Search

There is an ongoing effort to develop better methods for noninvasive detection and characterization of thrombi. Here we describe the synthesis and evaluation of three new fibrin-targeted positron emission tomography (PET) probes (FBP1, FBP2, FBP3). Three fibrin-specific peptides were conjugated as 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA)-monoamides at the C- and N-termini and chelated with (64)CuCl2. Probes were prepared with a specific activity ranging from 10 to 130 μCi/nmol. Both the peptides and the probes exhibited nanomolar dissociation constants (Kd) for the soluble fibrin fragment DD(E), although the Cu-DOTA derivatization resulted in a 2-3 fold loss in affinity relative to the parent peptide. Biodistribution and imaging studies were performed in a rat model of carotid artery thrombosis. For FBP1 and FBP2 at 120 min post injection, the vessel containing the thrombus showed the highest concentration of radioactivity after the excretory organs, that is, the liver and kidneys. This was confirmed ex vivo by autoradiography, which showed >4-fold activity in the thrombus-containing artery compared to the contralateral artery. FBP3 showed much lower thrombus uptake, and the difference was traced to greater metabolism of this probe. Hybrid MR-PET imaging with FBP1 or FBP2 confirmed that these probes were effective for the detection of an arterial thrombus in this rat model. A thrombus was visible on PET images as a region of high activity that corresponded to a region of arterial occlusion identified by simultaneous MR angiography. FBP1 and FBP2 represent promising new probes for the molecular imaging of thrombi.

Using a well-defined set of fragments/pharmacophores, a new methodology to calculate fragment/ pharmacophore descriptors for any molecule onto which at least one fragment/pharmacophore can be mapped is presented. To each fragment/pharmacophore present in a molecule, we attach a descriptor that is calculated by identifying the molecule's atoms onto which it maps and summing over its constituent atomic descriptors. The attached descriptors are named C-fragment/pharmacophore descriptors, and this methodology can be applied to any descriptors defined at the atomic level, such as the partition coefficient, molar refractivity, electrotopological state, etc. By using this methodology, the same fragment/pharmacophore can be shown to have different values in different molecules resulting in better discrimination power. As we know, fragment and pharmacophore fingerprints have a lot of applications in chemical informatics. This study has attempted to find the impact of replacing the traditional value of "1" in a fingerprint with real numbers derived form C-fragment/pharmacophore descriptors. One way to do this is to assess the utility of C-fragment/ pharmacophore descriptors in modeling different end points. Here, we exemplify with data from CYP and hERG. The fact that, in many cases, the obtained models were fairly successful and C-fragment descriptors were ranked among the top ones supports the idea that they play an important role in correlation. When we modeled hERG with C-pharmacophore descriptors, however, the model performances decreased slightly, and we attribute this, mainly to the fact that there is no technique capable of handling multiple instances (states). We hope this will open new research, especially in the emerging field of machine learning. Further research is needed to see the impact of C-fragment/pharmacophore descriptors in similarity/dissimilarity applications.

The ability to accurately predict biological affinity on the basis of in silico docking to a protein target remains a challenging goal in the CADD arena. Typically, "standard" scoring functions have been employed that use the calculated docking result and a set of empirical parameters to calculate a predicted binding affinity. To improve on this, we are exploring novel strategies for rapidly developing and tuning "customized" scoring functions tailored to a specific need. In the present work, three such customized scoring functions were developed using a set of 129 high-resolution protein-ligand crystal structures with measured Ki values. The functions were parametrized using N-PLS (N-way partial least squares), a multivariate technique well-known in the 3D quantitative structure-activity relationship field. A modest correlation between observed and calculated pKi values using a standard scoring function (r2 = 0.5) could be improved to 0.8 when a customized scoring function was applied. To mimic a more realistic scenario, a second scoring function was developed, not based on crystal structures but exclusively on several binding poses generated with the Flo+ docking program. Finally, a validation study was conducted by generating a third scoring function with 99 randomly selected complexes from the 129 as a training set and predicting pKi values for a test set that comprised the remaining 30 complexes. Training and test set r2 values were 0.77 and 0.78, respectively. These results indicate that, even without direct structural information, predictive customized scoring functions can be developed using N-PLS, and this approach holds significant potential as a general procedure for predicting binding affinity on the basis of in silico docking.

Potent and selective Aurora kinase inhibitors were identified from the combinatorial expansion of the 1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazole bi-cycle, a novel and versatile scaffold designed to target the ATP pocket of protein kinases. The most potent compound reported in this study had an IC(50) of 0.027 microM in the enzymatic assay for Aur-A inhibition and IC(50)s between 0.05 microM and 0.5 microM for the inhibition of proliferation of different tumor cell lines.

Solubility data for 930 diverse compounds have been analyzed using linear Partial Least Square (PLS) and nonlinear PLS methods, Continuum Regression (CR), and Neural Networks (NN). 1D and 2D descriptors from MOE package in combination with E-state or ISIS keys have been used. The best model was obtained using linear PLS for a combination between 22 MOE descriptors and 65 ISIS keys. It has a correlation coefficient (r2) of 0.935 and a root-mean-square error (RMSE) of 0.468 log molar solubility (log S(w)). The model validated on a test set of 177 compounds not included in the training set has r2 0.911 and RMSE 0.475 log S(w). The descriptors were ranked according to their importance, and at the top of the list have been found the 22 MOE descriptors. The CR model produced results as good as PLS, and because of the way in which cross-validation has been done it is expected to be a valuable tool in prediction besides PLS model. The statistics obtained using nonlinear methods did not surpass those got with linear ones. The good statistic obtained for linear PLS and CR recommends these models to be used in prediction when it is difficult or impossible to make experimental measurements, for virtual screening, combinatorial library design, and efficient leads optimization.

Novel scoring functions that predict the affinity of a ligand for its receptor have been developed. They were built with several statistical tools (partial least squares, genetic algorithms, neural networks) and trained on a data set of 100 crystal structures of receptor-ligand complexes, with affinities spanning 10 log units. The new scoring functions contain both descriptors generated by the QXP docking program and new descriptors that were developed in-house. These new descriptors are based on solvent accessible surface areas and account for conformational entropy changes and desolvation effects of both ligand and receptor upon binding. The predictive r(2) values for a test set of 24 complexes are in the 0.712-0.741 range and RMS prediction errors in the 1.09-1.16 log K(d) range. Inclusion of the new descriptors led to significant improvements in affinity prediction, compared to scoring functions based on QXP descriptors alone. However, the QXP descriptors by themselves perform better in binding mode prediction. The performance of the linear models is comparable to that of the neural networks. The new functions perform very well, but they still need to be validated as universal tools for the prediction of binding affinity.

Molecular recognition is the basis of rational drug design, and for this reason it has been extensively studied. However, the process by which a ligand recognizes and binds to its receptor is complex and not well understood. For the case in which the geometries (conformation and alignment) of the ligand and receptor are known from X-ray crystal structure data, the problem is simplified. The receptor-bound conformation and alignment of the ligand is assumed, and those of additional ligands are inferred. For the general case in which the geometries of the ligand(s) and receptor are unknown, no general treatment or solution is available and receptor-ligand geometries must be obtained indirectly from structure-activity studies or synthesis and evaluation of rigid analogs. A general treatment for solving for the receptor-bound geometry of a series of ligands is presented here. Using molecular shape analysis, for ligand description, tensor analysis of N-way arrays by partial least-squares (PLS) regression, and 3-way factor analysis, the receptor-bound geometries of trimethoprim and a series of trimethoprim-like dihydrofolate reductase inhibitors are correctly predicted.