Epithelial-to-mesenchymal transition (EMT) induced by inflammatory priming elicits mesenchymal stromal cell-like immune-modulatory properties in cancer cells.

BACKGROUND: Epithelial-to-mesenchymal transition (EMT) has a central role in cancer progression and metastatic dissemination and may be induced by local inflammation. We asked whether the inflammation-induced acquisition of mesenchymal phenotype by neoplastic epithelial cells is associated with the onset of mesenchymal stromal cell-like immune-regulatory properties that may enhance tumour immune escape. METHODS: Cell lines of lung adenocarcinoma (A549), breast cancer (MCF7) and hepatocellular carcinoma (HepG2) were co-cultured with T, B and NK cells before and after EMT induction by either the supernatant of mixed-lymphocyte reactions or inflammatory cytokines. RESULTS: EMT occurrence following inflammatory priming elicited multiple immune-regulatory effects in cancer cells resulting in NK and T-cell apoptosis, inhibition of lymphocyte proliferation and stimulation of regulatory T and B cells. Indoleamine 2,3-dioxygenase, but not Fas ligand pathway, was involved at least in part in these effects, as shown by the use of specific inhibitors. CONCLUSIONS: EMT induced by inflammatory stimuli confers to cancer cells some mesenchymal stromal cell-like immune-modulatory properties, which could be a cue for cancer progression and metastatic dissemination by favouring immune escape.

Role of stromal cell-mediated Notch signaling in CLL resistance to chemotherapy.

Stromal cells are essential components of the bone marrow (BM) microenvironment that regulate and support the survival of different tumors, including chronic lymphocytic leukemia (CLL). In this study, we investigated the role of Notch signaling in the promotion of survival and chemoresistance of human CLL cells in coculture with human BM-mesenchymal stromal cells (hBM-MSCs) of both autologous and allogeneic origin. The presence of BM-MSCs rescued CLL cells from apoptosis both spontaneously and following induction with various drugs, including Fludarabine, Cyclophosphamide, Bendamustine, Prednisone and Hydrocortisone. The treatment with a combination of anti-Notch-1, Notch-2 and Notch-4 antibodies or γ-secretase inhibitor XII (GSI XII) reverted this protective effect by day 3, even in presence of the above-mentioned drugs. Overall, our findings show that stromal cell-mediated Notch-1, Notch-2 and Notch-4 signaling has a role in CLL survival and resistance to chemotherapy. Therefore, its blocking could be an additional tool to overcome drug resistance and improve the therapeutic strategies for CLL.

Structure of chicken plasma retinol-binding protein.

The crystal structure of the specific carrier of retinol (retinol-binding protein, RBP) purified from chicken plasma has been determined (space group P2(1)2(1)2(1), with a=46.06(5) A, b=53.56(6) A, c=73.41(8) A, and one protein molecule in the asymmetric unit). Despite being obtained from a species phylogenetically distant from mammals, chicken holoRBP has an overall structure that closely resembles the previously determined structures of mammalian holoRBPs. The lack in chicken RBP of eight carboxy-terminal amino acid residues characteristic of mammalian RBPs does not significantly affect the protein structure. A distinctive feature of the avian protein is a better definition of the loop 63-67, close to the opening of the beta-barrel cavity accommodating the retinol molecule, which is rather disordered in the structures of mammalian RBPs.

Crystallization and preliminary X-ray data for the human transthyretin-retinol-binding protein (RBP) complex bound to an anti-RBP Fab.

A macromolecular complex of human transthyretin, human retinol-binding protein and an anti-retinol-binding-protein Fab was crystallized by vapour diffusion in sitting drops. Diffraction from these crystals at cryogenic temperatures was consistent with the space group C222, with cell parameters a = 159.34, b = 222.40 and c = 121.27 A. Crystals diffracted to a resolution limit of 3.36 A using synchrotron radiation. Based on a 2:2:1 stoichiometry for the Fab-retinol-binding-protein-transthyretin complex and the presence of one such complex per asymmetric unit, a reasonable Vm coefficient of 2.74 A3 Da-1 could be estimated.

Structure of pig plasma retinol-binding protein at 1.65 A resolution.

The crystal structure of pig plasma retinol-binding protein (RBP) has been determined at 1.65 A resolution. The space group is P212121, with a = 45.81 (4), b = 53.14 (5), c = 72.97 (8) A and one protein molecule in the asymmetric unit. The structure has been solved using the molecular replacement method and refined with restrained least squares to an R factor of 0.1844 and an Rfree of 0.237 for 18 874 and 1001 independent reflections, respectively. The relatively high resolution structure of pig holoRBP has revealed some new structural details. Moreover, it has provided a description of the binding site for Cd2+, a metal ion which is required for protein crystallization. The hepta-coordination of the RBP-bound cadmium ion involves different residues of two symmetry-related RBP molecules, consistent with the participation of the cation in intermolecular interactions that in turn promote protein crystallization.

Structure of the trigonal crystal form of bovine annexin IV.

The structure of a trigonal crystal form of N-terminally truncated [des-(1-9)] bovine annexin IV, an annexin variant that exhibits the distinctive property of binding both phospholipids and carbohydrates in a Ca2+-dependent manner, has been determined at 3 A (0.3 nm) resolution -space group: R3; cell parameters: a=b=118.560 (8) A and c=82.233 (6) A-. The overall structure of annexin IV, crystallized in the absence of Ca2+ ions, is highly homologous to that of the other known members of the annexin family. The trimeric assembly in the trigonal crystals of annexin IV is quite similar to that found previously in non-isomorphous crystals of human, chicken and rat annexin V and to the subunit arrangement in half of the hexamer of hydra annexin XII. Moreover, it resembles that found in two-dimensional crystals of human annexin V bound to phospholipid monolayers. The propensity of several annexins to generate similar trimeric arrays supports the hypothesis that trimeric complexes of such annexins, including annexin IV, may represent the functional units that interact with membranes.

Retinoid binding to retinol-binding protein and the interference with the interaction with transthyretin.

The retinol carrier retinol-binding protein (RBP) forms a complex with the thyroid hormone binding protein transthyretin in the plasma of a number of vertebrate species. The interactions of retinoid-RBP complexes, as well as of unliganded RBP, with transthyretin have been investigated by means of fluorescence anisotropy studies. The presence of two independent and equivalent RBP binding sites per transthyretin molecule has been established for proteins purified from species distant in evolution. Although the natural ligand retinol participates in the interaction between retinol-RBP and transthyretin, its binding to RBP is not a prerequisite for protein-protein interaction. The dissociation constants of human transthyretin binding liganded and unliganded forms of human RBP were determined to be: all-trans retinol-RBP, Kd approximately 0.2 microM; apoRBP, Kd approximately 1.2 microM; all-trans retinoic acid-RBP, Kd approximately 0.8 microM; all-trans retinyl methyl ether-RBP, Kd approximately 6 microM. The complex of RBP with the synthetic retinoid fenretinide, which bears the bulky hydroxyphenyl end group, exhibits negligible affinity for transthyretin. The replacement of RBP-bound retinol with synthetic retinoids affects RBP-transthyretin recognition to an extent that appears to be well correlated with the nature and steric hindrance of the groups substituting the retinol hydroxyl group, consistent with their location at the interface between the contact areas of RBP and transthyretin.

Crystal structure of the transthyretin--retinoic-acid complex.

Retinoids are quite insoluble and chemically unstable compounds in the aqueous medium, such that natural retinoids need to be bound to specific retinoid-binding proteins to be protected, solubilized and transported in body fluids. All-trans retinoic acid exhibits a relatively high affinity for thyroxine-binding transthyretin in vitro and this protein is a good candidate for the transport of retinoic acid administered as pharmacological or antitumor agent. To define structural features essential for the recognition by transthyretin of a ligand which is structurally unrelated to thyroxine, we have cocrystallized human transthyretin with retinoic acid and determined its structure at 0.18-nm resolution. The retinoid fits into the two chemically identical thyroxine-binding sites, which are located in the central channel that runs through the tetrameric transthyretin. The cyclohexene ring of the bound retinoid is innermost, occupying the same position of the phenolic ring of the bound 3,3'-diiodo-L-thyronine, whereas the carboxylate group, like the same group of the thyroid hormone, participates in an ionic interaction with the Lys15 side chain at the entrance of the channel. Despite the fact that transthyretin was cocrystallized with all-trans-retinoic acid, the isoprene chain of the bound retinoid has been found in a non-extended conformation. This feature, that allows the carboxylate to orient in a manner suitable for ion-pair association with the Lys15 side chain, is attributable to the conversion of all-trans-retinoic acid into cis-isomers or folded conformers. It is concluded that the presence, in an essentially hydrophobic molecular core of the appropriate size, of a negatively charged group at the correct position is a crucial requirement for ligand-transthyretin recognition. Whereas the binding of the ligand has no remarkable consequences for the protein structure, all-trans-retinoic acid undergoes structural changes such as to interact favorably with residues present in the thyroxine-binding sites, resembling roughly the natural ligand.

Interactions with retinol and retinoids of bovine cellular retinol-binding protein.

The interactions with retinol and retinol analogs of bovine cellular retinol-binding protein (CRBP) have been investigated, by means of fluorescence titrations, to obtain more information on the structural features of retinoid that may be required for their interaction with the binding protein. An approximately stoichiometric binding of retinol to bovine CRBP (K'd approximately 2 nM) has been found in direct binding assays. Although retinal exhibited relatively high binding affinity to bovine CRBP (K'd approximately 30 nM), a large excess of the retinoid could not compete with retinol for the carrier protein. On the assumption that retinol and retinal interact with the same binding site, this result indicates that the above-mentioned apparent dissociation constant for retinol.CRBP may be an overestimate and that its value may be as low as 0.1 nM. The finding of an exceedingly tight binding of retinol to CRBP provides further support for the possible role of CRBP-bound retinol, rather than its uncomplexed labile form, as substrate of enzymes involved in the metabolism of the vitamin. The results of these and previous studies indicate that CRBP is particularly sensitive to modifications of the retinol hydroxyl end group. Axerophthene, a retinol analog bearing a hydrogen atom in place of the hydroxyl end group, and beta-ionone exhibit rather low binding affinities for CRBP (K'd approximately 0.2 microM and approximately 4 microM, respectively), suggesting that the hydroxyl group and isoprene tail moieties contribute substantially to the retinol-binding affinity and specificity. These findings are consistent with the indications emerging from the three-dimensional structure determination of retinol.CRBP [Cowan, S. W., Newcomer, M. E. & Jones, T. A. (1993) J. Mol. Biol. 230, 1225-1246]. Additionally, the bulky end groups of fenretinide and N-ethyl retinamide replacing the retinol hydroxyl group have been found to prevent retinoid binding to CRBP. The primary structure of bovine CRBP has been determined and is highly similar to the structures of both human and rat CRBP (97% and 95% identical, respectively).

Crystallographic studies on complexes between retinoids and plasma retinol-binding protein.

The three-dimensional structures of complexes between bovine plasma retinol-binding protein (RBP) and three retinol analogs with different end groups (fenretinide, all-trans retinoic acid, and axerophthene) have been determined to 1.8-1.9-A resolution. Their models are very similar to that of the bovine retinol.RBP complex: the root mean square deviations between equivalent alpha-carbons in the two proteins range from 0.17 to 0.24 A. The retinoid molecules fit in the beta-barrel cavity assuming the same conformation of the vitamin, and the substitutions have no consequences on the overall protein structure. While confirming that an intact hydroxyl end group is not an absolute requirement for a correct retinoid binding to RBP, this study has shown the occurrence of conformational changes, although limited, in the rather flexible loop region at the entrance of the beta-barrel upon fenretinide and retinoic acid binding. These changes are suitable for accommodating the end groups of the above retinoids. Instead, no such changes have been revealed in RBP complexed with axerophthene, a retinol analog bearing a hydrogen atom in place of the hydroxyl end group. The protein conformational changes in the above loop region, the steric hindrance of bulky end groups of bound retinoids, and the lack of the retinol hydroxyl group appear to be responsible for the possible reduced affinity of retinoids for RBP relative to retinol and, at the same time, for the abolished or reduced affinity of retinoid.RBP complexes for transthyretin relative to retinol-RBP.

The Ile-84-->Ser amino acid substitution in transthyretin interferes with the interaction with plasma retinol-binding protein.

In plasma the thyroid hormone-binding protein transthyretin (TTR) forms a tight complex with the specific retinol carrier retinol-binding protein (RBP). The Ile-84-->Ser mutation and several other point mutations in TTR are associated with familial amyloidotic polyneuropathy, which is characterized by extracellular depositions of amyloid fibrils mainly consisting of mutated TTRs. The interactions with human RBP of recombinant human normal and Ser-84 TTRs were investigated by monitoring the fluorescence anisotropy of RBP-bound retinol. A nearly negligible affinity of the recombinant Ser-84 TTR for RBP was found. This result indicates the participation of a region on the outer surface of TTR that comprises Ile-84 in the recognition of RBP. In preliminary studies the Ser-84 TTR was the only one among several amyloidogenic variant TTRs to display negligible interaction with RBP. Therefore, in general a substantially altered binding of TTR to RBP is not associated with familial amyloidotic polyneuropathy. Instead, the altered binding of Ser-84 TTR to RBP appears to be responsible for an abnormal plasma transport of RBP. The recombinant normal TTR exhibits binding properties, in its interaction with human RBP, approximately similar to those of TTR purified from human plasma. Two independent and equivalent RBP binding sites on recombinant normal TTR are characterized by a dissociation constant of about 0.4 microM.

The interaction of N-ethyl retinamide with plasma retinol-binding protein (RBP) and the crystal structure of the retinoid-RBP complex at 1.9-A resolution.

The three-dimensional structure of bovine plasma retinol-binding protein (RBP) complexed with N-ethyl retinamide has been determined at 1.9-A resolution. The crystals of the complex (space group P2(1)2(1)2(1), a = 46.27, b = 49.11, c = 76.41 A) are isomorphous with those of bovine holo and apoRBP. The final crystallographic R factor is 0.172 for 11,261 observed reflections. The model of the retinoid-RBP complex is nearly identical to that of bovine retinol-RBP complex; the root mean square deviations between the alpha-carbons in the two proteins is 0.15 A. The N-ethyl retinamide molecule fits in the beta-barrel in the same position previously occupied by the vitamin. The ethyl group of the retinoid replaces the retinol hydroxyl group and a water molecule hydrogen bonded to it. This substitution has no consequence on the overall conformation of the protein. The modification of the functional end group of retinol does not lead to an apparent loss of affinity of the retinol analog for apoRBP, as established by means of fluorometric titrations with N-ethyl retinamide. However, the binding of the retinoid to RBP abolishes or greatly reduces the affinity of RBP for transthyretin. This behavior further supports the hypothesis that the area of the entrance of the beta-barrel, which includes the ethyl group of the retinoid bound to RBP, is involved in the interaction with transthyretin. Moreover, it indicates a high degree of complementarity of interacting surfaces between RBP and transthyretin. In fact, the replacement of the retinol hydroxyl group and quite small structural changes in the above region of the RBP molecule drastically affect the protein-protein recognition.

Retinoids: in vitro interaction with retinol-binding protein and influence on plasma retinol.

Studies have been conducted to investigate the structure-function relationships of retinoids in their in vitro interaction with plasma retinol-binding protein (RBP) and in their influence on plasma retinol concentration. Two classes of retinoids, one bearing modifications in the area of the retinol hydroxyl end group (fenretinide, N-(ethyl) retinamide, all-trans, and 13-cis retinoic acid) and the other one also bearing modifications in the area of the cyclohexene ring (etretinate, acitretin, and arotinoid Ro 13-7410), were investigated. Whereas substantial modifications of the retinol hydroxyl group do not prevent the binding to RBP, an intact trimethylcyclohexenyl group seems to be crucial for binding. Both classes of retinoids, administered orally at equimolar doses, reduce plasma retinol concentration in rats but with different kinetics. A marked lowering of plasma retinol occurs early (within 5 h) after administration of retinoids that interact with RBP in vitro, whereas it occurs at later times (24 h) after retinoids that do not interact with RBP. The concentrations of both classes of retinoids found in plasma do not account for the temporal difference in this effect. The early reduction of plasma retinol might be the consequence of in vivo specific binding of retinoids to RBP, as suggested by the in vitro results. The late reduction observed for retinoids lacking in vitro affinity for RBP is due to other mechanisms or to metabolism to retinoids binding to RBP.