Page 88 - Vibrational Spectroscopic Imaging for Biomedical Applications

64 Cha pte r T h ree

16
composition. 16,17 Using this method, Andrus and Strickland report an
−1
increasing ratio of peak areas corresponding to bands at 1121 cm and
−1
1020 cm (attributed to RNA/DNA), which were associated with
increased aggressiveness of malignant non-Hodgkin’s lymphomas.
17
Takahashi et al. used bulk tissue analysis to study glycogen levels in
tissues obtained from colorectal tumours, regions adjacent to the
tumour and regions of normal colorectum. The results indicated that
there was a statistically significant difference in glycogen levels (Peak
−1
−1
area ratio 1045 cm /1545 cm ) between cancer tissue and the other
two regions. 17
The use of ground tissue provides indiscriminate and composite
measurement of both epithelial and stroma tissue compartments.
However, this type of analysis must be treated with caution when
molecular assignments are made for discriminatory bands. In the
16
study by Andrus and Strickland, the influence of collagen absor-
bance was discussed; however, other confounding variables exist in
stroma tissue, namely, a variety of cell types such as endothelial cells
of blood vessels, fibroblasts, ganglion, and erythrocytes in addition to
possible bisecting nerves and muscle. This was clearly demonstrated
18
by Fernandez et al. who classified several prostate tissue compo-
nents for diagnostic FTIR imaging.
19
For microspectroscopic studies, a review by Faolain et al. reveals
that a number of approaches have been made to prepare tissues for
analysis that include fresh, frozen, air-dried, formalin-fixed, and
deparaffinized formalin-fixed tissue sections. A number of papers
have been published that compare the effects of these sample prepa-
ration protocols on Raman and FTIR tissue spectra 20–25 and these are
discussed in the following text.
Fresh and Cryo-Preserved Tissue
20
More than a decade ago, Shim and Wilson demonstrated that dehy-
dration at room temperature of fresh tissue specimens (subcutaneous
fat, smooth muscle, cheek pouch epithelium, esophagus) resulted in
−1
Raman spectra with a decrease in intensity of the 930 cm (C⎯C
stretch of proline and valine) peak relative to the peaks at 1655 cm −1
−1
and 1450 cm . Although this may be indicative of protein denatur-
21
ing the authors did not observe any shifts in the amide I peak. How-
ever, an increase in the lipid-protein signal was observed with
increasing drying times providing evidence that the protein vibra-
tional modes were perturbed by dehydration.
20
Interestingly, Shim and Wilson found that Raman spectra obtained
from OCT-freeze stored, snap-frozen tissue in phosphate-buffered
saline (PBS) were comparable to spectra obtained from fresh tissue in
−1
PBS. Additional peaks observed at lower frequency (764 cm and
−1
795 cm ) in the spectra of snap-frozen adipose tissue, were attributed
19
to the coagulation of erythrocytes. Faolain et al. also conducted a
comparative study of frozen and fresh tissue using parenchymal tissue