Computational Methods for Anticancer Drug Discovery; The MCT4 Paradigm.

Modern anticancer research has employed advanced computational techniques and artificial intelligence methods for drug discovery and development, along with the massive amount of generated clinical and in silico data over the last decades. Diverse computational techniques and state-of-the-art algorithms are being developed to enhance traditional Rational Drug Design pipelines and achieve cost-efficient and successful anticancer candidates to promote human health. Towards this direction, we have developed a pharmacophore- based drug design approach against MCT4, a member of the monocarboxylate transporter family (MCT), which is the main carrier of lactate across the membrane and highly involved in cancer cell metabolism. Specifically, MCT4 is a promising target for therapeutic strategies as it overexpresses in glycolytic tumors, and its inhibition has shown promising anticancer effects. Due to the lack of experimentally determined structure, we have elucidated the key features of the protein through an in silico drug design strategy, including for molecular modelling, molecular dynamics, and pharmacophore elucidation, towards the identification of specific inhibitors as a novel anti-cancer strategy.

A Holistic Evolutionary and 3D Pharmacophore Modelling Study Provides Insights into the Metabolism, Function, and Substrate Selectivity of the Human Monocarboxylate Transporter 4 (hMCT4).

Monocarboxylate transporters (MCTs) are of great research interest for their role in cancer cell metabolism and their potential ability to transport pharmacologically relevant compounds across the membrane. Each member of the MCT family could potentially provide novel therapeutic approaches to various diseases. The major differences among MCTs are related to each of their specific metabolic roles, their relative substrate and inhibitor affinities, the regulation of their expression, their intracellular localization, and their tissue distribution. MCT4 is the main mediator for the efflux of L-lactate produced in the cell. Thus, MCT4 maintains the glycolytic phenotype of the cancer cell by supplying the molecular resources for tumor cell proliferation and promotes the acidification of the extracellular microenvironment from the co-transport of protons. A promising therapeutic strategy in anti-cancer drug design is the selective inhibition of MCT4 for the glycolytic suppression of solid tumors. A small number of studies indicate molecules for dual inhibition of MCT1 and MCT4; however, no selective inhibitor with high-affinity for MCT4 has been identified. In this study, we attempt to approach the structural characteristics of MCT4 through an in silico pipeline for molecular modelling and pharmacophore elucidation towards the identification of specific inhibitors as a novel anti-cancer strategy.

Mapping Interactome Networks of DNAJC11, a Novel Mitochondrial Protein Causing Neuromuscular Pathology in Mice.

We previously identified DNAJC11, a mitochondrial outer membrane protein of unknown function, as a novel genetic cause in modeled neuromuscular disease. To understand the physiological role of DNAJC11, we employed a proteomic approach for the identification of the DNAJC11 interactome, through the expression of DNAJC11-FLAG in HEK293FT cells and transgenic mice. Our analysis confirmed known DNAJC11-interacting proteins including members of the MICOS complex that organize mitochondrial cristae formation. Moreover, we identified in both biological systems novel mitochondrial interactions including VDACs that exchange metabolites across the outer mitochondrial membrane. In HEK293FT cells, DNAJC11 preferentially interacted with ribosomal subunits and chaperone proteins including Hsp70 members, possibly correlating DNAJC11 with cotranslational folding and import of mitochondrial proteins in metabolically active cells. Instead, the DNAJC11 interactome in the mouse cerebrum was enriched for synaptic proteins, supporting the importance of DNAJC11 in synapse and neuronal integrity. Moreover, we demonstrated that the DUF3395 domain is critically involved in DNAJC11 protein-protein interactions, while the J-domain determines its mitochondrial localization. Collectively, these results provide a functional characterization for DNAJC11 domains, while the identified interactome networks reveal an emerging role of DNAJC11 in mitochondrial biogenesis and response to microenvironment changes and requirements.

Glutathione analogues as substrates or inhibitors that discriminate between allozymes of the MDR-involved human glutathione transferase P1-1.

Glutathione (GSH) structure-guided tripeptide analogues were designed and synthesized by solid phase technology, purified (≥95%) by RP and/or GF column chromatography, to identify those that, compared with GSH, exhibited similar or higher binding and catalytic efficiency toward the MDR-involved human GSTP1-1 isoenzyme, and could discriminate between the allozymic expression products of the polymorphic human GSTP1 gene locus, designated as hGSTP1*A (Ile(104) /Ala(113) ), hGSTP1*B (Val(104) /Ala(113) ), and hGSTP1*C (Val(104) /Val(113) ). The analogues bear single amino acid alterations as well as alterations in more than one position. Some analogues showed remarkable allozyme selectivity, binding catalytically to A (I, II, IV, XII), to C (V and XVI), to A and C (III, VII, XIV) or to all three allozymes (XV). A heterocyclic substituent at positions 1 or 2 of GSH favors inhibition of A, whereas a small hydrophobic/hydrophilic amide substituent at position 2 (Cys) favors inhibition of B and C. Heterocyclic substituents at position 1, only, produce catalytic analogues for A, whereas less bulky and more flexible hydrophobic/hydrophilic substituents, at positions 1 or 3, lead to effective substrates with C. When such substituents were introduced simultaneously at positions 1 and 3, the analogues produced have no catalytic potential but showed appreciable inhibitory effects, instead, with all allozymes. It is anticipated that when GSH analogues with selective inhibitory or catalytic binding, were conjugated to allozyme-selective inhibitors of hGSTP1-1, the derived leads would be useful for the designing of novel chimeric inhibitors against the MDR-involved hGSTP1-1 allozymes. © 2016 Wiley Periodicals, Inc. Biopolymers (Pept Sci) 106: 330-344, 2016.

Crystal and solution studies of the "Plus-C" odorant-binding protein 48 from Anopheles gambiae: control of binding specificity through three-dimensional domain swapping.

Much physiological and behavioral evidence has been provided suggesting that insect odorant-binding proteins (OBPs) are indispensable for odorant recognition and thus are appealing targets for structure-based discovery and design of novel host-seeking disruptors. Despite the fact that more than 60 putative OBP-encoding genes have been identified in the malaria vector Anopheles gambiae, the crystal structures of only six of them are known. It is therefore clear that OBP structure determination constitutes the bottleneck for structure-based approaches to mosquito repellent/attractant discovery. Here, we describe the three-dimensional structure of an A. gambiae "Plus-C" group OBP (AgamOBP48), which exhibits the second highest expression levels in female antennae. This structure represents the first example of a three-dimensional domain-swapped dimer in dipteran species. A combined binding site is formed at the dimer interface by equal contribution of each monomer. Structural comparisons with the monomeric AgamOBP47 revealed that the major structural difference between the two Plus-C proteins localizes in their N- and C-terminal regions, and their concerted conformational change may account for monomer-swapped dimer conversion and furthermore the formation of novel binding pockets. Using a combination of gel filtration chromatography, differential scanning calorimetry, and analytical ultracentrifugation, we demonstrate the AgamOBP48 dimerization in solution. Eventually, molecular modeling calculations were used to predict the binding mode of the most potent synthetic ligand of AgamOBP48 known so far, discovered by ligand- and structure-based virtual screening. The structure-aided identification of multiple OBP binders represents a powerful tool to be employed in the effort to control transmission of the vector-borne diseases.

2,2'-Dihydroxybenzophenones and their carbonyl N-analogues as inhibitor scaffolds for MDR-involved human glutathione transferase isoenzyme A1-1.

The MDR-involved human GSTA1-1, an important isoenzyme overexpressed in several tumors leading to chemotherapeutic-resistant tumour cells, has been targeted by 2,2'-dihydroxybenzophenones and some of their carbonyl N-analogues, as its potential inhibitors. A structure-based library of the latter was built-up by a nucleophilic cleavage of suitably substituted xanthones to 2,2'-dihydroxy-benzophenones (5-9) and subsequent formation of their N-derivatives (oximes 11-13 and N-acyl hydrazones 14-16). Screening against hGSTA1-1 led to benzophenones 6 and 8, and hydrazones 14 and 16, having the highest inhibition potency (IC50 values in the range 0.18 ± 0.02 to 1.77 ± 0.10 µM). Enzyme inhibition kinetics, molecular modeling and docking studies showed that they interact primarily at the CDNB-binding catalytic site of the enzyme. In addition, the results from cytotoxicity studies with human colon adenocarcinoma cells showed low LC50 values for benzophenone 6 and its N-acyl hydrazone analogue 14 (31.4 ± 0.4 µM and 87 ± 1.9 µM, respectively), in addition to the strong enzyme inhibition profile (IC50(6)=1,77 ± 0.10 µM; IC50(14)=0.33 ± 0.05 µM). These structures may serve as leads for the design of new potent mono- and bi-functional inhibitors and pro-drugs against human GTSs.

CrystTwiV: a webserver for automated phase extension and refinement in X-ray crystallography.

An important stage in macromolecular crystallography is that of phase extension and refinement when initial phase estimates are available from isomorphous replacement or anomalous scattering or other methods. For this purpose, an alternative method called the twin variables (TwiV) method has been proposed. The algorithm is based on alternately transferring the phase information between the twin variable sets. The phase extension and refinement is evaluated with the crystallographic symmetry test by deliberately sacrificing the space-group symmetry in the starting set, then using its re-appearance as a criterion for correctness. Here we present a software program (CrysTwiV) that runs on the web (freely available at: http://btweb.aua.gr/crystwiv/) implementing the above-mentioned method.

New Insights for RANKL as a Proinflammatory Modulator in Modeled Inflammatory Arthritis.

Receptor activator of nuclear factor-κB ligand (RANKL), a member of the Tumor Necrosis Factor (TNF) superfamily, constitutes the master regulator of osteoclast formation and bone resorption, whereas its involvement in inflammatory diseases remains unclear. Here, we used the human TNF transgenic mouse model of erosive inflammatory arthritis to determine if the progression of inflammation is affected by either genetic inactivation or overexpression of RANKL in transgenic mouse models. TNF-mediated inflammatory arthritis was significantly attenuated in the absence of functional RANKL. Notably, TNF overexpression could not compensate for RANKL-mediated osteopetrosis, but promoted osteoclastogenesis between the pannus and bone interface, suggesting RANKL-independent mechanisms of osteoclastogenesis in inflamed joints. On the other hand, simultaneous overexpression of RANKL and TNF in double transgenic mice accelerated disease onset and led to severe arthritis characterized by significantly elevated clinical and histological scores as shown by aggressive pannus formation, extended bone resorption, and massive accumulation of inflammatory cells, mainly of myeloid origin. RANKL and TNF cooperated not only in local bone loss identified in the inflamed calcaneous bone, but also systemically in distal femurs as shown by microCT analysis. Proteomic analysis in inflamed ankles from double transgenic mice overexpressing human TNF and RANKL showed an abundance of proteins involved in osteoclastogenesis, pro-inflammatory processes, gene expression regulation, and cell proliferation, while proteins participating in basic metabolic processes were downregulated compared to TNF and RANKL single transgenic mice. Collectively, these results suggest that RANKL modulates modeled inflammatory arthritis not only as a mediator of osteoclastogenesis and bone resorption but also as a disease modifier affecting inflammation and immune activation.

Thermodynamic, crystallographic and computational studies of non-mammalian fatty acid binding to bovine ß-Lactoglobulin.

The milk protein ß-lactoglobulin has been widely studied since its discovery, both as a purified protein and in mixtures with other milk proteins, where its effect on the processing properties is of importance to the dairy industry. The protein can bind a variety of small hydrophobic molecules, which may allow its use as an oral delivery vehicle. In the present study we have examined the binding of odd-numbered fatty acids by isothermal calorimetry (ITC), X-ray crystallography and computer modelling to provide a clearer picture of the extent and variability of the central binding pocket. The Kd values for the fatty acids C13, C15, C16, C17 and C19 as determined by ITC are 1.93, 2.91, 3.05, 4.11 and 8.67 × 10-7 M, respectively. The molecular structures revealed the ligands bound in the central cavity with generally well ordered lipophilic tails but significant positional variation at the carboxyl group end. In silico docking analyses identified the lipophilic interactions within the central cavity as the main driving force for binding with electrostatic interactions and H-bonds playing a minor role.

Identification of novel bioinspired synthetic mosquito repellents by combined ligand-based screening and OBP-structure-based molecular docking.

In this work we report a fast and efficient virtual screening protocol for discovery of novel bioinspired synthetic mosquito repellents with lower volatility and, in all likelihood, increased protection time as compared with their plant-derived parental compounds. Our screening protocol comprises two filtering steps. The first filter is based on the shape and chemical similarity to known plant-derived repellents, whereas the second filter is based on the predicted similarity of the ligand's binding mode to the Anopheles gambiae odorant binding protein (AgamOBP1) relative to that of DEET and Icaridin to the same OBP. Using this protocol, a chemical library containing 42,755 synthetic molecules was screened in silico and sixteen selected compounds were tested for their affinity to AgamOBP1 in vitro and repellence against A. gambiae female mosquitoes using a warm-body repellent assay. One of them showed DEET-like repellence (91%) but with significantly lower volatility (2.84 × 10-6 mmHg) than either DEET (1.35 × 10-3 mmHg) or its parental cuminic acid (3.08 × 10-3 mmHg), and four other compounds were found to exhibit repellent indices between 69 and 79%. Overall, a correlation was not evident between repellence and OBP-binding strength. In contrast, a correlation between binding mode and repellence was found.

Bidirectional Long Short-Term Memory Networks for predicting the subcellular localization of eukaryotic proteins.

An algorithm called Bidirectional Long Short-Term Memory Networks (BLSTM) for processing sequential data is introduced. This supervised learning method trains a special recurrent neural network to use very long ranged symmetric sequence context using a combination of nonlinear processing elements and linear feedback loops for storing long-range context. The algorithm is applied to the sequence-based prediction of protein localization and predicts 93.3 percent novel non-plant proteins and 88.4 percent novel plant proteins correctly, which is an improvement over feedforward and standard recurrent networks solving the same problem. The BLSTM system is available as a web-service (http://www.stepc.gr/~synaptic/blstm.html).

Analysis and interpretation of dynamic FDG PET oncological studies using data reduction techniques.

BACKGROUND: Dynamic positron emission tomography studies produce a large amount of image data, from which clinically useful parametric information can be extracted using tracer kinetic methods. Data reduction methods can facilitate the initial interpretation and visual analysis of these large image sequences and at the same time can preserve important information and allow for basic feature characterization. METHODS: We have applied principal component analysis to provide high-contrast parametric image sets of lower dimensions than the original data set separating structures based on their kinetic characteristics. Our method has the potential to constitute an alternative quantification method, independent of any kinetic model, and is particularly useful when the retrieval of the arterial input function is complicated. In independent component analysis images, structures that have different kinetic characteristics are assigned opposite values, and are readily discriminated. Furthermore, novel similarity mapping techniques are proposed, which can summarize in a single image the temporal properties of the entire image sequence according to a reference region. RESULTS: Using our new cubed sum coefficient similarity measure, we have shown that structures with similar time activity curves can be identified, thus facilitating the detection of lesions that are not easily discriminated using the conventional method employing standardized uptake values.

A survey of the availability of primary bioinformatics web resources.

The explosive growth of the bioinformatics field has led to a large amount of data and software applications publicly available as web resources. However, the lack of persistence of web references is a barrier to a comprehensive shared access. We conducted a study of the current availability and other features of primary bioinformatics web resources (such as software tools and databases). The majority (95%) of the examined bioinformatics web resources were found running on UNIX/Linux operating systems, and the most widely used web server was found to be Apache (or Apache-related products). Of the overall 1,130 Uniform Resource Locators (URLs) examined, 91% were highly available (more than 90% of the time), while only 4% showed low accessibility (less than 50% of the time) during the survey. Furthermore, the most common URL failure modes are presented and analyzed.

Concluding the trilogy: The interaction of 2,2'-dihydroxy-benzophenones and their carbonyl N-analogues with human glutathione transferase M1-1 face to face with the P1-1 and A1-1 isoenzymes involved in MDR.

A series of 2,2'-dihydroxybenzophenones and their carbonyl N-analogues were studied as potential inhibitors against human glutathione transferase M1-1 (hGSTM1-1) purified from recombinant E. coli. Their screening revealed an inhibition against hGSTM1-1 within a range of 0-42% (25 µM). The IC50 values for the two stronger ones, 16 and 13, were 53.5 ± 5.6 µΜ and 28.5 ± 2.5 µΜ, respectively. The results were compared with earlier ones for isoenzymes hGSTP1-1 and hGSTA1-1 involved in MDR. All but one bind more strongly to A1-1, than M1-1 and P1-1, the latter being a poor binder. An order of potency A1-1 > > M1-1 >  P1-1 meritted 13, 14 and 16 as the most potent inhibitors with hGSTM1-1. Enzyme kinetics with hGSTM1-1 (Km(CDNB) 213 ± 10 µΜ and Km(GSH) 303 ± 11 µΜ) revealed a competitive modality for 16 (Ki(16)  = 22.3 ± 1.1 µΜ) and a mixed one for 13 versus CDNB (Ki(13)  = 33.3 ± 1.6 µM for the free enzyme and Ki(13) ' = 17.7 ± 1.7 µM for the enzyme-CDNB complex). 5- or 5'-Bromo- or phenyl-substituted (but not in combination) inhibitors, having a H-bonded oxime weakly acidic group of a small volume, are optimal candidates for binding hGSTM1-1. The outcome of the isoenzyme trilogy identified good binder leads for the investigated GSTs involved in MDR.

MetaBasis: a web-based database containing metadata on software tools and databases in the field of bioinformatics.

UNLABELLED: We have developed an integrated web-based relational database information system, which offers an extensive search functionality of validated entries containing available bioinformatics computing resources. This system, called MetaBasis, aims to provide the bioinformatics community, and especially newcomers to the field, with easy access to reliable bioinformatics databases and tools. MetaBasis is focused on non-commercial and open-source software tools. AVAILABILITY: http://metabasis.bioacademy.gr/

Blind source separation for the computational analysis of dynamic oncological PET studies.

The analysis of dynamic positron emission tomography (PET) studies provides clinically useful parametric information, but often requires complex and time-consuming compartmental or non-compartmental techniques. Independent component analysis (ICA), a statistical method used for feature extraction and signal separation, is applied to dynamic PET studies to facilitate the initial interpretation and visual analysis of these large image sequences. ICA produces parametric images, where structures with different kinetic characteristics are assigned opposite values and readily discriminated, improving the identification of lesions and facilitating the posterior detailed kinetic analysis.

Isoenzyme- and allozyme-specific inhibitors: 2,2'-dihydroxybenzophenones and their carbonyl N-analogues that discriminate between human glutathione transferase A1-1 and P1-1 allozymes.

The selectivity of certain benzophenones and their carbonyl N-analogues was investigated towards the human GSTP1-1 allozymes A, B and C involved in MDR. The allozymes were purified from extracts derived from E. coli harbouring the plasmids pEXP5-CT/TOPO-TA-hGSTP1*A, pOXO4-hGSTP1*B or pOXO4-hGSTP1*C. Compound screening with each allozyme activity indicated three compounds with appreciable inhibitory potencies, 12 and 13 with P1-1A 62% and 67%, 11 and 12 with P1-1C 51% and 70%, whereas that of 15 fell behind with P1-1B (41%). These findings were confirmed by IC50 values (74-125 µm). Enzyme inhibition kinetics, aided by molecular modelling and docking, revealed that there is competition with the substrate CDNB for the same binding site on the allozyme (Ki(13/A)  = 63.6 ± 3.0 µm, Ki(15/B)  = 198.6 ± 14.3 µm, and Ki(11/C)  = 16.5 ± 2.7 µm). These data were brought into context by an in silico structural comparative analysis of the targeted proteins. Although the screened compounds showed moderate inhibitory potency against hGSTP1-1, remarkably, some of them demonstrated absolute isoenzyme and/or allozyme selectivity.

Iterative image reconstruction for clinical PET using ordered subsets, median root prior, and a web-based interface.

PURPOSE: The development, implementation and validation of simple, flexible and efficient iterative image reconstruction (IIR) methods for their take-up in routine clinical positron emission tomography (PET) static or dynamic studies. PROCEDURES: The ordered subsets (OS) technique applied for the acceleration of the maximum likelihood expectation maximization (MLEM) IIR algorithm is here extended to include the weighted least-squares (WLS), image space reconstruction algorithm (ISRA) and the space alternating generalized EM (SAGE). The median root prior (MRP) has been successfully applied as a Bayesian regularization to control the noise level in the reconstructed images. All methods are implemented on distributed Pentium systems and tested using simulated PET data from a brain phantom. A Javascript is used for the initiation of the reconstruction. RESULTS: Taking into consideration the image quality and the time required for the reconstruction, the MRP-OSEM (ordered subsets expectation maximization) seems to provide best results after four to eight iterations, with four subsets and a MRP coefficient of 0.2-0.4. Iterative reconstruction of the transmission images with OS-acceleration and MRP regularization with subsequent calculation of the attenuation correction factors (ACFs) is shown to effectively remove streak artifacts in the emission images, especially along paths of high attenuation. CONCLUSIONS: An efficient implementation using distributed processing principles and a web-based interface allows the reconstruction of one frame (with 63 cross-section slices) from a dynamic determination in few minutes. This work showed that regular PC systems can provide fast execution and produce results in clinically meaningful times. This eradicates the argument of the computational burden of the method that prevented the extensive use of IIR in today's modern PET systems.

Performance evaluation of principal component analysis in dynamic FDG-PET studies of recurrent colorectal cancer.

Performance evaluation of principal component analysis (PCA) of dynamic F-18-FDG-PET studies of patients with recurrent colorectal cancer. Principal component images (PCI) of 17 iteratively reconstructed data sets were visually and quantitatively evaluated. The F-18-FDG compartment model parameters were estimated using polynomial regression. All structures were present in PCI1. PCI2 was correlated with the vascular component and PCI3 with the tumor. The vessel density in the tumor was estimated with a correlation coefficient equal to 0.834. PCA supports the visual interpretation of dynamic F-18-FDG-PET studies, facilitates the application of compartment modeling and is a promising quantification technique.

Designer xanthone: an inhibitor scaffold for MDR-involved human glutathione transferase isoenzyme A1-1.

Glutathione transferases (GSTs) are cell detoxifiers involved in multiple drug resistance (MDR), hampering the effectiveness of certain anticancer drugs. To our knowledge, this is the first report on well-defined synthetic xanthones as GST inhibitors. Screening 18 xanthones revealed three derivatives bearing a bromomethyl and a methyl group (7) or two bromomethyl groups (8) or an aldehyde group (17), with high inhibition potency (>85%), manifested by low IC(50) values (7: 1.59 ± 0.25 µM, 8: 5.30 ± 0.30 µM, and 17: 8.56 ± 0.14 µM) and a competitive modality of inhibition versus CDNB (Ki(7) = 0.76 ± 0.18 and Ki(17) = 1.69 ± 0.08 µM). Of them, derivative 17 readily inhibited hGSTA1-1 in colon cancer cell lysate (IC(50) = 10.54 ± 2.41 µM). Furthermore, all three derivatives were cytotoxic to Caco-2 intact cells, with 17 being the least cytotoxic (LC(50) = 151.3 ± 16.3 µM). The xanthone scaffold may be regarded as a pharmacophore for hGSTA1-1 and the three derivatives, especially 17, as potent precursors for the synthesis of new inhibitors and conjugate prodrugs for human GSTs.