The incidence of gastrointestinal diseases doubles every decade, both in Russia and globally, posing a serious concern to public health and society [1]. Although there is great variability in the epidemiological data across studies, an increase in the incidence of ulcerative colitis (UC) and Crohn’s disease (CD), especially in young, socially active and employable individuals, is being reported by many authors [2, 3]. Worldwide, there are 3 to 62 and 50 to 70 incident cases of UC and CD, respectively, per 100,000 population a year [4]. Currently, as many as 1.6 million people are suffering from UC and CD in the USA [5, 6]. UC and CD are also associated with a variety of intestinal and extraintestinal complications that add to hospitalization, surgery and outpatient care costs.
The pathogenesis of UC and CD is not fully clear: there are a multitude of etiological factors that can activate cell-mediated and humoral components of the immune system by eliciting Тh1 and Тh2 immune responses and cause an imbalance between Тh17 and T_reg cells [7]. Just like cytokines, enzymes, immunoglobulins, and some other factors, reactive oxygen and nitrogen species (ROS, RNS) produced by activated neutrophils, monocytes/macrophages, endothelial and epithelial cells are involved in triggering and maintaining inflammation in the intestinal wall when there is deficit of antioxidant defense factors [8–10]. ROS and RNS, as well as products of their interaction with proteins and lipids, can act as markers of intestinal tissue damage, indicating the severity of the disease and the efficacy of treatment; they are also potential targets for novel personalized therapeutic and preventive approaches to the regulation of the local redox state in the large bowel affected by CD and UC [11–13]. Therefore, it would be interesting to study the aspects of free radical oxidation (FRO) and the relationship between the redox state of the intestinal lesion and the severity of clinical manifestations in UC and CD.
In our in vivo experiment described below, we aimed to investigate the aspects of FRO in the large intestine and the relationship between FRO and the clinical manifestations of UC and CD.

METHODS

The experiment was conducted in 35 male Wistar rats weighing 200–230 g. The animals were randomized into 3 groups: group 1 served as the intact control (n = 7); group 2 included animals with CD (n = 14); group 3 consisted of rats with UC (n = 14). The following procedure was applied to induce CD in the animals: 30 mg of 2,4,6-trinitrobenzenesulfonic acid (TNBS; Sigma- Aldrich; USA) was dissolved in 150 ml of 50% ethanol; 0.2– 0.3 ml of the solution (depending on the animal’s weight) was injected through a 2 mm polyurethane catheter (Sintez; Russia) inserted rectally to a depth of 8 cm [14]. A two-step oxazolone-based scheme was used to induce UC. In the first step, the animals were sensitized with 150 μl of 3% alcohol solution of oxazolone (Sigma-Aldrich; USA) applied onto the skin in the interscapular region; in the second step, 150 μl of 3% alcohol solution of oxazolone was rectally injected to a depth of 7–8 cm [15]. Before the procedure, the animals were anesthetized with 20 mg/kg Zoletil 100 (tiletamine hydrochloride; Virbac Sante Animale; France). CD and UC were evaluated based on the clinical symptoms and the morphology of the intestinal lesion on days 3 and 7. Symptoms were evaluated on a daily basis using a modified disease activity index (DAI) adapted to rats [16, 17]. Parameters assessed by DAI and the corresponding scores are provided in tab. 1. Liquid stools were defined as unformed fecal matter of mash-like or watery consistency. Diarrhea was defined as passage of unformed stools more than 3 times a day. Rectal bleeding was defined as the presence of fresh blood on the animal’s fur around the anus and in the feces. The presence of occult blood in the feces was evaluated using the benzidine test.

To prepare a sample of 10% intestinal mucosa homogenate, the proximal colon was removed and placed in a cooled 0.1M phosphate buffer solution (рН 7.4); then 100 mg of the tissue was homogenized in a mechanical glass tissue grinder (1 : 10) for 3 min at ≤ 4 °C. These conditions were applied to obtain 1 ml of homogenized intestinal mucosa. Lipid peroxidation (LPO) products contained in the homogenate were extracted in a mixture of heptane and isopropanol and measured using an SF-56 spectrophotometer (LOMO-Spectr; Saint-Petersburg) following the technique proposed by Volchegorsky et al. [18, 19]. Absorbance of heptane and isopropanol extracts was measured at 220 nm (isolated double bonds), 232 nm (diene conjugates, DC), 278 nm (ketodienes, KD, and conjugated trienes, CT), and 400 nm (Schiff bases, SB). The relative content of LPO products was expressed in the units of oxidation indices (UOI): E232/E220 (DC), Е278/Е220 (KD and СТ) and Е400/ Е220 (SB). To quantify products of oxidative protein modification (POM) in the homogenized tissue, carbonyl protein derivatives were allowed to react with 2,4-dinitrophenylhydrazine (two types of oxidation were tested: spontaneous and catalyzed by the Fenton reagent) and the resulting aldehyde/ketone DNPH derivatives were spectrophotometrically measured in the UV and visible regions of the spectrum [20, 21]. Measurement results were expressed in absorbance units per 1 mg of the protein (au/mg) and as percentage (%). Adaptive reserve potential (ARP, %) was calculated as a ratio of spontaneous oxidation products to induced oxidation products. The results were processed in IBM SPSS Statistics 19 (SPSS: An IBM Company; USA). The obtained values are presented below as a median (Ме) and an interquartile range [Q₁–Q₃]. Significance of differences between the groups was assessed using the Kruskal–Wallis, Mann–Whitney and Wald–Wolfowitz tests. Correlations between the studied parameters were evaluated using Spearmen’s rank correlation test. The differences were considered significant at р < 0.05.

RESULTS

On day 3 of the experiment, the animals with experimentally induced CD started exhibiting the following symptoms: frequent defecation, changes in stool consistency and blood in feces. On day 7, the symptoms deteriorated and were now accompanied by weight loss; those changes were reflected in the heightened DAI scores, as shown in tab. 2. In addition, the animals became less active. On day 7, the DAI scores were significantly higher than on day 3 (р < 0.05). Changes were also observed in the amount of LPO products in the lipid extract from the homogenized intestinal mucosa (tab. 3). On day 3, there was an increase in the amount of primary and secondary LPO products in the heptane phase and primary, secondary and end LPO products in the isopropanol phase. On day 7, there was an increase in the amount of primary, secondary and end LPO products in both heptane and isopropanol phases. Thus, the primary and end products of lipid peroxidation were building up between days 3 and 7 in the animals with CD, as seen from the isopropanol extracts (р < 0.01). In the next step, we quantified products of spontaneous and stimulated oxidative protein modification in the homogenates of large bowel tissue (tab. 3).

We found that the total amount of carbonyl protein derivatives was elevated on days 3 and 7 of observation. Carbonyl protein derivatives are irreversible products of oxidative stress. They are formed when a number of amino acid residues undergo oxidation and also when lipid peroxidation products interact with reducing sugars. The total amount of POM products was higher on day 7 than on day 3 (р < 0.01). When analyzing their relative content in the animals with CD, we discovered that on day 7 the share of primary products (aldehydes) shrank while the proportion of secondary products (ketones) increased. As neutral carbonyl derivatives absorb in the UV spectrum, whereas basic carbonyl derivatives absorb in the visible light region, the analysis of cumulative absorption in the UV and visible regions of the spectrum can shed light on the nature of POM products and their accumulation dynamics. On day 3, the proportion of neutral and basic POM products did not differ between the animals with CD and the intact group. On day 7, there was a shift towards accumulation of basic POM products, signaled by a decline in their total relative content in the UV spectrum and an increase in their total relative content in the visible region.

The adaptive reserve potential was estimated by calculating the total amount of POM products generated by spontaneous and stimulated oxidation. Briefly, protein oxidation in the intestinal mucosa homogenate was catalyzed by adding Fe²⁺ and Н2О2 to the reaction mix; the reaction yielded a highly reactive ОН ̇ radical. We found that the total amount of POM products yielded by stimulated oxidation was elevated on days 3 and 7 of the experiment in the animals with experimentally induced CD. On day 3, the proportion of primary and secondary POM products (both basic and neutral) did not differ from that in the intact animals; on day 7, the proportion of neutral aldehydes decreased and the proportion of basic ketones went up. So, changes in POM characteristics were similar for both spontaneous and stimulated oxidation modes. The total adaptive reserve potential was significantly increased on day 3 of the experiment and did not differ from that in the intact group on day 7. This phenomenon can be explained by the enhancement of the adaptive reserve potential of aldehyde/ ketone DNPH derivatives registered in the UV and visible light regions; however, ketone DNPH derivatives registered in the visible light region were the biggest contributor.

On day 3, the following symptoms were observed in the animals with UC: body weight loss, frequent defecation, loose stools, visible blood in feces, positive benzidine tests. On day 7, the symptoms deteriorated, as reflected in the heightened DAI scores (tab. 2). The DAI scores were significantly higher on day 7 than on day 3 (р < 0.05). On day 3, primary, secondary and end products of lipid peroxidation were significantly elevated in the heptane and isopropanol phases in this group of animals (tab. 3). Similarly, primary, secondary and end products of lipid peroxidation were significantly elevated in the heptane and isopropanol phases on day 7. Thus, primary and secondary LPO products were declining in the isopropanol phase between the days 3 and 7 in the animals with experimentally induced UC (р < 0.01).

In this group of animals, the total content of POM products yielded by stimulated oxidation was significantly increased on days 3 and 7 of the experiment (tab. 3). On days 3 and 7, the share of aldehyde DNPH derivatives shrank, whereas the share of ketone DNPH derivatives went up. On day 7, the relative amount of basic carbonyl derivatives was increased, in contrast to the reduced share of neutral carbonyl derivatives (this was inferred from an increase in the total relative content of POM products in the visible light region and their low total relative content in the UV region). The total amount of POM products yielded by stimulated oxidation was elevated on days 3 and 7. The share of aldehyde DNPH derivatives decreased whereas the share of ketone DNPH derivatives increased on days 3 and 7. The total relative content of POM products registered in the visible light spectrum was significantly elevated on day 3, suggesting a shift towards formation of basic aldehyde and ketone DNPH derivatives. The total adaptive reserve potential was significantly reduced on days 3 and 7.

The correlation analysis established correlations between DAI scores (the integral indicator of the clinical status) and the amounts of LPO/POM products in the homogenized intestinal mucosa of the large bowel (tab. 4). On day 3, there were moderate correlations in the CD group between DAI scores and the amount of secondary LPO products in the heptane and isopropanol phases, as well as the total amount of POM products yielded by spontaneous oxidation. In the UC group, DAI scores were strongly correlated with the amount of secondary LPO products in the heptane phase, primary and secondary LPO products in the isopropanol phase, and the total amount of POM products yielded by spontaneous and stimulated oxidation. On day 7, DAI scores correlated with the amount of primary LPO products in the heptane phase, primary, secondary and end LPO products in the isopropanol phase, and the total amount of POM products yielded by spontaneous and stimulated oxidation; for UC, DAI scores were strongly correlated with the amount of primary and end LPO products in the heptane phase, primary, secondary and end LPO products in the isopropanol phase, and the total amount of POM products yielded by spontaneous and stimulated oxidation. The largest number of strong correlations was observed for UC: 10 out of 12 correlations were strong; by contrast, only 1 of 9 correlations established in the CD group was strong.

DISCUSSION

In our experiment, clinical manifestations of experimentally induced CD and UC were consistent with the actual symptoms and signs of these diseases, suggesting that the proposed TNBS and oxazolone-based animal models can be used to study the pathogenesis of homeostatic changes characteristic of these conditions. On average, the rats did not lose more than 10% of their body weight. Such weight loss is traditionally associated with diarrhea and systemic inflammatory response and more specifically with the anorexigenic effect of some proinflammatory cytokines [22]. It is thought that TNBS acts as a hapten and, if administered rectally, induces a Th1 immune response, which involves proinflammatory cytokines and some mediators, against hapten-modified autologous proteins/ luminal antigens or against intestinal microbiota proteins, causing transmural infiltration by leukocytes and inflammation [23–25]. Oxazolone is also seen as a hapten that mediates a Th2 immune response typical to UC, and the majority of researchers prefer it as the most popular agent for modeling UC in rats [26]. The 50% ethyl alcohol solution used as a solvent for TNBS and oxazolone aggravates damage to the mucosa of the large bowel [27].

Experimental CD and UC are characterized not only by weight loss, frequent defecation and changes to stool consistency but also by the presence of blood in feces and the accumulation of LPO products in the large bowel mucosa. In our experiment, accumulation of LPO products was observed in the isopropanol phase, which primarily accumulates phospholipids of cell membranes, and in the heptane phase (in triglycerides). Besides, there was accumulation of POM products, mainly secondary basic ketone DNPH derivatives, following spontaneous and stimulated oxidation. This led us to hypothesize that combined effects of ОН ̇ and О2 ̇ cause accumulation of late markers of oxidative protein destruction and protein fragmentation [20, 28, 29]. Protein fragments are highly resistant to proteolysis, very toxic and can trigger apoptosis or necrosis, expanding the area of secondary alterations [30]. We have identified a few specific aspects of FRO occurring in the large bowel affected by CD and UC that are related to oxidative destruction of lipids and proteins. Firstly, comparative analysis reveals that the amount of end LPO products in the heptane phase and secondary products in the isopropanol phase measured on day 3 of the experiment, as well as the amount of end and secondary LPO products in the heptane phase and secondary LPO products in the isopropanol phase measured on day 7, was significantly higher in rats with UC (tab. 3). Secondly, comparison of POM products in the homogenized mucosa of the large bowel demonstrates that the adaptive reserve potential and the total content of POM products were higher in the spontaneous and stimulated oxidation modes on days 3 and 7. Besides, in the CD group, as opposed to the UC group, basic primary POM products yielded by spontaneous and stimulated oxidation prevailed on day 3 of observation; on day 7, basic POM products prevailed in the spontaneous oxidation mode.

We think that escalation of oxidative stress resulting from production of ROS by activated neutrophils, monocytes/ macrophages and endothelial cells in the primary lesion in a situation when the enzymes responsible for antioxidant defense are insufficiently active/scarce increases damage, dysfunction and death of large bowel cells, expands the area of secondary alteration and causes exacerbation of symptoms in patients with CD and UC. This conclusion is supported by the discovered correlations between DAI scores and the amount of LPO/POM products in the homogenized large bowel mucosa of rats with induced UC and CD.

CONCLUSIONS

We found that on days 3 and 7 of TNBS-induced CD and oxazolone-induced UC, DAI scores, the amount of primary, secondary and end LPO products in the heptane and isopropanol phases of homogenized intestinal mucosa and the total amount of POM products were increased. In UC, the proportion of secondary basic POM products increased on day 7. In the setting of UC, FRO was characterized by the accumulation of mostly LPO products, in contrast to CD characterized by the accumulation of mostly POM products. In UC, end LPO products were accumulated in the heptane phase of the lipid extract, whereas secondary products, in the isopropanol phase. CD was characterized by the accumulation of mostly secondary basic POM products. In CD and UC, there was a correlation between DAI scores and the amount of LPO products in the isopropanol phase, as well as the amount of POM products generated by spontaneous oxidation. The highest number of strong correlations was observed for UC. These findings broaden our knowledge about the role of redox state changes in the pathogenesis of inflammatory bowel diseases and encourage further research into FRO affecting the large bowel of patients with CD and UC. LPO and POM products should be investigated in the clinical setting as potential diagnostic markers of the disease and indicators of the efficacy of treatment aimed at reversing exacerbations and prolonging remission.