FULL PAPER www.advtherap.com Development of Innovative Antiatherosclerotic Peptides through the Combination of Molecular Modeling and a Dual (Biochemical-Cellular) Screening System Aleyda Benitez-Amaro, Chiara Pallara, Laura Nasarre, Ruben Ferreira, David de Gonzalo-Calvo, Roger Prades, Teresa Tarragó, and Vicenta Llorente-Cortés* 1. Introduction Cardiovascular disease (CVD) is a leading cause of death worldwide. Approximately 60% of patients treated with low-density lipoprotein Ischemic heart disease is the primary cause of death in Western countries, and myocar- (LDL)-lowering drug treatments, with on-target plasma cholesterol levels, are dial infarction accounts for ≈50% of deaths still suffering clinical acute ischemic events. Mechanisms, such as LDL from this disease. The Framingham study aggregation, underlie extracellular and intracellular cholesterol accumulation showed that cardiovascular risk positively in the vasculature. A peptide sequence (P3) of the low-density lipoprotein correlates with low-density lipoprotein receptor-related protein 1 (LRP1) efficiently protects LDL from (LDL)-cholesterol and inversely with high- density lipoprotein (HDL)-cholesterol.[1,2] sphingomyelinase (SMase-) and phospholipase A2 (PLA2 )-induced LDL High levels of LDL-cholesterol induce al- aggregation. The aim is to design families of peptide derivatives from P3 with terations in the endothelium permeability, enhanced potency and proteolytic stability. New peptides are designed which in turn promote LDL influx into the through in silico conformational sampling and ApoB-100 molecular docking, arterial intima.[3] Intimal LDL-cholesterol and are tested in dual (biochemical-cellular) screening assays. A total of 46 accumulation is a critical step in vascular new peptides including linear, fragment, cyclic, and alanine scanning cholesteryl ester (CE) deposition, a pro- cess that increases the tendency of the derivatives are generated through two consecutive optimization rounds. atherosclerotic plaque to rupture, trigger- Structurally and functionally optimized peptides contain hotspot residues that ing thrombosis and the development of are replaced by alanine. This strategy confers an increased capacity to form ischemic cardiomyopathy.[4–6] Cholesteryl prone alpha-helix conformations crucial for the electrostatic interaction with esters in atherosclerotic plaques are de- ApoB-100. These new compounds are highly efficient at inhibiting LDL posited both extra- and intracellularly. Ex- tracellular deposition of LDL-CEs, a central aggregation and human coronary vascular smooth muscle cell-cholesteryl initiating event in atherosclerosis, is medi- ester loading and should be studied in preclinical models of atherosclerosis. ated by proteoglycans in the extracellular matrix of the arterial intima. The electro- static interaction between proteoglycans A. Benitez-Amaro, L. Nasarre, Dr. D. de Gonzalo-Calvo, A. Benitez-Amaro, Dr. D. de Gonzalo-Calvo, Dr. V. Llorente-Cortés Dr. V. Llorente-Cortés Group of Lipids and Cardiovascular Pathology Institute of Biomedical Research of Barcelona (IIBB) Biomedical Research Institute Sant Pau (IIB Sant Pau) Spanish National Research Council (CSIC) Hospital de la Santa Creu i Sant Pau Barcelona 08036, Spain Barcelona 08025, Spain E-mail: vicenta.llorente@iibb.csic.es; Cllorente@santpau.cat Dr. C. Pallara, Dr. R. Ferreira, Dr. R. Prades, Dr. T. Tarragó Iproteos S.L Barcelona Science Park (PCB) Barcelona 08028, Spain The ORCID identification number(s) for the author(s) of this article Dr. D. de Gonzalo-Calvo, Dr. V. Llorente-Cortés can be found under https://doi.org/10.1002/adtp.202000037 CIBER enfermedades cardiovasculares (CIBERcv) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. Madrid 28029, Spain KGaA, Weinheim. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. DOI: 10.1002/adtp.202000037 Adv. Therap. 2020, 3, 2000037 2000037 (1 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com and LDL and the proteolytic/lipolytic actions of enzymes on LDL Table 1. Parameters for validation of biochemical (SMase-TB and PLA2 -TB) are enhanced in the arterial intima and promote LDL retention and cell-based (SMase-CE/FC) assays. These parameters were obtained from the mean and SD of the values obtained with positive (DP3) and and aggregation.[7,8] Two of the main enzymes that act on intimal negative control (P321) in ten different experiments performed by qua- retained LDL and play key roles in LDL aggregation in the arte- triplicate (SMase-TB assay) or duplicate (PLA2 -TB and SMase-CE/FC ra- rial intima during atherogenesis are sphingomyelinase (SMase) tio assays). SMase: Sphingomyelinase; PLA2 : phospholipase A2 ; TB: tur- and phospholipase A2 (PLA2 ).[9–12] Aggregated LDL (agLDL) has bidimetry; CE: cholesteryl esters; FC: free cholesterol. been detected and isolated from atherosclerotic plaques from an- imal models and humans.[10,13] Unlike native LDL, agLDL is a SMase-TB PLA2 -TB SMase-CE/FC potent inducer of massive intracellular CE accumulation both CV interassay [%] 11.60 7.97 12.79 in macrophages and human coronary vascular smooth muscle CV intra-assay [%] 6.24 6.32 8.47 cells (hcVSMCs).[14–18] In hcVSMCs, we reported that agLDL is Estimated Z-factor 0.86 0.60 0.75 actively taken up through the low-density lipoprotein receptor- related protein 1 (LRP1) which, in turn, induces LRP1 expres- sion, promoting a positive feedback loop that efficiently trans- forms hcVSMCs into foam cells.[19,20] hcVSMC-foam cells syn- 2. Results thesize and release high amounts of tissue factor, which is cru- cial for the prothrombotic transformation of the vascular wall and 2.1. Feasibility of Biochemical and Cell-Based Assays for thus for the progression of atherosclerosis to thrombosis.[21,22] Compound Screening The relevance of this mechanism in atherosclerosis is evident since vascular smooth muscle cells (VSMCs) are the main com- The scheme of the biochemical SMase and PLA2 -induced LDL ponent of the vascular wall and more than 50% of foam cells turbidimetry (SMase-TB and PLA2 -TB assays) and cell-based as- previously considered to be monocyte-derived macrophages in says (SMase-CE/free cholesterol (FC) assay) is described in Fig- human atherosclerotic plaques originate from VSMCs.[23] To- ure S1 of the Supporting Information. The potential of these as- gether, these findings support the notion that the generation of says as screening tools was evaluated by calculating the intra- hcVSMC-derived foam cells through LRP1-mediated agLDL up- assay and interassay coefficient of variation (CV) using a pos- take is a key mechanism underlying cholesterol accumulation in itive (DP3) and a negative control (P321) through several ex- vasculature susceptible to atherosclerosis. Our group has iden- periments performed in quadruplicate (SMase-TB) or duplicate tified LRP1 cluster II, in particular the region Gly1127 -Cys1140 (PLA2 -TB and SMase-CE/FC) (Figure S2, Supporting Informa- (peptide LP3: H2 N-GDNDSEDNSDEENC-NH2 ) that spans the tion). As shown in Table 1, the assays have an interassay CV C-terminal half of domain CR9, as pivotal for binding to agLDL < 15% and an intra-assay CV < 10%, showing the robustness of and subsequent internalization of agLDL into human VSMCs.[24] the assays and their feasibility to be used as evaluation tools. In Moreover, we have shown that anti-LP3 antibodies reduce high addition, we calculated the Z-factor, a pivotal parameter for com- fat diet-induced atherosclerosis in a rabbit model.[25] In addi- parison and evaluation of the quality of the assays that reflects tion, we have shown that LRP1-derived peptides (the original LP3 both the assay signal dynamic range and the data variation asso- and its retroenantiomer version-DP3) are protective against LDL ciated with the signal measurements and therefore suitable for aggregation, even in conditions of extreme lipolysis due to the assay quality assessment.[27] The calculated Z-factor for the three maintenance of ApoB-100 conformation.[26] We showed that DP3 assays was between 0.5 and 1.0 (Table 1), indicating the suitabil- forms a complex with ApoB-100 and this molecular interaction ity of these assays for screening of the effect of compounds on stabilizes ApoB-100 conformation. ApoB-100 conformation sta- LDL aggregation and foam cell formation. bilization might guarantee the structural preservation of surface cholesterol-enriched environments, where sphingomyelin (SM) is located. Structural preservation of cholesterol, a key regula- 2.2. Computational Design and Screening of Peptides from the tor of phospholipolysis, would protect SM from SMase activity. First-Round Optimization As a result, LDL complexed with DP3 remains unaltered when exposed to SMase. This scenario changes when LDL complexed 2.2.1. Computational Peptide Design with DP3 is exposed to PLA2 . The target for PLA2 is phosphatidyl- choline (PC), a phospholipid associated with low-cholesterol en- The computational approach used to design optimized peptides vironments. Therefore, PC is unprotected against the attack of yielded 20 exploratory compounds that were produced by solid- PLA2 , which hydrolyses PC producing lysoPC and nonesteri- phase peptide synthesis (summarized in Figure 1). The charac- fied fatty acids. Remarkably, LDL complexed with DP3 is pro- terization and sequence of designed peptides have been detailed tected against SMase and PLA2 -induced aggregation even in con- in Tables S1 and S2 of the Supporting Information, respectively. ditions of extreme phospholipolysis, indicating that the main- Considering the lack of structural information about the LRP1 tenance of ApoB-100 conformation is enough to prevent LDL receptor, the ApoB-100 protein and the low reliability of the cor- aggregation. responding homology models currently available, the computa- The main objective of the present study is to design and eval- tional design of CR9-based peptide analogues was performed by uate a series of antiatherosclerotic families of peptides with po- combining a ligand-based method that completely neglects re- tential for in vivo application. ceptor 3D information and relies only on the physicochemical Adv. Therap. 2020, 3, 2000037 2000037 (2 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Figure 1. Schematic representation of the designed peptide groups through the development of the first-round optimization. Ala: Alanine. and dynamic properties of the ligands, with a structure-based tire surface, it is known that in many cases only a small subset protocol, based on molecular docking approach. of individual amino acids, the so-called hotspots, contribute the The first molecular strategy was based on the hypothesis that most to the free energy of binding.[31] Homology modeling and unbound biomolecules (i.e., proteins, nucleic acids, peptides, molecular dynamics methods combined with molecular dock- small molecules) naturally adopt a variety of conformational ing provided the structural model of LP3 peptide in complex states, a subset of which are suitable to bind to their biologi- with ApoB-100 and offered a detailed insight into the mechanis- cal partner.[28,29] An exhaustive conformational sampling of the tic basis of LP3 recognition by ApoB-100 protein.[26] Moreover, LP3 peptide was performed by molecular dynamics (MD) sim- this result facilitated the identification of several specific residues ulations. The most populated low energy states were extracted, (namely Glu6 , Asp7 , and Asp10 ) found to be deeply buried upon considered to the ones most similar to the bound state, and then, binding. To confirm the key role of these residues, the suspected if possible, conformationally restricted in cyclic analogues by mu- hotspot residues in the P17 sequence were replaced by alanine tating two specific residues to cysteine, chosen as the most suit- residues because of its small and nonreactive side chain, thus able residues for creating rigid rings with relatively low synthetic obtaining several LP3 alanine scanning analogue peptides (P8, complexity. Finally, this strategy led to six cyclic analogues mim- P9, P10, P11, and P12). icking low-energy near-native bound conformations of LP3 pep- Shorter peptide sequences, derived from the parent peptide tide, namely peptides P1, P2, P3, P4, P5, and P6. (LP3) were also analyzed. These peptides (P13, P14, P15, P16, The second computational protocol emerged from the idea P17, P18, and P19) were useful to identify the minimal peptide that interacting partners that are involved in initial encounters sequence length responsible for the inhibitory activity of LDL ag- evolve toward the final specific complex by mutually adjusting gregation. their interfaces.[30] In accordance to this, multiple molecular Finally, the enantiomer version of LP3 peptide was explored docking simulations were performed taking into account several (P20 peptide), with an expectation of a lower inhibitory activity alternative conformations of both the receptor and the ligand, against LDL aggregation but on the other hand a theoretical en- namely, ApoB-100 protein and LP3 peptide, respectively. Then, hanced plasma stability given by the high proteolytic resistance the most favorable docking complex was submitted to molecular of d-amino acids containing peptides.[32] dynamic simulations in order to favor the mutual adjustment of the docking partners and explore active-like conformations of the peptide. The most LP3 active-like conformation was conforma- 2.2.2. Peptide Screening tionally restricted in a cyclic analogue that was named P7 peptide. An additional 13 LP3 derivatives were designed based on the SMase-TB: The inhibitory effect of these exploratory compounds careful evaluation of the final docking complexes, as described on SMase-induced LDL aggregation is shown in Figure 2A. Sur- below. The objective was to obtain a better understanding of the prisingly, P20 peptide, the enantiomer version of LP3, displayed mechanism of interaction between the original CR9 domain pep- similar activity compared to the original peptide, which is encour- tide and ApoB-100 protein. Although the protein–protein binding aging given its expected higher plasma stability compared to the interface involved many contacts distributed throughout the en- LP3 peptide (all l-amino acids). Additionally, three out of seven Adv. Therap. 2020, 3, 2000037 2000037 (3 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Figure 2. Inhibitory efficacy of the peptides from the first optimization round on A) SMase-TB, B) PLA2 -TB, and C) SMase-CE/FC assays. Data are presented as mean ± SD. N = 8 (SMase-TB), N = 4 (PLA2 -TB and SMase-CE/FC). P-values are calculated using the Mann-Whitney U test. *P < 0.01 versus negative control. Ala: alanine, der.: derivative. cyclic peptide analogues tested had a similar (P3) or higher (P4 dicating that P17 contains the essential motif to protect LDL from and P5) efficacy than LP3. Most of the smaller fragment com- SMase-induced LDL aggregation. This result opens the possibil- pounds showed lower inhibitory activity compared with the par- ity to generate a new library of compounds based on the struc- ent peptide. ture of P17. Most of the LP3 alanine scanning analogues (P9– We found that P17 maintained approximately the same activity P12) showed similar or slightly lower inhibitory activity than LP3, as LP3 despite of its smaller size (i.e., 9 instead of 14 residues), in- which suggest that none of the residues Asp7 , Asn8 , Glu11 , and Adv. Therap. 2020, 3, 2000037 2000037 (4 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Figure 3. Schematic representation of the designed peptide groups through the development of the second-round optimization. Ala: Alanine. Glu12 replaced by alanine have a key role in the binding of pep- these cells upon exposure to LDL derived exclusively from CE tide to LDL particles or that such mutations cause structural re- supplied by LDL, as hcVSMC unexposed to LDL did not contain arrangement of the peptide leading to an improvement in the intracellular CE. The TLC analysis shows the high efficacy of the binding, despite the loss of some specific intermolecular con- positive control and, in particular of DP3, to inhibit the intracel- tacts. However, the replacement of Glu6 residue by alanine led lular CE/FC ratio, an index of hcVSMC-foam cell formation. Sim- to a 30% reduction of the effects on LDL aggregation, as shown ilar to the result in the SMase-TB assay, the LP3 enantiomer ver- by P8 peptide. sion had the highest efficacy in the cell-based assay. PLA2 -TB: The inhibitory effect of the compounds on PLA2 - LP3 cyclic derivatives and LP3 alanine scanning derivatives induced LDL aggregation is shown in Figure 2B. LP3 fragments inhibited intracellular CE/FC to a similar extent to the SMase- were almost completely ineffective against PLA2 -induced LDL induced LDL aggregation, which is in line with the essential aggregation with the exception of P19. In contrast to P19, LDL condition of aggregation as cause of hcVSMC cholesterol which maintains the same inhibitory activity to LP3 despite a loading.[16,17,19–22] smaller size, P17 was ineffective against PLA2 -induced aggre- gation. Interestingly, P17 and P19 differ only in one aspartate residue located at the N-terminal that is present in P19 but 2.3. Computational Design and Screening of Peptides from the absent in P17. This suggests that this aspartate residue has a Second-Round Optimization key role in the inhibitory activity against PLA2 -induced LDL aggregation. Similar to SMase-induced LDL aggregation, LP3 2.3.1. Computational Peptide Design cyclic analogues and, in particular P3–P7, maintained a high inhibitory activity against PLA2 -induced LDL aggregation. While The computational strategy used to design peptides in the sec- all the LP3 alanine-scanning analogues showed a moderate ond round of optimization resulted into 26 peptides. This second to high inhibitory activity against SMase, only P10 and P12 set of compounds was synthesized by means of solid-phase pep- maintained high inhibitory activity against PLA2 modification. tide synthesis (summarized in Figure 3). The peptide sequences These results indicate that a higher number of residues involved and their characterization have been detailed in Tables S3 and S4 in intermolecular contacts are required to protect LDL against of the Supporting Information, respectively. Peptides were clas- PLA2 than against SMase. In this regard, Glu6 , Asp7 , and Glu11 sified into the following families according to their molecular residues appear to be essential for inhibitory activity against design: LP3 cyclic derivatives, LP3 alanine scanning derivatives, PLA2 but not against SMase-induced LDL aggregation. LP3 fragments, P17 linear d-derivatives, P17 cyclic derivatives, SMase-CE/FC: The effects of the compounds on hcVSMC- P17 linear l-derivatives, and P17 alanine-scanning derivatives. cholesterol loading were evaluated by analysis of the intracellu- P17 was one of the most promising peptides found in the first lar cholesteryl ester/free cholesterol ratio using thin layer chro- optimization round; therefore, it was chosen as the starting point matography (TLC). A representative TLC is showed in Figure for the second optimization round. P17 was optimized by ap- S3 of the Supporting Information. hcVSMC exposed to LDL plying both ligand-based and structure-based protocols similarly (nLDL or SMase-LDL) had similar free cholesterol (FC) levels to to the first optimization round. First, given the high inhibitory hcVSMC unexposed to LDL, indicating that LDL did not alter FC activity against LDL aggregation observed for P20 peptide (the content in these cells. Conversely, intracellular CEs detected in enantiomer version of LP3 peptide) within the first optimization Adv. Therap. 2020, 3, 2000037 2000037 (5 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com to the design of eight P21 analogues (namely, P26, P27, and P28) and five longer versions of P21 peptide (P29, P30, P31, P32, and P33). Additionally, multiple molecular docking calculations were also performed using several alternative conformations of both P17 and P20 peptides as ligand structures. MD simulations were run on the best-ranked docking poses to explore active-like con- formations and design proper analogues. Based on the analysis of the specific intermolecular interactions of P17 with ApoB-100 protein, one cysteine or one serine residue was added to the P17 peptide C-terminal end to obtain P34 and P35 peptides, respec- tively. In order to confirm the essential role of several specific residues found to be deeply buried upon binding, five P17 ala- nine scanning analogues were generated, thus obtaining P36, P37, P38, P39, and P40 peptides. However, four P21 analogues were generated by studying the favored orientation adopted by the peptide upon ApoB-100 pro- tein binding. Interestingly, virtually all the negatively charged side chains were oriented to the same face of the alpha helix and, thereby, directly interacted with several positively charged residues of ApoB-100. In this context, the optimization of elec- trostatic interactions between the peptide and ApoB-100 could reduce the free energy and favor the final binding. For this rea- son, P21 peptide residues exposed to the solvent upon bind- ing, namely, Ser5 and Asn9 , were mutated to alanine, positively charged residues or hydrophobic residues (as negative controls), leading to the design of six new peptides (namely, P41, P42, P43, P44, P45, and P46 peptides). 2.3.2. Peptide Screening SMase-TB: Both SMase- and PLA2 -induced assays clearly sug- Figure 4. In silico conformational sampling of A) P17 and B) P21 peptides. gested that the optimization of LP3 peptide by reducing its size Both the representative structures and the corresponding population dis- (P17) remained a promising approach. Nevertheless, additional tributions of the three most populated clusters are reported. modifications were needed in order to achieve similar inhibitory activity on both SMase- and PLA2 -induced LDL aggregation. In this regard the introduction of d-amino acids (second round) ap- round, the enantiomer version of peptide P17 was taken into peared to be a promising strategy. The effects of the compounds consideration for this second set of compounds (P21). Moreover, from the second optimization round on SMase-induced LDL ag- given the higher anti-LDL aggregation activity of P19 compared gregation are shown in Figure 5A. Like P20, P21 (the enantiomer to P17, longer versions of P21 were also obtained by adding one, version of P17) showed similar inhibitory activity compared to two, or three residues to both the N- and C-terminus P21 end, its parent peptide. Both the linear derivatives (P34 and P35) to give P22, P23, and P24. An exhaustive conformational sam- and the cyclic derivative (P25) had similar inhibitory effects pling of P17 and P21 peptides was performed by MD simula- to P17. tions and the most populated low energy conformations were ex- Unlike the results obtained on LP3 peptides, many of the P17 tracted. Similar to LP3, P17 peptide lacks well-defined secondary alanine scanning analogues, such as P36, P37, P38, P39, and structure elements; thus, the most populated conformation was P40, showed a significant decrease in the inhibitory activity com- constrained leading to one cyclic analogue, namely P25. Inter- pared with P17. The replacement of a single residue by alanine, estingly, unlike P17, P21 only assumed quasi-helical structures, led to a reduction of at least 30% in the peptide inhibitory efficacy suggesting that an alpha-helix conformation could favor the inter- against LDL aggregation. This clearly suggested that all these action with ApoB-100 due to minimal conformational rearrange- residues have a key role in the binding to LDL particles and that ments upon binding and the consequent reduced entropic cost the deletion of their side chain leads to the loss of some specific needed (Figure 4). On the basis of this hypothesis, the stabiliza- intermolecular contacts as well as unfavorable structural rear- tion of the helix could lead to a greater decrease in energy costs rangements of the peptide in solution or upon ApoB-100 protein and, therefore, favor the binding of the peptide to ApoB-100 pro- binding. tein. To favor low energy cost states, some of the P21 residues Interestingly, the attempt of enhancing P21 inhibitory effects were selectively mutated into analogues with similar physico- by stabilizing the expected alpha-helix conformation (P26, P27, chemical properties but a higher tendency to form alpha-helix and P28) proved to be successful since this increased the in- conformations (Ser to Ala and Asn to Gln).[33] This strategy led hibitory effects on LDL aggregation. We also observed that the Adv. Therap. 2020, 3, 2000037 2000037 (6 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Figure 5. Inhibitory efficacy of the peptides from the second optimization round on A) SMase-TB, B) PLA2 -TB, and C) SMase-CE/FC assays. Data are presented as mean ± SD. N = 8 (SMase-TB), N = 4 (PLA2 -TB and SMase-CE/FC). P-values are calculated using the Mann-Whitney U test. *P < 0.01 versus negative control. Ala: alanine, der.: derivative, scan: scanning, bind: binding, WT: wild. addition of one, two or three residues to the original sequence do PLA2-TB Assay: P17 alanine-scanning analogues (P36, not significantly change the overall inhibitory effects (as observed P37, P38, P39, and P40) were ineffective against PLA2 - in P22, P23, P24, P29, and P31) with the exception of P30, sug- induced LDL aggregation. However, some of the P17 d- gesting that P26 could have the optimal sequence to guarantee amino analogues (P22, P23, P24, P29, and P31) showed the highest inhibitory activity within LP3 analogues. Finally, the maximal inhibitory activity on PLA2 -induced aggregation introduction of positively charged or hydrophobic residues led (Figure 5B). to insignificant or negative effects, as shown by P41, P42, P43, Both SMase- and PLA2 -induced assays clearly pointed to P22 and P44 peptides. Collectively, the results from this biochemical and P31 peptides as the most promising LP3 derivatives com- screening shows that P21, P22, P23, P24, P26, P27, P28, P29, P31, bining significant inhibitory activity on both SMase- and PLA2 - and P41 peptides are the more promising compounds. induced LDL aggregation with a theoretically increased plasma Adv. Therap. 2020, 3, 2000037 2000037 (7 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Table 2. Pearson and Spearman’s correlations between SMase-TB and The strong correlation between the efficacy of the total com- PLA2 -TB for all peptides and peptides separated by control, first, and sec- pounds in the SMase-TB and SMase-CE/FC assays (Pearson r = ond optimization round compounds. SMase: Sphingomyelinase; PLA2 : phospholipase A2; TB: turbidimetry. 0.820, P = 0.000; Spearman’s r = 0.790, P = 0.000) is in line with the key role of LDL aggregation on hcVSMC-foam cell formation, Pearson Spearman’s previously described by our group.[16,17,19–22] Taken together, our results from molecular, biochemical and cell-based studies point Total compounds (N = 46) r = 0.686 r = 0.762 to P22 and P31 as the most promising peptides to be evaluated P < 0.001 P < 0.001 in vivo. Control r = 0.990 r = 1.000 P = 0.092 − First round (N = 20) r = 0.744 r = 0.750 P < 0.001 P < 0.001 3. Discussion Second round (N = 26) r = 0.660 r = 0.718 P < 0.001 P < 0.001 Currently, the prevention of atherosclerosis and other CVDs is mainly based on lipid-lowering agents (e.g., HMG-CoA reductase and PCSK9 inhibitors) that reduce blood cholesterol levels. Al- though this reduction of plasma cholesterol levels undoubtedly Table 3. Pearson and Spearman’s correlations between SMase-TB and affects the amount of cholesterol retained and accumulated in SMase-CE/FC for all peptides and peptides separated by control, first, and second optimization round compounds. SMase: Sphingomyelinase; TB: the vascular wall of the coronary vessels, it does not avoid the turbidimetry; CE: cholesteryl esters; FC: free cholesterol. risk and the mortality of CVD events. Indeed, the benefit of lipid- lowering drugs, such as statin-based therapies is often not re- Pearson Spearman’s lated to a sharp decrease in CVD mortality (acute myocardial in- farction and angina pectoris) in atherosclerosis patients.[34] Pre- Total compounds (N = 46) r = 0.820 r = 0.790 vious studies from different groups, including ours, have high- P < 0.001 P < 0.001 lighted the pathological relevance of blocking processes occur- Control r = 0.993 r = 1.000 ring locally in the arterial intima such as LDL aggregation and P = 0.074 − the uptake of aggregated LDL by smooth muscle cells.[20–22,25] We First round (N = 20) r = 0.924 r = 0.906 recently showed that LRP1-derived peptides, LP3, and, in par- P < 0.001 P < 0.001 ticular, its retroenantiomer version (DP3) efficiently preserved Second round (N = 26) r = 0.620 r = 0.629 LDL from SMase or PLA2 -induced LDL aggregation through the P = 0.001 P = 0.001 maintenance of ApoB-100 conformation.[26] We demonstrated that the protective effect of these peptides derives from their ca- pacity to establish electrostatic interactions with a specific highly stability if compared with LP3 peptide because of the use of positive sequence located in the ApoB-100 C-terminal region. d-amino acids. In this study, we developed LP3 peptide derivatives combining SMase-CE/FC: As shown in Figure 5C and Figure S3 (Sup- structural- and ligand-based computational designs with func- porting Information), most of the P17 (d-amino acids containing tional biochemical and cell-based screening systems to obtain op- analogues) were highly effective for the inhibition of hcVSMC timized peptides with similar activity to the original peptides. cholesterol loading. In particular, compounds P22 and P31, pre- viously selected as the more promising LP3 analogues because of their high efficiency of inhibition of aggregation induced by both SMase and PLA2 , also showed high inhibitory effects in the in- 3.1. Amino Acids with D-Chirality and a Tendency to Form tracellular cholesterol accumulation. P22 and P31 inhibited LDL Alpha-Helix Conformation Promote the Formation of aggregation and intracellular cholesterol loading to a similar ex- Peptide/ApoB-100 Stable Complexes that Are Required to tent than LP3 and substantially higher than P17 peptides. Maximize Peptide Functional Activity Peptides containing d-amino acids consistently showed far higher inhibitory activity with respect to their l-amino acid coun- 2.4. Analysis of the Correlation between the Efficacy of the New terparts, as with P17 and P21, LP3 peptide, and P20 and P19 and Compounds on SMase-TB, PLA2 -TB, and SMase-CE/FC Assays P22 compounds. This effect is hypothesized to be related to a higher metabolic stability derived from their stronger resistance The correlations between the efficacy of compounds in the to proteolytic degradation as well as to a higher binding affinity SMase-TB and PLA2 -TB assays and between the SMase-TB and to Apo-B100 protein or positive structural consequences of the SMase-CE/FC assays are shown in Figure 6A,B, respectively. binding. Given the distribution of the variables, we first analyzed Spear- Another structural feature that was observed to confer higher man’s correlations. However, Pearson’s correlation gave similar activity against LDL aggregation was related to the higher ten- qualitative and quantitative results for both the correlation be- dency to form alpha-helix conformations that could favor the tween the efficacy of compounds in the SMase-TB and PLA2 -TB binding of the peptide to ApoB-100 protein and increase its in- assays (Table 2) and between the efficacy of compounds in the hibitory activity against LDL aggregation, as was shown with P21, SMase-TB and SMase-CE/FC assays (Table 3). P26, P27, and P28. Adv. Therap. 2020, 3, 2000037 2000037 (8 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Figure 6. Graphs showing the scatterplot for the correlation A) between SMase-TB and PLA2 -TB and B) between SMase-TB and SMase-CE/FC. Blue squares label the compounds selected to perform studies in in vivo models. Adv. Therap. 2020, 3, 2000037 2000037 (9 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com 3.2. Additional Molecular Peptide Optimization Requirements mote agLDL formation.[9–12] In addition, targeting foam cell for- Are Needed to Reach Maximal Efficacy against Both SMase- and mation confers atheroprotection.[36,37] AgLDL also upregulates its PLA2 -Induced LDL Aggregation own receptor, LRP1, initiating a potent cycle that promotes foam cell formation and increases plaque cholesterol burden.[19,24,25] Despite having significant inhibitory activity on SMase-induced At a clinical level plaque burden determines the rapid progres- LDL aggregation, virtually all the 9-residue peptides, such as P17, sion of asyntomatic to syntomatic plaques, as reported in a its d-amino acid counterpart P21, and their analogues P26, P27, prospective observational study performed in 1.345 patients from and P28 were found to have a very limited or null activity upon 13 centers and seven countries.[38] We previously reported that LDL aggregation induced by PLA2 . By contrast, we found that aggregated LDL induces the production and secretion of tis- longer sequences with at least one additional residue in the C- sue factor, the main initiator of thrombosis, by human coro- terminal or the N-terminal end conferred higher inhibitory ac- nary VSMC.[21,22] Recent studies report that susceptibility of LDL tivity to the peptide in the PLA2 environment. This was notable to aggregate is associated with future coronary artery disease in P34 and P35, which were generated by adding a cysteine or a events.[39] We recently reported that circulating levels of soluble serine, respectively, to the C-terminal end of P17 sequence and LRP1 (sLRP1) predict future cardiac events at ten years in the co- in P19, which was generated by adding an aspartate to the N- hort REGICOR.[40] Together, these studies highlight the potential terminal end of P17 peptide. P19, in comparison to P17, showed relevance and clinical interest of these compounds in the treat- an inhibitory activity of more than 90% on both SMase- and PLA2 - ment of atherosclerosis and in the management of cardiovascular induced LDL aggregation. disease. Considering the limited structural data currently available on ApoB-100 protein and the dynamic arrangement upon LDL ag- 3.4. Essential Features of Optimized Peptides Selected to be gregation, it is difficult to provide a comprehensive explanation of Assayed in Preclinical Models of Atherosclerosis the specific structural requirements of the different peptides for the successful inhibition of LDL aggregation induced by SMase On the basis of both two computational strategies, namely, or by SMase and PLA2 . It may be reasonable to suggest that structure- and ligand-based drug design, 46 compounds were de- the binding of any of the active peptides to ApoB-100 protein signed according to the minimal energy status of the complex promotes a certain conformational change on ApoB-100 itself and maximal stability. This facilitated peptide cycling with the fi- and that it the extent of this conformational change that is re- nal aim of finding the minimal motif within LP3 peptide amino lated to the specific effect on LDL aggregation observed for each acid sequence able to inhibit both SMase- and PLA2 -induced LDL peptide. Additionally, shorter peptides, such as P17 or P21, may aggregation. A key structural feature that confers high activity be able to induce a limited conformational change sufficient to against LDL aggregation is the tendency to form alpha-helix con- block SMase, but not PLA2 affecting LDL particles. Conversely, formations that favor peptide/ApoB-100 protein complex forma- peptides composed of at least ten residues, (e.g., P19, P22 or tion and increase its inhibitory activity against LDL aggregation, P31) succeed in stimulating a more significant conformational as shown with P21, P26, P27, and P28. change on ApoB-100 leading to the inhibition of both SMase However, peptides composed of at least ten residues (e.g., P19, and PLA2 -induced LDL aggregation. It is also important to note P22, or P31) succeed in augmenting a more significant confor- that the different requirements for the peptide to be protective mational change on ApoB-100 leading to the inhibition of both against SMase and PLA2 -induced LDL aggregation are related to SMase- and PLA2 -induced LDL aggregation. the different locations of SM (the substrate of SMase) and PC Taking into account all these considerations, P22 and P31, (the substrate of PLA2 ) on the LDL surface.[35] SM is protected by which are P17-derived d-amino acid analogues are highly effi- these peptides against phospholipolysis as it is part of cholesterol- cient in inhibiting both SMase- and PLA2 -induced LDL aggre- enriched domains, whose structure, preserved through the for- gation. These enzymes are extremely active in the intimal ex- mation of peptide-ApoB-100 complexes, modulates SM lipolysis. tracellular matrix and these compounds are relatively small and By contrast, PC is located outside of these cholesterol-enriched expected to be highly stable against proteases. In this context, domains, and, therefore, is not protected by this mechanism. compounds P22 and P31 are highly promising to be tested in Thus, the preservation of LDL against PLA2 -induced LDL aggre- preclinical models of atherosclerosis. gation requires longer peptides that ensure ApoB-100 conforma- tion in conditions of extreme lipolysis. 3.5. Clinical Implications 3.3. Targeting LDL Aggregation as a Therapeutic Strategy to Although statin-based therapy is generally well tolerated and Inhibit hcVSMC-Cholesterol Loading highly effective in lowering blood cholesterol levels, it can be associated with various adverse events (e.g., intolerance, myal- This study demonstrated a strong correlation between the in- gia, myopathy, rhabdomyolysis, and diabetes mellitus)[41] and hibitory activity of peptides on SMase-induced LDL aggregation with an increased incidence of diabetes.[42] Indeed, as diabetic and hcVSMC-cholesterol loading. These results are in line with patients suffer a high incidence of atherosclerosis and cardio- the pivotal role of agLDL on foam cell formation from smooth vascular pathology, the development of innovative drugs for the muscle cells.[16–19] AgLDL is a key contributor of extracellular treatment of atherosclerosis in these patients, or those with cholesterol plaque burden since extracellular matrix proteogly- high susceptibility of diabetes development, is becoming of vi- cans and proteolytic enzymes of the arterial intima strongly pro- tal importance. In this context, inhibiting vascular cholesterol Adv. Therap. 2020, 3, 2000037 2000037 (10 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com accumulation by modulating not only LDL aggregation but also while a 2 fs simulation time step and a 12 Å cut off distance for long-range aggregated LDL internalization by vascular cells may be a promis- interaction were set. ing therapeutic strategy in the treatment of cardiovascular dis- Finally, the binding stability of each conformer was assessed by comput- ing its average root-mean-square deviation (LigRMSDavg ) along the last ease. 1.5 ns of the trajectory using ICM Browser Pro software (Molsoft, LLC, La Jolla, CA).[48] Conformers with LigRMSDavg values lower than 2 Å were pre- dicted as stable binders, close to the peptide active state and thus selected to be conformationally restrained, when possible, by mutating two specific residues to cysteine in order to obtain LP3 cycled derivatives. A similar pro- 4. Experimental Section tocol was applied to obtain derivatives of the P17 compound, one of the Molecular Modeling—In Silico Conformational Sampling: The in silico most promising LP3 derivatives obtained in the first optimization round, conformational sampling of LP3 peptide included a conjugate gradient within the second optimization round. As no feasible cycled derivatives minimization, an equilibration, and 100 ns long implicit solvent MD simu- were produced during the second round, linear derivatives were gener- lation using NAMD simulation package.[43] Thus, as the first preparation ated from the selected conformers based on the visual inspection of the step, an LP3 3D structure was created from scratch and parameterized specific intermolecular interactions of the docking partners. using AmberTool16 Leap module and ff12 AMBER force field.[44] Then, a Additionally, the most frequently buried residues upon binding, in 1000-cycle long minimization was performed applying harmonic restraints terms of accessible solvent area (ASA) within all the selected conformers to all nonhydrogen atoms with a force constant of 5 kcal mol−1 Å−2 in order were identified and used to design additional LP3 and P17 linear deriva- to remove initial intermolecular clashes. A 200 ps long equilibration was tives. The ASA value was calculated by ICM browser software.[48] run gradually heating the system to 56.85 °C and applying harmonic re- Peptide Synthesis: All the peptides were synthesized by Iproteos S.L straints with the same force constant as in the minimization but only to the (Barcelona, Spain) by means of solid-phase peptide synthesis following backbone atoms. During equilibration, the integration time step was set a 9-fluorenylmethoxycarbonyl/tert-butyl (Fmoc/tBu) strategy. Syntheses to 2 fs and the nonbonding cut off distance to 12 Å. Then, an unrestrained were performed manually on a 100 𝜇mol scale per peptide using Rink replica exchange molecular dynamics simulation in implicit water was run amide ChemMatrix resin. Peptide elongation and other solid-phase ma- consisting of 16 independent 100 ns long trajectories (replicas) performed nipulations were done manually in polypropylene syringes, each fitted with simultaneously at temperatures ranging from 26.85 to 339.85 °C. Finally, a polyethylene porous disk. Solvents and soluble reagents were removed 10 000 snapshots were extracted from the 300 K trajectory and were conse- by suction. Washings between synthetic steps were done with dimethylfor- quently clustered using the single linkage method by the cpptraj AMBER mamide and dichloromethane using 5 mL of solvents per gram resin each tool. A representative structure of each cluster was selected and confor- time. N𝛼 -Fmoc-protected amino acids (four equivalents) were coupled mationally constrained by mutating two specific residues to cysteine in using 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium (four equiva- order to obtain LP3 cyclic derivatives in the first optimization round, via in- lents), and N,N-diisopropylethylamine (eight equivalents). The extent of tramolecular disulphide bridge formation. The same strategy was used to the reaction was monitored using the Kaiser test. During couplings, the perform the conformational sampling of P17 and P21 compounds within mixture was allowed to react with intermittent manual stirring. The Fmoc the second optimization round. group was removed by treating the resin with 20% piperidine in DMF Molecular Modeling—I ApoB-100 Molecular Docking: The structural (3 mL g−1 resin). The peptides were cleaved from the resin using a mix- model of ApoB-100 in complex with LP3 peptide was built by rigid body ture of trifluoroacetic acid, triisopropylsilane, and water (95:2.5:2.5). The molecular docking simulation using the corresponding peptide con- crude products obtained were purified in reverse phase using a semiprep formers described above, as ligands, and ten different ab initio models HPLC instrument equipped with a C18 column. The purity and identity of of ApoB-100 LDL receptor binding domain, generated as previously the synthesized peptides were assessed by HPLC, HPLC-MS, and MALDI- described,[26] as receptors. TOF analysis. Starting from each ApoB-100 protein structure and LP3 peptide con- former previously obtained, different independent docking simulations In silico designed and synthesized compounds (N = 46) were evaluated in were performed using SMINA, an energy minimization optimized fork of a double screening system: The efficacy of the compounds to inhibit LDL the AutoDock Vina docking program.[45] All the ligands and the receptor aggregation was first monitored by turbidimetry. The turbidimetry assay structures were converted into input files suitable for SMINA using pre- has been optimized for screening by 1) using the enzymatic (SMase or pare_ligand4.py and prepare_receptor4.py scripts provided by AutoDock PLA2 ) induction of LDL aggregation and 2) using LDL from a unique nor- Tools.[46] A cuboid grid box of roughly 60 × 60 × 60 size with a grid space molipemic pool (Figure S1, Supporting Information). of 0.375 Å was adjusted around Apo-B100 3227 IKFDKYKAEK3236 region, as- Biochemical/Cell-Based Screening Assays for Peptide Efficacy—LDL Isola- sumed to be responsible for LP3 peptide recognition according to previous tion and Purification: Human LDL (d1.019 –d1.063 g mL−1 ) was obtained studies.[26] Exhaustiveness, number of modes and energy range were set from pooled normolipemic human plasma by sequential ultracentrifu- to 32, 100, and 50, respectively. gation in a KBr density gradient. Briefly, very-low density lipoproteins Once all the docking simulations were run, the lowest energy-docking (VLDLs) were first discarded after spinning plasma at 50 000 g for 18 h pose of each conformer in complex with ApoB-100 protein was selected at 4 °C using a fixed-angle rotor (50.2 Ti, Beckman) mounted on an Op- and submitted to a postdocking MD-based structural refinement. Thus, tima L-100 XP ultracentrifuge (Beckman). Subsequently, VLDL-free plasma each docking model underwent a conjugate gradient minimization, equili- was layered with 1.063 g mL−1 KBr solution and centrifuged at 50 000 g bration, and 3 ns long implicit solvent MD simulation, using NAMD simu- for 18 h at 4 °C. LDLs were dialyzed against 0.02 m Trizma, 0.15 m NaCl, lation package. As in the first preparation step, the overall protonation and 1 × 10−3 m EDTA, pH 7.5 for 18 h and then against normal saline for 2 h parameterization of the system was performed using the Leap module of at 4 °C. Finally, isolated LDLs were filter-sterilized (0.22 𝜇m Millex-GV fil- AmberTool16 and ff12 AMBER force field. A 1.000-cycle long minimization ter unit, Millipore). Protein concentration was determined using the BCA step was performed applying harmonic restraints to all backbone atoms protein assay (Thermo Scientific) and the cholesterol concentration was of the system (both receptor and ligand) with a force constant of 5 kcal determined with a commercial kit (IL test Cholesterol, Izasa). mol−1 Å−2 . Then, a 200 ps long equilibration step was performed gradu- Biochemical/Cell-Based Screening Assays for Peptide Efficacy—Exposure of ally heating the system to 300 K and applying harmonic restraints to the LDL to Sphingomyelinase (SMase) or Phospholipase A2 (PLA2 ) in the Pres- backbone atoms of the protein and the ligand with a force constant of 5 ence and Absence of Peptides under Strictly Controlled Conditions: Human and 2 kcal mol−1 Å−2 , respectively. Finally, a 3 ns long MD simulation was LDL (1.44 mg mL−1 ) were incubated with 40 U L−1 of Bacillus cereus SMase performed applying harmonic restraints only to the backbone atoms of the (Sigma-Aldrich, Schnelldorf, Germany) or with 50 𝜇g L−1 of type II secre- receptor protein with a force constant of 2 kcal mol−1 Å−2 . In all the sim- tory PLA2 from honey bee venom (Sigma-Aldrich, Schnelldorf, Germany) ulations, SHAKE[47] was applied to restrain all bonds to hydrogen atoms in 20 × 10−3 m Tris buffer (pH 7.0) containing 150 × 10−3 m NaCl, 2 × Adv. Therap. 2020, 3, 2000037 2000037 (11 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com 10−3 m CaCl2 , and 2 × 10−3 m MgCl2 at 37 °C. LDL incubation with SMase Loading: hcVSMC exposed to LDL (nLDL or SMase-LDL) showed similar (24 h) or PLA2 (36 h) was performed in absence or presence of peptides FC levels compared to hVSMC not exposed to LDL, indicating that LDL did at peptide concentrations of 10 × 10−6 m (peptide/ApoB-100 ratio: 5/1) not alter FC content in these cells. Conversely, intracellular CEs detected on the basis of the results obtained in dose- and time-course previous in these cells upon exposure to LDL derives exclusively from CE supplied experiments.[26] LDL lipolysis was stopped by addition of EDTA (final con- by LDL, as hcVSMC unexposed to LDL did not have intracellular CEs. The centration 10 × 10−3 m). inhibitory effect of the novel compounds on the intracellular cholesterol Biochemical/Cell-Based Screening Assays for Peptide Efficacy— accumulation was analyzed in terms of decrease in the ratio of intracellu- Quantification of the Inhibitory Efficacy of Each Compound against lar CE and FC content of hcVSMC exposed to LDL and SMase-LDL. LDL Aggregation Induced by SMase or PLA2 : The efficiency of peptides The efficacy of each peptide to decrease the ratio of intracellular CE and to inhibit LDL aggregation induced by SMase or PLA2 was estimated by FC content of hcVSMC exposed to LDL and SMase-LDL was calculated turbidimetry measuring the absorbance at a wavelength of 405 nm. The according to Equation (2) inhibitory activity was calculated according to Equation (1) Inhibition activity = [1 − (a − b∕c − b)] ∗ 100 (I) Efficacy = [1 − (a − b∕c − b)] ∗ 100 (2) wherein wherein a corresponds to the absorbance value in the presence of nLDL particles, LDL aggregating enzyme (i.e., SMase or PLA2 ) and test compound, b corresponds to the absorbance value in the presence of only nLDL par- a corresponds to the CE/FC ratio in hcVSMC exposed to SMase-LDL and ticles, and in the presence of the test compound, c corresponds to the absorbance value in the presence of nLDL particles b corresponds to the CE/FC ratio in hcVSMC exposed to nLDL in the ab- and LDL aggregating enzyme (i.e., SMase or PLA2 ). sence of the test compound, and c corresponds to the CE/FC ratio in hcVSMC exposed to SMase-LDL in the Biochemical/Cell-Based Screening Assays for Peptide Efficacy—Cell-Based absence of the test compound. Screening Assay: The efficacy of the peptides to inhibit LDL-induced intra- cellular cholesteryl ester accumulation was monitored by determination of intracellular CE/FC ratio. The cellular assay has been optimized for screen- Statistical Analysis: Data were described as the mean ± standard de- ing by 1) using the cells from a unique batch through all experiments and viation (SD). Comparisons among groups were performed using Mann- 2) testing LDL aggregation degree before adding LDL to the cell culture. Whitney U test. Correlation between continuous variables was analyzed Biochemical/Cell-Based Screening Assays for Peptide Efficacy—Quiescent using both Pearson’s linear correlation coefficient and the Spearman’s rho hcVSMCs: Cells of a single lot batch (61 646 600) were obtained from coefficient. Correlation was plotted using a scatterplot, identifying the val- ATCC (ATCC-PCS-100-021) to prevent variability derived from cell origin. ues of each round. The interassay CV was calculated as the average coef- Quiescence was induced by maintaining the cell culture for 24 h in a ficient of variation form plate control means and the intra-assay CV was medium with 0.2% foetal calf serum or for 48 h in a medium with 0.4% calculated as the average coefficient of variation between duplicates. The serum at 37 °C and 5% CO2 . Serum-deprived cells between passages 4 and Z-factor was calculated as previously described.[27] The two-tailed signif- 8 were used for experiments. Cells at these passages appeared as a rela- icance level was set at <0.05. The statistical software package IBM-SPSS tively homogeneous population with a hill-and-valley confluence pattern. (V25) was used for statistical analyses. Cell monolayers were grown in vascular cell basal medium (ATCC-PCS- 100-030) supplemented with vascular smooth muscle growth kit compo- nents (ATCC-PCS-100-042). Quiescent cells were exposed for 2 h to LDL previously treated with SMase for 18 h in the presence or absence of pep- Supporting Information tides at a concentration of 10 × 10−6 m. The degree of aggregation of the Supporting Information is available from the Wiley Online Library or from added LDL was assessed by turbidimetry measurements before the incu- the author. bation with the cells (Figure S4, Supporting Information). Cells were not exposed to PLA2 -treated LDL due to the high cytotoxicity of PLA2 . Cells were collected for lipid extraction followed by neutral intracellular lipid par- titioning through thin layer chromatography and lipid band analysis and Acknowledgements quantification. Biochemical/Cell-Based Screening Assays for Peptide Efficacy— A.B.-A. and C.P. contributed equally to this work. The authors thank Dr. Determination of Intracellular Cholesteryl Ester/Free Cholesterol Ratio: Ignasi Gich, Professor of Clinical Pharmacology and Therapeutics and Following the lipoprotein incubation period, hcVSMCs were washed Researcher of Sant Pau Biomedical Research Institute (IIB-SantPau) and exhaustively: twice with phosphate buffered saline (PBS), twice with PBS CIBER Epidemiología y Salud Pública (CIBERESP), for this help in sta- supplemented with 1% bovine serum albumin (BSA), and once with PBS tistical analysis and graph representations. The Ministry of Science and supplemented with both 1% BSA and 100 U mL−1 heparin); cells were Innovation of Spain, in the framework of the State Plan of Scientific then harvested into 1 mL of 0.15 m NaOH. Lipids were extracted using and Technical Innovation Investigation 2013–2016, awarded funding to the Bligh and Dyer method with minor modifications.[17,19] The lipid the project “DEVELOPMENT OF AN INNOVATIVE THERAPY FOR THE extract was dissolved in dichloromethane, applied to silica gel plates, TREATMENT OF THE ATHEROSCLEROSIS THROUGH INHIBITION OF and separated by thin layer chromatography. Cholesterol and cholesterol CHOLESTEROL VASCULAR ACCUMULATION” led by IPROTEOS SL with palmitate were run as standards of free and cholesteryl ester, respectively. File No. RTC-2016-5078-1. Support was also received from the Fundació A primary solvent combination of heptane/diethylether/acetic acid MARATÓ TV3 Project 201521-10 (to VLl-C), and FIS PI18/01584 (to VLl-C) (74:21:4, v/v/v) was used as chromatographic mobile phase followed by from the Instituto de Salud Carlos III (ISCIII) and cofinanced with ERDFs. heptane alone. After lipid separation, the plates were dried and stained Support was also received from Ministerio de Economía y Competitivi- as previously reported.[24] Finally, the spots corresponding to CE and dad to DdG-C (IJCI-2016-29393). CIBER Enfermedades Cardiovasculares FC were measured by densitometry against the standard curve using a (CIBERCV; CB16/1100403 (DdG-C, VLl-C) are projects run by the Instituto GS-800 Calibrated Densitometer (Bio-Rad). de Salud Carlos III (ISCIII). The authors also acknowledge the support Biochemical/Cell-Based Screening Assays for Peptide Efficacy—Calculation from “Secretaria d’Universitats i Recerca del Departament d’Economia I of the Inhibitory Efficacy of Each Compound against hcVSMC-Cholesterol Coneixement de la Generalitat de Catalunya” (2017SGR946 to VLl-C). Adv. Therap. 2020, 3, 2000037 2000037 (12 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 23663987, 2020, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adtp.202000037 by Readcube (Labtiva Inc.), Wiley Online Library on [13/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advtherap.com Conflict of Interest [24] P. Costales, P. Fuentes-Prior, J. Castellano, E. Revuelta-Lopez, M. Á. Corral-Rodríguez, L. Nasarre, L. Badimon, V. Llorente-Cortes, J. Biol. R.P., C.P., and T.T. are employees of Iproteos SL. T.T. is cofounder of Chem. 2015, 290, 14852. Iproteos SL and member of its board of directors. Some of the results [25] O. Bornachea, A. Benitez-Amaro, A. Vea, L. Nasarre, D. de Gonzalo- presented in this manuscript were previously described in the European Calvo, J. C. Escola-Gil, L. Cedo, A. Iborra, L. Martinez-Martinez, C. Patent Application EP19382335.8 Juarez, J. A. Camara, C. Espinet, M. Borrell-Pages, L. Badimon, J. Castell, V. Llorente-Cortés, Theranostics 2020, 10, 3263. [26] A. Benitez-Amaro, C. Pallara, L. Nasarre, A. Rivas-Urbina, S. Benitez, A. Vea, O. Bornachea, D. de Gonzalo-Calvo, G. Serra-Mir, S. Villegas, Keywords R. Prades, J. L. Sanchez-Quesada, C. Chiva, E. Sabido, T. Tarragó, aggregated LDL, ApoB-100, atherosclerosis, LRP1-derived peptides V. Llorente-Cortés, Biochim. Biophys. Acta – Biomembr. 2019, 1861, 1302. Received: February 26, 2020 [27] J.-H. Zhang, T. D. Y. Chung, K. R. Oldenburg, J. Biomol. Screen. 1999, Revised: April 9, 2020 4, 67. Published online: May 20, 2020 [28] F. B. Straub, G. Szabolcsi, Molecular Biology: Problems and Perspec- tives, Nauka, Moscow 1964, Ch. 182. [29] P. Závodszky, L. V Abaturov, Y. M. Varshavsky, Acta. Biochim. Biophys. Acad. Sci. Hung. 1966, 1, 389. [30] D. E. Koshland, Proc. Natl. Acad. Sci. USA 1958, 44, 98. [31] T. Clackson, J. A. Wells, Science 1995, 267, 383. [1] W. P. Castelli, K. Anderson, P. W. F. Wilson, D. Levy, Ann. Epidemiol. [32] R. Prades, B. Oller-Salvia, S. M. Schwarzmaier, J. Selva, M. Moros, M. 1992, 2, 23. Balbi, V. Grazú, J. M. De La Fuente, G. Egea, N. Plesnila, M. Teixidõ, [2] T. Gordon, W. P. Castelli, M. C. Hjortland, W. B. Kannel, T. R. Dawber, E. Giralt, Angew. Chem., Int. Ed. 2015, 54, 3967. Am. J. Med. 1977, 62, 707. [33] C. Nick Pace, J. Martin Scholtz, Biophys. J. 1998, 75, 422. [3] H.-J. Guretzki, K.-D. Gerbitz, B. Olgemöller, E. Schleicher, Atheroscle- [34] R. DuBroff, World. J. Cardiol. 2015, 7, 404. rosis 1994, 107, 15. [35] T. Hevonoja, M. O. Pentikäinen, M. T. Hyvönen, P. T. Kovanen, M. [4] M. Aikawa, P. Libby, Cardiovasc. Pathol. 2004, 13, 125. Ala-Korpela, Biochim. Biophys. Acta – Mol. Cell. Biol. Lipids 2000, 1488, [5] A. Mauriello, G. M. Sangiorgi, R. Virmani, S. Trimarchi, D. R. Holmes 189. Jr., F. D. Kolodgie, D. G. Piepgras, G. Piperno, D. Liotti, J. Narula, P. [36] S. H. Son, Y. H. Goo, M. Choi, P. K. Saha, K. Oka, L. C. Chan, A. Paul, Righini, A. Ippoliti, L. G. Spagnoli, Atherosclerosis 2010, 208, 572. Cardiovasc. Res. 2016, 109, 294. [6] R. Puri, S. E. Nissen, M. Shao, M. B. Elshazly, Y. Kataoka, S. R. Kapa- [37] A. Maier, H. Wu, N. Cordasic, P. Oefner, B. Dietel, C. Thiele, dia, E. M. Tuzcu, S. J. Nicholls, Arterioscler. Thromb. Vasc. Biol. 2016, A. Weidemann, K. U. Eckardt, C. Warnecke, FASEB. J. 2017, 31, 36, 2220. 4971. [7] K. Öörni, P. Posio, M. Ala-Korpela, M. Jauhiainen, P. T. Kovanen, Ar- [38] S. E. Lee, J. M. Sung, A. Rizvi, F. Y. Lin, A. Kumar, M. Hadamitzky, Y. terioscler. Thromb. Vasc. Biol. 2005, 25, 1678. J. Kim, E. Conte, D. Andreini, G. Pontone, M. J. Budoff, I. Gottlieb, B. [8] K. Oörni, M. O. Pentikäinen, M. Ala-Korpela, P. T. Kovanen, J. Lipid K. Lee, E. J. Chun, F. Cademartiri, E. Maffei, H. Marques, J. A. Leipsic, Res. 2000, 41, 1703. S. Shin, J. Hyun Choi, K. Chinnaiyan, G. Raff, R. Virmani, H. Samady, [9] S. Marathe, S. L. Schissel, M. J. Yellin, N. Beatini, R. Mintzer, K. J. P. H. Stone, D. S. Berman, J. Narula, L. J. Shaw, J. J. Bax, J. K. Min, H. Williams, I. Tabas, J. Biol. Chem. 1998, 273, 4081. J. Chang, Circ. Cardiovasc. Imaging 2018, 11, e007562. [10] S. L. Schissel, J. Tweedie-Hardman, J. H. Rapp, G. Graham, K. J. [39] M. Ruuth, S. D. Nguyen, T. Vihervaara, M. Hilvo, T. D. Laajala, P. K. Williams, I. Tabas, J. Clin. Invest. 1996, 98, 1455. Kondadi, A. Gisterå, H. Lähteenmäki, T. Kittilä, J. Huusko, M. Uusi- [11] M. Aviram, I. Maor, Biochem. Biophys. Res. Commun. 1992, 185, 465. tupa, U. Schwab, M. J. Savolainen, J. Sinisalo, M.-L. Lokki, M. S. [12] K. Öörni, J. K. Hakala, A. Annila, M. Ala-Korpela, P. T. Kovanen, J. Biol. Nieminen, A. Jula, M. Perola, S. Ylä-Herttula, L. Rudel, A. Öörni, M. Chem. 1998, 273, 29127. Baumann, A. Baruch, R. Laaksonen, D. F. J. Ketelhuth, T. Aittokallio, [13] M. W. Von Ballmoos, D. Dubler, M. Mirlacher, G. Cathomas, J. Muser, M. Jauhiainen, R. Käkelä, J. Borén, K. J. Williams, P. T. Kovanen, K. B. C. Biedermann, Arterioscler. Thromb. Vasc. Biol. 2006, 26, 359. Öörni, Eur. Heart J. 2018, 39, 2562. [14] J. C. Khoo, E. Miller, P. McLoughlin, D. Steinberg, Arterioscler. Thromb. [40] D. de Gonzalo-Calvo, R. Elosua, A. Vea, I. Subirana, S. Sayols- Vasc. Biol. 1988, 8, 348. Baixeras, J. Marrugat, V. Llorente-Cortés, Atherosclerosis 2019, 287, 93. [15] H. S. Kruth, Curr. Opin. Lipidol. 2002, 13, 483. [41] P. P. Toth, A. M. Patti, R. V. Giglio, D. Nikolic, G. Castellino, M. Rizzo, [16] P. Vijayagopal, D. L. Glancy, Arterioscler. Thromb. Vasc. Biol. 1996, 16, M. Banach, Am. J. Cardiovasc. Drugs 2018, 18, 157. 1112. [42] M. Barylski, D. Nikolic, M. Banach, P. Toth, G. Montalto, M. Rizzo, [17] T. Dissmore, C. I. Seye, D. M. Medeiros, G. A. Weisman, B. Bradford, Curr. Pharm. Des. 2014, 20, 3657. L. Mamedova, Atherosclerosis 2016, 252, 128. [43] J. C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, [18] V. Llorente-Cortés, M. Otero-Viñas, E. Hurt-Camejo, J. Martínez- C. Chipot, R. D. Skeel, L. Kalé, K. Schulten, J. Comput. Chem. 2005, González, L. Badimon, Arterioscler. Thromb. Vasc. Biol. 2002, 22, 387. 26, 1781. [19] V. Llorente-Cortés, M. Otero-Viñas, S. Sánchez, C. Rodríguez, L. Badi- [44] D. A. Case, T. E. Cheatham III, T. Darden, H. Gohlke, R. Luo, K. M. mon, Circulation 2002, 106, 3104. Merz Jr., A. Onufriev, C. Simmerling, B. Wang, R. J. Woods, J. Comput. [20] V. Llorente-Cortés, P. Costales, J. Bernués, S. Camino-Lopez, L. Badi- Chem. 2005, 26, 1668. mon, J. Mol. Biol. 2006, 359, 950. [45] O. Trott, A. J. Olson, J. Comput. Chem. 2010, 31, 455. [21] V. Llorente-Cortés, M. Otero-Viñas, S. Camino-López, O. Llampayas, [46] G. M. Morris, W. Huey, W. Lindstrom, M. F. Sanner, R. K. Belew, D. S. L. Badimon, Circulation 2004, 110, 452. Goodsell, A. J. Olson, A. J. Olson, J. Comput. Chem. 2009, 30, 2782. [22] S. Camino-López, L. Badimon, A. GonzáLez, D. Canals, E. Peña, V. [47] J.-P. Ryckaert, G. Ciccotti, H. J. C. Berendsen, J. Comput. Phys. 1977, Llorente-Cortés, J. Thromb. Haemost. 2009, 7, 2137. 23, 327. [23] S. Allahverdian, A. C. Chehroudi, B. M. McManus, T. Abraham, G. A. [48] R. Abagyan, W. H. Lee, E. Raush, L. Budagyan, M. Totrov, M. Sund- Francis, Circulation 2014, 129, 1551. strom, B. D. Marsden, Trends. Biochem. Sci. 2006, 31, 76. Adv. Therap. 2020, 3, 2000037 2000037 (13 of 13) © 2020 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim