Antibodies as Carrier Molecules: Encapsulating Anti-Inflammatory

Publication Date (Web): January 29, 2018. Copyright © 2018 American Chemical Society. *E-mail: [email protected]., *E-mail: [email protected]. Cite this:J. P...
0 downloads 0 Views 4MB Size
Article Cite This: J. Phys. Chem. B 2018, 122, 2064−2072

pubs.acs.org/JPCB

Antibodies as Carrier Molecules: Encapsulating Anti-Inflammatory Drugs inside Herceptine Published as part of The Journal of Physical Chemistry virtual special issue “Manuel Yáñez and Otilia Mó Festschrift”. José Pedro Cerón-Carrasco,*,† Horacio Pérez-Sánchez,† José Zúñiga,‡ and Alberto Requena*,‡ †

Bioinformatics and High Performance Computing Research Group (BIO-HPC), Universidad Católica San Antonio de Murcia (UCAM) Campus de los Jerónimos, 30107 Murcia, Spain ‡ Departamento de Química Física, Universidad de Murcia, 30100 Murcia, Spain S Supporting Information *

ABSTRACT: The human epidermal growth factor receptor 2 (HER2) is overexpressed in about a third of breast cancer patients, with a strong involvement of the cyclooxygenase-2 (COX-2) enzyme in the tumor progress. HER2 and COX-2 are consequently potential targets for inhibiting carcionogenesis. Herceptin (trastuzumab) is an antibody that partially blocks HER2positive cancers at their initial stage. Unfortunately, the overall response rate to the single treatment with this antibody is still modest, and therefore, it needs to be improved by combining several chemotherapeutic agents. On the other hand, nonsteroidal anti-inflammatory drugs (NSAIDs) are designed to halt COX-2 functionality, so they might also exhort an anticancer activity. In this contribution, dual Herceptin−NSAID drugs are designed using theoretical tools. More specifically, blind docking, molecular dynamics, and quantum calculations are performed to assess the stability of 14 NSAIDs embedded inside Herceptin. Our calculations reveal the feasibility of improving the antitumor activity of the parent Herceptin by designing a dual HER2-targeting with Etofenamate. That coupling mode might be used to further rationalize new clinical strategies beyond classical antibody dosages.



INTRODUCTION The human epidermal growth factor receptor 2 (HER2) is a transmembrane tyrosine kinase responsible for regulating normal cell growth and survival rate, along with other biological functions.1 However, it has been shown that about 30% of patients with breast cancer overexpress this protein.2 Cyclooxygenase-2 (COX-2) is another enzyme that contributes to many natural physiological processes such as hemostasis, platelet aggregation, and kidney and gastric function. Recent clinical findings have shown that COX-2 is also aberrantly expressed in inflammatory breast carcinomas developed by HER2-positive patients.3 Indeed, it has been shown that inflammation contributes largely to the tumor progression, probably due to the stimulation of prostanoid production.4 Furthermore, the erroneous activation of HER2 and COX-2 not only is associated with the propagation of cancerous tissues © 2018 American Chemical Society

but also correlates with more aggressive tumors and a poorer prognosis.5,6 Consequently, HER2 and COX-2 have been proposed as two of the main antitumor targets for treating breast cancer as well as other cases, including ovarian, gastric, and prostate tumors.7−13 Of all the possible novel anticancer drugs, antibodies are one of the most attractive, since they can be biologically designed to recognize malignant cells.14 As recently summarized by Ledford,15 one can use the immune system as a weapon against cancers resistant to conventional indications, a technique known as immunotherapy, which is expected to be used in more than 60% of the cancer treatments in the next Received: October 31, 2017 Revised: January 10, 2018 Published: January 29, 2018 2064

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B

Figure 1. Nonsteroidal anti-inflammatory drugs (NSAIDs) encapsulated into the Herceptin antibody.

drug cocktail.20 This is especially required in patients with high inflammatory biomarkers who respond less to the treatment with Herceptin.21 A first approach to enhance the Herceptin properties is to covalently conjugate a cytotoxic drug (payload) to the antibody structure by means of a stable linker, which eventually leads to the corresponding antibody−drug conjugate (ADC).22 The antibody then plays a double function: it must recognize the cancer cell by binding to the tumor antigen, while simultaneously favoring the in situ delivery of the embedded drug by acting as a “Trojan horse”. Although the ADC field has attracted the attention of many groups in the cancer therapy

decade. In this framework, the approval of Herceptin (trastuzumab) by the US Food and Drug Administration (FDA) in 1998 laid the cornerstone in molecular cancer therapeutics for HER2-positive patients.16 Herceptine is a humanized monoclonal antibody programmed to selectively interact with the HER2 oncoprotein,17 which has become standard in the treatment of metastatic breast cancer, with a decrease of a 50% of the recurrence cases over a 20-month period.18 Unfortunately, the long-term response to a singleHerceptin treatment is still modest because tumors become resistant after the initial benefit,19 so its activity needs to be enhanced by combining several chemotherapeutic agents in a 2065

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B field, very few Herceptin-based ACDs have been approved by the FDA for cancer treatment to date. The most successful case is the so-called T-DM1, an ADC that combines Herceptin with the microtubule-inhibitory drug DM1, a derivative of maytansine.23,24 Recent clinical trials have demonstrated that T-DM1 increases the survival rate and reduces the toxicity in patients with HER2-positive advanced breast cancer previously treated with trastuzumab.25 T-DM1 is, however, an exception rather than a general rule, and most of the trials with ADCs have failed so far at the last clinical stages. As recently reviewed by Jackson,26 preclinical data suggest that heterogeneity might be one of the main limiting factors in the therapeutic action of ADCs. The linkers used to attach the cytotoxic drug to the antibody usually react with lysine and cysteine residues located at the antibody surface. Since an antibody usually presents around 50 lysines and 12 cysteines, the use of nonspecific conjugation methods yields heterogeneous ADCs that vary in drug/antibody ratio and conjugation site structures.27 Such heterogeneous conjugation at the surface induces self-assembling, low solubility, and/or instability due to side reactions like thiol exchange, which eventually lead to premature release.26,28 It is possible to design site-specific conjugation by using more elaborate protocols based on engineered amino acids or enzyme mediated or linker modification processes, although all them are at a very initial clinical stage.29,30 An additional attractive approach is to use Herceptin as a noncovalent carrier, in which a small molecule is directly embedded in the antibody structure without the use of any additional molecule acting as a linker.31 In this line, Gao and Zingaro et al. synthesized a novel metallodrug with a large affinity for Herceptin, so it could be labeled without using a more complex stepwise synthesis.32 The metallodrug is effectively retained within the antibody structure, which protects the “cargo” from side reactions and early deactivation, and the resulting dual drug−antibody exhibits a potent synergistic anticancer effect.32 With the aim of enhancing currently available breast cancer treatments, in this contribution we assess the feasibility of combining the most used nonsteroidal anti-inflammatory drugs (NSAIDs, see Figure 1) with Herceptin by means of theoretical tools. Our approach is based on the strong correlation existing between inflammation and therapy efficiency, which suggests that NSAIDs may be combined to simultaneously target HER2 and COX-2.33,34 In a previous work, Higgins and co-workers35 studied the effects of two NSAID, Aspirin and Celecoxib, on breast cancer parameters, finding contradictory results on the efficiency of NSAIDs. Their clinical trial shows that neither Aspirin nor Celecoxib is associated with a positive effect in disease parameters.35 This outcome indicates that it is not easy to predict/rationalize the effectiveness of a combined Herceptin−NSAID treatment, so the selection of the NSAID should be attempted with care. Accordingly, we conduct herein docking, quantum mechanical (QM), and molecular dynamics (MD) simulations to systematically assess the stability of a set of commonly used NSAIDs embedded in Herceptin in order to rank the most promising dual drugs.

carboxamides, oxicams, sulphonanilides, and diaryl-substituted pyrazole and furanone derivatives. Their raw structural data were next refined by assigning bond orders, stereochemistry, hydrogen atoms, and protonation states at physiological pH with the LigPrep module37 and the OPLS3 force field,38 as implemented in the Small-Drug Design Suite of Schrödinger.39 As expected, all carboxylic groups are predicted to be deprotonated at pH = 7. The resulting structures are subsequently optimized at the M06-2X/6-31+G(d,p) level of theory.40 Additional vibrational calculations are performed to confirm that stationary points are true minima in the potential energy surface (no imaginary frequencies). Partial atomic charges are computed within the Merz−Singh−Kollman ESP scheme41,42 and subsequently implemented in docking and MD simulations. All QM calculations were carried out using Jaguar.43,44 The optimized NSAIDs are next embedded into the experimental crystal structure of Herceptin, deposited at the Protein Data Bank (PDB)45 with code 1N8Z (Figure 2).46 The

Figure 2. Representation of the experimentally resolved HER2− Herceptin (Fab) structure displayed as green and red−white cartoons, respectively.

PDB file contains the variable region of the antibody (Fab fragment) but not the constant region (Fc fragment is not included). However, the Fab fragment is enough to ensure that the encapsulation of NSAID does not disrupt the ability of the whole antibody to recognize the HER2 oncoprotein: if the NSAID is encapsulated far from the antigen binding site, the biological activity of the antibody is guaranteed. Furthermore, the Fab fragment might be used as a molecular carrier on its own, since it also reaches the cancer cell with a faster pharmacodynamic: Herceptin needs 1 day compared to just a few hours for the Fab fragment.47,48 Accordingly, the Herceptin (Fab) fragment extracted from the PDB data was subjected to further refinement using the Protein Preparation Wizard,49 to include missing hydrogen atoms and assign all bond orders. The protonation states of all residues were computed with the PROPKA code.50 A restricted optimization was finally carried out by minimizing all hydrogen atoms with the OPLS3 force field.38 The resulting refined Herceptin structure was then used as the host in SiteMap and blind docking (BD) approaches. The former code scanned the whole protein structure for identifying binding sites that possessed the suitable shape and chemical properties for embedding small molecules.51,52 This code was applied as implemented in Schrödinger,53 and used a library of



CHEMICAL MODELS AND COMPUTATIONAL DETAILS The computational protocol is initiated by obtaining the Cartesian coordinates of the selected NSAIDs as deposited at the PubChem chemical library.36 The training set was built with a wide panel of 14 NSAIDs including carboxylic acids, 2066

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B

Figure 3. SiteMap code identifies five binding pockets in the Herceptin structure: blue, red, and yellow surfaces highlight the H-bond donor, H-bond acceptor, and hydrophobic regions within the binding site, respectively.

and an entropy term (see ref 57 for further details). In our computational protocol, the MM-GBSA method is used to improve all poses obtained in the BD step by using the OPLS3 as implemented in the Small-Drug Design Suite of Schrödinger.39 All previous simulations account for the structure of the antibody as a rigid body. However, one might expect the Herceptin structure to readapt its spatial configuration once the NSAID molecule is encapsulated inside. To describe the Herceptin−NSAID binding mechanism better, the dynamic stability of the best-ranked drugs on the scale of MM/GBSA values was finally assessed by locating the corresponding adducts in an orthorhombic box, with a buffer distance of 10 Å in all directions, and by subsequently filling the box with water molecules with the simple point charge (SPC) model and sodium cations to keep the system electronically balanced. Additional sodium and chloride ions were incorporated into the system to simulate the physiological salt concentration of 0.15 M NaCl. An initial full minimization was conducted for 2000 steps using the steepest descent method, with a convergence threshold of 1.0 kcal/mol/Å. The solvated systems were then relaxed throughout several stages that included a soluterestrained minimization, free-restrain minimization, NVT simulation of 24 ps at T = 10 K, and NPT simulations at T = 10 K and P = 1 atm. For the production phase, the temperature was set to 300 K with the Nosé−Hoover algorithm, with a relaxation time of 1.0 ps.59,60 Pressure was controlled at 1 bar with the Martyna−Tobias−Klein barostat using isotropic coupling and a relaxation time of 2.0 ps.61 The RESPA integrator was used to integrate the motion equations with a 2.0 fs time step for bonded and near interactions, and a 6.0 fs time step for far interactions.62 A cutoff of 9 Å was applied to nonbonded interactions. van der Waals interactions

compounds to determine all possible biding pockets compatible with each drug. The BD simulations were carried out according to the procedure described by Navarro and co-workers,54 by searching for the global minimum of the potential energy surface with the genetic algorithm implemented in the Lead Finder program.55,56 The size of the grid box for ligand BD was set to 120 Å in each direction from the geometric center of the Herceptin model system. The dG score produced by Lead Finder was taken as the predicted value of the ligand binding energy, which accounts for the Lennard-Jones term (LJ), metal interactions, solvation term, hydrogen bonds (H-bonds), electrostatic interactions, internal energy of the ligand, contributions to entropy due to ligand torsions, and a solvation penalty term. In that BD approach, multiple BD runs started around the geometric centers of all residues within the selected threshold. A histogram with the resulting distribution of binding energies was obtained, and the 10 best poses ranked in BD were retained for further refinement. We have previously demonstrated that even if BD provides a valuable initial description of the Herceptin domain interacting with an embedded drug, the scoring dG energies need to be improved with a higher level of theory to reach the most grounded biological conclusions.31 Although the reliable prediction of any protein−ligand binding free energy is still a challenge for the computational chemistry community, the combined Molecular Mechanical/Generalized Born Surface Area (MM/GBSA) approach offers an affordable alternative for mimicking the binding reaction in large biosystems.57,58 The MM/GBSA method is based on the difference between the free energies of the protein, ligand, and the complex in solution. The free energy for each species involved in the reaction (ligand, protein, and ligand−protein complex) is described as a sum of a gas-phase energy, polar and nonpolar solvation terms, 2067

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B were evaluated using a cutoff radius of 9 Å, and the electrostatic part was computed using the Particle Mesh Ewald (PME)63 method with a tolerance of 10−9. To be consistent with all the previous computational steps, MD simulations were run for 10 ns using the OPLS3 force field as implemented in Desmond code.38,64



RESULTS AND DISCUSSION The search for the druggable binding sites on the Herceptin is first assessed with SiteMap. The refined crystal structure of Herceptin (i.e., the structure processed with the Protein Preparation Wizard workflow) is used as the starting model system. As discussed above, this algorithm scans the protein structure to identify pockets that can potentially accommodate a small molecule inside. The results obtained are summarized in Figure 3. As observed, SiteMap identifies five possible binding pockets in the Herceptin structure, which are displayed as cubic boxes with colored surface-fingerprints. In this color scheme, blue surfaces represent regions suitable for occupancy by drugs with H-bond donor groups, red surfaces highlight regions compatible with H-bond acceptors, and yellow surfaces localize areas in which hydrophobic groups can be located. A visual inspection reveals that sites 2, 3, and 5 are mainly placed in the outer region of the protein, which is not the optimal place for a “payload” since (i) it presents fewer anchoring points (smaller surfaces) to retain the drug inside the Herceptin structure, e.g., lower druggability, and (ii) the interaction with solvent or surrounding biomolecules can induce an early release. Site 4 should also be discarded for designing a dual carrier-cargo system. Effectively, a comparison of Figures 2 and 3 shows that site 4 matches the antigen binding site, and consequently, if a small drug is located in that region of Herceptin, its biological functionality cannot be guaranteed. Finally, therefore we focus on site 1, which is placed in the inner region of the antibody with no interaction with the antigen region. If a drug is embedded into that pocket, the interaction with solvent molecules or other surrounding biological molecules is avoided. In addition, site 1 possesses the largest H-bond acceptor and donor regions as well as the largest hydrophobic ability due to the inner core-place, with all those parameters supporting better druggability in this region. SiteMap offers valuable results as it identifies potential binding sites on the Herceptin structure. However, it does not account for any specific drug. In contrast, all pockets are characterized according to general parameters like size, shape, and chemical properties. Our next computational step was, therefore, to use a BD approach to explicitly dock the NSAIDs of Figure 1 into Herceptin. The best pose of each NSAID is collected in Figure 4 for a visual inspection. It is noticeable that the binding preference for embedding largely depends on the NSAID chemical structure. As discussed above, the series of selected NSAIDs accounts for carboxylic acids, carboxamides, oxicams, sulphonanilides, and diaryl-substituted pyrazole and furanone derivatives, which explains the dissimilar behavior. DB results allow us to perform a first drug screening and to prioritize the most suitable functional groups. As showed in Figure 4, one of the NSAIDs tested, Etodolac, is located in the region of site 4, so it interacts with the residues involved in the HER2 recognition function. Etodolac should therefore be discarded to preserve the properties of the native Herceptin antibody. Furthermore, two NSIADs (Celecoxib and Oxaprozin) are preferentially docked in the site 2 region. As mentioned above, a ligand can only be poorly anchored in that region, so it

Figure 4. Conformations of the top docked pose of each NSAID. Herceptin is displayed in cartoons, and small drugs are shown as sticks. Etodolac (shown in white blue) preferentially docks into the region of site 4, while Celecoxib and Oxaprozin (displayed as orange and yellow sticks, respectively) dock into site 2.

is expected to produce a quick release of any small drug embedded from Herceptin toward the solution. This is a noteworthy result that might shed light on previous trials in hospitals. According to the clinical study by Dang and coworkers, the combination of Celecoxib and Herceptin is not active in breast cancer patients.65 Our simulations indicate that Celecoxib cannot be efficiently loaded into the Herceptin structure, a computational result that might (at least partially) explain that poor synergic effect between both drugs. To the best of our knowledge, no equivalent clinical trails are available for Oxaprozin, although our predictions discard synergetic effects if that NSAID is combined with Herceptin. It is worth stressing that most BD poses lie on the inner region of Herceptin. This finding agrees with the better druggability of the region labeled as site 4 by the SiteMap code. Consequently, although BD calculations are performed without any bias (i.e., they do not depend on the SiteMap results), both techniques provide consistent conclusions and provide further support to our biological predictions. Let us move on now in the level of theory by computing the Herceptin−NSAID binding energy within the MM/GBSA framework. These calculations are performed on the 10 best poses of each NSAID arising from BD simulations. The lowest (more negative) energetic MM/GBSA values, e.g., the most likely binding spots, are listed in Table 1. According to the interaction energies computed, Etofenamate is the NSAID with the highest affinity for Herceptin, with an interaction energy of ca. −48 kcal mol−1. Indeed, three of its BD poses are on the top list. Nepafenac, Naproxen, Ketorocal, Meloxicam, and Nimesulife are also well-positioned candidates, although with a less intense interaction (ca. −34 kcal mol−1). As expected, Celecoxib is not among the best-ranked NSAIDs since it is located in the site 2 region. The computed MM/GBSA interaction energy is only ca. −20 kcal mol−1. It is also worth noticing that Aspirin, the most used NSAID, and probably the most prescribed drug, is not among the best candidates to be encapsulated either, since it has a predicted interaction energy 2068

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B

contacts with Lys45, the cation−π interactions between the −NH3+ group (the residue is protonated at the physiological pH, according to PROPKA predictions) and the CF3−phenyl ring of the Etofenamate are more flexible than the two H-bonds with the Leu46 residue. The former geometrical parameters vary in the range 4−9 Å, with an average value of about 5 Å. Figure 5 also shows that the H-bonds are somewhat more stable at ca. 2 Å, during the first 20 ns. However, both interactions seem to be lost at the last MD window. Although such structural changes might indicate a release of the NSAID, a visual inspection reveals that the Heceptin−Etofenamate adduct is retained during all the simulations at site 4 due to the rearrangement of the other close residues that interact with Etofoneamate. As illustrated in Figure 5, a new H-bond is established with the Gly57 (d4), which is allowed to flip toward the ligand after only 30 ns of trajectory, and consequently plays a pivotal role in the stability of the system. These results stress the importance of performing dynamic simulations in drug design, which is demonstrated to be a crucial step in the design of Herceptin bifunctional drugs. We therefore conclude that, among all the small drugs tested, Etofenamate is the NSAID most prone to be encapsulated in the Herceptin structure without altering its biological ability for HER2 recognition. Until now, Herceptin had been proven to carry metallodrugs and alkaloids.32,67 The reported data allow inclusion, for the first time, of an NSAID to the list of payloads compatible with Herceptin.

Table 1. Computed Herceptin−NSAID Binding Energies within the MM/GBSA Level of Theory NSAID

BD pose

MM/GBSA energy (kcal mol−1)

Etofenamate Etofenamate Etofenamate Nepafenac Nepafenac Naproxen Ketorolac Meloxicam Nepafenac Nimesulide

01 02 03 01 02 01 01 01 02 01

−47.92 −35.41 −34.61 −34.59 −34.48 −34.04 −33.71 −33.63 −33.37 −33.29

of −30 kcal mol−1 which is about 20 kcal mol−1 lower in magnitude than the energy predicted for Etofenamate. It should be noticed that the absolute MM/GBSA energies have to be cautiously analyzed as more refined methods, such as free energy perturbation (FEP) calculations, are to be used to provide more accurate binding energy predictions.66 Nevertheless, relative MM/GBSA energies can be used to prioritize NSAIDs in the design of stable bifunctional drugs, as they allow us to screen large libraries at affordable computational costs. In order to confirm the dynamic stability of the Herceptin− Etofenamate system, additional MD simulations were conducted during 50 ns of trajectory. The evolution of the rootmean-square-deviation (RMSD) on all residues is used to monitor the stability of the dual drug. As illustrated in the left panel of Figure 5, the RMSD value [black line] quickly reaches a stable value of about 2 Å during the initial 2 ns of the MD trajectory. This stability suggests that the embedding of Etofenamate does not induce a large perturbation in the overall structure of Herceptin, and it can therefore be easily accommodated within the carrier. An analysis of the ligand− protein contact along the trajectory shows that the interaction is initially governed by a cation−π interaction with Lys45 (d3) and two H-bonds (d2 and d1) established with Leu46 (see Figure 5, right panel). As illustrated by the evolution of the



CONCLUSIONS AND OUTLOOK In this work, we employ a wide panel of theoretical tools to ascertain the use of Herceptin, an antibody able to recognize the HER2 overexpressed protein in breast cancer cells, as a molecular carrier for nonsteroidal anti-inflammatory drugs (NSAIDs). Our approach is based on the host ability of Herceptin, which is supported by it having a number of binding pockets compatible with the encapsulation of small-drug molecules. Indeed, SiteMap calculations reveal the presence of an inner pocket located far from the antigen region, which might be used to charge the antibody without altering its

Figure 5. Summary of the molecular dynamics trajectory. Left panel: plots correspond to the evolution of the root-mean-square-deviation (RMSD) of all residues [black line], the two H-bonds established with Leu46 [d2 and d1 as red and blue lines, respectively], the cation−π interaction with Lys45 [d3, green line], and the H-bond which appears between Etofenamate and Gly57 residue [d4, orange line]. All distances are given in angstroms. Right panel: sketch of the binding site extracted from the last snapshot of the MD production and the monitored distances. 2069

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B

(2) Slamon, D.; Clark, G.; Wong, S.; Levin, W.; Ullrich, A.; McGuire, W. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 1987, 235, 177− 182. (3) Subbaramaiah, K.; Norton, L.; Gerald, W.; Dannenberg, A. J. Cyclooxygenase-2 is overexpressed in HER-2/neu-positive breast cancer. Evidence for involvement of AP-1 and PEA3. J. Biol. Chem. 2002, 277, 18649−18657. (4) Fouad, T. M.; Kogawa, T.; Reuben, J. M.; Ueno, N. T. In Inflammation and Cancer; Aggarwal, B. B., Sung, B., Gupta, S. C., Eds.; Springer: Basel, 2014; pp 53−73. (5) Slamon, D.; Godolphin, W.; Jones, L.; Holt, J.; Wong, S.; Keith, D.; Levin, W.; Stuart, S.; Udove, J.; Ullrich, A.; et al. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 1989, 244, 707−712. (6) Howe, L. R.; Dannenberg, A. J. COX-2 Inhibitors for the Prevention of Breast Cancer. J. Mammary Gland Biol. Neoplasia 2003, 8, 31−43. (7) Soslow, R. A.; Dannenberg, A. J.; Rush, D.; Woerner, B. M.; Khan, K. N.; Masferrer, J.; Koki, A. T. COX-2 is expressed in human pulmonary, colonic, and mammary tumors. Cancer 2000, 89, 2637− 2645. (8) Half, E.; Tang, X. M.; Gwyn, K.; Sahin, A.; Wathen, K.; Sinicrope, F. A. Cyclooxygenase-2 Expression in Human Breast Cancers and Adjacent Ductal Carcinoma in Situ. Cancer Res. 2002, 62, 1676−1681. (9) Holmes, M. D.; Chen, W. Y.; Schnitt, S. J.; Collins, L.; Colditz, G. A.; Hankinson, S. E.; Tamimi, R. M. COX-2 expression predicts worse breast cancer prognosis and does not modify the association with aspirin. Breast Cancer Res. Treat. 2011, 130, 657. (10) Nagaraja, V.; Eslick, G. D. HER2 expression in gastric and oesophageal cancer: a meta-analytic review. J. Gastrointest. Oncol. 2015, 6, 143−154. (11) Omar, N.; Yan, B.; Salto-Tellez, M. HER2: An emerging biomarker in non-breast and non-gastric cancers. Pathogenesis 2015, 2, 1−9. (12) Sharifi, N.; Salmaninejad, A.; Ferdosi, S.; Bajestani, A. N.; Khaleghiyan, M.; Estiar, M. A.; Jamali, M.; Nowroozi, M. R.; Shakoori, A. HER2 gene amplification in patients with prostate cancer: Evaluating a CISH-based method. Oncol. Lett. 2016, 12, 4651−4658. (13) Abrahao-Machado, L. F.; Scapulatempo-Neto, C. HER2 testing in gastric cancer: An update. World J. Gastroenterol. 2016, 22, 4619− 4625. (14) Trail, P.; Willner, D.; Lasch, S.; Henderson, A.; Hofstead, S.; Casazza, A.; Firestone, R.; Hellstrom, I.; Hellstrom, K. Cure of xenografted human carcinomas by BR96-doxorubicin immunoconjugates. Science 1993, 261, 212−215. (15) Ledford, H. Cancer treatment: The killer within. Nature 2014, 508, 24−26. (16) Vogel, C. L.; Cobleigh, M. A.; Tripathy, D.; Gutheil, J. C.; Harris, L. N.; Fehrenbacher, L.; Slamon, D. J.; Murphy, M.; Novotny, W. F.; Burchmore, M.; Shak, S.; Stewart, S. J.; Press, M. Effcacy and Safety of Trastuzumab as a Single Agent in First- Line Treatment of HER2-Overexpressing Metastatic Breast Cancer. J. Clin. Oncol. 2002, 20, 719−726. PMID: 11821453. (17) Hudis, C. A. Trastuzumab − Mechanism of Action and Use in Clinical Practice. N. Engl. J. Med. 2007, 357, 39−51. PMID: 17611206. (18) Stern, H. M. Improving Treatment of HER2-Positive Cancers: Opportunities and Challenges. Sci. Transl. Med. 2012, 4, 127rv2− 127rv2. (19) Slamon, D. J.; Leyland-Jones, B.; Shak, S.; Fuchs, H.; Paton, V.; Bajamonde, A.; Fleming, T.; Eiermann, W.; Wolter, J.; Pegram, M.; Baselga, J.; Norton, L. Use of Chemotherapy plus a Monoclonal Antibody against HER2 for Metastatic Breast Cancer That Overexpresses HER2. N. Engl. J. Med. 2001, 344, 783−792. PMID: 11248153. (20) Valabrega, G.; Montemurro, F.; Aglietta, M. Trastuzumab: mechanism of action, resistance and future perspectives in HER2overexpressing breast cancer. Ann. Oncol. 2007, 18, 977−984.

biological properties. Blind docking calculations are conducted to discard NSAIDs that are not compatible with this specific binding site. The resulting drug list is ranked according to their MM/GBSA interaction energies. Finally, molecular dynamics simulations are performed to ensure the dynamic stability of the new dual drug. All accumulated theoretical results hint that Herceptin can be combined with Etofenamate (a clinically approved NSAID) to give a stable bifunctional drug. This conclusion is supported by three findings: (i) Etofenamate’s preferred binding pocket is not located in the antigen binding site but is further inside the host structure; (ii) the Herceptin−Etofenamate complex has the largest interaction energy (ca. −48 kcal mol−1); and (iii) the dynamic stability of the cargo-carrier system with Etofenamate is retained inside the Herceptin. Other NSAIDs commonly used, such as Celecoxib, are shown to be nonoptimal in this task, since they are located at the solvent-exposed surface of antibody and can be quickly released to the solvent, whereas Aspirin presents a low interaction energy and should also be ruled out as a good choice. Finally, we would stress that the combined use of Herceptin and NSAIDs in immunological targeting needs to be designed with caution, and it is our hope that the results reported in this work pave the way for new routes for establishing synergic effects with Herceptin in the treatment of cancer.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.7b10749. (i) Cartesian coordinates for all NSAIDs at their QM optimized structures, (ii) refined Herceptin structure, and (iii) MM/GBSA of best 100 ranked poses (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

José Pedro Cerón-Carrasco: 0000-0003-0668-9227 Horacio Pérez-Sánchez: 0000-0003-4468-7898 Alberto Requena: 0000-0002-9408-9493 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work has been partially funded by the Fundación Séneca− Agencia de Ciencia y Tecnologiá de la Región de Murcia under Projects 19419/PI/14-1 and 19419/PI/14-2, and by the Ministerio de Economiá y Competitividad under Project CTQ2016-79345-P. The MetDrugs Network (CTQ201570371-EDT) is acknowledged for providing opportunities for discussion. The research used resources from the Plataforma Andaluza de Bioinformática at the Universidad of Málaga, the CEN1 and Cabezon clusters installed at Universidad de Murcia, and the local Galileo cluster installed at Universidad Católica San Antonio de Murcia.



REFERENCES

(1) Yarden, Y.; Sliwkowski, M. X. Untangling the ErbB signalling network. Nat. Rev. Mol. Cell Biol. 2001, 2, 127−137. 2070

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B (21) Alkhateeb, A. A.; Leitzel, K.; Ali, S. M.; Campbell-Baird, C.; Evans, M.; Fuchs, E.-M.; Köstler, W. J.; Lipton, A.; Connor, J. Elevation in In ammatory Serum Biomarkers Predicts Response to Trastuzumab-Containing Therapy. PLoS One 2012, 7, e51379. (22) Chari, R. V. J.; Miller, M. L.; Widdison, W. C. Antibody-Drug Conjugates: An Emerging Concept in Cancer Therapy. Angew. Chem., Int. Ed. 2014, 53, 3796−3827. (23) Lewis Phillips, G. D.; et al. Targeting HER2-Positive Breast Cancer with Trastuzumab-DM1, an Antibody-Cytotoxic Drug Conjugate. Cancer Res. 2008, 68, 9280−9290. (24) Junttila, T. T.; Li, G.; Parsons, K.; Phillips, G. L.; Sliwkowski, M. X. Trastuzumab-DM1 (T-DM1) retains all the mechanisms of action of trastuzumab and efficiently inhibits growth of lapatinib insensitive breast cancer. Breast Cancer Res. Treat. 2011, 128, 347−356. (25) Verma, S.; Miles, D.; Gianni, L.; Krop, I. E.; Welslau, M.; Baselga, J.; Pegram, M.; Oh, D.-Y.; Diéras, V.; Guardino, E.; Fang, L.; Lu, M. W.; Olsen, S.; Blackwell, K. Trastuzumab Emtansine for HER2Positive Advanced Breast Cancer. N. Engl. J. Med. 2012, 367, 1783− 1791. PMID: 23020162. (26) Jackson, D. Y. Processes for Constructing Homogeneous Antibody Drug Conjugates. Org. Process Res. Dev. 2016, 20, 852−866. (27) Harris, L. J.; Larson, S. B.; Hasel, K. W.; McPherson, A. Refined Structure of an Intact IgG2a Monoclonal Antibody. Biochemistry 1997, 36, 1581−1597. PMID: 9048542. (28) Hamblett, K. J.; Senter, P. D.; Chace, D. F.; Sun, M. M. C.; Lenox, J.; Cerveny, C. G.; Kissler, K. M.; Bernhardt, S. X.; Kopcha, A. K.; Zabinski, R. F.; Meyer, D. L.; Francisco, J. A. Effects of Drug Loading on the Antitumor Activity of a Monoclonal Antibody Drug Conjugate. Clin. Cancer Res. 2004, 10, 7063−7070. (29) Hofer, T.; Thomas, J. D.; Burke, T. R.; Rader, C. An engineered selenocysteine defines a unique class of antibody derivatives. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 12451−12456. (30) Farias, S. E.; Strop, P.; Delaria, K.; Galindo Casas, M.; Dorywalska, M.; Shelton, D. L.; Pons, J.; Rajpal, A. Mass Spectrometric Characterization of Transglutaminase Based Site-Specific AntibodyDrug Conjugates. Bioconjugate Chem. 2014, 25, 240−250. PMID: 24359082. (31) Cerón-Carrasco, J. P.; Cerezo, J.; Requena, A.; Zúñiga, J.; Contreras-García, J.; Chavan, S.; Manrubia-Cobo, M.; Pérez-Sánchez, H. E. Labelling Herceptin with a novel oxaliplatin derivative: a computational approach towards the selective drug delivery. J. Mol. Model. 2014, 20, 2401−2409. (32) Gao, J.; Liu, Y. G.; Liu, R.; Zingaro, R. Herceptin-Platinum(II) Binding Complexes: Novel Cancer-Cell-Specific Agents. ChemMedChem 2008, 3, 954−962. (33) Dannenberg, A. J.; Lippman, S. M.; Mann, J. R.; Subbaramaiah, K.; DuBois, R. N. Cyclooxygenase-2 and Epidermal Growth Factor Receptor: Pharmacologic Targets for Chemoprevention. J. Clin. Oncol. 2005, 23, 254−266. PMID: 15637389. (34) Crawford, S. Anti-inflammatory/antioxidant use in long-term maintenance cancer therapy: a new therapeutic approach to disease progression and recurrence. Ther. Adv. Med. Oncol. 2014, 6, 52−68. (35) Higgins, M.; Chapman, J.-A.; Ingle, J.; Sledge, G.; Budd, G.; Ellis, M.; Pritchard, K.; Clemons, M.; Badovinac, C. T.; Han, L.; Gelmon, K.; Rabaglio, M.; Elliott, C.; Shepherd, L.; Goss, P. Abstract P2−13−02: Effect of aspirin (ASP) or celecoxib (CC) use on outcomes in postmenopausal breast cancer patients randomized to adjuvant exemestane or anastrozole: NCIC CTG MA.27. Cancer Res. 2012, 72, P2-13-02−P2-13-02. (36) National Center for Biotechnology Information, https:// pubchem.ncbi.nlm.nih.gov (accessed June 19, 2017). (37) Schrödinger Release 2017-2: LigPrep.; Schrödinger, LLC: New York, 2017. (38) Harder, E.; et al. OPLS3: A Force Field Providing Broad Coverage of Drug-like Small Molecules and Proteins. J. Chem. Theory Comput. 2016, 12, 281−296. (39) Schrödinger Release 2017-2: Small-Molecule Drug Discovery Suite; Schrödinger, LLC: New York, 2017.

(40) Zhao, Y.; Truhlar, D. G. Density Functionals with Broad Applicability in Chemistry. Acc. Chem. Res. 2008, 41, 157−167. (41) Singh, U. C.; Kollman, P. A. An approach to computing electrostatic charges for molecules. J. Comput. Chem. 1984, 5, 129− 145. (42) Besler, B. H.; Merz, K. M.; Kollman, P. A. Atomic charges derived from semiempirical methods. J. Comput. Chem. 1990, 11, 431− 439. (43) Bochevarov, A. D.; Harder, E.; Hughes, T. F.; Greenwood, J. R.; Braden, D. A.; Philipp, D. M.; Rinaldo, D.; Halls, M. D.; Zhang, J.; Friesner, R. A. Jaguar: A high-performance quantum chemistry software program with strengths in life and materials sciences. Int. J. Quantum Chem. 2013, 113, 2110−2142. (44) Schrödinger Release 2017-2: Jaguar; Schrödinger, LLC: New York, 2017. (45) Sussman, J. L.; Lin, D.; Jiang, J.; Manning, N. O.; Prilusky, J.; Ritter, O.; Abola, E. Protein Data Bank (PDB): database of threedimensional structural information of biological macromolecules. Acta Crystallogr., Sect. D: Biol. Crystallogr. 1998, 54, 1078−1084. (46) Cho, H.-S.; Mason, K.; Ramyar, K. X.; Stanley, A. M.; Gabelli, S. B.; Denney, D. W., Jr; Leahy, D. J. Structure of the extracellular region of HER2 alone and in complex with the Herceptin Fab. Nature 2003, 421, 756−760. (47) Scollard, D. A.; Chan, C.; Holloway, C. M.; Reilly, R. M. A kit to prepare 111In-DTPA-trastuzumab (Herceptin) Fab fragments injection under GMP conditions for imaging or radioimmunoguided surgery of HER2-positive breast cancer. Nucl. Med. Biol. 2011, 38, 129−136. (48) Hermanto, S.; Haryuni, R. D.; Ramli, M.; Mutalib, A.; Hudiyono, S. Preparation of F(ab′)2 trastuzumab fragment for radioimmunoconjugate synthesis of 177Lu-DOTA-F(ab′)2-trastuzumab. J. Pharm. 2012, 2, 12−18. (49) Madhavi Sastry, G.; Adzhigirey, M.; Day, T.; Annabhimoju, R.; Sherman, W. Protein and ligand preparation: parameters, protocols, and influence on virtual screening enrichments. J. Comput.-Aided Mol. Des. 2013, 27, 221−234. (50) Olsson, M. H. M.; Søndergaard, C. R.; Rostkowski, M.; Jensen, J. H. PROPKA3: Consistent Treatment of Internal and Surface Residues in Empirical pKa Predictions. J. Chem. Theory Comput. 2011, 7, 525−537. PMID: 26596171 (51) Halgren, T. New method for fast and accurate binding-site identification and analysis. Chem. Biol. Drug Des. 2007, 69, 146−148. (52) Halgren, T. A. Identifying and Characterizing Binding Sites and Assessing Druggability. J. Chem. Inf. Model. 2009, 49, 377−389. PMID: 19154148. (53) Schrödinger Release 2017-2: SiteMap; Schrödinger, LLC: New York, 2017. (54) Navarro-Fernández, J.; Pérez-Sánchez, H.; Martínez-Martínez, I.; Meliciani, I.; Guerrero, J.; Vicente, V.; Corral, J.; Wenzel, W. In silico discovery of a compound with nanomolar affinity to antithrombin causing partial activation and increased heparin affinity. J. Med. Chem. 2012, 55, 6403−6412. (55) Stroganov, O. V.; Novikov, F. N.; Stroylov, V. S.; Kulkov, V.; Chilov, G. G. Lead Finder: An Approach To Improve Accuracy of Protein?Ligand Docking, Binding Energy Estimation, and Virtual Screening. J. Chem. Inf. Model. 2008, 48, 2371−2385. (56) Navarro-Fernandez, J.; Pérez-Sánchez, H.; Martinez-Martinez, I.; Meliciani, I.; Guerrero, J. A.; Vicente, V.; Corral, J.; Wenzel, W. In silico discovery of a compound with nanomolar affinity to antithrombin causing partial activation and increased heparin affinity. J. Med. Chem. 2012, 55, 6403−6412. (57) Kollman, P. A.; Massova, I.; Reyes, C.; Kuhn, B.; Huo, S.; Chong, L.; Lee, M.; Lee, T.; Duan, Y.; Wang, W.; Donini, O.; Cieplak, P.; Srinivasan, J.; Case, D. A.; Cheatham, T. E. Calculating Structures and Free Energies of Complex Molecules:? Combining Molecular Mechanics and Continuum Models. Acc. Chem. Res. 2000, 33, 889− 897. PMID: 11123888. (58) Greenidge, P. A.; Kramer, C.; Mozziconacci, J.-C.; Wolf, R. M. MM/GBSA Binding Energy Prediction on the PDBbind Data Set: 2071

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072

Article

The Journal of Physical Chemistry B Successes, Failures, and Directions for Further Improvement. J. Chem. Inf. Model. 2013, 53, 201−209. PMID: 23268595. (59) Nosé, S. A Molecular Dynamics Method for Simulations in the Canonical Ensemble. Mol. Phys. 1984, 52, 255−268. (60) Hoover, W. G. Canonical Cynamics: Equilibrium Phase-Space Distributions. Phys. Rev. A: At., Mol., Opt. Phys. 1985, 31, 1695−1697. (61) Martyna, G. J.; Tobias, D. J.; Klein, M. L. Constant Pressure Molecular Dynamics Algorithms. J. Chem. Phys. 1994, 101, 4177− 4189. (62) Tuckerman, M.; Berne, B. J.; Martyna, G. J. Reversible Multiple Time Scale Molecular Dynamics. J. Chem. Phys. 1992, 97, 1990−2001. (63) Darden, T. A.; York, D. M.; Pedersen, L. G. Particle Mesh Ewald: An N-log(N) Method for Ewald Sums in Large Systems. J. Chem. Phys. 1993, 98, 10089−10092. (64) Desmond Molecular Dynamics System, version 4.4; D. E. Shaw Research: New York, 2015. Maestro-Desmond Interoperability Tools, version 4.4; Schrödinger: New York, 2015. (65) Dang, C. T.; Dannenberg, A. J.; Subbaramaiah, K.; Dickler, M. N.; Moasser, M. M.; Seidman, A. D.; D’Andrea, G. M.; Theodoulou, M.; Panageas, K. S.; Norton, L.; Hudis, C. A. Phase II Study of Celecoxib and Trastuzumab in Metastatic Breast Cancer Patients Who Have Progressed after Prior Trastuzumab-Based Treatments. Clin. Cancer Res. 2004, 10, 4062−4067. (66) Ciordia, M.; Pérez-Benito, L.; Delgado, F.; Trabanco, A. A.; Tresadern, G. Application of Free Energy Perturbation for the Design of BACE1 Inhibitors. J. Chem. Inf. Model. 2016, 56, 1856−1871. (67) Yadav, A.; Sharma, S.; Yadav, V. K. Non-covalent carriage of anticancer agents by humanized antibody trastuzumab. J. Mol. Model. 2016, 22, 112−126.

2072

DOI: 10.1021/acs.jpcb.7b10749 J. Phys. Chem. B 2018, 122, 2064−2072