Design, Pharmacological Characterization, and Molecular Docking of Minimalist Peptidomimetic Antagonists of α4β1 Integrin

Integrin receptors mediate cell–cell interactions via the recognition of cell-adhesion glycoproteins, as well as via the interactions of cells with proteins of the extracellular matrix, and upon activation they transduce signals bi-directionally across the cell membrane. In the case of injury, infection, or inflammation, integrins of β2 and α4 families participate in the recruitment of leukocytes, a multi-step process initiated by the capturing of rolling leukocytes and terminated by their extravasation. In particular, α4β1 integrin is deeply involved in leukocyte firm adhesion preceding extravasation. Besides its well-known role in inflammatory diseases, α4β1 integrin is also involved in cancer, being expressed in various tumors and showing an important role in cancer formation and spreading. Hence, targeting this integrin represents an opportunity for the treatment of inflammatory disorders, some autoimmune diseases, and cancer. In this context, taking inspiration from the recognition motives of α4β1 integrin with its natural ligands FN and VCAM-1, we designed minimalist α/β hybrid peptide ligands, with our approach being associated with a retro strategy. These modifications are expected to improve the compounds’ stability and bioavailability. As it turned out, some of the ligands were found to be antagonists, being able to inhibit the adhesion of integrin-expressing cells to plates coated with the natural ligands without inducing any conformational switch and any activation of intracellular signaling pathways. An original model structure of the receptor was generated using protein–protein docking to evaluate the bioactive conformations of the antagonists via molecular docking. Since the experimental structure of α4β1 integrin is still unknown, the simulations might also shed light on the interactions between the receptor and its native protein ligands.


Introduction
The outer surface of the cell membrane presents receptors that mediate interactions between cells or with components of the extra cellular matrix (ECM). These interactions are fundamental in determining the scope and activities of the cell, providing physical, biochemical, and mechanical signals. These surface receptors are also connected to cytoplasmic proteins via their internal end, and can thus transmit messages bi-directionally between cells and their environment.
This peptide owes its tremendous affinity to the α4-targeting diphenylurea pharmacophore. However, BIO1211 is scarcely stable in a biological environment, as tested in plasma; heparinized blood; and in homogenates of rat liver, lungs, and intestines [15,16], and is subjected to rapid clearance in vivo [17].
On the other hand, the stability of peptidic sequences can be improved by adopting a peptidomimetic strategy [18,19]. Several orthosteric peptidomimetic α4β1 integrin ligands have garnered much interest as antagonists for the treatment of asthma, [13] allergic conjunctivitis [20], or age-related macular degeneration (AMD) [21,22]. In addition, selective α4 integrin peptidomimetics were utilized to prepare monolayers of biofunctionalized nanoparticles and furnished cell-adhesive surfaces capable of detecting and quantifying leucocytes expressing active integrins, potentially useful for monitoring the severity and progression of correlated diseases [23][24][25].
From this perspective, in recent years we have explored the use of β-amino acids for improving the enzymatic stability of peptide integrin ligands [26]. Interestingly, while linear α/β-hybrid sequences were generally found to be antagonists [27][28][29], cyclic LDV pentapeptides including a β-residue proved themselves to be potent agonists, being capable of activating intracellular signalling and promoting integrin-mediated cell adhesion [30]. Among the linear structures containing a β-Pro scaffold, we found the potent α4 integrin antagonist DS-70 ( Figure 1, Table 1) [20]. In contrast to BIO1211, the hybrid and minimalist α/β-peptidic structure conferred DS-70 noteworthy stability in mouse serum and significant in vivo efficacy [20,22].
From this perspective, in recent years we have explored the use of β-amino acids for improving the enzymatic stability of peptide integrin ligands [26]. Interestingly, while linear α/β-hybrid sequences were generally found to be antagonists [27][28][29], cyclic LDV pentapeptides including a β-residue proved themselves to be potent agonists, being capable of activating intracellular signalling and promoting integrin-mediated cell adhesion [30]. Among the linear structures containing a β-Pro scaffold, we found the potent α 4 integrin antagonist DS-70 ( Figure 1, Table 1) [20]. In contrast to BIO1211, the hybrid and minimalist α/β-peptidic structure conferred DS-70 noteworthy stability in mouse serum and significant in vivo efficacy [20,22]. Table 1. Effect of straight hybrid peptides 1-8 [25] and the reference compounds BIO1211 and DS-70 [20] on α 4 β 1 integrin-mediated Jurkat E6.1 cell adhesion to the endogenous ligand FN (10 µg/mL). Data are presented as IC 50  Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design. Table 1. Effect of straight hybrid peptides 1-8 [25] and the reference compounds BIO1211 and DS-70 [20] on α4β1 integrin-mediated Jurkat E6.1 cell adhesion to the endogenous ligand FN (10 µg/mL). Data are presented as IC50 (nM) a .
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 (Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1, 5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design. Table 1. Effect of straight hybrid peptides 1-8 [25] and the reference compounds BIO1211 and DS-70 [20] on α4β1 integrin-mediated Jurkat E6.1 cell adhesion to the endogenous ligand FN (10 µg/mL). Data are presented as IC50 (nM) a .
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 (Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1, 5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 (Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 (Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 (Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 ( Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 ( Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 ( Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α4β1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 ( Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Materials and Methods and Supporting Information). Herein, we pursue the potential therapeutic use of the α/β-hybrid peptidomimetic α 4 β 1 integrin ligands containing diverse simple or functionalized β 2 , or β 3 -amino acid cores (β-alanine, diaminopropionic acid [31], iso-aspartic acid), also in combination with the retro-sequence strategy [32]. The ability of the ligands to interfere with integrin functions was determined via cell adhesion assays, competitive binding assays, and by analyzing intracellular signaling.
Additionally, we discuss the plausible ligand-receptor interactions using molecular modelling studies. Since the three-dimensional structure of α 4 β 1 integrin is not yet available, the study of a ligand's structural determinants for agonism/antagonism at this integrin is of great interest to the development of a reliable pharmacophoric model for future drug design.
β-amino acids carrying the side chain adjacent to the carboxylic acid group are termed β 2 , or β 3 when the side chain is located adjacent to the β-amino group ( Figure 1) [33,34]. The β-residue consents to maintain the same span of 14 bonds from the urea carbonyl to the carboxylate of Gly as the distance between the urea carbonyl and the carboxylate of Asp in BIO1211 (Figure 1). The presence of a carboxylate group is generally regarded as a mandatory requisite for a large majority of integrin ligands [1,5,23] (see also Section 2.6).
The peptide mimetics were prepared in solution or in solid phase under standard conditions, as reported in Materials and Methods and Supporting Information. The sequences 1, 2, 5, 6, and 9-16 were prepared in solution, while the remaining sequences were prepared with solid-phase extraction. Although the syntheses of sequences 1-8 have been already described [25], the fundamental synthetic steps are outlined for thoroughness. The MPUPA moiety was prepared (Supplementary Materials) via condensation of o-tolylisocyanate and 2-(4-aminophenyl)acetic acid. AMPUMP was prepared via a similar protocol and from Boc-4-(aminomethyl)aniline (Supplementary Materials, Scheme S1). As for the β-residues, we prepared simple or functionalized β 2 or β 3 -amino acids in-house (see Section 4 and Supporting Information).
In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H 2 O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α 4 β 1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α 4 β 1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α 4 β 1 -mediated cell adhesion, behaving as antagonists, with IC 50 values in the submicromolar range (Tables 1 and 2). Table 2. Effect of partially retro hybrid sequences 9-16 and DS-23 [20] on α 4 β 1 integrin-mediated Jurkat cell adhesion to the endogenous ligand FN (10 µg/mL). Data are presented as IC 50 (nM) a .

DS-23
In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3). In general, after synthesis the crude peptides were purified using semipreparative RP HPLC over a C18 column (General Methods) using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was confirmed to be >95% via RP HPLC; structure identity was assessed via ESI-MS and NMR spectroscopy.

Effects of Peptidomimetics on Integrin-Mediated Cell Adhesion
To investigate the ability of the new peptides to modulate integrin-mediated cell adhesion, we employed cell adhesion assays using a Jurkat E6.1 cell line, which endogenously expresses α4β1 integrin and the natural ligand FN. Cell adhesion experiments allowed us to identify compounds defined as antagonists, i.e., those capable of reducing the number of adherent cells bound to an integrin endogenous ligand. On the contrary, ligands able to increase cell adhesion are defined as agonists.
DS-70 and BIO1211 were considered as reference antagonists [20]. Regarding the new peptides, the potencies of the retro peptides 9-16 (Table 2) were found to be comparatively inferior to those of the ligands with straight sequences 1-8 (Table 1) [25]. Concentrationresponse curves are shown in Supporting Information ( Figure S1). Straight sequence 8 was shown to be an antagonist of α4β1 integrin with an excellent potency, comparable to those of the reference compounds DS-70 and BIO1211 (Table 1). In addition, peptidomimetic 3 and retro sequence 13 were able to reduce α4β1-mediated cell adhesion, behaving as antagonists, with IC50 values in the submicromolar range (Tables 1 and 2).
The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α4β1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin αMβ2, and on K562 cells, mainly expressing α5β1, which shares a β1 subunit with the heterodimer α4β1. Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by αMβ2 integrin, behaving as an agonist. Regarding α5β1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α5β1 integrin. The reference compound DS-70 was ineffective toward both αMβ2 and α5β1 integrins under the same experimental conditions (Table 3).  The peptides that showed the most relevant activity (Tables 1 and 2) were selected for determining their selectivity to different integrins (Table 3; concentration-response curves are shown in Supporting Information, Figures S2-S4). To this purpose, the ability of the best peptidomimetics to modulate cell adhesion mediated by integrins other than α 4 β 1 was evaluated in HL60 cells, mainly expressing the leukocyte integrin α M β 2 , and on K562 cells, mainly expressing α 5 β 1 , which shares a β 1 subunit with the heterodimer α 4 β 1 . Interestingly, peptide 3 strongly reduced the adhesion of HL60 cells to the natural ligand Fg, showing excellent potency. In contrast, 13 was a moderate promoter of cell adhesion mediated by α M β 2 integrin, behaving as an agonist. Regarding α 5 β 1 integrin, none of the peptidomimetics employed were able to modulate K562 cell adhesion, thus being shown to be completely ineffective toward α 5 β 1 integrin. The reference compound DS-70 was ineffective toward both α M β 2 and α 5 β 1 integrins under the same experimental conditions (Table 3). Table 3. Selectivity of the hybrid peptides and the reference compound DS-70 [20] toward α M β 2 and α 5 β 1 integrin, evaluated using cell adhesion assays. Data are presented as IC 50 (nM) a for antagonists b and as EC 50 for agonists c .

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Evaluation of Binding Affinity to α 4 β 1 Integrin
To better characterize α 4 β 1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α 4 β 1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC 50 for ligand binding [30], being K i 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Effects of Peptidomimetics on α4β1 Integrin Conformation
Integrin functions are regulated by a number of conformational changes in the protein itself. Three major conformations have been described: an inactive or bent

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Evaluation of Binding Affinity to α4β1 Integrin
To better characterize α4β1 integrin-peptidomimetic interaction, competitive binding experiments on intact Jurkat E6.1 cells were performed; thus, measuring the ability of increasing concentrations of the new compounds to displace the labelled ligand LDV-FITC from α4β1 integrin ( Table 4). The measured binding affinity of the reference compounds BIO1211 and DS-70 was comparable to the previously published IC50 for ligand binding [30], being Ki 6.9 ± 3.1 nM and 0.94 ± 0.32 nM, respectively.
Regarding the straight hybrid peptides, only compounds 3 (Ki = 1.3 ± 0.7 nM) and 8 (Ki = 1.9 ± 0.8 nM) were able to bind to α4β1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (Ki = 16.7 ± 5.2 nM).

Effects of Peptidomimetics on α4β1 Integrin Conformation
Integrin functions are regulated by a number of conformational changes in the protein itself. Three major conformations have been described: an inactive or bent 16.7 ± 5.2 a Values represent the mean ± SD of three independent experiments carried out in quadruplicate.
Regarding the straight hybrid peptides, only compounds 3 (K i = 1.3 ± 0.7 nM) and 8 (K i = 1.9 ± 0.8 nM) were able to bind to α 4 β 1 integrin, showing excellent affinity. The competitive binding experiments performed on partially retro hybrid peptides showed that only peptide 13 binds with very good potency to the receptor (K i = 16.7 ± 5.2 nM).

Effects of Peptidomimetics on α 4 β 1 Integrin Conformation
Integrin functions are regulated by a number of conformational changes in the protein itself. Three major conformations have been described: an inactive or bent conformation, an intermediate-activity conformation, and an open high-activity conformation [35]. Conformational changes can be evaluated by using conformationally sensitive antibodies which specifically recognize epitopes exposed only in a defined structural conformation. To better depict peptidomimetics-α 4 β 1 integrin interactions, we employed PE-conjugated HUTS-21 mAb that recognizes an epitope mapped to the hybrid domain of the β 1 subunit [36]. This epitope is a LIBS (ligand-induced binding site epitope) which is hidden in the inactive bent conformation but is exposed in the intermediate-activity and open high-activity conformation, upon agonist binding or partial integrin activation.
To evaluate the effects of the most active peptidomimetics (3, 8, and 13), Jurkat E6.1 cells were pre-incubated with different concentrations (1-100 nM) of the ligands, then PE-HUTS-21 mAb was added and fluorescence was measured using flow cytometry. As expected, the α 4 β 1 endogenous ligand FN induced a conformational change in α 4 β 1 integrin, leading to the exposure of HUTS-21 epitope and thus to the increased binding of the HUTS-21 mAb, with respect to vehicle-treated cells that have α 4 β 1 integrin mainly in the inactive conformation ( Figure 2). Conversely, antagonists 3, 8, and 13, although binding to α 4 β 1 integrin as previously shown, were not able to induce any conformational rearrangement, showing very low binding of HUTS-21 mAb.  [35]. Conformational changes can be evaluated by using conformationally sensitive antibodies which specifically recognize epitopes exposed only in a defined structural conformation. To better depict peptidomimetics-α4β1 integrin interactions, we employed PE-conjugated HUTS-21 mAb that recognizes an epitope mapped to the hybrid domain of the β1 subunit [36]. This epitope is a LIBS (ligand-induced binding site epitope) which is hidden in the inactive bent conformation but is exposed in the intermediateactivity and open high-activity conformation, upon agonist binding or partial integrin activation.
To evaluate the effects of the most active peptidomimetics (3, 8, and 13), Jurkat E6.1 cells were pre-incubated with different concentrations (1-100 nM) of the ligands, then PE-HUTS-21 mAb was added and fluorescence was measured using flow cytometry. As expected, the α4β1 endogenous ligand FN induced a conformational change in α4β1 integrin, leading to the exposure of HUTS-21 epitope and thus to the increased binding of the HUTS-21 mAb, with respect to vehicle-treated cells that have α4β1 integrin mainly in the inactive conformation ( Figure 2). Conversely, antagonists 3, 8, and 13, although binding to α4β1 integrin as previously shown, were not able to induce any conformational rearrangement, showing very low binding of HUTS-21 mAb.

Effects of Peptidomimetics on Integrin-Mediated Intracellular Signaling
Integrins relay signals across cell membrane, functioning as integrators of the ECM or other cells-driven cues. Since integrin cytoplasmic tails are short and lack kinase activity, integrin transduces signals within the cell, facilitating the assembly of cytoplasmic adaptor and/or signaling proteins in large complexes named focal adhesions (FAs). FAs are intracellular signaling platforms that enable cells to respond to changing extracellular signals by modulating downstream long-term events, including cell proliferation, differentiation, and migration [37]. Among the different signaling pathways activated by integrins, the MAPK and AKT pathways play an important role.
To further characterize the most promising peptidomimetics, 3, 8, and 13, their effects on integrin-mediated AKT and MAPK signaling pathways were evaluated using a Western blot (Figure 3).  3, 8, and 13 were not able to increase the binding of HUTS-21 mAb. Results are expressed as mean fluorescence intensity (MFI) ± SD of three independent experiments carried out in duplicate. MFI values for respective isotype control mAb were set to 0. **** p < 0.0001 versus vehicle (Tukey's test after ANOVA).

Effects of Peptidomimetics on Integrin-Mediated Intracellular Signaling
Integrins relay signals across cell membrane, functioning as integrators of the ECM or other cells-driven cues. Since integrin cytoplasmic tails are short and lack kinase activity, integrin transduces signals within the cell, facilitating the assembly of cytoplasmic adaptor and/or signaling proteins in large complexes named focal adhesions (FAs). FAs are intracellular signaling platforms that enable cells to respond to changing extracellular signals by modulating downstream long-term events, including cell proliferation, differentiation, and migration [37]. Among the different signaling pathways activated by integrins, the MAPK and AKT pathways play an important role.
To further characterize the most promising peptidomimetics, 3, 8, and 13, their effects on integrin-mediated AKT and MAPK signaling pathways were evaluated using a Western blot (Figure 3).
To this purpose, Jurkat E6.1 cells were pre-treated with different concentrations (10 −7 , 10 −8 , 10 −9 M) of the compounds and then exposed to FN (10 µg/mL) for 30 min. As shown in Figure 3, the FN was able to induce integrin-mediated intracellular signaling activation through the phosphorylation of ERK1/2, AKT, and JNK. When Jurkat E6.1 cells were preexposed to 3, 8, or 13, all three α 4 β 1 ligands significantly prevented FN-induced ERK1/2 and AKT phosphorylation. In addition, the effects observed on intracellular signaling were concentration-dependent, except for compound 13 on AKT activation.
Conversely, regarding JNK activation, only peptidomimetic 8 was able to significantly prevent FN-induced JNK phosphorylation, whereas 3 and 13 did not counteract FN-activating effects on JNK intracellular signaling pathways.
Overall, these data confirm that peptidomimetics 3, 8, and 13 behave as α 4 β 1 integrin antagonists, being able to bind to the receptor with excellent affinity but without inducing any conformational switch. Therefore, through their direct interaction with the receptor, they are able to inhibit cell adhesion to integrin endogenous ligands and to prevent FN-induced activation of intracellular signaling pathways, thus blocking α 4 β 1 integrin functions that are involved in the pathogenesis of several diseases, including different types of cancer and inflammatory diseases. Nevertheless, further studies are needed to better unravel the detailed mechanisms by which peptidomimetic antagonists block α 4 β 1 integrin functions.

Molecular Docking
The 3D structure of α 4 β 1 integrin has not been experimentally disclosed yet. However, both the α 4 and β 1 subunits have been determined in several separate heterodimeric complexes throughout X-ray crystallography and cryo-EM. Consequently, an α 4 β 1 integrin structure model can be obtained by employing molecular modelling techniques. In this work, the model structure of human α 4 β 1 integrin was obtained through protein-protein docking, using α 4 and β 1 monomers taken from experimental dimeric structures of integrins α 4 β 7 (PDB ID 3V4V) and α 5 β 1 (PDB ID 4WK0), respectively. The docking calculation was guided using the conserved residues found at the interfaces in the experimental integrin dimeric structures ( Figure S5). The resulting model was in good agreement with other integrin structures and shows a large interaction surface between the two subunits (1733 Å 2 ). The metal ions binding sites were optimized using a known modelling procedure [38,39] in order to be as close as possible to equivalent sites present in the experimental structures.
The α 4 subunit appears remarkably different from the other α subunits of RGDbinding integrins, since it completely lacks in the cavity deputed to host the ligand's arginine. In our homology model, the supposed binding site is delimited by the βI domain of the β subunit and the β-propeller in the α subunit ( Figure 4A). In the β 1 subunit, the βI domain contains a Mg 2+ ion in the metal ion dependent adhesion site (MIDAS), fundamental to the interaction with negatively charged ligands ( Figures S6 and S7, and Table S1).
Most of the known integrin ligands share a common carboxylate group as the fundamental pharmacophore. In addition, two other metal centers are present ( Figures 4A and S6) that contain one Ca 2+ metal ion each, the "adjacent to metal ion-dependent adhesion site" (ADMIDAS) and the "synergistic metal ion-binding site" (SyMBS) (for ADMIDAS: Figure S8 and Table S2; for SyMBS, Figure S9 and Table S3). In α 5 β 1 integrin, these metal centers play a regulatory role in ligand binding. Additionally, the inspection of the solventaccessible surface of the model revealed that the putative binding site is formed by subpockets A-E, characterized by diverse dimensions, shapes, and delimiting residues ( Figure 4B). . The representative Western blot shows that Jurkat E6.1 cells plated on FN had a stronger signal for pERK1/2, pAKT, and pJNK than vehicle-treated cells (vehicle). The graphs represent densitometric analysis of the bands (mean ± SD; three independent experiments). The amount of phosphorylated kinases (pERK1/2 or pAKT) was normalized to that of the corresponding total kinase (totERK1/2 or totAKT); phospho-JNK (pJNK) was normalized to actin. * p < 0.05, ** p < 0.01 vs. vehicle; # p < 0.05, ## p < 0.01, ### p < 0.001, #### p < 0.0001 vs. FN (Tukey's test after ANOVA).
In the so-obtained model of the α4β1 integrin, the coordination sphere of Mg 2+ ion at the MIDAS includes four water molecules (W) and the residues Ser132(β) and Glu229(β). As reported [1], an orthosteric, negatively charged ligand is expected to form a salt bridge with Mg 2+ ion at the MIDAS by substituting one of the four water molecules ( Figure S10). In the X-ray crystallographic structures of integrin α5β1, which are used as templates for the β subunit (PDB ID: 4WK0), as well as in the analogous structures (PDB IDs: 4WK2 and 4WK4), the ligands always occupy the position labelled as W1 ( Figure S10). The first series of docking experiments was carried out to understand whether that position is the only one of the first metal coordination spheres accessible to the ligand. Different docking calculations were performed with the same parameters, each time eliminating a different water molecule from the first coordination sphere of Mg 2+ to make it accessible to the ligand. In particular, the MIDAS water molecules at positions W1, W2, and W3 were alternatively removed, while the water molecule at position W4 was maintained because it is not accessible to the bulk of the solvent (Supporting Information). The simulation was repeated using different ligands, a process already reported in the literature. In general, the removal of W1 allowed the ligands to coordinate Mg 2+ ion, while in the other cases the ligands could not get close enough to the metal center. Consequently, the docking calculations were conducted by eliminating water and leaving position W1 accessible to potential ligands ( Figure 4A). In the so-obtained model of the α 4 β 1 integrin, the coordination sphere of Mg 2+ ion at the MIDAS includes four water molecules (W) and the residues Ser132(β) and Glu229(β). As reported [1], an orthosteric, negatively charged ligand is expected to form a salt bridge with Mg 2+ ion at the MIDAS by substituting one of the four water molecules ( Figure S10). In the X-ray crystallographic structures of integrin α 5 β 1 , which are used as templates for the β subunit (PDB ID: 4WK0), as well as in the analogous structures (PDB IDs: 4WK2 and 4WK4), the ligands always occupy the position labelled as W1 ( Figure S10).
The first series of docking experiments was carried out to understand whether that position is the only one of the first metal coordination spheres accessible to the ligand. Different docking calculations were performed with the same parameters, each time eliminating a different water molecule from the first coordination sphere of Mg 2+ to make it accessible to the ligand. In particular, the MIDAS water molecules at positions W1, W2, and W3 were alternatively removed, while the water molecule at position W4 was maintained because it is not accessible to the bulk of the solvent (Supporting Information). The simulation was repeated using different ligands, a process already reported in the literature. In general, the removal of W1 allowed the ligands to coordinate Mg 2+ ion, while in the other cases the ligands could not get close enough to the metal center. Consequently, the docking calculations were conducted by eliminating water and leaving position W1 accessible to potential ligands ( Figure 4A).
For the computations, we selected the hybrid peptides exhibiting at least a measurable efficacy, i.e., DS-23, DS-70, 8, 3, 13, and the reference antagonist BIO1211, as well as structurally correlated compounds with modest-to-null activity (7, 5, 1, 2, 4, 14, 10), for comparison. The aim of the computations was to highlight the specific structural determinants of ligand binding.
Each ligand was minimized and parametrized (see Section 4, and Figure S13); special attention was paid to modelling and optimizing the structures of the diphenylurea and β-Pro fragments (Figures S11 and S12). Then, the ligands were docked into the α 4 β 1 integrin model achieved in the previous modelling step.
The best docking poses of each compound exhibiting good potency during in vitro studies were used to define the main interactions between these antagonist ligands and α 4 β 1 integrin, and thus the main amino acid residues involved. The complexes between each one of these poses and α 4 β 1 were analyzed with the web server PacDOCK [40] and the program LigPlot+ [41]. The main integrin residues involved in the binding of these compounds are reported in Figure S14. In all the best docked poses, every compound interacts with both residues from the β and α subunit. The amino acid residues more often involved in the interaction are Ser132, Gly223, Asn224, Leu225, Asp226, Ser227, and Glu229 of the β subunit, and Tyr187 and Phe214 of the α subunit.
To improve the accuracy in the definition of the docked poses and to identify the best pose for each good ligand, a re-scoring of all the most representative docked poses was performed using the AMBER score implemented in the program UCSF DOCK 6 and based on the MM-GBSA method (see Section 4 for more details on the procedure). The best docked poses according to the AMBER score for all the compounds with significant potency, BIO1211, DS-70, DS-23, 8, 3, and 13, are reported in Figure 5.
In the best pose predicted for BIO211, MPUPA-Leu-Asp-Val-ProOH, the peptidic sequence lies vertically along the α4β1 crevice ( Figure 5), using subpockets B-E as defined in Figure 4B. Ligand's Asp is well inserted into subpocket C, so that the main interaction is the salt bridge between Asp carboxylate and Mg 2+ . AspCOO − also interacts with other residues of the MIDAS region, Ser132 and Tyr133.
As for the rest of the ligand's structure, the peptidic portion attains interactions only with the β 1 subunit; in particular, the side chains of Leu and Val occupy subpockets D and B, respectively. In subpocket B, Pro carboxylate makes a salt bridge with Lys182(β1) and also interacts with Thr188(β1). The diphenylurea is longitudinally inserted within subpocket D, showing stabilizing contacts with residues of the α subunit, i.e., Gln244, Lys213, and Phe214, as well as residues from the β subunit, Ser227, Pro228.
Very recently, the docking of BIO1211 into a homology model of the α 4 β 1 integrin was simulated by da Silva et al. [42]. The proposed bioactive conformation has something in common with the structure shown in Figure 5, the main difference being the orientation of the C-terminal Val-Pro dipeptide, which in da Silva's geometry is rotated towards the A subpocket, so that Pro cannot reach Lys182(β 1 ).
Concerning the much shorter sequences of the hybrid peptides, for the compounds with a significant activity the docked poses reproduced the position of the diphenylurea-Leu-Asp portion of BIO1211, occupying subpockets C, D, and E. In particular, for each compound the negatively charged carboxylate entered into C and formed a salt bridge with Mg 2+ at the MIDAS, while the diphenylurea moiety found a place in lower subpocket E of the α/β crevice ( Figure 5).
Interestingly, simulations of compounds exhibiting scarce potency in vitro failed to reproduce the fundamental interaction between the ligands' carboxylate and the MIDAS. For some of these compounds, i.e., 2, 5, 9, and 14, poses have been identified in which the diphenylurea fits subpocket E and the carboxylate is oriented towards the MIDAS; however, in general the distance between the carboxylate and Mg 2+ is greater than 3 Å (Supporting information, Figure S15). Apparently, the cyclic β 2 -Pro scaffolds allow DS-70 and DS-23 to optimally direct the carboxylate and diphenylurea pharmacophores across the binding site, allowing the backbones to pass over the saddle that separates subpockets C and E ( Figure 5).
The simulation of potent antagonist 8 revealed that the (R)-isoAsp β 3 -core nicely fits subpocket D (see Figure 4B) and is delimited by residues of the β1 subunit, i.e., Phe321, Ala260, and Ser227. The profound insertion of the propylamide side chain into pocket D seems to pull the entire structure against the β1-subunit, so that both urea NHs of MPUPA Apparently, the cyclic β 2 -Pro scaffolds allow DS-70 and DS-23 to optimally direct the carboxylate and diphenylurea pharmacophores across the binding site, allowing the backbones to pass over the saddle that separates subpockets C and E ( Figure 5).
The simulation of potent antagonist 8 revealed that the (R)-isoAsp β 3 -core nicely fits subpocket D (see Figure 4B) and is delimited by residues of the β 1 subunit, i.e., Phe321, Ala260, and Ser227. The profound insertion of the propylamide side chain into pocket D seems to pull the entire structure against the β 1 -subunit, so that both urea NHs of MPUPA in subunit E make a bifurcated hydrogen bond with E320, while Gly carboxylate is tightly bound to MIDAS (salt bridge) and to the β 1 residues Tyr133, Gly223, and Asn224.
The computations suggest that the comparatively inferior affinity of 3 might depend on the modest contribution to complex stabilization from N-acetyl (S)-Dap (diaminopropionic acid) scaffold access into subpocket D. GlyCOO − can still bind to the MIDAS and interact with other elements of subpocket C, Asn224, and Ser132, but the diphenylurea seems to lose most of the above-mentioned stabilizing contacts within the E pocket. This interpretation seems in line with the complete lack of receptor affinity of 1, topologically correlated to 3, since the former addresses subpocket D with a simple methyl ( Figure S15).
Regarding the calculated pose of 13, although the pharmacophores are still able to attain the necessary positions within the binding site, the (R)-β 3 -Ala scaffold completely lacks any interactions with the D subpocket, plausibly explaining the very low efficacy of this ligand.

Discussion
Integrins are involved in many biological processes, such as cell adhesion and migration, and are responsible for the survival of the cell and the control of the cell cycle. In addition, integrins play a fundamental role in several diseases, including cancer progression and metastasis, coronary disease, thrombosis, fibrosis, inflammatory, and autoimmune pathologies. It is precisely their involvement in the main human biological process and pathobiological processes of various diseases that render integrins very attractive therapeutic targets. To date, six different integrin-targeting drugs have been marketed, and many others are under investigation [3][4][5].
A recent research topic concerns the development of inhibitor peptides, associated with many advantages, such as higher specificity and efficacy with respect to small molecules, and lower immunogenicity and costs with respect to macromolecules [3,5,13,22,23]. Nevertheless, native peptides are associated with many limitations, such as low metabolic stability toward proteolysis and poor absorption. To improve these limited properties, natural peptides can be structurally modified, obtaining compounds called peptidomimetics. Most of the peptides entering clinical trials today are, in fact, peptidomimetics.
To increase the bioavailability of integrin ligands, we opted for small hybrid α/βpeptidomimetics. In particular, the compound DS-70, MPUPA-β 2 Pro-Gly, and its partially retro analogue DS-23, were found to be potent antagonists of this receptor [20]. Starting from these ligands, derivatives 1-16 were designed by introducing diverse β 2 -or β 3residues, in either an (S)-or (R)-configuration, and straight sequences 1-8 [25] and partially retro-sequences 9-16 were tested in vitro in integrin-mediated cell adhesion experiments to understand their effects.
As it turned out, the ligands DS-70, DS-23, 3, 8, and 13 were able to bind with excellent affinity to the receptor. All these compounds strongly reduced the adhesion of integrinexpressing cells to the natural ligands, without inducing receptor open conformation nor activation of intracellular signaling pathways. Intriguingly, DS-70, DS-23, and 8 showed very high potency, with IC 50 values in a low nanomolar range, similar to that of the reference peptide BIO1211, despite the much smaller structure.
This observation provoked interest in the potential mechanisms that form the basis of the efficacy toward α 4 β 1 integrin. To investigate the chemical and structural determinants that lead to integrin binding and inhibition, some representative compounds were analyzed via molecular docking. The α 4 β 1 integrin dimer was obtained from the structures of the separate monomers, and the resulting binding site appeared to be partitioned in subpockets A-E, which were of different sizes and compositions ( Figure 4B). For the reference BIO1211, the simulations revealed a perfect fit into subpockets B-E of the α/β crevice ( Figures 4B and 5), giving rise to many interactions, including salt bridges involving Asp and Pro carboxylates. As for the minimalist hybrid ligands, the calculations were suggestive of bioactive poses that in part reproduce the binding of BIO1211 by occupying subsites C, D, and E. In particular, the latter nicely hosts the diphenylurea group, confirming the efficacy of this α 4 integrin-targeting motif [13].
Interestingly enough, the simulations highlighted the fundamental role of the βresidues, their configuration, and the specific side chains. Indeed, for DS-70 and DS-23, high affinity for the receptor seems to be correlated to the ability of the cyclic β 2 -Pro scaffold to pass over the saddle between pockets C and E and to orient the pharmacophores toward the respective subpockets.
For the bioactive compounds which include the linear β 2 -or β 2 -residues, the results in general were inferior in respect to DS-70 and DS-23, with the exception of 8, characterized by a low nanomolar affinity. The comparison of the binding poses of 8, 3, and 13 support that good receptor affinity seems to correlate to the ability of the central β-residue to place its side chains into subpocket D. This would consent the ligands to maintain at least in part the network of interactions of the reference LDVP peptide BIO1211, including the fundamental interaction, i.e., the ionic bond between negatively charged carboxylate with the MIDAS Mg 2+ ion of the β subunit, with a distance comparable to those found in X-ray crystallographic structures (between 1.92 and 2.04 Å).
In contrast, this direct coordination was not maintained for compounds associated with an inferior in vitro potency. Although docked poses in which the carboxylate is oriented toward the Mg 2+ of the MIDAS have been identified for some of the low potency compounds, the distance was >3 Å ( Figure S15).
Although the computational and experimental evidence seems in good agreement, it must be stressed that, as with any theoretical study, our model has some limitations. The accuracy that can be expected from homology modeling is highly dependent on the sequence identity between the target and templates. In the present case, the model structure of α4β1 integrin was derived from the experimental structures of the α 4 and β 1 monomers bound to different partner subunits. We assumed that the tertiary structure of each monomer was the best available, while the relative orientation of the two monomers and the conformations adopted by the side-chains of the residues at the interface may be less accurate. Fortunately, the largest part of the binding pocket was maintained unaltered during the protein-protein docking procedure. The metal binding sites were the subject of an accurate refinement in order to reproduce the experimental coordination geometries.
For these reasons, we felt reasonably confident to perform protein-ligand docking. For the subsequent docking calculations, we relied on scoring functions to rank and select the best binding poses. Diverse docking/scoring methods are available to provide reasonable predictions of ligand binding modes, but their performances are often systemdependent [43]. Hence, we conducted preliminary redocking calculations to establish the best docking procedure to be used in the case of integrins.
To conclude, on the basis of the experimental and theoretical evidence, we believe that our findings support that α 4 β 1 integrin antagonists may represent useful research tools to develop more effective drugs to fight cancer and inflammatory diseases. Nevertheless, further studies are needed to better unravel the mechanisms by which integrin peptidomimetic antagonists block integrin functions, including the analysis of adhesion kinetics, the effects under repulsive forces, and the formation of focal adhesion. Put in perspective, more functional assays need to be employed to better investigate the anti-inflammatory and/or anti-cancer effects of α 4 β 1 integrin peptidomimetic antagonists.

General Methods
Chemicals and solvents were purchased from commercial sources and used without further purification. Peptides were purified via semipreparative RP HPLC under the following conditions: Agilent 1100 series apparatus; reverse-phase column Waters XSelect Peptide CSH C18 OBD column, 19 × 150 mm 5 µm (column description: stationary phase octadecyl carbon chain-bonded silica, double-end-capped, particle size of 5 µm, pore size of 130 Å, length of 150 mm, internal diameter of 19 mm; DAD of 210 nm, DAD of 254 nm); mobile phase isocratic 1:1 H 2 O/CH 3 CN/0.1% TFA, flow rate 10 mL/min −1 .
ESI MS analysis was carried out under the following conditions: MS single quadrupole HP 1100 MSD detector, drying gas flow of 12.5 L/min −1 , nebulizer pressure of 30 psig, drying gas temp. of 350 • C, capillary voltages of 4500 (+) and 4000 (−), scan range of 50-2600 amu. NMR spectra were recorded using Varian Gemini apparatus, for 1 H NMR, at 400 MHz, and for 13 C NMR, at 100 MHz, at 298 K in 5-mm tubes, using 0.01 M peptide. Solvent suppression was carried out via the solvent presaturation procedure implemented in Varian (PRESAT). Chemical shifts are reported in ppm (δ) with the following internal standards:

Synthetic Procedures
The synthesis and isolation of compounds 1-8 has been reported elsewhere [25]. For convenience, the synthetic steps are resumed in Supporting Information, together with the synthetic schemes for 9-16.

Peptide Synthesis, General Procedures
Sequences 1, 2, 5, 6, and 9-16 were prepared in a solution using EDC . HCl/HOBt/TEA, in a DCM/DMF ratio of 4:1, at RT for 12 h, or TBTU/HOBt/DIPEA, in a DCM/DMF ratio of 4:1, at RT for 12 h, as the activating agents (Schemes S3-S6). The removal of acid-labile protecting groups was performed with TFA/DCM (1:1) at RT for 1 h. The removal of benzyl ester protecting groups was performed via catalytic hydrogenation with H 2 /Pd/C, in EtOH, at RT for 12 h.
The remaining sequences were prepared in solid phase using polypropylene syringes fitted with a polyethylene porous disc, Wang resin preloaded with Fmoc-Gly, and Fmoc protecting amino acids. For peptide bond formation, the Fmoc-amino acid (2.5 eq.) was activated with TBTU/HOBt/ DIPEA or DCC/HOBt in DCM/DMF at RT for 3 h. Fmoc group deprotection was performed with 20% piperidine/DMF (2x). Peptide cleavage from Wang resin was conducted with a mixture of TFA and scavengers, i.e., TFA/H 2 O/TIS/PhOH (80:10:10 v/v/v), at RT for 2.5 h.

Modelling of α 4 β 1 Integrin Structure
To define the secondary and the tertiary structure of the integrin, homology modelling was performed with the program MODELLER 10.4 [50], using as templates the experimental dimeric structures of integrins α 4 β 7 and α 5 β 1 (PDB IDs: 3V4V and 4WK0, respectively). In total, 100 different models were generated and evaluated with the DOPE score. The model with the most favourable score was retained and used in the subsequent step. Details concerning structures used as templates and considered as potential templates are shown in Supporting Information.
To define the correct interface between the α and β subunits, protein-protein docking was performed with the web server HADDOCK 2.4 [51]. Different models were obtained, and the model from the most populated cluster with the favorable HADDOCK score was selected for the next step. Finally, the metal ions containing sites of the β subunit were refined by performing several optimization steps with the program MODELLER 10.4. In each refinement step, 100 models were generated and evaluated with the DOPE score. The model with the best DOPE score was retained and used as the input in the subsequent step. Thereby, it was possible to define the coordination for each metal ion with surrounding residues and water molecules (Supporting Information). The resulting model was used as the α4β1 structure in protein-ligand docking studies.

Protein-Ligand Docking
The initial geometry of each ligand was generated using UCSF Chimera 1.16 [52] and then optimized at the Hartree-Fock 6-31G* level using ORCA 4 [53]. Using a re-docking procedure and analyzing the results with the web server PacDOCK, it was possible to identify UCSF DOCK 6 [54] as the best docking program for this specific system (see Supporting Information). As a result, UCSF DOCK 6 was used to conduct all the docking using the compounds selected as potential inhibitors of α 4 β 1 integrin.
To prepare these ligands for docking calculation, UCSF Chimera 1.16 was used in combination with the antechamber software (AmberTools23 version) [55]. In particular, with the DockPrep tool included in UCSF Chimera, hydrogens and charges (AMBER ff14SB [56] for standard residues and AM1-BCC [57] for non-standard residues) were added. Chimera and this tool were also used to prepare the receptor input file.
The program UCSF DOCK 6 explores a 3D region defined by a cluster of spheres reproducing the negative image of the binding pocket. To define several sets of overlapping spheres, the sphgen tool was used, considering all points outside the receptor surface, a maximum sphere radius of 4.0 A, and a minimum of 1.4 Å. To select the spheres that represent the binding site, all the spheres within 10 Å from the MIDAS extending toward the α subunit were selected with the sphere_selector tool. The aim is to direct the ligand toward the binding site, rather than all over the receptor.
Initially, the simulations were performed using the grid-based score of UCSF DOCK 6, thus it was necessary to define the receptor grid (a series of grid points with an overall size defined by a box). The program showbox was used to enclose the selected spheres in the grid box, setting the box length value at 5 Å, while the program grid was used to generate the grid on the basis of these spheres.
Different docking experiments with rigid protein and flexible ligands were performed for each ligand, using different conditions to check for convergence in docked poses, in particular docking experiments, where 1000 different orientations were generated for each anchor and pruned on the basis of clustering, keeping 100 different orientations. To match the receptor spheres with the ligand heavy atoms, an automatic matching method was performed. All these docking experiments were processed using the grid-based score as the scoring function. From these first experiments, for each selected compound, 1000 docked poses were generated. Through a cluster analysis using the tool ClusDOCK from the web server PacDOCK with the gromos algorithm, it was possible to identify a reduced number of clusters for each promising ligand. In conclusion, the most representative poses of the most populated and most favorable energy clusters were identified. The best poses of most of the different docking experiments belong to these clusters; a convergence of the docking experiments can therefore be assumed.
To identify the best pose for each ligand that would be able to exhibit a good potency during the in vitro studies and bind to the receptor during the simulations, a re-scoring was performed. In particular, all the most representative docked poses of each ligand were re-scored using the AMBER score implemented in the program UCSF DOCK 6. In addition, to ameliorate the accuracy of these poses, re-scoring simulations were conducted, treating both the ligand and the residues within a distance of 3.5 Å from the ligands as flexible. For each pose, 3000 MD steps and 100 minimization cycles both before and after the MD simulations were conducted.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The calculated model of α4β1 integrin is freely available at https://site.unibo. it/bioinorgchem/en/downloads.