Characterization of the Immune Microenvironment in Inflammatory Breast Cancer Using Multiplex Immunofluorescence.

INTRODUCTION: Inflammatory breast cancer (IBC) is an aggressive form of breast cancer with a poorly characterized immune microenvironment. METHODS: We used a five-colour multiplex immunofluorescence panel, including CD68, CD4, CD8, CD20, and FOXP3 for immune microenvironment profiling in 93 treatment-naïve IBC samples. RESULTS: Lower grade tumours were characterized by decreased CD4+ cells but increased accumulation of FOXP3+ cells. Increased CD20+ cells correlated with better response to neoadjuvant chemotherapy and increased CD4+ cells infiltration correlated with better overall survival. Pairwise analysis revealed that both ER+ and triple-negative breast cancer were characterized by co-infiltration of CD20 + cells with CD68+ and CD4+ cells, whereas co-infiltration of CD8+ and CD68+ cells was only observed in HER2+ IBC. Co-infiltration of CD20+, CD8+, CD4+, and FOXP3+ cells, and co-existence of CD68+ with FOXP3+ cells correlated with better therapeutic responses, while resistant tumours were characterized by co-accumulation of CD4+, CD8+, FOXP3+, and CD68+ cells and co-expression of CD68+ and CD20+ cells. In a Cox regression model, response to therapy was the most significant factor associated with improved patient survival. CONCLUSION: Those results reveal a complex unique pattern of distribution of immune cell subtypes in IBC and provide an important basis for detailed characterization of molecular pathways that govern the formation of IBC immune landscape and potential for immunotherapy.

DockIT: a tool for interactive molecular docking and molecular complex construction.

SUMMARY: DockIT is a tool that has a unique set of physical and graphical features for interactive molecular docking. It enables the user to bring a ligand and a receptor into a docking pose by controlling relative position and orientation, either with a mouse and keyboard, or with a haptic device. Atomic interactions are modelled using molecular dynamics-based force-fields with the force on the ligand being felt on a haptic device. Real-time calculation and display of intermolecular hydrogen bonds and multipoint collision detection either using maximum force or maximum atomic overlap, mean that together with the ability to monitor selected intermolecular atomic distances, the user can find physically feasible docking poses that satisfy distance constraints derived from experimental methods. With these features and the ability to output and reload docked structures it can be used to accurately build up large multi-component molecular systems in preparation for molecular dynamics simulation. AVAILABILITY AND IMPLEMENTATION: DockIT is available free of charge for non-commercial use at http://www.haptimol.co.uk/downloads.htm. It requires a windows computer with GPU that supports OpenCL 1.2 and OpenGL 4.0. It may be used with a mouse and keyboard, or a haptic device from 3DSystems.

Interactive Flexible-Receptor Molecular Docking in Virtual Reality Using DockIT.

Interactive docking enables the user to guide and control the docking of two biomolecules into a binding pose. It is of particular use when the binding site is known and is thought to be applicable to structure-based drug design (SBDD) and educating students about biomolecular interactions. For SBDD, it enables expertise and intuition to be brought to bear in the drug design process. In education, it can teach students about the most basic level of biomolecular function. Here, we introduce DockIT for virtual reality (VR) that uses a VR headset and hand-held controllers. Using the method of linear response on explicit solvent molecular dynamics simulations, DockIT can model both global and local conformational changes within the receptor due to forces of interaction with the ligand. It has real-time flexible molecular surface rendering and can show the real-time formation and breaking of hydrogen bonds, both between the ligand and receptor and within the receptor itself as it smoothly changes conformation.

The role of the half-turn in determining structures of Alzheimer's Aß wild-type and mutants.

Half-turns are shown to be the main determinants of many experimental Alzheimer's Aß fibril structures. Fibril structures contain three half-turn types, ßαRß, ßαLß and ß뵧 which each result in a ∼90° bend in a ß-strand. It is shown that only these half-turns enable cross-ß stacking and thus the right-angle fold seen in fibrils is an intrinsic feature of cross-ß. Encoding a strand as a conformational sequence in ß, αR, αL and ε(ßL), pairwise combination rules for consecutive half-turns are used to decode this sequence to give the backbone path. This reveals how structures would be dramatically affected by a deletion. Using a wild-type Aß(42) fibril structure and the pairwise combination rules, the Osaka deletion is predicted to result in exposure of surfaces that are mutually shielding from the solvent. Molecular dynamics simulations on an 11-mer ß-sheet of Alzheimer's Aß(40) of the Dutch (E22Q), Iowa (D23N), Arctic (E22G), and Osaka (E22Δ) mutants, show the crucial role glycine plays in the positioning of ßαRß half-turns. Their "in-phase" positions along the sequence in the wild-type, Dutch mutant and Iowa mutant means that the half-folds all fold to the same side creating the same closed structure. Their out-of-phase positions in Arctic and Osaka mutants creates a flatter structure in the former and an S-shape structure in the latter which, as predicted, exposes surfaces on the inside in the closed wild-type to the outside. This is consistent with the gain of interaction model and indicates how domain swapping might explain the Osaka mutant's unique properties.

Determination of amino acids that favour the αL region using Ramachandran propensity plots. Implications for α-sheet as the possible amyloid intermediate.

In amyloid diseases an insoluble amyloid fibril forms via a soluble oligomeric intermediate. It is this intermediate that mediates toxicity and it has been suggested, somewhat controversially, that it has the α-sheet structure. Nests and α-strands are similar peptide motifs in that alternate residues lie in the αR and γL regions of the Ramachandran plot for nests, or αR and αL regions for α-strands. In nests a concavity is formed by the main chain NH atoms whereas in α-strands the main chain is almost straight. Using "Ramachandran propensity plots" to focus on the αL/γL region, it is shown that glycine favours γL (82% of amino acids are glycine), but disfavours αL (3% are glycine). Most charged and polar amino acids favour αL with asparagine having by far the highest propensity. Thus, glycine favours nests but, contrary to common expectation, should not favour α-sheet. By contrast most charged or polar amino acids should favour α-sheet by their propensity for the αL conformation, which is more discriminating amongst amino acids than the αR conformation. Thus, these results suggest the composition of sequences that favour α-sheet formation and point towards effective prediction of α-sheet from sequence.

The CD151-midkine pathway regulates the immune microenvironment in inflammatory breast cancer.

The immune microenvironment in inflammatory breast cancer (IBC) is poorly characterised, and molecular and cellular pathways that control accumulation of various immune cells in IBC tissues remain largely unknown. Here, we discovered a novel pathway linking the expression of the tetraspanin protein CD151 in tumour cells with increased accumulation of macrophages in cancerous tissues. It is notable that elevated expression of CD151 and a higher number of tumour-infiltrating macrophages correlated with better patient responses to chemotherapy. Accordingly, CD151-expressing IBC xenografts were characterised by the increased infiltration of macrophages. In vitro migration experiments demonstrated that CD151 stimulates the chemoattractive potential of IBC cells for monocytes via mechanisms involving midkine (a heparin-binding growth factor), integrin α6ß1, and production of extracellular vesicles (EVs). Profiling of chemokines secreted by IBC cells demonstrated that CD151 increases production of midkine. Purified midkine specifically stimulated migration of monocytes, but not other immune cells. Further experiments demonstrated that the chemoattractive potential of IBC-derived EVs is blocked by anti-midkine antibodies. These results demonstrate for the first time that changes in the expression of a tetraspanin protein by tumour cells can affect the formation of the immune microenvironment by modulating recruitment of effector cells to cancerous tissues. Therefore, a CD151-midkine pathway can be considered as a novel target for controlled changes of the immune landscape in IBC. © 2020 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Investigation of sequence features of hinge-bending regions in proteins with domain movements using kernel logistic regression.

BACKGROUND: Hinge-bending movements in proteins comprising two or more domains form a large class of functional movements. Hinge-bending regions demarcate protein domains and collectively control the domain movement. Consequently, the ability to recognise sequence features of hinge-bending regions and to be able to predict them from sequence alone would benefit various areas of protein research. For example, an understanding of how the sequence features of these regions relate to dynamic properties in multi-domain proteins would aid in the rational design of linkers in therapeutic fusion proteins. RESULTS: The DynDom database of protein domain movements comprises sequences annotated to indicate whether the amino acid residue is located within a hinge-bending region or within an intradomain region. Using statistical methods and Kernel Logistic Regression (KLR) models, this data was used to determine sequence features that favour or disfavour hinge-bending regions. This is a difficult classification problem as the number of negative cases (intradomain residues) is much larger than the number of positive cases (hinge residues). The statistical methods and the KLR models both show that cysteine has the lowest propensity for hinge-bending regions and proline has the highest, even though it is the most rigid amino acid. As hinge-bending regions have been previously shown to occur frequently at the terminal regions of the secondary structures, the propensity for proline at these regions is likely due to its tendency to break secondary structures. The KLR models also indicate that isoleucine may act as a domain-capping residue. We have found that a quadratic KLR model outperforms a linear KLR model and that improvement in performance occurs up to very long window lengths (eighty residues) indicating long-range correlations. CONCLUSION: In contrast to the only other approach that focused solely on interdomain hinge-bending regions, the method provides a modest and statistically significant improvement over a random classifier. An explanation of the KLR results is that in the prediction of hinge-bending regions a long-range correlation is at play between a small number amino acids that either favour or disfavour hinge-bending regions. The resulting sequence-based prediction tool, HingeSeek, is available to run through a webserver at hingeseek.cmp.uea.ac.uk.

Haptic-Assisted Interactive Molecular Docking Incorporating Receptor Flexibility.

Haptic-assisted interactive docking tools immerse the user in an environment where intuition and knowledge can be used to help guide the docking process. Here we present such a tool where the user "holds" a rigid ligand via a haptic device through which they feel interaction forces with a flexible receptor biomolecule. To ensure forces transmitted through the haptic device are smooth and stable, they must be updated at a rate greater than 500 Hz. Due to this time constraint, the majority of haptic docking tools do not attempt to model the conformational changes that would occur when molecules interact during binding. Our haptic-assisted docking tool, "Haptimol FlexiDock", models a receptor's conformational response to forces of interaction with a ligand while maintaining the required haptic refresh rate. In order to model receptor flexibility we use the method of linear response for which we determine the variance-covariance matrix of atomic fluctuations from the trajectory of an explicit-solvent molecular dynamics simulation of the ligand-free receptor molecule. The key to satisfying the time constraint is an eigenvector decomposition of the variance-covariance matrix which enables a good approximation to the conformational response of the receptor to be calculated rapidly. This exploits a feature of protein dynamics whereby most fluctuation occurs within a relatively small subspace. The method is demonstrated on glutamine binding protein in interaction with glutamine and maltose binding protein in interaction with maltose. For both proteins the movement that occurs when the ligand is docked near to its binding site matches the experimentally determined movement well. It is thought that this tool will be particularly useful for structure-based drug design.

Geometrical principles of homomeric ß-barrels and ß-helices: Application to modeling amyloid protofilaments.

Examples of homomeric ß-helices and ß-barrels have recently emerged. Here we generalize the theory for the shear number in ß-barrels to encompass ß-helices and homomeric structures. We introduce the concept of the "ß-strip," the set of parallel or antiparallel neighboring strands, from which the whole helix can be generated giving it n-fold rotational symmetry. In this context, the shear number is interpreted as the sum around the helix of the fixed register shift between neighboring identical ß-strips. Using this approach, we have derived relationships between helical width, pitch, angle between strand direction and helical axis, mass per length, register shift, and number of strands. The validity and unifying power of the method is demonstrated with known structures including α-hemolysin, T4 phage spike, cylindrin, and the HET-s(218-289) prion. From reported dimensions measured by X-ray fiber diffraction on amyloid fibrils, the relationships can be used to predict the register shift and the number of strands within amyloid protofilaments. This was used to construct models of transthyretin and Alzheimer ß(40) amyloid protofilaments that comprise a single strip of in-register ß-strands folded into a "ß-strip helix." Results suggest both stabilization of an individual ß-strip helix and growth by addition of further ß-strip helices can involve the same pair of sequence segments associating with ß-sheet hydrogen bonding at the same register shift. This process would be aided by a repeat sequence. Hence, understanding how the register shift (as the distance between repeat sequences) relates to helical dimensions will be useful for nanotube design.

Virtual Environment for Studying the Docking Interactions of Rigid Biomolecules with Haptics.

Haptic technology facilitates user interaction with the virtual world via the sense of touch. In molecular docking, haptics enables the user to sense the interaction forces during the docking process. Here we describe a haptics-assisted interactive software tool, called Haptimol_RD, for the study of docking interactions. By utilizing GPU-accelerated proximity querying methods very large systems can now be studied. Methods for force scaling, multipoint collision response and haptic navigation are described that address force stability issues that are particular to the interactive docking of large systems. Thus, Haptimol_RD expands, for the first time, the use of interactive biomolecular haptics to the study of protein-protein interactions. Unlike existing approaches, Haptimol_RD is designed to run on relatively inexpensive consumer-level hardware and is freely available to the community.

Quantitative method for the assignment of hinge and shear mechanism in protein domain movements.

MOTIVATION: A popular method for classification of protein domain movements apportions them into two main types: those with a 'hinge' mechanism and those with a 'shear' mechanism. The intuitive assignment of domain movements to these classes has limited the number of domain movements that can be classified in this way. Furthermore, whether intended or not, the term 'shear' is often interpreted to mean a relative translation of the domains. RESULTS: Numbers of occurrences of four different types of residue contact changes between domains were optimally combined by logistic regression using the training set of domain movements intuitively classified as hinge and shear to produce a predictor for hinge and shear. This predictor was applied to give a 10-fold increase in the number of examples over the number previously available with a high degree of precision. It is shown that overall a relative translation of domains is rare, and that there is no difference between hinge and shear mechanisms in this respect. However, the shear set contains significantly more examples of domains having a relative twisting movement than the hinge set. The angle of rotation is also shown to be a good discriminator between the two mechanisms. AVAILABILITY AND IMPLEMENTATION: Results are free to browse at http://www.cmp.uea.ac.uk/dyndom/interface/. CONTACT: sjh@cmp.uea.ac.uk. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Rings and ribbons in protein structures: Characterization using helical parameters and Ramachandran plots for repeating dipeptides.

Helical parameters displayed on a Ramachandran plot allow peptide structures with successive residues having identical main chain conformations to be studied. We investigate repeating dipeptide main chain conformations and present Ramachandran plots encompassing the range of possible structures. Repeating dipeptides fall into the categories: rings, ribbons, and helices. Partial rings occur in the form of "nests" and "catgrips"; many nests are bridged by an oxygen atom hydrogen bonding to the main chain NH groups of alternate residues, an interaction optimized by the ring structure of the nest. A novel recurring feature is identified that we name unpleated ß, often situated at the ends of a ß-sheet strand. Some are partial rings causing the polypeptide to curve gently away from the sheet; some are straight. They lack ß-pleat and almost all incorporate a glycine. An example is the first glycine in the GxxxxGK motif of P-loop proteins. Ribbons in repeating dipeptides can be either flat, as seen in repeated type II and type II' ß-turns, or twisted, as in multiple type I and type I' ß-turns. Hexa- and octa-peptides in such twisted ribbons occur frequently in proteins, predominantly with type I ß-turns, and are the same as the "ß-bend ribbons" hitherto identified only in short peptides. One is seen in the GTPase-activating protein for Rho in the active, but not the inactive, form of the enzyme. It forms a ß-bend ribbon, which incorporates the catalytic arginine, allowing its side chain guanidino group to approach the active site and enhance enzyme activity.

Protein folding, protein dynamics and the topology of self-motions.

It has long been recognized that segments of the protein main chain are like robotic manipulators and inverse kinematics methods from robotics have been applied to model loops to bridge gaps in protein comparative modelling. The complex internal motion of a redundant manipulator with fixed ends is called a self-motion and its character is determined by the relative position of its ends. Self-motions that are topologically equivalent (homotopic) occupy the same continous region of the configuration space. Topologically inequivalent (non-homotopic) regions are separated by co-regular surfaces and crossing a co-regular surface can result in a sudden dramatic change in the character of the self-motion. It is shown, using a five-residue type I ß-turn, that these concepts apply to protein segments and that as the ends of the five-residue segment come closer together, a co-regular surface is crossed, and the structure is locked in to becoming either a type I or type I' turn. It is also shown that the type II turn is topologically equivalent to the type I' turn, not the type I turn. These results have implications for both native-state protein dynamics and protein folding.

A Retrospective on the Development of Methods for the Analysis of Protein Conformational Ensembles.

Analysing protein conformational ensembles whether from molecular dynamics (MD) simulation or other sources for functionally relevant conformational changes can be very challenging. In the nineteen nineties dimensional reduction methods were developed primarily for analysing MD trajectories to determine dominant motions with the aim of understanding their relationship to function. Coarse-graining methods were also developed so the conformational change between two structures could be described in terms of the relative motion of a small number of quasi-rigid regions rather than in terms of a large number of atoms. When these methods are combined, they can characterize the large-scale motions inherent in a conformational ensemble providing insight into possible functional mechanism. The dimensional reduction methods first applied to protein conformational ensembles were referred to as Quasi-Harmonic Analysis, Principal Component Analysis and Essential Dynamics Analysis. A retrospective on the origin of these methods is presented, the relationships between them explained, and more recent developments reviewed.

Tspan6 stimulates the chemoattractive potential of breast cancer cells for B cells in an EV- and LXR-dependent manner.

The immune microenvironment in breast cancer (BCa) is controlled by a complex network of communication between various cell types. Here, we find that recruitment of B lymphocytes to BCa tissues is controlled via mechanisms associated with cancer cell-derived extracellular vesicles (CCD-EVs). Gene expression profiling identifies the Liver X receptor (LXR)-dependent transcriptional network as a key pathway that controls both CCD-EVs-induced migration of B cells and accumulation of B cells in BCa tissues. The increased accumulation oxysterol ligands for LXR (i.e., 25-hydroxycholesterol and 27-hydroxycholesterol) in CCD-EVs is regulated by the tetraspanin 6 (Tspan6). Tspan6 stimulates the chemoattractive potential of BCa cells for B cells in an EV- and LXR-dependent manner. These results demonstrate that tetraspanins control intercellular trafficking of oxysterols via CCD-EVs. Furthermore, tetraspanin-dependent changes in the oxysterol composition of CCD-EVs and the LXR signaling axis play a key role in specific changes in the tumor immune microenvironment.

Simulation of the ß- to α-sheet transition results in a twisted sheet for antiparallel and an α-nanotube for parallel strands: implications for amyloid formation.

α-sheet has been proposed to be the main constituent of the toxic amyloid intermediate. Molecular dynamics simulations on proteins known to be involved in amyloid diseases have demonstrated that ß-sheet can, under certain conditions, spontaneously convert to α-sheet via ßß→α(R)α(L) peptide-plane flipping. Using torsion-angle driving to simulate this flip the transition has been investigated for parallel and antiparallel sheets. Concerted and sequential flipping processes were simulated, the former allowing direct calculation of helical parameters. For antiparallel sheet, the strands tend to splay apart during the transition. This can be understood by consideration of the geometry of repeating dipeptide conformations. At the end of the transition antiparallel α-sheet is slightly twisted, comprising gently curving strands. In parallel sheet, the strands maintain identical conformations and stay hydrogen bonded during the transition as they curl up to suggest a hitherto unseen structure, the multi-helix α-nanotube. Intriguingly, the α-nanotube has some of the characteristics of the parallel ß-helix, a single-helix structure also implicated in amyloid. Unlike the ß-helix, α-nanotube formation could involve identical strands aligning with each other in register as in most amyloids.

The effect of end constraints on protein loop kinematics.

Despite the prevalent involvement of loops in function little is known about how the constraining of end groups influences their kinematics. Using a linear inverse-kinematics approach and assuming fixed bond lengths, bond angles, and peptide bond torsions, as well as ignoring molecular interactions to assess the effect of the end-constraint only, it is shown that the constraint creates a closed surface in torsion angle space. For pentapeptides, the constraint gives rise to inaccessible regions in a Ramachandran plot. This complex and tightly curved surface produces interesting effects that may play a functional role. For example, a small change in one torsion angle can radically change the behavior of the whole loop. The constraint also produces long-range correlations, and structures exist where the correlation coefficient is 1.0 or -1.0 between rotations about bonds separated by >30 A. Another application allows some torsion angles to be targeted to specified values while others are constrained. When this application was used on key torsions in lactate dehydrogenase, it was found that the functional loop first folds forward and then moves sideways. For horse liver alcohol dehydrogenase, it was confirmed that the functional loop's Pro-Pro motif creates a rigid arm in an NAD-activated switch for domain closure.

The interwinding nature of protein-protein interfaces and its implication for protein complex formation.

MOTIVATION: Structural features at protein-protein interfaces can be studied to understand protein-protein interactions. It was noticed that in a dataset of 45 multimeric proteins the interface could either be described as flat against flat or protruding/interwound. In the latter, residues within one chain were surrounded by those in other chains, whereas in the former they were not. RESULTS: A simple method was developed that could distinguish between these two types with results that matched those made by a human annotator. Applying this automatic method to a large dataset of 888 structures, chains at interfaces were categorized as non-surrounded or surrounded. It was found that the surrounded set had a significantly lower folding tendency using a sequence based measure, than the non-surrounded set. This suggests that before complexation, surrounded chains are relatively unstable and may be involved in 'fly-casting'. This is supported by the finding that terminal regions are overrepresented in the surrounded set. AVAILABILITY: http://cib.cf.ocha.ac.jp/DACSIS/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

DTA: dihedral transition analysis for characterization of the effects of large main-chain dihedral changes in proteins.

MOTIVATION: The biological function of proteins is associated with a variety of motions, ranging from global domain motion to local motion of side chain. We propose a method, dihedral transition analysis (DTA), to identify significant dihedral angle changes between two distinct protein conformations and for characterization of the effect of these transitions on both local and global conformation. RESULTS: Applying DTA to a comprehensive and non-redundant dataset of 459 high-resolution pairs of protein structures, we found that a dihedral transition occurs in 82% of proteins. Multiple dihedral transitions are shown to occur cooperatively along the sequence, which allows us to separate a polypeptide chain into fragments with and without transitions, namely transition fragments (TFs) and stable fragments (SFs), respectively. By characterizing the magnitude of TF conformational change and the effect of the transition on the neighboring fragments, flap and hinge motions are identified as typical motions. DTA is also useful to detect protein motions, subtle in RMSD but significant in terms of dihedral angle changes, such as the peptide-plane flip, the side-chain flip and path-preserving motions. We conclude that DTA is a useful tool to extract potential functional motions, some of which might have been missed using conventional methods for protein motion analysis. AVAILABILITY: http://dynamics.iam.u-tokyo.ac.jp/DTA/

A method for the analysis of domain movements in large biomolecular complexes.

A new method for the analysis of domain movements in large, multichain, biomolecular complexes is presented. The method is applicable to any molecule for which two atomic structures are available that represent a conformational change indicating a possible domain movement. The method is blind to atomic bonding and atom type and can, therefore, be applied to biomolecular complexes containing different constituent molecules such as protein, RNA, or DNA. At the heart of the method is the use of blocks located at grid points spanning the whole molecule. The rotation vector for the rotation of atoms from each block between the two conformations is calculated. Treating components of these vectors as coordinates means that each block is associated with a point in a "rotation space" and that blocks with atoms that rotate together, perhaps as part of the same rigid domain, will have colocated points. Thus a domain can be identified from the clustering of points from blocks that span it. Domain pairs are accepted for analysis of their relative movements in terms of screw axes based upon a set of reasonable criteria. Here, we report on the application of the method to biomolecules covering a considerable size range: hemoglobin, liver alcohol dehydrogenase, S-Adenosylhomocysteine hydrolase, aspartate transcarbamylase, and the 70S ribosome. The results provide a depiction of the conformational change within each molecule that is easily understood, giving a perspective that is expected to lead to new insights. Of particular interest is the allosteric mechanism in some of these molecules. Results indicate that common boundaries between subunits and domains are good regions to focus on as movement in one subunit can be transmitted to another subunit through such interfaces.