FTIR microspectroscopic analysis of the effects of certain drugs on oxidative stress and brain protein structure

Experimental Animals

Four groups (A–D) of 16 3-month-old, male, Wistar albino rats, weighing 300 g, bred at the Cajal Institute, were housed in standard cages. Food and water were made available ad libitum. The temperature was controlled at 22°C and a 12 h/12 h light/dark cycle maintained. All the procedures were conducted in accordance with NIH20 and Spanish21 welfare guidelines. The experimental protocols were approved by the CSIC animal welfare committee. Drugs were administered daily via intraperitoneal injections containing 0.5 ml of the corresponding solution in 0.9% (w/v) NaCl. No behavioral or weight differences were observed between control and experimental animals during normal handling.

The treatments of the different rat groups were as follows: (A) saline (NaCl 0.9%) control solution; (B) nicotine ([−] nicotine hydrogen tartrate salt; Sigma Chem, St. Louis, MO), at a daily dose of 1 mg/kg—equivalent to 0.35 mg of (−)-nicotine free base/kg/day; (C) amphetamine (D(+)-amphetamine sulfate salt; Uquifa Lab, Barcelona, Spain), at a daily dose of 4 mg/kg—equivalent to 3.92 mg of D(+)-amphetamine free base/kg/day; and (D) nicotine and D(+)-amphetamine (both drugs administered simultaneously at the above doses). The nicotine dose administered was the same as that used in previous work showing it to be responsible for dehydrogenase activation and NGF induction, but not the induction of apoptosis.13, 14 We have previous information on the minimum amphetamine dose inducting apoptosis in the hippocampus using a similar treatment schedule as here. Here, the selected dose was 75% lower than in that study in order to prevent apoptosis.

Tissue Handling for Microspectroscopy and Immunohistochemistry for Apoptosis

Two weeks after initiating the experimental treatments (1 h after the last injection), 12 of the animals from each group were deeply anesthetized with 60 mg/kg of i.p. Equitesin (Jansen Lab, Piscataway, NJ) and then transcardially perfused via the aorta with 100 ml of 0.9% (w/v) NaCl followed by 400 ml of a fixative solution containing 4% (w/v) formaldehyde (prepared from paraformaldehyde) in 0.1M phosphate-buffered saline solution (PBS, pH 7.4). Brains were dissected out and postfixed for 3 h at 4°C in the same fixative solution, and then kept overnight at 4°C in 0.1M phospate buffer (PB, pH 7.4).

Six brains were cryoprotected in ascending sucrose solutions (15–30%), rapidly frozen with dry ice, and 12 coronal 15-μm thick serial sections, at the level of 3.60 mm anteroposterior to the bregma in the brain atlas of Paxinos and Watson22 were obtained using a Leitz sledge cryotome at −20°C (Leitz, Germany). The even cuts were placed between two ZnSe or CaF₂ discs and kept at −20°C until FTIR microspectroscopical analyses in the 4000–700 cm⁻¹ and 4000–1000 cm⁻¹ ranges, respectively. The odd cut numbers were processed of habitual way, hematoxylin/eosin stained, to verify the anatomical structures.

The remainder six fixed brains were washed in 0.1M PBS (pH 7.4) for 1 h, dehydrated and embedded in paraffin wax. Serial coronal sections of 5-μm thick at coronal level −3.60 mm of the above atlas were obtained using a Shandon rotary microtome (Shandon, Cheshire, UK). Sections separated by 30 μm were used in the histochemical study of apoptosis. The brain atlas of Paxinos and Watson22 was used to define the anatomical areas. The areas studied were as follows: (a) the frontoparietal cortex (motor and sensory areas) and (b) the hippocampus (CA-1, CA-2, CA-3, and the gyrus dentatus).

Apoptosis

Tissue Handling for Lipid Peroxidation Analysis

The remaining four rats from each group were sacrificed by decapitation, and two anatomical regions were obtained by microdissection: (a) the frontoparietal cortex [the motor and somatosensory areas were taken together from the limits of the gyrus cingularis and retrospenial agranular cortex to the rhinal fissure (upper and lower limits), and from the 1.00 mm to −4.50 mm (anterior and posterior limits)] coronal planes according to the atlas of Paxinos and Watson22 and (b) the hippocampus (CA1, CA2, CA3, and the gyrus dentatus; from the fimbria to the above-mentioned posterior limit). Lipid peroxidation levels at the time of the assay, which are closely related to the production of free radicals, were determined using the Lipid Hydroperoxide Assay Kit manufactured by the Caiman Chemical (MI) based on the redox reactions of the hydroperoxides existing in the tissue with ferrous ions. The preparation of the homogenates, the chloroform/methanol extraction, and the measurement of ferric ions generated during the reaction were performed according to the manufacturer's instructions. The lipid peroxidation levels (mean values ± SD; n = 4) were calculated, and the results for the treatment groups were compared with those of the control group. ANOVA and the Student's t-test were used to analyze the results; significance was set at P < 0.05.

FTIR Microspectroscopic Analysis

Acquisition of spectral images and measurement of spectra were carried out on a Perkin-Elmer Spectrum Spotlight 300 infrared imaging spectrometer, as described elsewhere.19 Briefly, infrared images were acquired with a liquid nitrogen cooled mercury cadmium telluride (MCT-A) line detector composed of 16 pixel elements, which can be operated either at 6.25 or 25 μm/pixel resolution. In this work, we used the 25 μm/pixel spatial resolution. Each absorbance spectrum composing the IR images, and resulting from 10 scans at 4 cm⁻¹ spectral resolution, was recorded for each pixel in the transmission mode. To allow meaningful comparisons, all the infrared data were uniformly pretreated. About 12 spectra in each brain area were selected and further assessed by calculating band area and curve fitting. The Grams/AI software (ThermoGalactic) was used to fit the 1600–1500 cm⁻¹ region to a sum of Lorentzian functions by a nonlinear least-squares procedure, to determine protein β-sheet contents as described elsewhere.19

After atmospheric correction, all the spectra (in the 1000–700 cm⁻¹ region) were converted to their second derivative using an 11-point Savitzky-Golay algorithm to minimize the influence of background artifacts. The resulting spectra in the above region were treated by principal component analysis (PCA). This analysis has been performed on the data at different locations in the spectra to find the independent sources of variation in all spectra and to reduce the number of variables describing the data set. Maps based on principal component scores have been used to find the independent sources associated with controls and drug-treated tissues.

Infrared Spectra from Controls and Drug-Treated Tissues

Reactive oxygen species (ROS) are thought to be produced in living tissues as a result of oxidative stress acting over a long period of time. Accordingly, free radical reactions and the formation of lipid peroxides in tissues lead to age-related damage and eventually to various age-related pathological processes. Oxidative stress has been demonstrated to occur also in response to high doses of substituted amphetamines such as methamphetamine (METH) and 3,4-methlyene-dioxymethamphetamine (MDMA) leading to long-term neurochemical changes.19 Therefore, the spectroscopic measurement of peroxides in brain tissue samples is of interest to determine levels of the oxidative stress. Peroxide determinations can be carried out using various different ways. One of them is the use of the νOH band of hydroperoxides located in the 3500–3000 cm⁻¹ region. Although this procedure is reliable for dried samples, this band is masked by the analogous band of water when studying fresh brain tissues (see Figure 1). Another alternative could be measurement of the CH stretching bands, because oxidation can result in disordering of the hydrocarbon lipid chains and subsequent increase of the corresponding half-band widths.19 However, the νCH bands can be influenced by water to some extent, and their spectral changes are not necessarily produced just by lipid oxidation, but also by conformational transitions of membrane proteins and other membrane-related phenomena. Therefore, in fresh tissue samples it is necessary to find other peroxide bands that are free from water interferences. Only very few studies have been reported on organic peroxides, mainly because of their unstable nature.23 Despite the fact that the OO stretch appears in a region where carbon skeletal modes may interfere, Vacque et al.23 determined it using a narrower range of frequencies for this vibration, i.e., 845–875 cm⁻¹. This has prompted us to carry out PCA for the 900–800 cm⁻¹ region to extract, in a multiple variable system, one, two, or more PCs that carry maximum information. The first one generally represents the maximum variance and discrimination of tissue samples. In conclusion, the spectra of brain tissue samples (see Figure 1) are dominated by bands from water, lipids, and proteins in the region above 1000 cm⁻¹, whereas the bands below this frequency being of weak intensity. This is particularly visible in Figure 2 where an expanded scale is shown in the 1000–700 cm⁻¹ range. Although the 873 cm⁻¹ band appears with weak intensity in the original spectrum (Figure 2A), the height of its second-derivative peak is enhanced (Figure 2B). This band can be assigned to the νOO vibrational mode from peroxides as described earlier.23

The second-derivative analysis of infrared spectra permits direct quantitative analysis of peroxides through this second-derivative peak. In fact, the intrinsic shape of an infrared absorbance is approximated by a Lorentzian function.24 In the second-derivative spectrum, the peak frequency of an absorbance is practically identical with the original peak frequency, but the half-bandwidth is reduced.25, 26 The height of a second-derivative peak is proportional to the original peak height with an opposite sign, and the half-bandwidth of the second-derivative peak is proportional to the original half-bandwidth.25, 26 On this basis, the 873 cm⁻¹ peak height (Figure 2B) relative to that of the 720 cm⁻¹ peak (lipid chain CH₂ rocking) was used here as spectroscopic quantitative evaluation of peroxide content (Table I). The use of the 873 cm⁻¹ second-derivative peak with this aim is supported by its assignment to peroxides on the basis of group frequency considerations,23 and on the fact that the intensity of this peak increases upon amphetamine treatment, which is known to generate higher hydroperoxide levels relative to controls19 (Table I). In fact, PCA of second-derivative spectra in the 900–800 cm⁻¹ region shows the possibility of discriminating amphetamine-treated and control tissues through the two-dimensional (2D) scatter plot displayed in Figure 3. Most part of the control samples are located on the positive part of PC1 axis, unlike amphetamine-treated brain samples that tend to fall on the PC1 negative part. The spectral significance of this negative part of the PC1 is shown in the loading plots (see Figure 4) of this principal component for the score plot of Figure 3. Thus, the most significant contribution to the negative part of PC1 axis, where amphetamine-treated samples tend to be located, stems from the predominant second-derivative peak located at 873 cm⁻¹. This can be explained by considering higher peroxide content in the amphetamine-treated samples located on the negative part of PC1.

Changes in Lipid Peroxidation and Protein β-Sheet Content Induced by Nicotine and D(+)-Amphetamine

On the basis of the above assignment, we have measured the height of the 873 cm⁻¹ second-derivative peak relative to that of the 720 cm⁻¹ peak which is attributable to CH₂ rocking motions in lipid chains. The results are included in Table I and reveal that, compared with the untreated brain samples, the 873 cm⁻¹ intensity is relatively higher in amphetamine-treated brain sections, both in the cortex and hippocampus area. Analogous results have been found through the biochemical procedure for evaluation of peroxides. Interestingly, however, the above intensity increase is attenuated when administering amphetamine plus nicotine (see Figure 5), and the same can be said in relation to hydroperoxide content as determined through conventional biochemical methods. Particularly, in control animals, the lipid hydroperoxide level at the time of the assay in the frontoparietal cortex was double more than that in the hippocampus (Table I). In nicotine-treated rats, only the hippocampus showed a statistical significant increase (P < 0.05) in lipid hydroperoxides. However, in the D(+)-amphetamine-treated group, the two brain regions showed a very statistical significant increase (P < 0.001). When both the drugs were administered simultaneously, lipid peroxidation returned to levels with no statistical significance against the control values.

Oxidative stress manifested by lipid and protein oxidation among other indices is observed in most of the neurodegenerative and neurotoxic disorders, both as a cause and a consequence of the pathogenic pathways. In AD, ROS, the primary mediators of oxidative stress, seem to be either directly or indirectly involved in the production of amyloid β-peptides,27 a keystone in AD. Simultaneously, ROS generated during or after aggregation of these amyloid proteins are putative responsible of the cellular and molecular alterations in the cascade of pathogenic events.28, 29 In a previous article,19 it was suggested that ROS generated by amphetamine-mediated oxidative stress can induce formation of β-sheet rich proteins, which can be of amyloid β-like character. In this work, we have also found that protein β-sheet content increases upon amphetamine treatment. This increase, however, attenuates through combined amphetamine plus nicotine treatment, as occurs also with hydroperoxide levels. Thus, there seems to be a correlation between oxidative stress as reflected by peroxide levels and increasing of protein β-structure in the drug-treated brain tissues. On the other hand, the above results suggest that nicotine may play the role of inhibiting hydroperoxide and ROS formation, whereby inhibiting the appearance of protein β-sheet structure.

The ability of nicotine to regulate the oxidative stress induced by D(+)-amphetamine (and probably by other substances) could be of interest in different pathological processes such as AD, in which oxidative stress is a component of the pathological cascade leading to neurodegeneration. Nicotinergic treatment in AD may also reduce the production of β-amyloid peptides, a putative source of oxidative stress.30 The results obtained with the nicotine and D(+)-amphetamine-treated animals are of great interest. According to unifying theories on the role of psychostimulant drugs,31 this treatment should have theoretically increased the response seen with each treatment on its own. However, this was not the case, as we found, by Western blot analysis, that 50–100% increase in the dose of nicotine or D(+)-amphetamine leads to a more intense induction of cyclooxygenase 2 (protein related to neurodegeneration) than that promoted by the double drug treatment. This suggests the existence of different response mechanisms to these two psychostimulants which are not complementary. One important reason for these unexpected results might lie in the observation that nicotine reduced lipid peroxidation (free radical production or neutralization).