Protein microarray analysis of disease activity in pediatric inflammatory bowel disease demonstrates elevated serum PLGF, IL-7, TGF-beta1, and IL-12p40 levels in Crohn's disease and ulcerative colitis patients in remission versus active disease

Abstract

Objectives: Cytokines and growth factors play a major role in the dysregulated immune response in inflammatory bowel disease (IBD). We hypothesized that significant differences exist between the serum cytokine and growth factor profiles of pediatric IBD patients with active disease (AD) and those in remission, and that levels of some of these soluble mediators may be used to define regulators in IBD and determine disease activity.

Methods: Eighty-eight consecutive patients with confirmed Crohn's disease (CD) and ulcerative colitis (UC) seen at the Duke Children's Hospital were prospectively enrolled and a serum sample was obtained. Data were recorded at the time of serum collection to calculate disease activity indices. The relative expression of 78 cytokines, growth factors, and soluble receptors was determined using proprietary antibody-based protein microarrays amplified by rolling circle amplification. SPSS 8 (SPSS Inc., Chicago, IL) was used to compare protein profiles for CD and UC patients in clinical remission (CR) versus AD.

Results: Sixty-five CD patients and 23 UC patients were enrolled. Forty-one CD patients had available samples and PCDAI results. Twenty-two patients were in remission PCDAI < or = 12.5 (median 5), 19 patients had disease activity >15 (median 30). Univariate analysis revealed that PLGF, IL-7, IL-12p40, and TGF-beta1 cytokine levels were significantly elevated for patients in CR versus AD (p < 0.01). Twelve UC serum samples had Seo/Truelove Witt AI for analysis. Five patients were in remission by TW AI and Seo AI < or =110 and 7 patients had active mild-to-severe disease by TW and Seo AI >110. Only one cytokine, IL12p40, showed significance between CR versus AD (p < 0.02).

Conclusions: Surprisingly, we found no differences in circulating levels of proinflammatory cytokines but found that pediatric IBD patients in remission compared to those with AD had higher levels of specific circulating cytokines, including the regulatory cytokines IL-12p40 and TGF-beta1. It may be that these cytokines directly regulate intestinal inflammation in IBD or reflect the activity of T regulatory cells in negatively regulating the inflammatory response. Further studies will be needed to validate our results to define the molecular pathways involved in the intestinal immune response in man.

Figures

Figure 1

Schematic representation of a sample protein microarray slide with 16 subarrays. Subarrays refer to the 16 wells (A–L and 4 control subarrays), or circular analysis sites, on the slide. Array refers to the antibody content printed in a well. Spots with capture antibodies are deposited in each subarray. The number of spots is for illustration purpose only.

Figure 2

Group separation based on canonical function values obtained by linear discriminant analysis (LDA) for comparing CD patients in clinical remission versus CD patients with active disease. Each square represents one individual. Left panel, prediction model based on four cytokines (TRAIL,TARC,PLGF,SCD23). Right panel, prediction model based on 10 cytokines (BDNF, I-309, IL-17, MCP-1, MPIF-1, PLGF, TARC, TRAIL, SCD23, UPAR). Abbreviations identified in Materials and Methods section.

Figure 3

PPV, sensitivity, cross-validation rate, and prediction rate (saturation capacity) of LDA-based models of comparing CD patients in clinical remission versus CD patients with active disease. X-axis represents number of cytokines used as classifiers; Y-axis represents the percentage value of the corresponding parameter, PPV, cross-validation rate, sensitivity, or saturation capacity. The number of cytokine predictors used in building a prediction model influences the model’s correct prediction confidence level.