Evaluation of premalignant lesions of rat colon by [1]H nuclear magnetic resonance and infrared spectroscopy

Thumbnail Image
Krupnik, Eduardo
Journal Title
Journal ISSN
Volume Title
1H magnetic resonance spectroscopy (1H MRS) and Fourier transform infrared (FTIR) spectroscopy are being widely used to study the biochemical changes associated with cancer. In particular, based upon the hypothesis that biochemical changes associated with cancer precede morphological manifestations of the disease, these spectroscopic techniques are being evaluated as potential diagnostic and prognostic tool. In the current study, 1H MRS and FTIR spectroscopy were applied to the study of colonic tissue from rats treated with the specific colon carcinogen azoxymethane (AOM) to determine whether tumor induction was associated with identifiable spectroscopic changes in the colon. In particular the attention was centered on aberrant crypt foci or ACF, small lesions of the colon, which are hypothesized to represent an early preneoplastic stage in the adenoma-carcinoma sequence. The 1H MRS analysis of control mucosa taken from healthy male Sprague Dawley rats, ACF, colonic mucosa and tumor samples taken from AOM treated rats revealed that the peak intensities and areas of ACF lie between those from normal and normal-appearing carcinogen-treated mucosa samples and tumors. Application of pattern recognition algorithms such as linear discriminant analysis (LDA) allowed non-subjective classification of the spectra into four groups with a high degree of accuracy (81.5%). The results place the ACF mainly as an independent group, but with 1H MRS features that put them between those of the normal mucosa samples and the tumors, supporting the hypothesis that ACF represent an intermediate step in colon carcinogenesis. For the FTIR analysis, and due to the heterogeneity of the colon, we first studied the mid-infrared spectroscopic characteristics of cryostat sections of the different layers of the colon with the aid of an infrared microscope. The spectra of the muscularis, submucosa and mucosa were significantly different, allowing identification of mucosal tissue with relative ease. The spectroscopic characteristics of healthy mucosa, AOM-treated mucosa, and tumors were then studied. Spectra of the three tissues were similar. However, subtle differences in the absorbance bands assigned to nucleic acids were seen, indicative of alterations in the structure of the nucleic acids. Spectra indicate a progressive alteration in nucleic acid structure from normal mucosa through treated mucosa to tumors. LDA allowed non-subjective classification of spectra as arising from normal mucosa, treated mucosa and tumors with an accuracy of 85%. Similarly, LDA allowed classification of spectra of AOM-treated mucosa, ACF and tumor biopsies with an accuracy of 79.2%. Misclassification of spectra arising from ACF as arising from both tumor and mucosa suggests that the ACF exhibit biochemical characteristics intermediate between the control and AOM-mucosa samples and the tumor groups. Several biochemical characteristics of the ACF, detected by 1 H MRS and FTIR, suggest that these lesions are an early event in the development of colon carcinogenesis. The application of pattern recognition algorithms to classify the different classes of colon tissues supports this conclusion, since ACF samples were grouped between the normal tissue samples and the tumors. Given the importance of early diagnosis, the multidisciplinary approach presented in this thesis could serve as an important complement to the routine histopathological assessment for a better and more efficient diagnosis of colorectal cancer.