Background Growing evidence links high ultra-processed food (UPF) intake to various diseases, including colorectal cancer (CRC). However, the metabolomic changes underlying this relationship remain unknown.

Methods Integrating dietary and high throughput metabolomic profiling data from large US cohorts – Nurses’ Health Study (NHS) and Health Professionals Follow-up Study (HPFS), we derived a metabolomic signature of UPF intake among 1,715 participants and prospectively evaluated the signature in pre-diagnostic blood specimens among a separate group of 686 matched pairs of incident CRC cases and controls. Specifically, we estimated intakes of total UPFs and major UPF subgroups (serving/d) using validated food frequency questionnaires administered at the time of blood collection (1989-90 for NHS and 1993-95 for HPFS). Liquid chromatography-mass spectrometry was used to measure plasma metabolites. We used elastic net regression in a 10-fold cross-validation framework to identify UPF-correlated metabolites and derive a metabolomic signature, calculated as the weighted sum of these metabolites. We conducted Spearman correlation analysis to assess the relationship between metabolite levels and the intake of total UPF and UPF subgroups. Multivariable-adjusted conditional logistic regression was used to estimate the odds ratios (ORs) and 95% CIs for the association of UPF signature and its component metabolites with CRC risk.

Results Among 231 metabolites, we constructed a UPF signature comprising 39 metabolites, of which 17 were positively and 22 inversely associated with total UPF intake (Spearman rho range: -0.15 to 0.14). These were primarily lipids and amino acids. Several UPF subgroups showed a strong correlation with metabolite levels, including fats/sauces, savory snacks, mixed dishes, grain products, sweets, meat products, and beverages. The metabolomic signature was associated with a higher risk of CRC (highest vs. lowest quintile: multivariable OR [95%CI] = 1.52 [1.05 to 2.18], p-trend = 0.02). No differences were observed between men and women or according to tumor subsites. Among individual metabolites, C4-OH carnitine, trimethylbenzene, and N2, N2-dimethylguanosine showed a positive association while 21-deoxycortisol and proline betaine showed an inverse association with CRC risk. Of note, N2, N2-dimethylguanosine, a potential biomarker of meat/poultry intake, showed a particularly strong positive association with CRC risk (1.56 [1.08 to 2.26], p-trend=0.05), especially in men (2.30 [1.27 to 4.19], p-trend=0.004).

Conclusion We developed a metabolomic signature as a potential biomarker of UPF intake. The signature and several individual metabolites showed a strong association with CRC risk. Our findings provide biological insight into the potential metabolic pathways linking UPF intake to CRC risk.

<b>Figure 1. Associations between the 39 Metabolites Constituting the Metabolomic Signature and the Consumption of Total UPFs and UPF Subgroups</b>

Figure 1. Associations between the 39 Metabolites Constituting the Metabolomic Signature and the Consumption of Total UPFs and UPF Subgroups

Table 1. Associations of Metabolomic Signature of UPFs and its Components with Colorectal Cancer Risk

Danielle Haslam

Appearances