The investigation of the molecular changes during lipopolysaccharide‐induced systemic inflammation on rat hippocampus by using FTIR spectroscopy

The aim of this study is to reveal the molecular changes accompanying the neuronal hyper‐excitability during lipopolysaccharide (LPS)‐induced systemic inflammation on rat hippocampus using Fourier transform infrared (FTIR) spectroscopy. For this aim, the body temperature of Wistar albino rats administered LPS or saline was recorded by radiotelemetry. The animals were decapitated when their body temperature began to decrease by 0.5°C after LPS treatment and the hippocampi of them were examined by FTIR spectroscopy. The results indicated that systemic inflammation caused lipid peroxidation, an increase in the amounts of lipids, proteins and nucleic acids, a decrease in membrane order, an increase in membrane dynamics and changes in the secondary structure of proteins. Principal component analysis successfully separated control and LPS‐treated groups. In conclusion, significant structural, compositional and functional alterations occur in the hippocampus during systemic inflammation and these changes may have specific characteristics which can lead to neuronal hyper‐excitability.


| INTRODUCTION
Infectious diseases may cause systemic inflammation that severely affects the brain with the neurological symptoms, changing people's cognition and behavior.In recent years, both experimental and clinical data have supported that inflammation in central nervous system (CNS) plays a crucial role in the pathogenesis of epilepsy, particularly in the mechanisms underlying the initiation of seizures [1] and involves in the pathophysiology of various brain diseases including Alzheimer's, Parkinson's diseases and Multiple sclerosis [2].
Bacterial endotoxin lipopolysaccharide (LPS), which can be used for the development of animal model of systemic inflammation, triggers a number of biological events including immunological, neurobehavioral, metabolic and endocrine responses [3].These events are collectively called the acute phase response and this is an adaptation strategy developed by the host organism against pathogens [4].Fever or hypothermia, activation of the hypothalamohypophyseal-adrenal cortex and inhibition of spontaneous locomotor activity are among the acute phase responses generated by the brain in LPS-induced systemic inflammation [5,6].
In the literature, it is also possible to find some studies indicating that inflammation exacerbates the results of epilepsy in immature and mature brains.For example, it has been shown that the injection of LPS intraperitoneally (ip) or intracerebroventricularly increased seizure susceptibility by increasing cyclooxygenase (COX)-2, interleukin-1β (I-1β), prostaglandins (PGs) or nitric oxide (NO) in different epilepsy model [3,[12][13][14].Riazi et al. [7] showed that microglia-derived tumor necrosis factor-α (TNF-α) contributes to neuronal excitability and worsens seizures induced by pentylenetetrazol (PTZ), a well-known convulsant agent, following systemic inflammation.There are some suggested mechanisms for inflammation-induced seizure susceptibility in the brain.For example, it has been known that the actions of LPS in CNS are related to the activation of astrocytes, microglial cells and the influx of monocytes into the brain.The activation of astrocytes and microglial cells causes some structural and functional alterations in neuronal networks through the excessive generation and secretion of inflammatory cytokines, such as IL-1β and TNF-α and different types of gliotransmitters [15][16][17][18].It has been suggested that these cytokines and gliotransmitters can induce an increase in neuronal excitability and seizure susceptibility [19].In addition, it has been reported that LPS causes a rapid release of glutamate in the cortex, potentially leading to an imbalance between excitation and inhibition of neuronal circuits [20].
Although systemic inflammation has been known to affect neurochemical structure and alters neuronal networks, the molecular changes accompanying the neuronal hyper-excitability during systemic inflammatory response in the brain are still unknown.In the last decades, Fourier Transform Infrared (FTIR) spectroscopy has appeared as a novel method for monitoring structure, composition and function of tissues and membranes [21][22][23][24][25][26].Variations in the functional groups of biomolecules such as lipids, nucleic acids and proteins can be sensitively determined by using FTIR spectroscopy without introducing any foreign disturbing probes into the system [27].Differences in the band areas/area ratios of spectral bands and bandwidths, shifts of the peak positions provide valuable information about the composition and structure of the biological materials.Since these infrared parameters are sensitive to molecular changes induced by pathological conditions, FTIR spectroscopy has been commonly used in the classification and differentiation of neurological [28,29] and other disease states [24,30,31].It has also been utilized to detect the alterations in the secondary structure of proteins, which is important for the diagnosis of certain neurodegenerative diseases, in aqueous or dry media [27].Although FTIR spectroscopy has previously been employed to assess inflammatory processes in various tissues [32][33][34], to our knowledge, there are no studies reporting the structural and functional alterations in CNS during systemic inflammation by utilizing this technique.
Hippocampus is one of the critical brain sections controlling neuronal excitability during systemic inflammatory response and any changes in its biomolecules are associated with important cognitive, affective and neurological disorders in humans.In addition, it has been known that hippocampus is one of the primary sections of the brain involved in epilepsy [35].Therefore, understanding the molecular changes occur in the hippocampus during the inflammatory response will reveal whether these changes have specific characteristics that cause an increase in neuronal excitability or not.Since inhibiting inflammation might be helpful in reducing the severity of seizures, it is very critical to reveal the molecular changes occur in the hippocampus during systemic inflammation.
As shown in one of our previous studies along with findings from other studies in the literature, a tendency to convulsions was facilitated in rodents a few hours after the administration of LPS.In other words, various convulsant agents administered during this time period can induce convulsions more easily [9,14,36].On the other hand, we also showed that a hypothermic response occurred in the same temporal period after the administration of LPS.This hypothermic response is one of the hallmarks of LPS-induced systemic inflammation in rats [5].Therefore, by considering this hypothermic response as a marker of LPS-induced systemic inflammation, we previously investigated its relationship to convulsions [37].To this end, we investigated the changes in the parameters of PTZ-induced convulsions on different phases of LPS-induced hypothermia including the initial phase, the plateau (including the nadir) and the end of the response in rats.In that study, we showed that the latency of PTZ-induced seizures was facilitated in the initial phase of the hypothermia, corresponding to a decrease of 0.5 C in the body temperature of the animals, by EEG spectra changes [37].Therefore, that study indicated that systemic inflammation induced by LPS may cause a tendency to convulsion at the initial phase of hypothermia.In the light of our previous studies, in the current study, we aimed to determine the changes in the hippocampus accompanying neuronal hyperexcitability during systemic inflammation.The hippocampal molecular changes were determined at the time point that we observed LPS-induced hypothermia and neuronal hyper-excitability state in rats by using FTIR spectroscopy.

| Animals
Animal experimental protocol was approved by the Local Ethical Committee of Ankara University, Faculty of Medicine (2014/13/77).Male Wistar albino rats (8-12 weeks old, 220-280 g) were used.They were kept in a controlled environment with a 12-h cycle of light and darkness at a temperature of 21 ± 1 C and provided with unrestricted access to standard food and water.The animals were separated into two groups by block randomization, each containing seven rats, as a control and LPS-treated group.
A radiotelemetry (Mini-Mitter, Bend, OR, USA) was used to record the abdominal temperature of the rats.A temperature transmitter (VM-FM 3000), whose output signal frequency is regulated by its own temperature, was inserted into the peritoneal cavity of each animal.At least 1 week before the experiment, this procedure was carried out under general anesthesia using xylazine (10 mg/kg, ip) and ketamine (80 mg/kg, ip).The signals emitted by the transmitter were picked up by a receiver (series TR-3000), which was positioned beneath the rats' cages, and were subsequently transferred to a computer via an interface.Seven days after surgery, animals with normal (pre-operative) body weight and no abnormalities were considered healthy and suitable for the experiment.The surgically treated animals, which were housed individually in cages, were placed on radio-signal receivers (series: TR-3000) to transmit data signals to a data matrix which was managed by Vital View software (version 4.1, Mini-Mitter, Bend, OR, USA) [37].Following a stabilization period of approximately 3 h, during which the animals were acclimatized to the laboratory conditions at a temperature of 25 ± 1 C and their body temperature was stabilized, the experiment started.

| Control group
The rats were injected with saline (5 mL/kg; ip).The animals were decapitated 1 h after injection (a time period that corresponds to the induction of hypothermia), the hippocampi of them were extracted and stored at À80 C until the experiment day.

| LPS group
The rats were injected with 250 μg/kg body weight LPS (ip).The animals were decapitated once their body temperature displayed a decrease of 0.5 C, indicating the initial phase of hypothermia, which occurred approximately 40-60 min after the intraperitoneal administration of LPS (the period during which the body temperature begins to fall is considered an indicator of systemic inflammation).The hippocampi of them were removed and stored at À80 C until the experiment day.

| Sample preparation
The hippocampus tissues were incubated in a lyophilizer (Labconco FreeZone ® , Freeze Dry System Model 77 520, USA) for 24 h to get rid of water.Then dried sample was ground in a liquid N 2 containing agate mortar to have a tissue powder. 1 mg of hippocampus powder was mixed with 100 mg of KBr and dried again in the lyophilizer for 24 h to get rid of remaining water.The mixture was exposed to a pressure of 100 kg/cm 2 within an evacuated mold for having a transparent KBr pellet suitable for FTIR studies.

| FTIR spectroscopic study and data analysis
The spectra of the hippocampus tissue were obtained using a Perkin Elmer Spectrum Two FTIR spectrometer (Perkin Elmer, Ltd.UK) with 100 scans in the 4000-450 cm À1 at a resolution of 4 cm À1 .Three different pellets were made from each sample, these pellets were scanned with under the identical conditions.The spectra obtained from these three pellets were averaged and the averages were utilized in data analysis.The analyses of spectra were carried out utilizing Perkin Elmer Spectrum 100 software.The bandwidth values and band positions were calculated from 0.75 Â height of the absorption band on the basis of wavenumbers from raw spectra [21].The baseline correction and normalization processes were applied only for visual demonstration.For the accurate determination of the values of band areas, peak positions, and bandwidths, the original raw spectrum belonging to each individual in the groups was taken into consideration.
The secondary structures of proteins were analyzed utilizing the OPUS NT software (Bruker Optics, Reinstetten, Germany).For this analysis, the second derivative vector normalization method was employed in the Amide I (1700-1600 cm À1 ) region as explained in detail in our previous study [23].
To discriminate the control and LPS-treated group, Principal Component Analysis (PCA) was used by utilizing Unscrambler X 10.4 (Camo, NO) multivariate analysis (MVA) software.This analysis was conducted on the vector normalized second derivative spectra in the 3025-2800 cm À1 region.The results were given as scores and loadings plots [28,38].PCA serves as a highly convenient data reduction method to interpret complex multivariate datasets.In practice each spectrum consists of thousands of absorption values and can be considered as a data point (or vector) within a high-dimensional space, where each coordinate corresponds to one of the variables (wavenumber).PCA accomplishes a linear transformation on this multidimensional space such a way that most of the variations in the original space are preserved in the first few dimensions of the transformed space.This enables more straightforward visualization and identification of spectral clusters that correspond to different classes of samples.The coordinates in the transformed space are referred to as principal components (PCs) and the resulting graphical representation is known as a scores plot.PCs are typically expressed in terms of percentage of variables explained.They are ordered in such a way that the first PC (PC1) has the highest % variation, followed by PC2, PC3 and so on.The transformation matrix is composed of a set of orthogonal vectors referred to as loadings.Loading scaled by the score value of the respective PC provides an indication of that PC's contribution to the variation observed in the original spectrum.Thus, the loadings plot shows the importance of the variables and gives information about which variables provide the highest contribution to the components [39,40].

| Statistical analysis
The significance of the differences between the LPS-treated and control groups was assessed using the Mann-Whitney U test.A p-value less than 0.05 considered statistically significant.When deciding the sample size, we benefited from our previous studies on systemic inflammation [5,14,37] and FTIR spectroscopy studies on the brain [29,41,42].The minimum number of rats that should be used in each experimental protocol was seven in order to detect a 0.5 C body temperature change under our experimental conditions, with a 5% type I error and a 20% type II error probability.In addition, to justify the sample size, a post hoc power analysis was carried out based on the obtained quantitative spectral results.Power values ranging from 73% to 99% were obtained.These results imply that the sample numbers used in each group are sufficient to draw conclusions about the findings of the present study.Power values for each analysis was calculated using G*Power 3.1 [38].

| RESULTS
In this study, the hippocampal molecular changes accompanying neuronal hyper-excitability observed after LPS administration were investigated by using FTIR spectroscopy.Figure 1 depicts an FTIR spectrum obtained from the hippocampus tissue in the 4000-450 cm À1 wavenumber range.In this figure, the essential infrared bands are marked and the general band assignment is presented in Table 1.As can be seen from Figure 1 and Table 1, the hippocampus gives a complex spectrum which contains various bands belonging to various functional groups of biomolecules.These functional groups belong, for example, to nucleic acids, proteins and lipids.
Figure 2A-C displays the spectra of control and LPSadministered rat brain hippocampus in the regions of 3800-3025, 3025-2800 and 1800-900 cm À1 .As seen from these figures, there are significant variations in the spectral parameters of the control and LPS-administered An infrared spectrum of the control rat brain hippocampus in the 4000-450 cm À1 region.
groups.The numerical comparisons of the band area values, band area ratios, peak positions and bandwidths of FTIR bands associated with lipids, proteins and nucleic acids are given in Figures 3-5, respectively, as bar graphs.
The area under the infrared bands originating from particular species is directly proportional to the concentration of those molecules [47].Therefore, the band areas were analyzed in this study to obtain relative information about the changes in the concentration of biomolecules in hippocampus tissue after LPS treatment.In addition, in order to remove any artifacts that may arise from the experimental conditions, such as sample thickness, band area ratios were also utilized to obtain information about the variations in the concentrations of the molecules [29,48].The band area ratios employed in the current study, the functional groups utilized to obtain these ratios and the information provided by the calculated ratios are given in the Table 2.
The areas of the infrared bands in the 3025-2800 cm À1 range (Figure 2B) were employed to detect changes in lipid content since the protein band at 2871 cm À1 is a weak band.In this wavenumber range, the olefinic=CH band (3014 cm À1 ), provides information  about unsaturated lipids while the CH 2 antisym.
(2923 cm À1 ) and sym.str.(2851 cm À1 ) bands provide information about saturated lipids [22].As can be seen from Figures 2B and 3A,B, the area under the olefinic=CH band and olefinic=CH/saturated lipid ratio, known as lipid peroxidation index, increased significantly (29.5% in olefinic=CH band, 12% in olefinic=CH/saturated lipid ratio).These increases suggest an increase in the unsaturated lipid content in the LPS-treated hippocampus.It has been known that lipid peroxidation takes play primarily at the double bond sites of polyunsaturated fatty acids and after lipid peroxidation, the amount of unsaturated lipids decreases.Consequently, lipid peroxidation causes a reduction in olefinic bonds [49].However, in this study, we determined an increase in the olefinic groups instead of a decrease and we attributed this increase to the lipid peroxidation end products which include olefinic groups [44].This finding suggested that the loss of unsaturated lipid during lipid peroxidation reactions was compensated by double bonds present in the lipid peroxidation end products [44].
As seen from Figures 2B and 3C in CH 2 sym.band) indicating an increment in the saturated lipid amount in the LPS-treated tissue.Changes in the concentration of saturated lipids were also examined by obtaining the CH 2 antisym.str./CH 2 sym.str.band area ratio [48].As can be seen from Figure 3E, this ratio significantly increased in the LPS-administered group (0.7%).This observation was also supported by the results obtained from the examination of the 1800-900 cm À1 region (Figure 2C).A statistically significant increment in the area of the carbonyl ester str.band, appearing at 1738 cm À1 (16%), was detected in the LPS-treated group implying an increment in the concentration of carbonyl groups in triglycerides or cholesterol esters in this group (Figure 3F) [29,43].The ratio of the carbonyl ester/ saturated lipid also increased (13.5%) in the LPS-treated group supporting an increment in the carbonyl amount in the system (Figure 3G).Furthermore, the area under the CH 2 bending band (at 1461 cm À1 ), that also arises from saturated lipids [43], increased significantly (17%) in LPS-treated group (Figure 3H).The CH 2 /CH 3 and CH 2 /saturated lipid ratios were employed to obtain information about the chain length and the amount of the CH 2 groups of the saturated lipids, respectively, in the hippocampus tissue (Table 2) [24,48].As seen from Figure 3I,J, these ratios decreased (1.4% in CH 2 /CH 3 ratio, 9% in CH 2 /saturated lipid ratio) in the LPS-treated tissue.The PO 2 À antisym.(1236 cm À1 ) and PO 2 À sym.
(1074 cm À1 ) str. bands give information about phosphate containing molecules such as phospholipids and nucleic acids.Therefore, significant increases (16% in PO 2 À antisym.band, 14% in PO 2 À sym.band.)detected in the areas of these bands (Figure 3K,L) indicated that there is an increment in the phospholipid and nucleic acid amounts in the LPS-treated group [41,43].
In this study, the changes in the wavenumbers and bandwidths of the CH 2 antisym.str.band were also analyzed since these changes reflect the variations in the order parameters and fluidity of the lipid acyl chains, respectively [22].As can be seen from Figure 3M, the peak position of this band significantly shifted to higher wavenumbers in the LPS-administered group.This shifting indicates the presence of more disordered lipids in the system after LPS treatment [48].A significant increment detected in the bandwidth of this band shows that there is an increment in the dynamics of lipids in the LPS-treated hippocampus (Figure 3N) [29].
Increases in the areas of Amide I and II bands, which are a well-known protein bands (Figure 4A,B), as well as a weak CH 3 sym.str.band (Figure 4C) were recorded in the LPS-administered group (21% in Amide I band, 20% in Amide II band, 18% in CH 3 sym.str.band).These findings show that there is an increment in the concentration of proteins in the LPS-administered group.Moreover, significant increases determined in the area values of the Amide A band (3370 cm À1 ) which arises mainly from N H str. modes of proteins and Amide B band (3069 cm À1 ) which arises from N H and C H str. modes of proteins show an increment in the protein content after LPS treatment (19% in Amide A band, 18% in Amide B band) (Figures 2A and 4D,E).Similarly, an increase (16%) detected in the area under the Amide III band (1306 cm À1 ) which is due to mainly proteins, in the LPS-administered group supported the increment in the concentration of proteins in the system (Figure 4F) [41].Variations in the protein amounts were also tested by calculating the ratio of Amide I/Amide I + Amide II [48].As seen from Figure 4G, this ratio increased (2.5%) after LPS treatment.In addition, the area under the COO À sym.str.band (1395 cm À1 ), dominantly originating from amino acids and fatty acids, increased (17%) in the LPS-treated group (Figure 4H) [41].This increase supported the results that there is an increment in the amounts of both proteins and lipids.
To understand the variations in the conformation and structure of proteins, the variations of the peak position and bandwidth of Amide I band and in the ratio of Amide I/Amide II were analyzed [46,48].As seen from Figure 4I,J, the peak position of Amide I band shifted to higher values (from 1653.74 ± 0.40 cm À1 to 1654.20 ± 0.266 cm À1 ) and the bandwidth of this band decreased in the LPS-treated group.Furthermore, Amide I/Amide II ratio was elevated significantly in the LPS-administered group (0.8%) (Figure 4K).The increase in the Amide I/Amide II ratio, the narrowing and shifting in the Amide I band to higher wavenumbers could be interpreted as the result of the changes occurred in the conformation and structure of proteins in LPS-treated hippocampus tissue.The ratio of saturated lipid/protein was evaluated to compare the relative alterations in lipid and protein metabolism.As can be seen from Figure 4L, this ratio decreased significantly in the LPS-administered tissue (2.2%).
The infrared bands appeared in the 1300-900 cm À1 are mainly due to various bands of several functional groups of macromolecules such as nucleic acids, phospholipids, and carbohydrates.As depicted in Figure 5A,B, significant increases were detected in the areas of the main nucleic acid bands appeared at 973 cm À1 , which arises from mainly C N + C str. vibrations of nucleic acids (12.5%), and at 927 cm À1 , which is due to z-form DNA (21%).These results showed that there is an increase in the nucleic acid amount in LPS-treated group [22].The increment noted in the amount of nucleic acids was supported by the increases detected in the areas of phosphate antisym.and sym.str.bands (Figure 3K,L), which give information about the amount of nucleic acids as well as phospholipids.Moreover, the peak position of the C N + C str. band shifted to a lower value showing that there are some conformational changes in the nucleic acids (Figure 5C).The ratios of nucleic acid/protein (7%) and DNA/protein (15%) increased in the LPS-treated group (Figure 5D,E).This finding shows an increment in the nucleic acid amounts and the production of RNA and DNA in the system [29].
In order to have more detailed information about the changes on the secondary structure of proteins, the Amide I band range was analyzed by measuring the intensity of the second-derivative spectra.Figure 6A,B shows the peaks under Amide I and II bands in second derivative spectra for control and LPS-treated groups and the changes in the intensities of characteristic components of Amide I mode, respectively.As seen from Figure 6A, the second derivative spectra of proteins exhibit distinct peaks.The peak at around 1684 cm À1 is associated with beta turns, the peak at around 1658 cm À1 is due to unordered alpha helix, the peak at around 1655 cm À1 is assigned to alpha helix, the peak at around 1648 cm À1 is associated to random coil, the peak at 1638 cm À1 is due to native beta sheet, and the peak at 1629 cm À1 is associated with aggregated beta-sheet structures [46,[50][51][52][53].As seen from Figure 6B, there is a significant reduction in the intensity of turns and a significant increase in the intensity of random coil structures.The increase in random coil structure may indicate protein denaturation [41].
Ultimately, to discriminate the control and LPStreated group, PCA was applied to the infrared spectra.Since 3025-2800 cm À1 gave the best discrimination of control and LPS-treated samples, this spectral region was used for PCA. Figure 7 shows the scores and loadings plots of PCA.As seen from Figure 7A, control group located on the positive part while LPS-administered group located on the negative part of the PC1 axis and a considerable maximum variation value is observed (PC1 + PC2 = 98%).The loadings plot indicated a high dissimilarity between PC1 and PC2 and supported the presence of differentiation between control and LPS-treated groups (Figure 7B).Thus, according to the PCA results LPS-treated and control groups were successfully segregated from each other with a high accuracy of success.
T A B L E 2 FTIR band area ratios utilized in this study, the functional groups used to obtain these ratios and their assignments [29,48].spectroscopy studies [22,54,55].Since dried samples don't precisely display actual biological systems, which are in aqueous solutions, the analyses with dried tissues are not convenient for making quantitative assessments, but instead they could be utilized to gather relative information.In this study, it may not cause an important problem because the essential concern was to find out the relative alterations between molecular structure and composition of two identical preparations of un-treated and LPS-treated tissues.Furthermore, in the current study, free and unbound water was eliminated from the system during the drying process while intra-and intermolecular water, which is necessary for the stability of biomolecules such as lipids, was retained in the system [54,55].It has been previously demonstrated that the structure and dynamics of lipid bilayers are controlled by this inter-bilayer water [56].

Ratio
In the current study, a significant increase was detected in the amount of unsaturated lipids and shortchained lipid peroxidation end products in the LPS-treated hippocampus.A similar effect was also observed in our previous FTIR study on the effects of PTZ-induced convulsions on rat brain tissue [29].Consistent with the previous studies, the higher carbonyl ester/ saturated lipid and lower CH 2 /saturated lipid and CH 2 / CH 3 ratios detected in the LPS-treated group showed that lipids were broken down into smaller fragments which include more carbonyl and less CH 2 groups [43,57].This result supported the conclusion that the shorter chained lipid peroxidation end products accumulated following the injection of LPS in the hippocampus tissue.It has been known that free radicals, which are the most important reasons of lipid peroxidation, are included in the formation of inflammatory response.The microglias, astrocytes and neurons, which are activated during any inflammatory response occurring in the CNS, secrete cytokine, chemokine, NO, and reactive oxygen species (ROS) [58].In some previous studies, high levels of radical production and increase in lipid peroxidation level have been reported in various regions of brain during systemic inflammation after LPS application [59][60][61].It has also been previously shown that several inflammatory mediators, including TNF-α, I-1β, NO, and ROS generated by microglia and astrocytes, increase Principal component analysis scores (A) and loadings (B) plots for infrared spectra of hippocampus tissues of control and lipopolysaccharide-administered groups in the 3025-2800 cm À1 region.
neuronal excitability, and facilitate the epileptic process [19,62].Hence, according to our results, the inflammatory medium in the hippocampus lead to the formation of free radicals, these radicals induced lipid peroxidation and this process might be related to an increased neuronal hyper-excitability.
A significant increase detected in the saturated lipid amount after LPS treatment showed that LPS caused an alteration in the metabolism of lipids in the hippocampus tissue.Some lipid mediators, such as platelet-activating factor (PAF) and PGs, are also released during inflammatory response.The PGs are lipid mediators produced by the action of COX-1 and 2 on arachidonic acid (AA), which is an unsaturated fatty acid.These enzymes are located mainly in astrocytes and microglia.Marked induction of COX-2 and increased biosynthesis of PGs or PAF were reported in the brain during inflammation [63].Thus, the accumulation of lipids might be resulted from the increased amount of these lipid mediators, probably shorter chained lipids, in the brain during inflammation.Impaired lipid metabolism has been reported in the pathogenesis of various neurological diseases including epilepsy [64].It has also been reported that prostaglandin E2 (PGE2) stimulated its receptor (EP3) on astrocytes, increased astrocytic glutamate release, and induced hyper-excitability [65].The other pro-inflammatory lipid mediator, PAF, has also been shown to cause neuronal hyper-excitability by stimulating glutamate release and activating COX-2 gene expression [66].It has been known that PGs increase markedly after seizures and they may participate in epileptogenesis and decrease seizure threshold.It has been shown by Cole-Edwards and Bazan [67] that epileptic seizures cause the accumulation of fatty acids derived from membrane lipids in the synapses.Therefore, in accordance with the literature, our findings indicated that inflammation causes the accumulation of lipids by inducing the generation of lipid mediators in the brain and these lipid mediators might contribute to decrease seizure threshold.
Due to the associations occurring between membrane lipids and proteins, physical features of these molecules may affect mutually their conformation and consequently, their activities [42].For example, changes in the structure of membrane lipids cause alterations in the activity and kinetics of membrane proteins, including crucial membrane channels that play important roles in neuronal activity.Therefore, it is very crucial to maintain the order and fluidity of membrane at optimum levels for the proper functioning of these membrane proteins.In this study, a decrement in lipid order and an increment in the lipid dynamics were observed in the hippocampus tissue after LPS treatment.Moreover, saturated lipid/ protein ratio, which is another parameter influencing the dynamics and structure of the membranes, significantly decreased in the hippocampus tissue after LPS treatment.Changes in the ratio of saturated lipid to protein indicate alterations in the metabolism of lipids and proteins in the tissue.Additionally, this ratio provides information about lipid and/or protein asymmetry, which is crucial for cellular functions [68].It has been known that the alterations occurred in the lipid asymmetry induce important changes in the intra and inter-cellular ion amounts and therefore may cause some changes in the membrane potential by disturbing the kinetics of ion channels [69].Various types of membrane proteins, including receptors (e.g., TNF receptor 1, IL-1 receptor), ion channels (e.g., voltage-gated Na, K, Ca, Cl channels) and membranebound enzymes (e.g., COXs) are associated with the generation of inflammatory response in the brain [2,63].It has been known that cytokines, which can modulate the ion channels, regulate neuronal activity not only by stimulating the release of some inflammatory molecules but also by activating their receptors expressed by neurons [70].On the other hand, excessive production or prolonged exposure to these cytokines may lead to some pathological consequences, such as epilepsy [1,8].For example, transient epileptic episodes have been observed in infectious diseases and it has been suggested that it was induced by excessive release of cytokine from activated microglia.In addition, it has been reported that epileptiform neuronal excitability might be stimulated by alterations in the intrinsic neuronal excitability controlled by the membrane density of voltage-gated Na + channels.Two important cytokines, TNF-α and IL-18 released from microglia, are activated by inflammation, up regulate sodium current density and in this way excitability in hippocampal neurons [2].Therefore, it can be suggested that alterations in the order, fluidity and lipid/ protein ratio may result in significant changes in the activity of important membrane channels which govern neuronal excitability.
The increase observed in protein amount may be resulted from the elevated levels of the inflammatory mediators within the protein structure, which are increased in the brain as a part of immune response initiated against the stimuli during systemic inflammation.These well-known inflammation mediators can be categorized as proinflammatory enzymes (COX, NOX, iNOS), cytokines (IL-1β, TNF-α and IL-6), growth factors (TGFβ) and brain-derived neurotrophic factor (BDNF) [71].Furthermore, LPS induces the rapid release of glutamate which is the principal excitatory amino acid neurotransmitter in the CNS [20].The increase in protein concentrations was also supported by the increment observed in the ratio of nucleic acid/protein, which shows an increase in nucleic acid production, confirming elevated level of protein synthesis.Previously, it has been found an increase in mRNA levels for TNFα, IL-1, NF-kB, COX-2, iNOS, and NMDA receptor subunit after LPS injection [72][73][74].Upon LPS administration, the increase observed in the areas of the main and complementary nucleic acid bands and the area ratios of nucleic acid/protein and DNA/protein were also due to an increase in the amounts of nucleic acids [22].As argued above, in the literature, there are evidences for an increase in the number of receptors (up regulation) of many mediator protein genes such as TNF-α, IL-1, IL-2, HMGB1, and so forth during the process of adaptation to systemic inflammation.In addition, the shifts detected in the wavenumber of C N + C str. band show that there is an alteration in the conformation and structure of the nucleic acids in the LPS group.The increases observed in the amounts and the alterations in the conformations of nucleic acids may be because of the increased level of replication, transcription and translation activity in LPS-treated group.Although the DNA and RNA amount are constant in the same tissues of individuals of the same species under normal conditions, in some special circumstances such as, increased metabolic activity, cell proliferation or in a pathological condition, the amount of nucleic acids may change.In the previous studies, it has been reported that the increased DNA signal can be attributed to differences in the morphology, organization and architecture of the cell nucleus.These differences include an increased nuclear/cytoplasm ratio, hyperchromicity, chromatin aggregation, cell division, and possibly decreased DNA condensation [75].It has been known that the absorbance of DNA bands varies depending on the stage of cell division and there is an increase in the absorbances of the DNA bands in actively dividing or metabolically active cells [76,77].Similarly, the strength of RNA bands may vary significantly, depending on the cell's activity.In actively dividing or metabolically active cells, RNA spectral features, most likely originating from ribosomal RNA, were found in the cytoplasm [78].After LPS treatment the increases appeared in the amounts (due to cell proliferation and migration to the hippocampus) and in the rate of protein synthesis of microglias and astrocytes might have caused increases in the amounts of nucleic acids in the rat hippocampus [79,80].In previous studies, it has been suggested that some inflammatory cytokines, such as IL-1β, TNF-α, or IL-6, can directly increase neuronal excitability and lead to the development of spontaneous seizures or epileptogenesis [18].Indeed, it has been found that transgenic mice with increased expression of these inflammatory cytokines showed an increase in seizure intensity and elevated levels of these inflammatory factors were also detected in both in patients and animals suffering from seizures [62].These findings showed that LPS-induced inflammation in the brain affects the level of gene expression in a similar manner of increased neuronal excitability.
In this study, the observed increase in the Amide I/Amide II ratio, variation in bandwidth and shifting in the Amide I band could be interpreted as the results of changes in the conformation and structure of proteins in LPS-treated hippocampus tissue.In addition, a significant increase observed in the random coil content suggests a denaturation of proteins which may affect protein function [81].The disorder of protein structure has been observed in certain pathological conditions and protein misfolding and aggregation are indication of a number of neurological diseases [29].For example, it has been shown that ROS cause protein oxidation in different epileptic models [82][83][84] and the protein oxidation causes functional alterations or deactivation of some important enzymes [85].These structural and functional changes occurred in enzymes may be involved in the increase of neuronal hyper-excitability.Consistent with our results, it has been reported that the astrocyte response to inflammation which enhances susceptibility to epileptic seizures causes some conformational alterations in some enzymatic (e.g., glutamine synthetase) and cytosolic (e.g., glial fibrillary acidic protein) proteins.
According to the PCA results applied to the spectral data, LPS-treated and control groups were successfully differentiated from each other with a high accuracy of.Two different clusters in the PCA corresponding to control and LPS-treated group confirmed that LPS administration induced significant changes in the hippocampus in terms of structure, composition and function of tissue components.
In summary, we found that systemic inflammation caused an increase in the concentrations of lipid peroxidation end products, saturated lipids, proteins, and nucleic acids, a decrease in membrane order, an increase in membrane dynamics and protein denaturation in the hippocampus.These changes might be due to microglia and astrocytes being activated and increased in number in the hippocampus during systemic inflammation.For example, lipid peroxidation and protein denaturation might be induced by high amounts of oxygen and nitrogen-free radicals produced by these cells during systemic inflammation.The increase observed in the amount of saturated lipids might be due to lipid mediators such as PAF and PGs, which are released in high amounts during inflammatory response.Similarly, the increase in the amount of proteins might be due to increased biosynthesis of protein mediators such as some proinflammatory enzymes, cytokines, growth factors, and BDNF in the hippocampus during systemic inflammation.The increased amounts of nucleic acids may be attributed to the increased rate of cell proliferation and protein synthesis of microglia and astrocytes in the hippocampus after LPS administration.Changes in membrane order and dynamics resulting from the increased amounts of cytokines during the inflammatory response might cause important alterations in the activity of some ion channels, receptors and enzymes in the membrane, which govern neuronal excitability.As argued above in detail, similar changes were also reported in the hippocampus when neuronal excitability increased in the previous studies.Therefore, these findings emphasize the potential significance of inflammation in the context of increased neuronal excitability.

| CONCLUSIONS
In this study, for the first time, we reported the molecular changes in the hippocampus during the LPS-induced systemic inflammation by taking the advantage of FTIR spectroscopy.Our results indicated that an inflammatory response occurs in the hippocampus as a part of systemic inflammation and significant changes take place in the structure and composition of lipids, proteins and nucleic acids during the inflammatory response.These changes were in the direction of explaining the increased neuronal excitability we had previously observed in the same time period.
In conclusion, the results of the current study revealed that changes occurring in the hippocampus during the systemic inflammatory response observed after LPS injection may have specific characteristics which can lead to neuronal hyper-excitability.This study also showed that FTIR spectroscopy could be used as a fast and sensitive technique to monitor structural and functional changes induced by a toxic substance in the brain.
DNAF I G U R E 2 The average Fourier transform infrared spectra of control and lipopolysaccharide-administered rat hippocampus in the (A) 3800-3025 cm À1 , (B) 3025-2800 cm À1 , and (C) 1800-900 cm À1 region.The spectra were normalized with respect to the Amide I band (A and B) and to the Amide A band (C).
,D, the areas under the saturated lipid bands (CH 2 antisym.and sym.bands) increased significantly (18.5% in CH 2 antisym.band, 20% F I G U R E 3 Alterations in the band area, band area ratio, peak position and bandwidth values of functional groups related to lipids in the control and lipopolysaccharide-administered groups.(The values represent mean ± standard deviation.*p< 0.05 represents the degree of significance.)

F I G U R E 4
Alterations in the band area, band area ratio, peak position, and bandwidth values of functional groups related to proteins in the control and lipopolysaccharide-administered groups.(The values represent mean ± standard deviation.*p< 0.05 represents the degree of significance.)

F I G U R E 5
Alterations in the band area, band area ratio and peak position of functional groups related to nucleic acids in the control and lipopolysaccharide-administered groups.(The values represent mean ± standard deviation.*p< 0.05 represents the degree of significance.)

4 |
DISCUSSIONSince we observed neuronal hyper-excitability in the brain at the 0.5 C drop point in the body temperature of the animals after LPS injection, we selected this specific time point to examine the brain hippocampus.We utilized dried hippocampus tissue in the experiments as it has been extensively used previously in FTIR F I G U R E 6 (A) Representative second derivative spectra of control and LPS-administered groups in the region between 1700 and 1500 cm À1 .The numbered peaks are attributed to: 1, beta turns; 2, unordered alpha helix; 3, alpha helix; 4, random coil; 5, native beta sheet; 6, aggregated beta sheet.(B) Alterations in the intensities of characteristic components of Amide I mode for control and LPS-administered groups.(The values represent mean ± standard deviation.*p < 0.05; **p < 0.01 represent the degree of significance.) [22,29,[43][44][45][46]ns of hippocampus tissue[22,29,[43][44][45][46].