Ernest V. Curto

Regulation of gene expression in inflammatory bowel disease and correlation with IBD drugs. Screening by DNA microarrays

Potential biomarkers for Crohns disease (CD) and ulcerative colitis (UC) were identified from two sets of full thickness pathologic samples utilizing DermArray® and PharmArray® DNA microarrays relative to uninvolved (Un) colon or normal colon. Seven of the over-expressed genes were verified using quantitative RT-PCR (i.e., TMPT, FABP1, IFI27, LCN2, COL11A2, HXB, and metallothionein). By correlating gene expression profiles between inflammatory bowel disease (IBD) tissue samples and IBD drug-treated cell cultures it might be possible to identify new candidate molecular target genes for IBD therapy and drug discovery. Potential biomarkers for CaCo2 cell cultures, which are routinely used as a GI tract surrogate model for in vitro pharmacokinetic studies, treated with azathioprine, 5-aminosalicylic acid, metronidazole, and prednisone were also identified from another experiment. Metallothionein mRNA expression was found to be down-regulated in azathioprine-treated CaCo2 cells, and was coincidentally up-regulated in the CD sample, thus resulting in an anti-correlation. These results suggest that this new screening methodology is feasible, that metallothioneins might be biomarkers for azathioprine therapy in vivo in CD, and that azathioprine might mechanistically down-regulate metallothionein gene expression. Correlations were also observed between IBD samples and either metronidazole- or 5-aminosalicylic acid-treated CaCo2 cells. Similar comparisons of disease tissue samples in vivo vs drug-treated cell cultures in vitro might reveal new mechanistic insights concerning established or experimental drug therapies. This affordable in vitro methodology is promising for expanded studies of IBD and other diseases.

A method to improve selection of molecular targets by circumventing the ADME pharmacokinetic system utilizing PharmArray DNA microarrays.

DNA microarrays may be used to identify potential molecular targets for drug discovery. Yet, DNA microarray experiments provide massive amounts of data. To limit the choice of potential molecular targets, it may be desirable to eliminate genes coincidentally up-regulated in tissues implicated in absorption, distribution, metabolism, and excretion (ADME) pharmacokinetics. DNA microarray experiments were performed to demonstrate a gene-exclusion approach using as an example RNA samples of neural origin, i.e., a human neuroblastoma cell line (SK-N-SH) and brain tissue, as the intended hypothetical site(s) of drug action. Biomarkers were identified using PharmArray DNA microarrays. The lists of neuroblastoma and neural biomarkers were constrained by limiting selection to the subset of genes that were not highly expressed in three transformed cell lines from liver, colon, and kidney (HepG2, Caco-2, and 786-O, respectively) that are routinely used as representatives of the ADME system during in vitro pharmacology and toxicology experiments. Principal component analysis methods with likelihood ratio-related bioinformatic tools were utilized to identify robust potential biomarker genes for the three ADME-related cell lines, neuroblastoma, and normal brain. Biomarkers of each sample were identified and selected genes were validated by qRT-PCR. Hundreds of biomarkers of the three ADME-related cell types, representing hepatocytes, kidney epithelium, and gastrointestinal tract, may now be used as a valuable database to restrict selection of biomarkers as potential molecular targets from the intended samples (e.g., neuroblastoma in this work). In addition to biomarker discovery per se, this demonstration suggests that our model method may be viable to help restrict gene lists during selection of potential molecular targets for subsequent drug discovery.