable to excess adiposity, colorectum and liver account for 55% of cancer among men and 48% among women, excluding reproductive cancers. Although most epidemiologic studies of obesity as a cancer risk factor evaluated body mass index, accumulating evidence for colorectal and liver cancers implicates viscerally located adiposity (and its closely related glycemic metabolic dysregulation) as the likely direct causal component. How visceral adiposity mechanistically predisposes its proximal organs to cancer is largely unknown. Inflammation undoubtedly plays a role in development and progression of malignancies, including colorectal and liver cancers; however, the large body of evidence for general inflammation processes and systemic markers like C-reactive protein in relation to digestive cancers are underwhelming. Thus, distilling the inflammatory pathways and markers to identify those most reflective of the metabolically unhealthy obese state has immense potential to uncover key mechanisms underlying obesity’s role in the development of colorectal and liver cancer.

Characterizing metabolically unhealthy obesity is critical to investigate obesity-related mechanisms of inflammation with colorectal and liver cancer risk but is often prohibitively costly and logistically infeasible in the context of large population-based studies. Therefore, we propose an innovative approach to address these gaps by (i) deriving novel proteomic-based inflammation signatures of metabolically unhealthy obesity (“Inflammotypes”) in cohorts with visceral adipose tissue quantified via dual-energy X-ray absorptiometry (DXA) and traits of glycemic metabolic function; then (ii) prospectively investigating these novel Inflammotypes in longitudinal cohorts with stored blood samples in relation to incident colorectal and liver cancer risk. Our deeply-phenotyped cohorts for derivation and validation of Inflammotypes include general population adults without cancer from the US-based Vitamin D and Omega-3 Trial and Cocoa Supplement and Multivitamin Outcomes Study. Colorectal and liver cancer cases and matched controls will be included from prospective cohorts with bloods collected at study baselines prior to diagnosis, including the Nurses’ Health Study, Health Professionals Follow-up Study, Prostate, Lung, Colorectal, and Ovarian, Southern Community Cohorts Study, Physicians’ Health Studies, and the Women’s Health Study. Collectively, these data will be used to address these specific aims:

AIM 1. To characterize patterns of chronic inflammation (Inflammotypes) via state-of-the-art Olink proteomic panel (384 inflammation-related proteins) that describe metabolically unhealthy obesity (i.e., higher DXA visceral adiposity, together with homeostatic model assessment for insulin resistance [HOMA-IR], hemoglobin A1c [HbA1c], or lipoprotein insulin resistance score [LPIR]). Machine learning analyses to identify the Inflammotypes among highly phenotyped adults without cancer will be performed in VITAL (N=767) and replicated in COSMOS (N=426)

AIM 2. To investigate the prospective relationship between proteomic Inflammotypes with long-term risk of incident colorectal (1000 cases/1000 controls) and liver cancer (500 cases/500 controls), adjusting for competing cancer risk factors. Analyses will combine across eligible nested case-control pairs from longitudinal cohorts (NHS, HPFS, PLCO, SCCS, PHS, WHS) with stored baseline bloods and long-term follow-up (median ranges 6.1-16.7 years).

Xuehong Zhang (Harvard Medical School and Brigham and Women's Hospital)
Katherine McGlynn (NCI)
Edward Giovannucci (Harvard Chan School)
Tobias, Deirdre K (Harvard Medical School and Brigham and Women's Hospital)