INTRODUCTION:

Obesity is defined as the excessive or abnormal accumulation of fatty tissue in the body, and is commonly diagnosed using the Body Mass Index (BMI > 30 kg/m²) ¹, and it is estimated that in 2022, 8% of children and 16% of adults worldwide were obese². It is the second most common cause of preventable death after smoking, as an estimated 5 million noncommunicable disease deaths were caused by high BMI in 2019 alone ^{1, 2}. For every 5 increments increased in BMI above 25 kg/m², overall mortality increases by 29%, vascular mortality by 41%, and diabetes-related mortality by 210%. Moreover, obesity is implicit in multiple other conditions such as hypertension, chronic kidney disease, hyperlipidemia, nonalcoholic fatty liver disease, metabolic syndromes, certain types of cancer, obstructive sleep apnea, osteoarthritis, and depression ^{3, 4}.

Obesity also incurs a massive economic burden exceeding 700 billion dollars annually worldwide, and 100 billion in the USA alone ¹. Many efforts were made to manage obesity through various strategies. However, none have been successful in halting its prevalence, as the number of people suffering from obesity has almost doubled since 1980, with no general effective treatment in sight⁴. Multiple therapeutic compounds were developed in hope to incentivize weight loss, and there are currently 15 drugs used worldwide, with 6 being approved by the FDA, namely: Bupropion-naltrexone combination, Liraglutide, Orlistat, Phentermine-topiramate combination, Semaglutide, and Setmelanotide. However, apart from Orlistat, they all work through reducing appetite to varying degrees, which could lead to inhibition of nutritional intake with long term use ⁵. Orlistat (figure 1) on the hand works through irreversible inhibition of gastric and pancreatic lipase in the GI tract, which results in lowered fatty acid absorption (average of 30% lower), with only a small percentage (3-5%) of the drug is absorbed, and as such is a safer option for long term use ⁶. However, Orlistat does not present the same effectiveness seen with other medication such as Liraglutide, and takes longer to produce statistically significant weight loss⁷. Furthermore, a few clinical cases of orlistat induced hepatotoxicity have been reported recently^6-8. Thus, a new effective and safe alternative is in high demand for treating obesity and inducing weight loss. This study will use a molecular modelling approach in order to design novel potent potential pancreatic lipase inhibitors with a more favorable drug profile and low oral absorption to avoid potential systemic side effects.

Figure 1: Chemical formula of Orlistat

A. Pancreatic lipase (PL):

Lipases are a family of enzymes that break down triglycerides into free fatty acids and glycerol. The major human lipases include the gastric, pancreatic and bile salt stimulated lipase that aid in the digestion and of dietary fats in the GI tract, while the hepatic, lipoprotein and endothelial lipase aid in the metabolism of serum lipoproteins ⁹. Pancreatic lipase (PL) is the primary dietary lipase, as it hydrolyses 50-70% of all dietary triglycerides, which is essential for the absorption of the constituting fatty acids by enterocytes, and as such, is the main target for inhibition^10,11. PL is a serine hydrolase, and requires a cofactor – named colipase – to exhibit its enzymatic activity, and it functions through the catalytic serine 152 residue (part of a catalytic triad with histidine 263 and aspartic acid 176), situated in the active site in the N-terminal domain, and is surrounded by a large hydrophobic surface (figure 2). The access to the active site is controlled through a loop (called the lid), the motion of which reveals the catalytic serine residue and the large hydrophobic surface^{12, 13}.

Figure 2: The structure of pancreatic lipase

A. Lipase inhibitors:

Many lipase inhibitors were discovered or developed throughout the years, which work either through covalently or non-covalently binding to the active site pocket, and so far, covalent bonding has proven to be the more effective way of inhibiting PL. The reactive β-lactones, Alkyl phosphonate (figure 3), and carbonyl groups are some of the most widely studied serine hydroxyl binders ^{13, 14}. However, these reactive groups suffer from several drawbacks, such as sub-bar clinical effectiveness with carbonyls, and low selectivity due to the high reactivity of phosphonates. On the hand, β-lactones such as Orlistat have proven to be effective serine binders, but require specific stereochemistry for the reactive group ¹³.

Figure 3: General structure of β-lactones and Alkyl phosphonate

B. Selecting a new serine binding group:

In order to design potent PL inhibitors that avoid the aforementioned drawbacks, a new serine binder was used in the form of sulfonyl fluorides, as this group has been proven to be an effective inhibitor of serine hydrolases, by covalently and irreversibly binding to the hydroxyl group of serine residues through a sulfonylation reaction, forming a stable sulfonyl ester (figure 4)¹⁵. Furthermore, sulfonyl fluorides are easily and readily synthesized, and contain no chiral centers, and as such do not rely on specific stereochemistry ¹⁶.

Figure 4: Reaction of sulfonyl fluorides to serine residues

C. Improving binding to the hydrophobic surface surrounding the active site:

To further improve binding affinity, increased interactions with the hydrophobic surface was sought after. Multiple strategies were implemented, the most effective of which turned out to be the inclusion of flexible, long-chain hydrocarbons and fatty acid substituents within the structure of the potential inhibitors¹³, which formed multiple favorable hydrophobic interactions with the surrounding surface ^17-19.

METHODOLOGY:

A pharmacophore based virtual screening approach was implemented to define a suitable lead compound containing the serine binder sulfonyl fluoride group. The resulting lead compound was then optimized into a new scaffold, which was in turn used to design the novel inhibitors by adding and/or substituting various long-chain hydrocarbons and fatty acids.

The Molecular Operating Environment (MOE) 2022 ²⁰ and BIOVIA Discovery Studio 2016 ²¹ software were used in conducting this study.

D. Developing the pharmacophore:

1) Data set selection:

Selecting the appropriate compounds is a crucial step in developing a new pharmacophore, as such, the selection process should follow a set of general rules: compounds having different structures with similar activity, similar structures with different activities, and compounds spanning multiple orders of magnitude. Following the previous rules, 49 compounds were retrieved from the literature, and were drawn and prepared using the builder panel in the MOE interface.

2) Pharmacophore model generation:

The prepared compounds were first aligned using the flexible alignment protocol, iteration limit was set to 200, failure limit was set to 20, and energy cutoff was set to 15. The consensus function within the pharmacophore query editor was used on the aligned molecules to determine the shared pharmacophoric features, with a set tolerance of 1.2 A^o, and a similarity threshold of 70% or higher. A pharmacophore model was generated using the features with the highest similarity among the selected aligned compounds.

E. Virtual screening:

The developed pharmacophore was used to screen a suitable database in order to determine the lead compound.

1) Database selection and preparation:

The ENAMINE ²² sulfonyl fluoride curated database – containing 3953 compounds - was retrieved and prepared using Discovery Studio ligand preparation module. The prepared compounds were then subjected to a conformation search using MOE in order to generate all possible conformations for each compound. Iteration and conformation limits were set to 10000, rejection limit was set to 100, RMS gradient was set to 0.005 Kcal/mol/A^o2, and RMSD limit was set to 0.25 A^o.

2) Database screening:

The developed pharmacophore was used to screen the compound conformation database using MOE query editor. Only compounds matching all pharmacophore features were retained.

F. Molecular docking studies:

In order to evaluate binding affinity towards the PL active site, the resulting compounds were subjected to docking studies using the MOE covalent docking protocol.

1) Protein preparation:

The crystal structure of pancreatic lipase was retrieved from the RCSB protein data bank ^{23, 24} (PDB ID: 1lpb), which has good resolution (2.46 A^o), and a suitable ligand in the active site. The protein was then prepared using the structure preparation module in MOE software. The protein structure was corrected, and was thereafter protonated within a pH level of 6.5, which is the optimal pH level for PL enzymatic activity in vivo ²⁵. Lastly, the Amber10 forcefield ²⁶ was used to perform energy minimization with a gradient of RMS = 0.1 Kcal/mol/A^o2, and water molecules were deleted after completion.

2) Docking:

The covalent docking protocol was used to dock the resulting compounds into PL active site. Reaction type was set to sulfonyl halide sulfonylation, and the reactive site was set to active site residues to determine all possible reactions. Refinement was set to rigid receptor, and scoring methodology was set to GBVI/WSA dG.

G. Scaffold optimization and designing novel inhibitors:

The compound with the best docking score and the best predicted affinity was selected to be the lead compound. This compound was then optimized by studying its interactions with active site residues. The resulting scaffold was used to design the new inhibitors through addition/substitution of various hydrophobic long chain hydrocarbons and fatty acid chains. The novel inhibitors were then tested through docking studies to determine PL binding affinity using the same protocol mentioned previously. Finally, the ‘calculate molecular descriptor’ module in Discovery Studio was used to predict human intestinal absorption.

RESULTS AND DISCUSSION:

H. Developing the pharmacophore:

1) Data set selection:

The selected compounds are shown in table I.

2) Pharmacophore model generation:

Figure 5: The new pharmacophore, the blue spheres represent the hydrogen bond acceptor groups, and the orange spheres represent the Aromatic/Hydrophobic groups

The consensus function returned multiple similar features, the top four ranking features, comprising of two hydrogen bond acceptors (98% similarity), and two Aromatic/Hydrophobic groups (92% and 98% similarity), were selected to form the new pharmacophore. Figure 5 illustrates the 3D representation of the new pharmacophore.

A. Virtual screening:

Out of the 3953 compounds screened, 100 compounds matched all pharmacophore features within the set threshold. Figure 6 demonstrates the compound with the highest fitness value.

Figure 6: The compound with the highest fitness overlayed on the new pharmacophore

B. Molecular docking:

100 compounds were docked into the PL active site, 67 compounds successfully docked and bonded with ser152. Table II shows the results of the top 5 compounds compared to Orlistat which was docked using the same parameters (reaction type was lactone ring opening esterification). The docking results are represented with the value S, which is an estimate of the free energy of binding of the ligand from a given pose, and as such, the lower the value of S (larger negative number) the higher the estimated affinity will be. It is evident from the table that Orlistat exhibits better binding affinity than the 5 compounds, and upon inspection through comparing binding poses shown in figure 7, we see that both the first compound and Orlistat bind covalently to ser152, and form two hydrogen bonds with phe77 and leu153. However, where it comes to it, is the long hydrophobic chains of Orlistat, which significantly increase the favorable interactions with the hydrophobic surface of the active site. As such, optimizing the chosen compound is a must in order to improve its characteristics.

A. Scaffold optimization and designing novel inhibitors:

As was previously shown, hydrophobic interactions play a significant role in increasing binding affinity, it can also be incurred that hydrophobic group flexibility could play a role in insuring the right placement within the active site and around the surface. The first compound was chosen as a lead compound for designing the novel inhibitors since it exhibited the highest predicted affinity. The compounds structure and interactions were studied in order to shave off all redundant groups and improve upon others.

Table II: Top 5 potential lead compounds

Compound	S
	-6.2153
	-6.2074
	-6.2010
	-6.1836
	-6.1548
Orlistat	-7.8123

Figure 8 presents the new scaffold obtained from the lead compound, and it was found that a more flexible large group would contribute more favorably through hydrophobic interactions compared to the rigid β-phenyl quinolyl substituent, and was therefore substituted with an acyl group.

Figure 7: 2D and 3D representation of Orlistat (A) and the first compound (B) docked into PL active site

Figure 8: The new scaffold obtained from the lead compound

Furthermore, it was found that the inclusion of a long alkyl chain in the ortho position relative to the sulfonyl fluoride mimics the hexyl substituent formed after Orlistat binds to ser152, and helps improve hydrophobic interactions. In an effort to simplify the structure of the novel inhibitors, and reduce potential synthesis costs, the selected acyl groups were almost all naturally occurring and abundant fatty acids, such as oleic, stearic, and palmitic acids. Table III presents the novel potential inhibitors and their corresponding docking score.

Table III: the novel inhibitors and their docking scores

compound	R₁	Acyl	S
1	decyl	Docosanoyl	-8.722
2	decyl	Arachidyl	-8.626
3	decyl	α-Linolenyl	-8.261
4	nonyl	Arachidonyl	-8.241
5	decyl	Elaidyl	-8.179
6	nonyl	Myristyl	-8.171
7	decyl	Oleoyl	-8.159
8	nonyl	Elaidyl	-8.13
9	nonyl	Linoleoyl	-8.073
10	decyl	Myristyl	-8.024
11	nonyl	Palmityl	-7.986
12	decyl	Arachidonyl	-7.774
13	decyl	Palmityl	-7.769
14	nonyl	Arachidyl	-7.758
15	nonyl	Docosanoyl	-7.758
16	nonyl	α-Linolenyl	-7.731
17	decyl	Lauryl	-7.526
18	nonyl	Stearyl	-7.498
19	nonyl	Oleoyl	-7.48
20	nonyl	Lauryl	-7.356
21	decyl	Linoleoyl	-7.311
22	decyl	Stearyl	-6.485

It is apparent from the results that on average longer hydrophobic chains increase predicted affinity, as was speculated. All compounds managed to bind to ser152 through sulfonylation, and out of the 22 compounds, 11 exhibited higher predicted affinity than Orlistat. Moreover, most compounds formed two hydrogen bonds with phe77 and leu153 via the sulfonyl oxygen, however this did not correlate directly to increased binding affinity. Figure 9 demonstrates the poses of both compound 1 and 4 within the active site. Finally, the molecular descriptors of all novel inhibitors were calculated, and the human oral absorption was estimated. All compounds returned a value of 3, which indicates very poor absorption, indicating minimal potential for systemic side effects.

Figure 9: 2D and 3D representation of compound 1 (A) and compound 4 (B) docked into PL active site

CONCLUSION:

Obesity more than doubled in its prevalence in the last 40 years, and it has been implicated in a multitude of dangerous diseases. This study aimed to make use of molecular modelling strategies to design potential pancreatic lipase inhibitors for the treatment of obesity and inducing weight loss in obese patients.

A new pharmacophore was developed and was used to scan a database of compounds containing the selected serine binder sulfonyl fluoride group, followed by docking studies to determine a suitable lead compound. The lead was then optimized, and the new scaffold was introduced to long-chain alkyls and acyls to improve hydrophobic bonding with the active site. Finally, oral absorption was predicted for the novel compounds. 11 compounds exhibited higher predicted affinity than the control drug Orlistat, in addition to minimal oral absorption.

REFERENCES:

1. Panuganti, K.K., M. Nguyen, and R.K. Kshirsagar. Obesity, in StatPearls. 2024, StatPearls Publishing

2. The World Health Organization. Feb 2024]; Available from: https://www.who.int/.

3. Apovian, C.M. Obesity: definition, comorbidities, causes, and burden. Am J Manag Care. 2016. 22(7 Suppl): p. s176-85.

4. Lin, X. and H. Li. Obesity: Epidemiology, Pathophysiology, and Therapeutics. Front Endocrinol (Lausanne). 2021; 12: 706978. https://doi.org/10.3389/fendo.2021.706978

5. S. N. C, S., S. Palawat, and A.T. Paul. Design, synthesis, biological evaluation and molecular modelling studies of conophylline inspired novel indolyl oxoacetamides as potent pancreatic lipase inhibitors. New Journal of Chemistry. 2020; 44(28): 12355-12369. https://doi.org/10.1039/D0NJ02622K

6. Bansal, A.B. and Y. Al Khalili. Orlistat, in StatPearls. 2024; StatPearls Publishing

7. Valladales-Restrepo, L.F., et al. Effectiveness, persistence of use, and safety of orlistat and liraglutide in a group of patients with obesity. Expert Opin Pharmacother. 2023; 24(4): 535-543. https://doi.org/10.1080/14656566.2023.2178900

8. Diseases, B.M.N.I.o.D.a.D.a.K. LiverTox: Clinical and Research Information on Drug-Induced Liver Injury [Internet]. 2012-2024.

9. Mukherjee, M. Human digestive and metabolic lipases—a brief review. Journal of Molecular Catalysis B: Enzymatic. 2003; 22(5): 369-376. https://doi.org/https://doi.org/10.1016/S1381-1177(03)00052-3

10. Zhang, J., et al. Design, Synthesis, and Structure-Activity Relationship Study of Pyrazolones as Potent Inhibitors of Pancreatic Lipase. ChemMedChem. 2021; 16(10): 1600-1604. https://doi.org/https://doi.org/10.1002/cmdc.202000850

11. Birari, R.B. and K.K. Bhutani. Pancreatic lipase inhibitors from natural sources: unexplored potential. Drug Discov Today. 2007; 12(19-20): 879-89. https://doi.org/10.1016/j.drudis.2007.07.024

12. Colin, D.Y., et al. Exploring the active site cavity of human pancreatic lipase. Biochemical and Biophysical Research Communications. 2008; 370(3): 394-398. https://doi.org/https://doi.org/10.1016/j.bbrc.2008.03.043

13. Kumar, A. and S. Chauhan. Pancreatic lipase inhibitors: The road voyaged and successes. Life Sciences. 2021; 271: 119115. https://doi.org/https://doi.org/10.1016/j.lfs.2021.119115

14. Lunagariya, N.A., et al. Inhibitors of pancreatic lipase: state of the art and clinical perspectives. Excli j. 2014; 13: 897-921.

15. Narayanan, A. and L.H. Jones, Sulfonyl fluorides as privileged warheads in chemical biology. Chem Sci. 2015; 6(5): 2650-2659. https://doi.org/10.1039/c5sc00408j

16. Carneiro, S.N., et al. Sulfur(vi) fluorides as tools in biomolecular and medicinal chemistry. Organic and Biomolecular Chemistry. 2023; 21(7): 1356-1372. https://doi.org/10.1039/D2OB01891H

17. Tian, Y., et al. Efficient screening of pancreatic lipase inhibitors from cod meat hydrolysate through ligand fishing strategy. Front Nutr. 2022; 9: 969558. https://doi.org/10.3389/fnut.2022.969558

18. Thomas, A., et al. Role of the Lid Hydrophobicity Pattern in Pancreatic Lipase Activity. Journal of Biological Chemistry. 2005; 280(48): 40074-40083. https://doi.org/https://doi.org/10.1074/jbc.M502123200

19. Samual, T. A comprehensive review on the pancreatic lipase inhibitory peptides: A future anti-obesity strategy. Electronic Journal of General Medicine. 2023; 20: 2516-3507. https://doi.org/10.29333/ejgm/12943

20. Group, C.C. Molecular Operating Environment. 2022.

21. Biovia, D.S., Discovery Studio 2016, San Diego: Dassault Systèmes. 2016.

22. Shivanyuk, A., et al. Enamine real database: Making chemical diversity real. Chimica Oggi. 2007; 25: 58-59.

23. Berman, H.M., et al. The Protein Data Bank. Nucleic Acids Research. 2000; 28(1): 235-242. https://doi.org/10.1093/nar/28.1.235

24. The RCSB Protein database. 19 October 2023]; Available from: http://www.rcsb.org/.

25. Colin, D.Y., et al. Modification of pancreatic lipase properties by directed molecular evolution. Protein Engineering, Design and Selection. 2010; 23(5): 365-373. https://doi.org/10.1093/protein/gzq008

26. Case, D.A., et al. The Amber biomolecular simulation programs. J Comput Chem. 2005; 26(16): 1668-88. https://doi.org/10.1002/jcc.20290

Received on 23.03.2024 Revised on 12.09.2024

Accepted on 25.12.2024 Published on 28.01.2025

Available online from February 27, 2025

Research J. Pharmacy and Technology. 2025;18(2):831-838.

DOI: 10.52711/0974-360X.2025.00123

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License.