Olaparib: A Clinically Applied PARP Inhibitor Protects from Experimental Crohn's Disease and Maintains Barrier Integrity by Improving Bioenergetics through Rescuing Glycolysis in Colonic Epithelial Cells

Crohn's disease (CD) is an inflammatory disorder of the intestines characterized by epithelial barrier dysfunction and mucosal damage. The activity of poly(ADP-ribose) polymerase-1 (PARP-1) is deeply involved in the pathomechanism of inflammation since it leads to energy depletion and mitochondrial failure in cells. Focusing on the epithelial barrier integrity and bioenergetics of epithelial cells, we investigated whether the clinically applied PARP inhibitor olaparib might improve experimental CD. We used the oral PARP inhibitor olaparib in the 2,4,6-trinitrobenzene sulfonic acid- (TNBS-) induced mouse colitis model. Inflammatory scoring, cytokine levels, colon histology, hematological analysis, and intestinal permeability were studied. Caco-2 monolayer culture was utilized as an epithelial barrier model, on which we used qPCR and light microscopy imaging, and measured impedance-based barrier integrity, FITC-dextran permeability, apoptosis, mitochondrial oxygen consumption rate, and extracellular acidification rate. Olaparib reduced the inflammation score, the concentration of IL-1β and IL-6, enhanced the level of IL-10, and decreased the intestinal permeability in TNBS-colitis. Blood cell ratios, such as lymphocyte to monocyte ratio, platelet to lymphocyte ratio, and neutrophil to lymphocyte ratio were improved. In H2O2-treated Caco-2 monolayer, olaparib decreased morphological changes, barrier permeability, and preserved barrier integrity. In oxidative stress, olaparib enhanced glycolysis (extracellular acidification rate), and it improved mitochondrial function (mitochondrial coupling efficiency, maximal respiration, and spare respiratory capacity) in epithelial cells. Olaparib, a PARP inhibitor used in human cancer therapy, improved experimental CD and protected intestinal barrier integrity by preventing its energetic collapse; therefore, it could be repurposed for the therapy of Crohn's disease.


Introduction
Inflammatory bowel disease (IBD) is a chronic and remitting inflammatory disease of the gut. More than 1 million inhabitants in the USA and approximately 2.5 million in Europe suffer from IBD, and its incidence is permanently rising [1].
IBD exhibits two main forms, namely, ulcerative colitis (UC) and Crohn's disease (CD), and it appears in flare-up and remission phases [2]. Although UC and CD are two distinct forms of IBD, they share the phenomenon of epithelial barrier dysfunction. Barrier failure often results in increased intestinal permeability, a condition called "leaky gut" [3]. In this disorder, the gut microbiota can directly enter the colonic tissue and induce the activation of immune cells causing chronic inflammation [4,5]. The initiators of increased gut permeability are not clearly elucidated, but it is often suggested that increased permeability is a consequence of altered energy metabolism and mitochondrial dysfunction of intestinal epithelial cells (IEC) [6]. For example, investigations with conplastic mouse strains, which share the same nuclear genome but have different mitochondrial genomes, demonstrated that those mice with high mucosal respiratory chain activity and elevated concentration of ATP develop less intense colitis than those that produce a smaller amount of mucosal ATP [7]. In CD patients, increased mucosal permeability in the ileum was accompanied by mitochondrial swelling and decreased ATP concentration [8]. In addition, the activity of complex II (CII), a part of the mitochondrial electron transport chain (ETC), was found to be abolished in the colon of UC patients [9]. Another group found lower levels of CI and CIV in IBD patients compared to control subjects and also measured lower ATP concentrations [10]. Furthermore, enhanced lactate levels were found in CD patients in comparison with healthy individuals, which correlated with the disease activity [11]. All these results suggest mitochondrial dysfunction, disturbed oxidative phosphorylation, and enhanced glycolytic activity in the mucosa of IBD patients.
Under physiological conditions, IECs use butyrate as a primary energy source [12]. Butyrate is produced by several species of the microbiota, and it is catabolized in IECs via βoxidation and citric acid cycle (CAC) [13][14][15]. In addition, dehydrogenases of these catabolic pathways reduce NAD + and FAD to NADH+H + and FADH 2 which promote the reduction of the mitochondrial respiratory chain CI and CII [16]. Thereafter, CI, CIII, and CIV pump protons across the inner membrane from the matrix to the intermembrane space raising a proton gradient [16]. At the end of ETC, CIV consumes O 2 and reduces it to H 2 O. Finally, the proton gradient drives F O F 1 -ATPase, which produces ATP from ADP and P i [16].
However, in inflammation, mitochondrial dysfunction and mitochondria-derived ROS increase. Under these circumstances, IECs switch their metabolism from oxidative phosphorylation (OXPHOS) to aerobic glycolysis [13,17]. In aerobic glycolysis, glucose transforms to lactate without oxygen consumption, although sufficient amount of oxygen is present in the cells [18]. In this situation, glycolysis produces ATP and, as a by-product, lactate is synthesized from pyruvate by lactate dehydrogenase [17]. Since the mitochondria are a major source of ROS [19], the catabolic pathway via glycolysis and lactate dehydrogenase bypasses the mitochondria and do not feed mitochondrial ROS generation [20]. Thus, the cell shuts down the mitochondria to protect itself from mitochondrial ROS [21]. This concept is strengthened by the findings that proinflammatory cytokines (TNF-α, IL-1β, and IFN-γ) increased the rate of glycolysis in rat enterocytes and also triggered ATP turnover [22]. Also C. rodentium infection in mice induced aerobic glycolysis and enhanced the level of sodium-glucose transporter 4 and lactate dehydrogenase A. At the same time, enzymes of CAC and OXPHOS were downregulated [23]. Most importantly, a strong expression of glycolytic enzymes was found in the colon of IBD patients [24]. In active CD, lactate levels were significantly higher compared to the control subjects [11]. Therefore, in colitis, aerobic glycolysis becomes the main source of ATP. Nevertheless, in severe inflammation, activation of the enzyme poly(ADP-ribose)polymerase-1 (PARP-1) blocks glycolysis [25], i.e., it terminates the "last safe way" of energy production and forces the cells along the death pathway causing strong mucosal damage with severe ulceration and compromised barrier function.
PARP-1 has been long involved in cancer development and inflammation. Accordingly, PARP-1 -/mice were protected in 2,4,6-trinitrobenzene sulphonic acid-(TNBS-) induced colitis [26] and pharmacological inhibitors of PARP-1 improved dextran sodium sulfate-induced [27] and TNBS-induced colitis [28] in rodents. PARP-1 is activated by DNA damage and catalyzes polyADP-ribosylation (PARylation) of numerous nuclear proteins using NAD + as a substrate [29]. This process is a part of the DNA damage response leading to activation of the DNA repair enzymes [30]. However, excessive PARP activation can totally deplete NAD + pools, which makes cellular energy metabolism impossible [31]. Several lines of evidence demonstrate that PARP activation not only depletes NAD + pools but also inhibits the enzyme hexokinase, which catalyzes the first step of glycolysis [25]. As a result, repressed glycolysis cannot feed CAC with Acetyl-CoA (produced by pyruvate dehydrogenase from the glycolytic end-product pyruvate), and CAC is not able to reduce NAD + and FAD to feed mitochondrial ETC and OXPHOS [32], so PARP-induced mitochondrial dysfunction originates, at least partially, from the decreased substrate flow from glycolysis to CAC and ETC [25]. Since, in severe colitis, glycolysis is the main source of ATP (because of mitochondrial shutdown) [21] and also glycolysis is inhibited by PARP [25], IECs have to face with energetic collapse and they lose the ability to form a strong and continuous barrier [7].
In the present study, we investigated whether olaparib, a PARP inhibitor used in human cancer therapy, has a beneficial effect in a CD mouse model and, accordingly, whether it could be repurposed for CD treatment. To answer this question, we applied olaparib during a TNBS-induced experimental colitis model. Additionally, since IECs are the first line of defence in the colon and barrier interruption is a hallmark of IBD, we used Caco-2 colonic epithelial cells and investigated barrier function and energy production in vitro.  (Figure 1(a)). On day 1, olaparib treatment started (pretreatment), and thereafter, we administered it daily once for 3 times (thus, in total, we performed 4 olaparib treatments). On day 0, animals were treated with TNBS (1 bolus), and on day 3, mice were anesthetized and euthanized. Olaparib (AZD2281, MedChemExpress, New Jersey, USA) was administered intraperitoneally (single injection) on the day before TNBS challenge, followed by daily administration for 3 days at the dose of 20 or 50 mg/kg bodyweight. The applied dose of olaparib was selected based on literature data [33]. The vehicle group received sterile distilled water containing 4% DMSO and 30% PEG300. After 12 hrs fasting, mice were anesthetized with 5% isoflurane (Baxter Hungary Ltd., Budapest, Hungary) in 100% oxygen in an anaesthetic chamber. Colitis was induced by a single intracolonic injection of TNBS (4 mg in 100 μl of 30% ethanol; Sigma-Aldrich, Missouri, USA) through a catheter inserted 3 cm into the colon. The VEH group received an equal volume of 30% ethanol. Animals were weighed daily during the experiment and sacrificed 72 hrs after TNBS administration. Mice were anesthetized with 5% isoflurane and decapitated gently by a dedicated surgical scissor to collect the highest possible amounts of trunk blood. This technique was approved by the Animal Research Review Committee of the University of Pécs. Trunk blood was collected; the colons were removed, measured, weighted, and opened longitudinally to detect the macroscopic colon damage. Tissue samples were processed for further analyses. Treatments and macroscopical scoring were carried out blind.

Intestinal Permeability Measurement.
Intestinal permeability was determined by measuring the concentration of fluorescein isothiocyanate (FITC)-dextran (40 kDa; Sigma-Aldrich Missouri, USA) in serum. 3 days after TNBS treatment, FITC-dextran solution (100 μl of a 60 mg/ml solution) was administered intrarectally. Serum was collected 1 hour after the administration, and fluorescence intensities were detected by a Promega GloMax plate reader (excitation, 490 nm; emission, 510-570 nm). A standard curve was generated from a serial dilution of FITC-dextran in PBS.

Hematological Analysis.
At the endpoint of the TNBS model, mice were anesthetized with 5% isoflurane and decapitated gently by a dedicated surgical scissor, and trunk blood was collected directly into microtainer tubes (Becton Dickinson, Hungary) containing EDTA as an anticoagulant. Hematological parameters were determined by a Sysmex XN-1000-V Multispecies Hematology Analyzer (Sysmex America Inc., USA) within 2 hours of sampling. Lymphocyte to monocyte ratio (LMR), platelet to lymphocyte ratio (PLR), neutrophil to lymphocyte ratio (NLR), and neutrophil to monocyte ratio (NMR) were calculated from the absolute cell counts for each animal separately.
2.4. Macroscopic Scoring. Colonic tissue damage score was assessed by a macroscopic scoring system described previously [34]. Briefly, individual points were added for ulcers (0.5 points for each 0.5 cm), adhesions (0 points = absent, 1 point = 1 adhesion, and 2 points = 2 or more adhesions or adhesions to organs), colon shortening, based on a mean length of a healthy colon (1 point = >15%, 2 points = >25%), wall thickness (measured in mm), consistency of the stool, and the presence of blood in the stool (hemorrhage, fecal blood, or diarrhea increase the total points by 1).

Histology of Colon
Tissue. Segments of the distal colon were stapled flat onto a cardboard with the mucosal side up and fixed for at least 24 hrs in 10% neutral-buffered formalin. Tissue was then dehydrated and embedded in paraffin, and standard hematoxylin staining was performed on 5 μm thick sections. To this end, slides were deparaffinized, cleared in xylol, rehydrated in a descending ethanol series, stained with hematoxylin solution according to Gill II, and cleared in tap water. Images were taken with an Olympus DP50 camera and processed with cellSens imaging software (Olympus, Vienna, Austria).
2.6. Cytokine Levels of Colon Tissue. Levels of inflammatory cytokines IL-1β, IL-6, TNF-α, and IL-10 were measured in colon tissues. Tissue was homogenized mechanically in an extraction buffer supplemented with protease inhibitor cocktail (Sigma-Aldrich, Missouri, USA

Determination of Apoptosis. Mouse Annexin V & Dead
Cell Kit (Merck Millipore, Massachusetts, USA) was used for the quantitative analysis of live, early, and late apoptotic and necrotic cells. Caco-2 cells were seeded onto 6-well plates at a density of 10 6 cells/well. Treatments and treatment groups were exactly the same as described above at the FITCdextran assay. 24 hrs after treatment, cells were trypsinized and collected; sample preparation was performed as suggested by the manufacturer. Briefly, 100 μL of cell suspension was incubated with 100 μl of Muse Annexin V & Dead Cell reagent for 20 minutes, in the dark at room temperature. After staining, the assay was performed with a Muse Cell Analyzer (flow cytometer).  at a density of 1:5 × 10 4 cells/well. After reaching 100% confluence, cells were treated exactly as described at the FITCdextran assay. After the treatment, a complete growth medium was replaced with an unbuffered, serum-free Agilent XF Base assay medium, pH 7.4. XFp Mito Stress Test Kit was used to test mitochondrial function. Injection of oligomycin, carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP), and the mix of rotenone and antimycin A allows determining the key bioenergetic parameters: basal respiration, ATP production-linked respiration (ATP production), maximal respiration, spare respiratory capacity, nonmitochondrial respiration, proton leak, and coupling efficiency. Oligomycin inhibits the F O subunit of the F O F 1 -ATP synthase, thereby indicating ATP-linked OCR, i.e., level of ATP synthesis. ATP-linked respiration is calculated by the difference between baseline OCR and OCR after oligomycin injection. Distracting nonmitochondrial respiration from the OCR after FCCP injection represents maximal respiration. FCCP is a mitochondrial uncoupler, which separates the activity of phosphorylation and oxidation. Under these circumstances, ETC might work with its maximum rate and consumes higher amounts of O 2 without developing membrane potential between the two sides of the mitochondrial inner membrane. Spare respiratory capacity is defined by the difference between maximal and basal respiration. The mixture of rotenone and antimycin A inhibits CI and CIII, respectively; thus, mitochondrial ETC and O 2 consumption are blocked. The final concentrations of the modulators were 1 μM. OCR after rotenone/antimycin A injection represents nonmitochondrial respiration. ATPlinked respiration divided by basal respiration reveals coupling efficiency.

Seahorse
2.13. Light Microscopy Imaging. Caco-2 cells were seeded at a density of 10 6 cells/well on 6-well plates. After reaching confluency, the monolayers were treated exactly as described at the FITC-dextran assay. 24 hours later, monolayers were visualized by EVOS XL Core Cell Imaging System (Thermo Fisher Scientific, USA) using a 20× objective.
2.14. Statistical Analysis. Experimental data were analyzed by using GraphPad Prism Software (GraphPad Software Inc., California, USA). Statistical difference between groups was established by Student's t-test, with Bonferroni correction; P values less than 0.05 were considered statistically significant.

Olaparib Improved TNBS-Colitis in Mice.
To evaluate the effect of olaparib in experimental colitis, we used the TNBS-colitis model (Figure 1(a)), a mouse model of CD [35]. Olaparib was used as a pretreatment in 20 and 50 mg/kg bodyweight dose. On the one hand, olaparib failed to significantly ameliorate weight loss in TNBS-challenged animals (Figure 1(b)). But on the other hand, it decreased inflammation scores by more than~50% in 50 mg/kg (n = 21), but not in 20 mg/kg dosage (n = 9) (Figure 1(d)). Hence, we used 50 mg/kg dose in the further experiments.
Olaparib impeded histological injury in the colon (Figure 1(c)), reduced the number of ulcers (n = 27) (Figure 1(e)) and their lengths (n = 27) (Figure 1(f)), and most importantly, diminished FITC-dextran permeability (n = 8) (Figure 1(g)) compared to the CTRL group (n = 5). Levels of inflammatory cytokines were also modulated. Olaparib diminished IL-1β (Figure 2(a)) and IL-6 ( Figure 2(b)) proinflammatory cytokine levels, but enhanced antiinflammatory IL-10 production (Figure 2(c)) in the colon (n = 13). Interestingly, we could not find statistically significant alteration in the TNF-α level (Figure 2(a)). We also evaluated numerous hematological parameters in colitic mice (Figure 3(a)). We found only 2 parameters, namely, the amounts of lymphocytes and monocytes, which were significantly modulated by the treatments. In agreement with others' findings on colitis models, TNBS substantially reduced lymphocyte number in mice (n = 9), while olaparib counteracted this effect (n = 20). In contrast, monocyte number was higher in the TNBS group (n = 9), whereas it was significantly less elevated in the TNBS+olaparib group ( Figure 3(a)) (n = 20). We calculated specific blood cell ratios, which were previously shown to be changed in CD [36] based upon the individual blood cell counts. Similarly to CD, neutrophil to lymphocyte ratio (NLR) (Figure 3(e)) and platelet to lymphocyte ratio (PLR) (Figure 3(c)) were both increased in TNBS-colitis and they were markedly reduced by olaparib treatment. Again, as in CD, lymphocyte to monocyte ratio (LMR) (Figure 3(b)) was reduced in experimental colitis, and it was amended by the PARP inhibitor. Unfortunately, TNBS-induced changes in the neutrophil to monocyte ratio (NMR) (Figure 3(d)) did not reach statistical significance compared to the vehicle. However, olaparib improved NMR related to the TNBS-treated group.

Olaparib Improved Barrier Function of Epithelial
Monolayer in Oxidative Stress. Caco-2 monolayers are widely used as a model for intestinal epithelial barrier [38].
Since the activity of PARP-1, PARP-2, and PARP-3 isoforms can be induced by DNA-damage [39], and as oxidative stress induces mucosal injury in IBD [40,41], we tested different H 2 O 2 concentrations (100-1000 μM) on Caco-2 monolayers. We assessed barrier integrity by an impedance-based technique (Figure 4(d)).  (Figure 4(d)). Consequently, in the further experiments, we applied 1 mM concentration of H 2 O 2 to challenge the barrier. Olaparib pretreatment, 30 min before H 2 O 2 exposure, improved CI compared to H 2 O 2 -treated cells and protected monolayer integrity ( Figure 5(b)). To confirm these findings, we also performed FITC-dextran trans-epithelial permeability assay in the same model at the endpoint of the impedance-based measurement, after 24 hrs incubation ( Figure 5(a)). We detected about~20-fold increase in FITC-dextran fluorescent intensity after H 2 O 2 treatment (i.e., FITC-dextran could pass the monolayer) in relation to CTRL. In contrast, olaparib reduced H 2 O 2 -induced FITC-dextran permeability near to the level of control ( Figure 5(a)). To even further refine our results, we performed microscopic imaging and observed morphological changes in the structure of epithelial monolayer after H 2 O 2 treatment. We realized compromised, in some places broken monolayer, with presumably dying cells, which lost their connection to neighbors in the monolayer. Olaparib prevented these morphological changes and kept the cells as an integral part of the barrier in their normal, epithelial phenotype ( Figure 5(c)).

Olaparib Protected against Oxidative Stress-Induced Cell
Death in Epithelial Barrier. To assess whether oxidative stress-induced barrier dysfunction involves epithelial cell death we performed flow cytometry analysis using Annexin V/7-AAD labeling ( Figure 5(d)).

PARP Inhibition Recovered Glycolytic Activity
Compromised by H 2 O 2 Treatment. In inflammation, colonocytes switch their oxidative metabolism (butyrate consumption) to aerobic glycolysis and produce lactate [13,17] ( Figure 6(a)). Thus, we investigated the glycolytic activity by measuring extracellular acidification rate (ECAR), i.e., lactate production ( Figure 6(b)), two hours after H 2 O 2 treatment in the early phase of oxidative stress. H 2 O 2 caused a dramatic collapse in basal ECAR (w/o oligomycin) compared to CTRL, which was markedly enhanced by olaparib ( Figure 6(b) (1-3 points of the measurement) and Figure 6(c)). Oligomycin treatment increased ECAR both in the CTRL and H 2 O 2 +olaparib-treated cells compared to the untreated (w/o oligomycin) group but failed to stimulate acidification in the H 2 O 2 -damaged monolayer ( Figure 6(b) (4-6 points of measurement) and Figure 6(d)). FCCP, rotenone, and antimycin A did not influence ECAR significantly in either treatment groups ( Figure 6(b) (7-12 points of measurement)).     Oxidative Medicine and Cellular Longevity

PARP Inhibitor Olaparib Preserved Mitochondrial
Respiration in H 2 O 2 -Induced Stress. Under physiologic conditions, butyrate is the main source of ATP in colonocytes [12], and butyrate metabolism involves dynamic mitochondrial ETC activity and continuous OXPHOS [42]. Therefore, we investigated the activity of ETC and OXPHOS by measuring the oxygen consumption rate in our epithelial barrier model (Figure 7(a)). First, the basal respiration (OCR w/o oligomycin, green field on Figure 7(a)) was determined. H 2 O 2 reduced basal respiration in epithelial cells compared to CTRL, and olaparib did not modulate this effect (Figure 7(b) (1-3 points of measurement and Figure 7(c)) indicating that olaparib had no effect on basal respiration. After oligomycin treatment, the ATP production-linked OCR (OCR with oligomycin, yellow field on Figure 7(a)) can be measured that reflects the activity of OXPHOS and ATP generation. Oligomycin reduced OCR and OXPHOS xoverall in all three experimental groups (Figure 7(b) (4-6 points of measurement]) and Figure 7(d)), but H 2 O 2 -treated cells consumed O 2 even to a lesser extent than CTRL, which suggested a reduced ATP production. Olaparib had no significant effect on the ATP-linked OCR in H 2 O 2 -treated cells (Figure 7(d)). Furthermore, olaparib ameliorated the H 2 O 2induced decline in coupling efficiency (Figure 7(e)). In contrast, FCCP, an uncoupling agent that induces maximal respiration in the mitochondria (OCR with FCCP, beige field on Figure 7(a)), enhanced OCR in all three groups in different extents (Figure 7(b) (7-10 points of measurement)). We detected the highest OCR in CTRL, the lowest in the H 2 O 2induced cells while olaparib counteracted the effect of H 2 O 2 (Figure 7(b) (7-10 points of measurement) and Figure 7(f)). FCCP application also determined spare respiratory capacity (blue field on Figure 7(a)). Spare respiratory capacity was intensely reduced by H 2 O 2 compared to CTRL, but olaparib       4) and expressed as mean ± SD; ns: not significant; * P < 0:05, * * P < 0:01, and * * * P < 0:001. attenuated this reduction (Figure 7(g)). H 2 O 2 reduced proton leak (orange field on Figure 7(a)) compared to CTRL, and it was further reduced by olaparib in Caco-2 cells (Figure 7(h)).

Discussion
The anti-inflammatory role of PARP inhibition is thoroughly established; however, introduction of PARP inhibitors into clinical therapy of anti-inflammatory diseases has not been initiated yet because of the potential risk in longterm use of the drugs [33]. In this study, we intended to provide experimental support for repositioning the PARP inhibitors (which are successfully applied in human cancer therapy) for the clinical management of the acute flare-up periods of CD. For that purpose, we used a TNBS-induced experimental colitis model in mice, which is widely accepted for studying CD, since they share many pathological (clinical, histological, and biochemical) characteristics [35]. Furthermore, enhanced PARP-1 expression was found in the colon of rodents [43,44] in experimental colitis models, as well as in IBD patients [45], which makes the model more valuable for studying PARP inhibitors. In this report, we explicitly focused on the epithelial barrier function, cell survival, and bioenergetics; hence, we used a Caco-2 monolayer as an in vitro model of intestinal epithelial barrier [46]. We demonstrate that olaparib improves inflammation in TNBS-colitic mice and that it protects Caco-2 epithelial barrier in oxidative stress by rescuing glycolytic activity and by protecting some aspects of mitochondrial function.
In a recent review about repurposing PARP inhibitors for the therapy of nononcological diseases [33], Berger et al. did not consider IBD among those chronic diseases, in which the assumed benefits vs. the risks justify first priority of repurposing. However, the available preclinical data on IBD models successfully utilized outdated PARP inhibitors such as 3-aminobenzamide [47]. Our findings of in vivo anti-inflammatory effects of olaparib, a PARP inhibitor approved for human cancer therapy, may justify initiation of clinical trials for repurposing this drug for IBD therapy. We hypothesize that PARP inhibition might be beneficial in the acute flare-ups of severe CD, where detrimental

12
Oxidative Medicine and Cellular Longevity ulceration and tissue damage are caused by the energetic collapse of mucosal cells. It could be especially true in the severe cases of drug nonresponders, e.g., those one-third of IBD patients who primarily do not respond to infliximab (anti-TNF-α mAb; commonly prescribed drug in IBD) [48] or to other pharmacological therapies. We demonstrated that olaparib improved TNBSinduced colitis in mice, reduced histological damage of the colon, diminished the number and length of the ulcers, inhibited proinflammatory cytokine production (IL-1β, IL-6), but it enhanced the level of anti-inflammatory cytokine IL-10. Inflammatory cytokines participating in the generation of colon damage are predominantly produced by activated leukocytes [49]. The fact that PARP inhibition reduces leukocyte infiltration into the colon in experimental colitis is well characterized [50][51][52]. Accordingly, we investigated hematological parameters from peripheral blood in our TNBS-colitis model. Several types of blood cell ratios were recently highlighted as possible diagnostic parameters in IBD [36,53]. Specifically in CD, NLR and PLR were suggested to be valuable diagnostic factors [36]. Moreover, NLR might predict disease severity [54]; however, this notion is debated [55]. In detail, elevated NLR, PLR, and reduced LMR were found in CD patients compared to control subjects, while NMR was not modified [36]. In our experiments, alterations in NLR, PLR, LMR ,and NMR in TNBS-treated mice followed the observed changes in CD patients. In addition, olaparib effectively reversed CD-specific alterations in PLR and LMR, and most importantly, it reduced NLR. NLR was found to be a significant predictor of infliximab drug response in CD patients [56]. That is, the antiinflammatory efficacy of infliximab correlated with decreased NLR in CD. Olaparib's identical antiinflammatory effect in our TNBS-colitis model underlines the drug's potential in CD therapy.
Elevated NLR can also refer to oxidative stress in CD patients [57], which is an important inducer of PARP activation [58]. In active CD, excessive amounts of ROS are produced by the immune cells. The main source of ROS in immune cells is the H 2 O 2 production [59], and H 2 O 2 causes DNA damage in colonocytes [60] that triggers PARP activation. High concentrations of ROS result in apoptosis of IECs leading to disruption of epithelial barrier integrity in the colon, which is a definite hallmark of IBD. Accordingly, we treated Caco-2 monolayer with high concentration of H 2 O 2 (1 mM) to imitate a strong oxidative stress-injured barrier, in vitro. We demonstrated that Caco-2 monolayer cells expressed PARP-1 mRNA in a high extent similarly to colonic mucosa [45]. In addition, we detected PARP-2 and PARP-3 expressions, however, in a decreasingly lower extent. That is, Caco-2 cells express the mRNA of olaparib's target enzymes (PARP-1, PARP-2, and PARP-3); furthermore, in the monolayer, these cells mimic the intestinal barrier [46]. Therefore, the Caco-2 monolayer seemed to be an appropriate model to investigate the effect of PARPinhibition on barrier integrity, in vitro. Our results showed that olaparib preserved the Caco-2 monolayer integrity in oxidative stress and protected the epithelial cells from apoptosis. These findings indicated that colonic epithelial cells might be direct targets of olaparib in TNBS-colitis, and barrier protection might be one of the key components of its anti-inflammatory action.
Reduced barrier integrity and increased gut permeability are typical signs of IBD. They have been recently associated with epithelial cell death, mitochondrial dysfunction, and depleted energy metabolism in IECs. IECs produce ATP predominantly by aerobic glycolysis in colitis. However, overactivation of PARP might fully block glycolysis [25]. To study the metabolic effect of olaparib in epithelial cell death, we induced powerful oxidative stress (1 mM H 2 O 2 , 2 hrs) in Caco-2 monolayer and determined various parameters of the energy metabolism ( Figure 8) ee -e -energetic collapse in Caco-2 cells. Lower proton leak rate in H 2 O 2 -treated cells compared to control also reflected this metabolic failure, because ATP demand reduces proton motive force and diminish proton leakage [61,62]. In contrast, olaparib significantly enhanced ECAR in H 2 O 2 -treated cells. Our results are strongly supported by the finding that poly(ADP-ribose) (PAR) binds to the PAR-binding motif in hexokinase and inhibits it thereby reducing glycolysis [25]. These data indicate that olaparib exerted its effect by inhibiting PAR production, thereby preventing PARmediated inhibition of hexokinase and the blockade of aerobic glycolysis (Figure 8).
Oligomycin is a F O F 1 -ATPase inhibitor, which blocks OXPHOS. As expected, oligomycin had no effect on ECAR in the H 2 O 2 treatment group; however, it intensely increased ECAR in H 2 O 2 +olaparib-treated cells. An explanation for this finding might be that oligomycin blocked ATP synthesis and cells could not compensate for the lack of ATP by boosting glycolysis in H 2 O 2 treatment group because of the strong PARP-mediated repression of hexokinase. Olaparib, however, prevented PARP activation and glycolytic collapse in H 2 O 2 +olaparib-treated cells (Figure 8) even in the presence of oligomycin. One may say that oxidative stress per se might affect glycolytic enzymes and not only PARP activation regulates glycolytic activity. By using FCCP (mitochondrial uncoupler), rotenone, and antimycin A (inhibitors of ETC) mitochondrial energy production is totally abrogated. Under these circumstances, glycolysis remains the ultimate source of ATP ( Figure 8). The fact that ECAR could reach the level of CTRL cells in the H 2 O 2 +olaparib group after FCCP, rotenone and antimycin A treatment clearly indicated that glycolytic enzymes were not significantly affected by the oxidative stress in our system, and glycolytic energy production was controlled by PARP activation. Thus, olaparib protected from PARP-induced energetic collapse by improving aerobic glycolysis in oxidative stress ( Figure 8).
Previous studies indicated that the mitochondria are the primary source of ROS in IECs during inflammation [63] and PARP-1 activation in oxidative stress causes mitochondrial dysfunction [25]. Accordingly, we wanted to know whether olaparib can prevent mitochondrial failure in our model. We found that basal respiration, mitochondrial ATP production, and nonmitochondrial oxygen consumption were strongly impeded in oxidative stress, and olaparib could not reverse these changes. However, it increased maximal respiration, spare respiratory capacity, and coupling efficiency and reduced proton leak. For understanding these results, we should consider the direct and indirect effects of H 2 O 2 on CAC enzymes, ETC complexes, and F O F 1 -ATPase and the effects of PARP activation on glycolysis and ETC complexes.
Previous studies using the DNA-alkylating agent Nmethyl-N-nitroso-N-nitroguanidine (MNNG) for PARP activation reported mitochondrial dysfunction [25]. MNNG treatment resulted in decreased basal OCR, maximal OCR, ATP synthesis, and ECAR, while PARP inhibition significantly reversed these changes. Most importantly, they found that administration of pyruvate completely prevented MNNG-induced mitochondrial failure [25]. That is, the mitochondrial dysfunction was a direct consequence of downregulated glycolysis, i.e., it compromised fuel supply for the CAC and ETC. Our results led to the same conclusion, namely, PARP activation reduced glycolysis and caused mitochondrial dysfunction (Figure 8).
In contrast to MNNG-induced PARP activation, we found that inhibition of the enzyme did not prevent the H 2 O 2 -induced reduction of basal respiration and mitochondrial ATP production. On the other hand, olaparib enhanced ECAR, i.e., it effectively prevented glycolytic collapse ( Figure 8). Additionally, it increased maximal respiration, spare respiratory capacity, coupling efficiency, and reduced proton leak, i.e., it increased efficacy of OXPHOS. The discrepancy between our results and the previous ones [25] can be resolved by considering that MNNG alkylates the DNA leading to DNA breaks and PARP activation. H 2 O 2 , however, induces oxidative DNA damage-mediated PARP activation but can cause direct structural impairment to CAC, ETC, and F O F 1 -ATPase components as collateral damage. It is well established that H 2 O 2 deteriorates CAC activity and reduces proton motive force, which is a prerequisite for mitochondrial pyruvate transport [64]. Also, F O F 1 -ATPase was reported to be susceptible to oxidative stress [65,66]. In addition, ROS and especially H 2 O 2 can directly block many components of the respiratory chain, such as NADH dehydrogenase or cytochrome c oxidase [65]. Our findings that maximal OCR of the H 2 O 2 +olaparib treatment group could not reach the maximum OCR level of control cells are in line with the notion that ETC is sensitive toward oxidative stress (Figure 8). The observed difference between maximal OCRs of the two groups could be the result of oxidative damage to CAC, ETC, and F O F 1 -ATPase components in the H 2 O 2 +olaparib group. Because the glycolytic fuel supply pathway was unimpeded in both groups, thanks to olaparib's inhibitory effect on PARP in the H 2 O 2 +olaparib group. Furthermore, the observed increase of maximal OCR in the H 2 O 2 +olaparib group vs. the H 2 O 2 group could result from direct control of the ETC by PARylation. Studies found excessively PARylated mitochondrial proteins, including components of ETC. In addition, PARP-inhibitors such as 3-aminobenzamide and nicotinamide prevented the H 2 O 2 -induced electron transport blockade on CIV in isolated mitochondria [67]. These results suggest that PARP could regulate ETC activity on CIV, and it also proposed a mitochondrial target for PARP inhibitors. Another study suggested a pivotal role for PARP-1 in mitochondrial energy homeostasis and demonstrated CI as a mitochondrial target of PARP-1 activation [68]. However, it should be noted that the existence of PARP-1 or other PARP isoforms in the mitochondria is debated. But whether present or not in the mitochondria, PARP has a clear effect on mitochondrial function [69].

Conclusion
In conclusion, olaparib, a PARP inhibitor used in human oncotherapy, restored bioenergetics by glycolytic reactivation of colonic epithelial cells, and it decreased cell death. Epithelial cell protection might be a cause of improved barrier function that eventually resulted in reduced incidence and severity of CD-like symptoms in an experimental rodent IBD model. However these findings provide experimental evidence for repurposing olaparib for IBD treatment and highlight its potential in the therapy of CD; clinical application of the drug in IBD needs further investigations.

Data Availability
The data underlying this article are available in the article.

Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.