Noninvasive detection of candidate molecular biomarkers in subjects with a history of insulin resistance and colorectal adenomas

Abstract

We have developed novel molecular methods using a stool sample, which contains intact sloughed colon cells, to quantify colonic gene expression profiles. In this study, our goal was to identify diagnostic gene sets (combinations) for the noninvasive classification of different phenotypes. For this purpose, the effects of a legume-enriched, low glycemic index, high fermentable fiber diet was evaluated in subjects with four possible combinations of risk factors, including insulin resistance and a history of adenomatous polyps. In a randomized crossover design controlled feeding study, each participant (a total of 23; 5-12 per group) consumed the experimental diet (1.5 cups of cooked dry beans) and a control diet (isocaloric average American diet) for 4 weeks with a 3-week washout period between diets. Using prior biological knowledge, the complexity of feature selection was reduced to perform an exhaustive search on all allowable feature (gene) sets of size 3, and among these, 27 had (unbiased) error estimates of 0.15 or less. Linear discriminant analysis was successfully used to identify the best single genes and two- to three-gene combinations for distinguishing subjects with insulin resistance, a history of polyps, or exposure to a chemoprotective legume-rich diet. These results support our premise that gene products (RNA) isolated from stool have diagnostic value in terms of assessing colon cancer risk.

Conflict of interest statement

Disclosure of Potential Conflicts of Interest: No potential conflicts of interest were disclosed.

Figures

Figure 1

Overall study design. BL, baseline measurement; DP, diet period.

Figure 2

The concept of intrinsically multivariate predictive (IMP) genes is shown where expression profiles of a group of genes predict the phenotype. Results represent a linear classification of (+IR, +Polyps) subjects (○) versus (−IR, −Polyps) subjects (Δ) at BL1. UCP2 and HOXA3 were used as individual one-feature sets (A and B) as compared with both genes together as a two-feature set (C). The bolstered error is 0.2784, 0.4882, and 0.1415 for A, B, and C, respectively.

Figure 3

Effective classification of clinical phenotype or diet. A, linear (LDA) classification of (+IR, +Polyps) subjects (○) versus (−IR, −Polyps) subjects (Δ) at BL1; B, linear (LDA) classification of (−IR, −Polyps) subjects on the control diet (○) versus (−IR, −Polyps) subjects on the legume diet (Δ) using the crossover design and combining the microarrays from samples collected at the end of the two diet periods DP1 and DP2.

Figure 4

Potential design problems and importance of the experimental design factors IR and history of adenomas. A, increased error in the LDA classification of (+IR, +Polyps) subjects (○) versus (−IR, −Polyps) subjects (Δ) when both baselines BL1 and BL2 were included. B, (+Polyps) subjects (○) versus (−Polyps) subjects (Δ) at baselines BL1 and BL2. C,(+IR) subjects (○) versus (−IR) subjects (Δ) at all time points.