Drug Repurposing for Targeting ISL LIM Homeobox 2 in Treatment of Endometriosis: A Computational Study

Endometriosis is a prevalent gynaecological disorder characterised by active lesions of endometrial tissue, including glandular cells and stroma, located outside the uterus that may affect fertility. The overall prevalence of this chronic inflammatory disease may be around 10-15% in reproductive age females ( 1 ). The exact pathogenesis of endometriosis has not been completely determined; however, it is believed to be influenced by factors such as retrograde menstruation, benign metastasis, hormonal imbalance, immune dysregulation, and genetic and epigenetic changes ( 2 , 3 ). Genes play key roles in female reproductive biology, and the role of the homeobox ( HOX ) family genes has been shown from the formation of the uterus to endometrial receptivity ( 4 ). HOX genes express in the Müllerian duct during the development of the female reproductive tract and are regulated by cyclic hormonal changes during each reproductive cycle. This regulation is essential for the growth, differentiation, and implantation of the endometrium to facilitate successful embryo implantation ( 4 , 5 ). One of the cofactors for HOX genes is ISL LIM homeobox 2 (ISL2). Our previous study results revealed significantly higher expression of the ISL2 gene in endometriotic tissues compared to normal endometrium ( 6 ). On the other hand, it has been reported that ISL2 is overexpressed in glioma tumours and promotes the angiogenesis, proliferation, and invasion of human brain microvessel endothelial cells through the vascular endothelial growth factor A (VEGFA)-mediated ERK signalling pathway ( 7 ). Angiogenesis is a principal factor in the aberrant growth of endometrial tissue outside the uterus. VEGFA plays an essential role in promoting angiogenesis due to its higher expression in endometriosis patients compared to healthy women ( 8 ). Based on the results of our previous study ( 9 ) and considering the association of ISL2 with HOX genes and VEGF , as well as its role in promoting angiogenesis, it appears that inhibiting ISL2 could serve as an effective treatment for endometriosis. In the current study, we use drug repurposing to target ISL2 in an attempt to identify the most suitable food and drug administration (FDA)-approved drug for endometriosis treatment. To substantiate the selection of ISL2 as a viable target for this research, we evaluated its expression in the endometrial tissues of patients with endometriosis compared to a control group using reverse transcription-polymerase chain reaction (RTPCR) and real-time PCR. Subsequently, computational methods that included molecular dynamics (MD) simulations and molecular docking were utilised to identify the optimal inhibitor for ISL2 from among the FDAapproved drugs.

qRT-PCR analysis was conducted to evaluate the expression levels of the ISL2 gene in both eutopic and ectopic endometrial tissues. Our data revealed a significant increase in ISL2 expression levels (P=0.0036) in the eutopic tissues of patients with endometriosis compared to the control and ectopic groups. ISL2 expression in ectopic endometrial tissues also increased (P=0.0019) compared to the control group ( Fig .2 ). Comparative expressions of ISL2 in eutopic and ectopic endometrial tissues versus normal endometrial tissues. ISL2 expression levels increased in eutopic (n=13) and ectopic (n=13) endometrial tissues in endometriosis patients compared to normal endometrium from the control group (n=8, **; P0.05) in eutopic versus ectopic tissues. Each of these tests were repeated twice. ISL2 ; ISL LIM homeobox 2. As depicted in Figure 2, the ISL2 expression showed a significant and several hundredfold increase in endometriotic tissues of patients compared to the control group. Therefore, we chose ISL2 as the target for drug repurposing by computational methods. The results obtained from computational methods were utilised to identify the most suitable FDA-approved drug to inhibit ISL2. Initially, we assessed the ability of 2471 FDA-approved drugs to interact with the identified binding pocket of ISL2. The preliminary docking results indicated that, out of the pool of 2471 FDA-approved drugs, six drugs (Dactinomycin, Paritaprevir, Ivermectin, Ergotamine, Alectinib, and Simeprevir) exhibited the most favourable binding energy (∆G ≤-8 kcal.mol -1 ) with the identified binding pocket on ISL2. Two-dimensional and three-dimensional representations of the complexes formed between these drugs and ISL2 are shown in the ( Figes.S1-S6 , See Supplementary Online Information at www.ijfs.ir ). As depicted, Ivermectin established the greatest number of hydrogen bonds with ISL2. Next, we used MD simulations to assess the stability of these complexes by calculating RMSD, RMSF, the number of hydrogen bonds, and the number of contacts between the drugs and the protein. Overall, the RMSD and RMSF analyses provide insights into the stability of the complexes (Figes .3 , 4 ). As shown, the RMSD value converges at 80 ns. The RMSD value for Ergotamine was higher than the other complexes, which indicated that other complexes were more stable than this complex. RMSF analysis examines the fluctuation of various parts of the structure from their mean positions. Among the complexes, the Paritaprevir complex had the highest RMSF value, which indicated a greater flexibility of the protein in this complex compared to the other complexes. RMSD graphs for protein in complexes of ISL2 /drugs during 100 ns of the MD simulation period. The RMSD value converges at 80 ns. The RMSD value for Ergotamine was higher than the other complexes. ISL2; ISL LIM homeobox 2, RMSD; Root mean square deviation, and MD: Molecular dynamics. RMSF graphs for protein in complexes of ISL2/drugs during 100 ns of the MD simulation period. The RMSF analysis examines the fluctuation of various parts of the structure from their mean positions. ISL2; ISL LIM homeobox 2, RMSF; Root mean square fluctuation, and MD: Molecular dynamics. The most appropriate bonding between ISL2 and the drugs selected in the simulation analyses can be evaluated by considering the number of hydrogen bonds and contacts. Our findings indicate that Ivermectin, Paritaprevir, and Dactinomycin formed the highest number of hydrogen bonds with ISL2 ( Fig .5 ). The Ivermectin and Paritaprevir complexes had the highest number of contacts with ISL2 ( Fig .6 ). The number of H-bonds between the drugs and ISL2. Ivermectin, Paritaprevir, and Dactinomycin formed the highest number of hydrogen bonds with ISL2. ISL2; ISL LIM homeobox 2. The number of contacts between the drugs and ISL2. Ivermectin and Paritaprevir formed the highest number of contacts with ISL2. ISL2; ISL LIM homeobox 2.

Endometriosis is a female health disorder for which the exact cause and definitive treatment remain uncertain. Consequently, extensive studies have been conducted to select key genes implicated in the development of this disease and to identify the most effective drugs for its treatment ( 23 , 24 ). Computational methods allow for identifying structures that have a high likelihood of binding to a drug target, with the intent to reduce cost and time, and increase accuracy. One of the strategies in drug discovery for disease treatment is drug repurposing ( 25 , 26 ). This approach, which has previously yielded numerous promising candidates, may rationally repurpose drugs that have already undergone the necessary stages for FDA approval for treatment of diseases that have uncertain treatments ( 27 ). Recent studies have been conducted to determine the role of ISL2 in cancer development. In tumours like oligodendrogliomas, ISL2 expression showed a significant increase. One potential pathway for disease development due to ISL2 involves the elevated expression of angiopoietin 2 ( ANGPT2 ), which occurs through ISL2 binding to the promoter region of the ANGPT2 gene. Increased ANGPT2 expression results in cancer cell proliferation, invasion, and metastasis ( 28 ). Additionally, angiopoietins are integral to angiogenesis, and play a crucial function in the pathogenesis of endometriosis ( 29 - 31 ). These findings allowed researchers to consider ISL2 as a critical target for drug design to treat these conditions. ISL2 induces VEGF A expression, which is one of the pivotal angiogenesis factors that promotes tumour growth and development ( 32 ). VEGF A has a substantial impact on the development of endometriosis ( 33 ). Consequently, the increased expression of ISL2 could lead to heightened levels of ANGPT2 and VEGF expressions in endometriosis patients. We observed a significant and several hundredfold increase in ISL2 gene expression in endometriotic tissues of endometriosis patients compared to the control group. Based on the outcomes derived from the initial step of this study, ISL2 was selected as the target for drug repurposing by computational methods. First, we observed the interactions of 2471 FDA-approved drugs with the identified binding pocket within the LIM domain of ISL2. The preliminary docking results revealed that, among these 2471 FDA-approved drugs, six (Dactinomycin, Paritaprevir, Ivermectin, Ergotamine, Alectinib, and Simeprevir) exhibited the most favourable binding energies (∆G ≤-8 kcal.mol -1 ) with the identified pocket within the LIM domain of the ISL2 protein. Twodimensional representations of the docking results for these six drugs show that Ivermectin formed the highest number of hydrogen bonds with the LIM domain MD simulations were then employed to assess stability of the six drug complexes. The results indicated that, from these six complexes, Ivermectin exhibited the lowest RMSD and RMSF, and highest number of hydrogen bonds and contacts, which indicated a higher potential for forming a stable complex with ISL2. Our findings indicate that Ivermectin is more likely to effectively inhibit ISL2. Currently, numerous studies have used in vivo and in vitro methods to investigate the role of Ivermectin as a cancer treatment ( 34 - 36 ). Li and Zhan ( 37 ) observed that this drug significantly suppressed the growth of various cancer cell types, including ovarian cancer. Ivermectin also appears to be an effective glioma treatment ( 38 ). Although the precise molecular mechanism of its anti-cancer action has not been fully elucidated, the study outcomes suggest that its effectiveness might be linked to inhibition of ISL2. The outcome of our research could improve diagnostic methods and enable development of new therapies for endometriosis. Limitations of this study included challenges in sample collection, as the inclusion and exclusion criteria narrowed the pool of eligible participants.

ISL2 plays a role in enhancing angiogenesis and it has elevated expression in the endometrial tissues of patients diagnosed with endometriosis. Therefore, inhibiting this protein could be efficacious in modulating endometriosis. Based on the results obtained from the docking and molecular simulation conducted in this study, Dactinomycin, Paritaprevir, Ivermectin, Ergotamine, Alectinib, and Simeprevir all have the potential to inhibit ISL2. It is recommended that the efficacy of these drugs be assessed in laboratory environments and animal models to evaluate their impact.

The medical Ethics Committee at Royan Institute, Tehran, Iran approved this study (IR.ACECR.ROYAN. REC.1400.157, March 2, 2022). Written informed consent was obtained from all participants, in accordance with the guidelines of the Declaration of Helsinki 2000. Participants’ consent was received before collecting tissue samples. The participants were 20-40 years of age and exhibited no signs of endometrial hyperplasia, endometriosis, visible endometrial hyperplasia or neoplasia, inflammatory diseases, myoma, polyps, or reproductive inflammatory diseases. All women were in the secretory and proliferative phases of their menstrual cycles and had not received hormone therapy for a period of three months ( Table S1 , See Supplementary Online Information at www.ijfs.ir ). The women with endometriosis were diagnosed with stages III or IV disease, as classified by the revised American Society for Reproductive Medicine classification. This computational study assessed ectopic, eutopic, and normal endometrial tissue samples. Ectopic (n=6) and eutopic (n=12) tissue samples were collected from women diagnosed with endometriosis who were undergoing laparoscopy at Royan Institute (Tehran, Iran). Ectopic endometrial lesions were obtained through laparoscopic procedures, while eutopic endometrium biopsy specimens were collected from the uterine cavity in the same procedure using Pipelle sampling. Normal endometrial tissues (control group) were obtained from eight healthy women during diagnostic laparoscopy, and tissue samples were collected from the uterine cavity using Pipelle sampling. All tissues were gently stirred in Dulbecco’s phosphate-buffered saline (Cat. No. 21600, Life Technologies, USA) for 20 minutes to remove excess blood and debris. Total RNA was extracted from the samples with TRIzol reagent (Cat. No. 15596, Life Technologies, USA) according to the manufacturer’s protocol, followed by DNase I treatment (DNase I, RNase-free, Cat. No. EN0521, Thermo Fisher Scientific, USA) to eliminate any genomic DNA contamination. Subsequently, cDNA synthesis was performed using a cDNA synthesis kit (ExcelRT™ Reverse Transcription Kit, SMOBIO Technology, Cat. No. RP1300, Taiwan) following the manufacturer’s protocol. The expression of the target gene in all samples was assessed by quantitative real-time PCR (qRT-PCR). The SYBR Green q-PCR master mix (RealQ Plus 2x Master Mix Green with high ROX™, Ampliqon, Cat. No: A323402, Denmark) was utilised for the qRTPCR reactions. Table S2 (See Supplementary Online Information at www.ijfs.ir ) lists the primer sets, lengths and product sizes. The primer was designed using PerlPrimer software (version 1.1.21) and verified by Gene Runner software (version 3.05, Informer Computer Terminals, USA). The qRT-PCR reaction consisted of three stages: a holding stage (10 minutes at 95°C for one cycle), a cycling stage (15 seconds at 95°C, followed by one minute at 60°C for 40 cycles), and the melting curve stage (15 seconds at 95°C, one minute at 60°C, and 15 seconds at 95°C). The relative expression of ISL2 in the sample was calculated using the 2 -ΔΔCt technique with GAPDH as the internal control. The LIM domain is a conserved region within the ISL2 protein, and its malfunction can lead to pathological effects such as tissue detachment, embryonic lethality, and cancer development ( 10 ). The LIM domain plays a significant role in forming protein complexes and binding ISL2 to promoter genes ( 11 ). Therefore, targeting the LIM domain could be a strategy to inhibit ISL2. In this study, the LIM domain of the ISL2 protein was selected as the target for molecular docking and MD simulations. The crystal structure of the human LIM domain was not available in the protein data bank; therefore, human and mouse LIM sequences were compared using the Basic Local Alignment Search Tool program (National Centre for Biotechnology, USA). We confirmed that the mouse LIM domain sequence has 100% identity with the human LIM domain amino acid sequence. Consequently, the crystal structure of the LIM domain with the PDB code 3MMK was downloaded from the PDB database ( 12 ). Next, water molecules were eliminated from the coordinate data, and hydrogen atoms were added to optimise hydrogen bonding interactions. The GROMACS 5.1.4 package was employed to perform energy minimisation of the protein structure ( 13 ). The DrugBank database was searched to obtain 2471 FDA-approved drugs ( 14 ). In this stage, the pocket identification was performed using the CB-Dock web server ( 15 ). The cavity detection is an effective method to enhance molecular docking and a representation of this pocket is shown in Figure 1. The cavity volume was 444 Å3, with a cavity centre located at coordinates 34, 38, 21. The cavity size was 9, 13, 9 with a grid spacing of 0.375Å. In terms of ligand preparation, the PyRx tool was employed to modify energy minimisation, and ligands were converted into the PDBQT format. The AutoDock Vina tool integrated in PyRx was utilised to conduct virtual screening ( 16 , 17 ) and the docking results were saved to evaluate binding affinities. Eventually, ligand-target bonding interactions were examined using Discovery Studio software. Cavity detection using the CB-DOCK web server. The cavity volume is 444 Å3, with a cavity centre located at coordinates 34, 38, 21. The cavity size is 9, 13, 9, with a grid spacing of 0.375Å. Once all the FDA-approved drugs were docked with ISL2, the medications that had a desirable delta G (∆G ≤-8 kcal/mol) with ISL2 were chosen for further analyses. We aimed to identify drugs that form a more stable complex and have the strongest binding to the protein. Therefore, root mean square fluctuation (RMSF), root mean square deviation (RMSD), the number of hydrogen bonds, and the number of contacts between drugs and the protein were calculated using MD simulations. The GROMACS 5.1.4 software (Netherlands) was employed to accomplish all MD simulations in this study ( 18 ) within the GROMOS 54a7 force field. The ATB server was used to prepare ligand coordinates and topology. The system was neutralised by adding the necessary quantities of chloride ions and sodium using modelling packages. The periodic boundary conditions were implemented along the axial directions of the simulation box in each modelling framework. The SPC water simulation was applied for systemic solubilisation ( 19 ). All covalent bonds were restricted and Van der waals interactions were cut off at 1.2 nm ( 20 ). Energy minimisation of the structures was performed with 50000 steps. MD simulations were conducted at a temperature of 310 K for a duration of 100 nanoseconds ( 21 , 22 ). The Prism Graphpad was used for all statistical analyses (Insight Partners, version 9.5.1, USA). The data are presented as mean ± standard error of the mean (SEM) and 99% confidence (**; P<0.01). Gene expression in the three tissue groups was compared using the KruskalWallis test.