Inflammation Modulation by Vitamin D and Calcium in the Morphologically Normal Colorectal Mucosa of Patients with Colorectal Adenoma in a Clinical Trial

Abstract

Increased COX-2 and decreased 15-hydroxyprostaglandin dehydrogenase (15-HPGD) expression promote prostaglandin-mediated inflammation and colorectal carcinogenesis. Experimental studies suggest that vitamin D and calcium may inhibit these pathways, but their effects on colorectal tissue COX-2 and 15-HPGD expression in humans are unknown. We tested the effects of supplemental vitamin D (1,000 IU/day) and/or calcium (1,200 mg/day) on COX-2 and 15-HPGD expression in the morphologically normal rectal mucosa from 62 paients with colorectal adenoma in a placebo-controlled chemoprevention trial. We measured biomarker expression using automated IHC and quantitative image analysis at baseline and 1-year follow-up, and assessed treatment effects using mixed linear models. The primary outcome was the COX-2/15-HPGD expression ratio, because these enzymes function as physiologic antagonists. After 1 year of treatment, the mean COX-2/15-HPGD expression ratio in full-length crypts proportionately decreased 47% in the vitamin D group (P = 0.001), 46% in the calcium group (P = 0.002), and 34% in the calcium + vitamin D group (P = 0.03), relative to the placebo group. Among individuals with the functional vitamin D-binding protein isoform DBP2 (GC rs4588*A), the COX-2/15-HPDG ratio decreased 70% (P = 0.0006), 75% (P = 0.0002), and 60% (P = 0.006) in the vitamin D, calcium, and combined supplementation groups, respectively, relative to placebo. These results show that vitamin D and calcium favorably modulate the balance of expression of COX-2 and 15-HPGD-biomarkers of inflammation that are strongly linked to colorectal carcinogenesis-in the normal-appearing colorectal mucosa of patients with colorectal adenoma (perhaps especially those with the DBP2 isoform). PREVENTION RELEVANCE: Supplemental calcium and vitamin D reduce indicators of cancer-promoting inflammation in normal colorectal tissue in humans, thus furthering our understanding of how they may help prevent colorectal cancer.

Figure 1.

Measurement of COX-2 expression in crypts and stroma of normal-appearing rectal mucosa using custom-designed quantitative image analysis software. The scoring process entailed (A) finding and tracing a full-length hemicrypt and then (B) automated sectioning and quantification of biomarker labeling optical density, overall, and within each segment of the hemicrypt; (C) the stroma adjacent to previously scored hemicrypts was also outlined and scored.

Figure 2.

Distributions of COX-2 and 15-HPGD labeling optical densities in normal rectal mucosa crypts and adjacent stroma, by treatment arm, at 1-year follow-up among adjunct biomarker study participants (n = 62). Mean biomarker labeling optical densities across all measured crypts and adjacent mucosa from all participants, in each of the 50 equal-height segments from crypt bases (position 0) to the colorectal lumen (position 50), plotted for: (A) COX-2 in crypts, (B) COX-2 in stroma, (C) 15-HPGD in crypts, and (D) 15-HPGD in stroma. The dashed outlines in the bottom-right corner of each panel depict the areas (i.e., the crypt or adjacent stroma) in which biomarker expression was quantified using image analysis software. Distributions of COX-2 and 15-HPGD at baseline (similar across treatment groups, as expected) are shown in Supplementary Figure S2.