Clinical implementation of deep-learning based auto-contouring tools-Experience of three French radiotherapy centers.

Deep-learning (DL)-based auto-contouring solutions have recently been proposed as a convincing alternative to decrease workload of target volumes and organs-at-risk (OAR) delineation in radiotherapy planning and improve inter-observer consistency. However, there is minimal literature of clinical implementations of such algorithms in a clinical routine. In this paper we first present an update of the state-of-the-art of DL-based solutions. We then summarize recent recommendations proposed by the European society for radiotherapy and oncology (ESTRO) to be followed before any clinical implementation of artificial intelligence-based solutions in clinic. The last section describes the methodology carried out by three French radiation oncology departments to deploy CE-marked commercial solutions. Based on the information collected, a majority of OAR are retained by the centers among those proposed by the manufacturers, validating the usefulness of DL-based models to decrease clinicians' workload. Target volumes, with the exception of lymph node areas in breast, head and neck and pelvic regions, whole breast, breast wall, prostate and seminal vesicles, are not available in the three commercial solutions at this time. No implemented workflows are currently available to continuously improve the models, but these can be adapted/retrained in some solutions during the commissioning phase to best fit local practices. In reported experiences, automatic workflows were implemented to limit human interactions and make the workflow more fluid. Recommendations published by the ESTRO group will be of importance for guiding physicists in the clinical implementation of patient specific and regular quality assurances.

Automation in radiotherapy treatment planning: Examples of use in clinical practice and future trends for a complete automated workflow.

Modern radiotherapy treatment planning is a complex and time-consuming process that requires the skills of experienced users to obtain quality plans. Since the early 2000s, the automation of this planning process has become an important research topic in radiotherapy. Today, the first commercial automated treatment planning solutions are available and implemented in a growing number of clinical radiotherapy departments. It should be noted that these various commercial solutions are based on very different methods, implying a daily practice that varies from one center to another. It is likely that this change in planning practices is still in its infancy. Indeed, the rise of artificial intelligence methods, based in particular on deep learning, has recently revived research interest in this subject. The numerous articles currently being published announce a lasting and profound transformation of radiotherapy planning practices in the years to come. From this perspective, an evolution of initial training for clinical teams and the drafting of new quality assurance recommendations is desirable.

Structures of a legume lectin complexed with the human lactotransferrin N2 fragment, and with an isolated biantennary glycopeptide: role of the fucose moiety.

BACKGROUND: Lectins mediate cell-cell interactions by specifically recognizing oligosaccharide chains. Legume lectins serve as mediators for the symbiotic interactions between plants and nitrogen-fixing microorganisms, an important process in the nitrogen cycle. Lectins from the Viciae tribe have a high affinity for the fucosylated biantennary N-acetyllactosamine-type glycans which are to be found in the majority of N-glycosylproteins. While the structures of several lectins complexed with incomplete oligosaccharides have been solved, no previous structure has included the complete glycoprotein. RESULTS: We have determined the crystal structures of Lathyrus ochrus isolectin II complexed with the N2 monoglycosylated fragment of human lactotransferrin (18 kDa) and an isolated glycopeptide (2.1 kDa) fragment of human lactotransferrin (at 3.3 A and 2.8 A resolution, respectively). Comparison between the two structures showed that the protein part of the glycoprotein has little influence on either the stabilization of the complex or the sugar conformation. In both cases the oligosaccharide adopts the same extended conformation. Besides the essential mannose moiety of the monosaccharide-binding site, the fucose-1' of the core has a large surface of interaction with the lectin. This oligosaccharide conformation differs substantially from that seen in the previously determined isolectin I-octasaccharide complex. Comparison of our structure with that of concanavalin A (ConA) suggests that the ConA binding site cannot accommodate this fucose. CONCLUSIONS: Our results explain the observation that Viciae lectins have a higher affinity for fucosylated oligosaccharides than for unfucosylated ones, whereas the affinity of ConA for these types of oligosaccharides is similar. This explanation is testable by mutagenesis experiments. Our structure shows a large complementary surface area between the oligosaccharide and the lectin, in contrast with the recently determined structure of a complex between the carbohydrate recognition domain of a C-type mammalian lectin and an oligomannoside, where only the non-reducing terminal mannose residue interacts with the lectin.

Visualization of lactotransferrin brush-border receptors by ligand-blotting.

The uptake of iron (III) mediated by lactotransferrin to human biopsies from upper intestine has suggested the presence of specific receptors for human lactotransferrin at the brush border (Cox, T., Mazurier, J., Spik, G., Montreuil, J. and Peters, T.J. (1979) Biochim. Biophys. Acta 588, 120-128). In the present data, using 125I-radiolabeled transferrins, we have demonstrated that a preparation of microvillous membrane vesicles, from rabbit jejunal brush-border specifically binds human lactotransferrin. This binding is specific, saturable and calcium dependent. Scatchard plots analysis of lactotransferrin binding indicates 1.5 X 10(13) sites per mg of membrane proteins with an equilibrium constant of 1.2 X 10(6) M-1. Sodium dodecyl sulfate solubilization of the brush-border proteins allows the lactotransferrin receptor to retain its binding activity. Moreover, the ligand blotting of the detergent solubilized membrane proteins on nitrocellulose sheet and after incubation with 125I-labeled lactotransferrin, has shown that the receptor is a protein of about 100 kDa. In the same experimental conditions, the rabbit microvillous membrane vesicles do not specifically bind rabbit serotransferrin indicating the absence of serotransferrin receptors at the brush border.

Ultraviolet difference spectral studies of human serotransferrin and lactotransferrin.

In order to determine the conformational relationship of iron binding of human serotransferrin and lactotransferrin, ultraviolet difference spectral studies were performed in the presence of guanidine chloride and perturbants as deuterium oxide, ethylene glycol, glycerol and polyethylene glycol. In the presence of guanidine chloride solution the molar absorption differences at 292 nm of iron-saturated forms versus iron-free forms of human serotransferrin and lactotransferrin are respectively -16000 +/- 1000 and -14000 +/- 775. These modifications may be attributed to the involvement of tryptophan residues in the iron-binding sites of the two proteins. However, the results do not demonstrate that these tryptophan residues are bound directly to iron. Difference spectral studies in the presence of perturbants show that the apparent exposed tryptophan and tyrosine residues are higher with shorter range perturbants in iron-free forms of both transferrin molecules. The most important modification of exposed tyrosine residues has been noticed upon removing iron from human lactotransferrin than from serotransferrin.

Iron binding proteins and influx of iron across the duodenal brush border. Evidence for specific lactotransferrin receptors in the human intestine.

The ability of a range of homologous transferrin-like proteins to donate iron to pieces of human duodenal mucosa, was examined with an in vitro incubation technique. In contrast to serum transferrin and ovotransferrin, only lactotransferrin was able to yield its iron to intestinal tissue, but in an autologous system this protein was unable to donate iron to human reticulocyte preparations. Studies with 125I-labelled lactotransferrin and lactotransferrin dual-labelled with 59Fe and 125I, indicated that the intact protein is excluded from entry into the enterocytes. The experiments suggest that iron may be transported across the brush border after delivery to specific protein binding sites at the cell surface.

Comparative study of the iron-binding properties of human transferrins. II. Electron paramagnetic resonance of mixed metal complexes of human lactotransferrin.

Human lactotransferrin is able to bind two vanadyl(IV) ions in specific metal-binding sites. The EPR signals of the two vanadyl bound ions, however, appear as one. This result suggests that the environments of the binding sites of human lactotransferrin are similar. The binding activity is promoted to pH 4 using carbonate or bicarbonate as synergistic anion. This unusual stability of the anion-binding site, which is destroyed below pH 6 for other transferrins, can explain in part the great stability of the metallic complexes of human lactotransferrin. However, the different sensitivities of the two metal-binding sites towards protonation permit the preparation of mixed vanadyl(IV), iron(III) complexes with VO2+ bound either on the N-terminal (acid-labile or B site) or on the C-terminal (acid-stable or A site) site. Analysis of the spectra of such mixed complexes shows the presence of a third nonspecific VO2+-binding site termed A'. The nonspecific A' site seems to be located on the outer surface of the protein close to the C-terminal site.

Characterization and localization of an iron-binding 18-kDa glycopeptide isolated from the N-terminal half of human lactotransferrin.

Mild treatment of iron-saturated human lactotransferrin by trypsin at pH 8.2 cleaves the molecule into a N-tryptic (Mr approximately equal to 30000) and a C-tryptic (Mr approximately equal to 50000) fragment, which have been isolated. Each of them carries a glycan moiety and keeps the property to bind reversibly one Fe3+. The N-tryptic fragment has been submitted to a second tryptic digestion which led to an iron-binding glycopeptide fragment with a molecular weight of about 18500. This fragment, the smallest iron-binding peptide isolated up to now from a transferrin, includes the ND2 domain of human lactotransferrin.

Amino acid sequence, location and phylogenetic aspects of the glycopeptides of human lactotransferrin.

Human lactotransferrin contains two prosthetic sugar groups situated in two different cyanogen bromide fragments: the amino acid sequences around the two polysaccharide attachment sites were established. The location of the prosthetic groups was quite different in serum and lactotransferrins and no sequence homology could so far be characterized between the glycopeptides of these two transferrins, whereas a close relationship between a glycopeptide of human lactotransferrin and a glycopeptide of hen ovotransferrin was noted.

The present state of the human lactotransferrin sequence. Study and alignment of the cyanogen bromide fragments and characterization of N- and C-terminal domains.

Human lactotransferrin contains six residues of methionine per mol. Seven different fragments were characterized after treatment with cyanogen bromide (CNBr) and large parts of their sequences were determined. The alignment of the CNBr fragments was established by the determination of N- and C-terminal sequences, by the study of the C-terminal domain obtained by peptic digestion of the protein and by taking into account the internal homology as well as homology with human serum transferrin. The two glycopeptides were situated in the N- and C-terminal parts of the protein, respectively, a situation quite different from that encountered in serum transferrin. The sequence studies allowed us to suggest a 4- and perhaps 6-fold internal homology.

Effects of human lactoferrin on NK cell cytotoxicity against haematopoietic and epithelial tumour cells.

Lactoferrin is an iron-binding glycoprotein implicated in particular in the control of immune functions and cell proliferation. We have investigated its involvement, at inflammatory concentrations, in cancer progression. We report that lactoferrin has a significant effect on natural killer (NK) cell cytotoxicity against haematopoietic and breast epithelial cell lines. Lactoferrin increases cytolysis at a low concentration (10 micrograms/ml) while at a high concentration (100 micrograms/ml) it modulates cytolysis depending on the target cell phenotype. By pre-treatment of either NK cells or target cells with lactoferrin, we have demonstrated that the lactoferrin effect is due both to a modulation of NK cell cytotoxicity and the target cell sensitivity to lysis. Lactoferrin binds to 91% of the naturally heterogeneous CD56dim/bright NK cell population and increases the NK cell cytotoxic activity at low concentrations. High concentrations of lactoferrin seem to be toxic for the CD56bright NK cells and decrease NK cell cytotoxicity. Lactoferrin also exerts an effect on target cells depending on the cell phenotype. It does not modify the susceptibility to lysis of haematopoietic cells such as Jurkat and K-562 cells, but does significantly increase that of the breast and colon epithelial cells. We have also demonstrated that lactoferrin inhibits epithelial cell proliferation by blocking the cell cycle progression.

Crystallization and preliminary X-ray diffraction study of Lathyrus ochrus isolectin II complexed to the human lactotransferrin N2 fragment.

Isolectin II (LOL II) isolated from the seeds of Lathyrus ochrus has been crystallized in the presence of the N2 fragment (18,500 Da) isolated from human lactotransferrin, which contains an N-acetyllactosamine type biantennary glycan linked to Asn137. This is the first example of a legume lectin crystallized with an N-glycosylprotein. Crystals of the LOL II-N2 complex belong to the tetragonal space group (P4(1)2(1)2 or the enantiomorph) with cell dimensions: a = b = 63.5 A, c = 251.9 A. They diffract well up to at least 3.5 A resolution and more weakly up to 2.8 A resolution. Assuming one functional half-entity in the asymmetric unit, an alpha, beta monomer complexed to one N2 fragment (24,500 Da + 18,500 Da) would give a Vm of 2.95 A3/Da and a solvent content of approximately 58%. SDS/polyacrylamide gels of the dissolved crystals show the presence of both the LOL II and N2 fragment.

[Implementing new technology in radiation oncology: The French agency for nuclear safety (ASN) report]. / Conditions de mise en œuvre des « nouvelles techniques et pratiques ¼ en radiothérapie. Synthèse du rapport du groupe de travail issu du groupe permanent d'experts en radioprotection médicale de l'Autorité de sûreté nucléaire.

In August 2013, the French nuclear safety agency (ASN) requested the permanent group of experts in radiation protection in medicine (GPMED) to propose recommendations on the implementation of new technology and techniques in radiation oncology. These recommendations were finalized in February 2015 by the GPMED. In April 2015, the ASN sent a letter to the French ministry of health (DGS/DGOS), and its national health agencies (ANSM, INCa, HAS). In these letters, ASN proposed that, from the 12 recommendations made by the GPMED, an action plan should be established, whose control could be assigned to the French national cancer institute (INCa), as a pilot of the national committee for radiotherapy and that this proposal has to be considered at the next meeting of the national committee of radiotherapy.

Internalization of human lactoferrin by the Jurkat human lymphoblastic T-cell line.

Binding of either iron-saturated or iron-free lactoferrin to the Jurkat human lymphoblastic T-cell line was saturable with a dissociation constant Kd of 40 nM. The total number of binding sites was estimated to be approximately 300,000. Non-specific binding did not exceed 30% of the total binding. Removal of the 4 clustered arginine residues of lactoferrin at position 2 to 5, which are involved in the interactions with heparan sulfate, did not modify the binding parameters. Therefore, the high number of low affinity binding sites previously described as responsible for the interaction of lactoferrin with either hepatocytes, enterocytes or the U937 monocytic cell line, is not involved in the binding of lactoferrin to Jurkat cells. After binding at 4 degrees C, a shift to 37 degrees C causes cell to internalize lactoferrin, with the maximum intracellular concentration found at 3 to 8 and 5 to 15 min for iron-saturated and iron-free forms, respectively. Addition of colchicine had no effect on binding or internalization. These results suggest that endocytosis of lactoferrin by Jurkat cells occurs through a receptor-mediated process. Jurkat cells internalize lactoferrin monophasically with a first-order endocytic constant K(in) of 0.060 min-1 at 37 degrees C. Confocal microscopic analysis, using fluorescein-carbohydrate-labeled lactoferrin showed that lactoferrin was mainly localized in intracellular vesicles. Following uptake, the endocytic path utilized by fluorescein-carbohydrate-labeled lactoferrin was shown to diverge from that of rhodamine-labeled serum transferrin; after internalization, lactoferrin and serum transferrin did not fully colocalize. Intracellular lactoferrin was found in endosome vesicles as assessed by electron microscopy. Raising the pH in endosomes using chloroquine led to the accumulation of lactoferrin into endosomes (acidic compartment). After internalization, Jurkat cells released both degraded and intact lactoferrin into the culture medium, suggesting that a fraction (30-40%) of the ligand is degraded at each round of endocytosis.

Immunolocalization of the lactotransferrin receptor on the human T lymphoblastic cell line Jurkat.

Monoclonal antibodies have been raised against the soluble lactotransferrin binding protein purified from the cell culture supernatant of Jurkat cell line, a human T-lymphoblastic cell. All monoclonal antibodies were able to specifically bind to the membrane of Jurkat cells. One of the monoclonal antibodies, DP5B3G10, recognized both the soluble lactotransferrin-binding protein and the membrane lymphocyte lactotransferrin receptor after SDS-PAGE in presence of 2-mercaptoethanol and electrotransfer on nitrocellulose. The monoclonal antibody DP5B3G10 inhibited the binding of lactotransferrin to Jurkat cells and human peripheral activated lymphocytes. In addition, lactotransferrin inhibited the binding of the monoclonal antibody to the cell surface. These results suggest that the 95 kDa lactotransferrin-binding protein isolated from the cell culture medium corresponds to the soluble form of the 105 kDa lymphocyte lactotransferrin receptor. Corresponding proteins of 105 kDa molecular mass were identified in Jurkat and CEM T-cells and Raji B-cells. Finally, the monoclonal antibody DP5B3G10 was used to immunolocalize the lactotransferrin receptor on the Jurkat cells. Using fluorescence and electron microscopy, the receptor was localized both inside and at the cell surface. The cell membrane receptor was associated into clusters. After permeabilization of the plasma membrane, the staining was positive in the peri-membrane area. The region near the nucleus was devoid of receptor.

Role of heparan sulphate proteoglycans in the regulation of human lactoferrin binding and activity in the MDA-MB-231 breast cancer cell line.

We previously demonstrated that lactoferrin increases breast cell sensitivity to natural killer cell cytotoxicity whereas haematopoietic cells are unaffected by lactoferrin. It has been described that lactoferrin binds to various glycosaminoglycans. Compared to haematopoietic cells, breast cancer cells and particularly the breast cell line MDA-MB-231, possess a high level of proteoglycans. Scatchard analysis of 125I-lactoferrin binding to MDA-MB-231 cells revealed the presence of two classes of binding sites: a low affinity site with a Kd of about 700 nM and 3.9 x 10(6) sites and a higher affinity class with a Kd of 45 nM and 2.9 x 10(5) sites per cell. To investigate the potential regulation of lactoferrin activity by proteoglycans expressed on the MDA-MB-231 cells, we treated these cells with glycosaminoglycan-degrading enzymes or sodium chlorate, a metabolic inhibitor of proteoglycan sulphation. We showed that chondroitinase treatment has no effect, while heparinase or chlorate treatment significantly reduces both the binding of lactoferrin to cell surface sulphated molecules such as heparan sulphate proteoglycans (HSPG) and the affinity of lactoferrin for the higher affinity binding sites. The modulation of the lactoferrin binding was correlated with a decrease in lactoferrin activities on both MDA-MB-231 cell sensitisation to lysis and proliferation. Taken together, these results suggest that the presence of adequately sulphated molecules, in particular HSPG, is important for lactoferrin interaction and activity on the breast cancer cells MDA-MB-231.

Lactoferrin inhibits the binding of lipopolysaccharides to L-selectin and subsequent production of reactive oxygen species by neutrophils.

The activation of leukocytes by lipopolysaccharides (LPS), resulting in the oxidative burst, contributes to the pathogenesis of septic shock. The binding of LPS to L-selectin, which was reported as a serum-independent LPS receptor on neutrophils, induces the production of oxygen free radicals. Human lactoferrin (hLf), an anti-inflammatory glycoprotein released from neutrophil granules during infection, binds to LPS. In this study, we investigated the capacity of hLf to inhibit the L-selectin-mediated activation of neutrophils. Our experiments revealed that hLf prevents the binding of LPS to L-selectin in a concentration-dependent manner. Inhibition was maximum (87.7+/-0.5%) at a concentration of 50 microg/ml of hLf. Furthermore, hLf inhibited up to 55.4+/-0.5% of the intracellular hydrogen peroxide production induced by LPS in neutrophils. These findings suggest that the anti-inflammatory properties of hLf are due, at least in part, to their ability to prevent the binding of LPS to neutrophil L-selectin.

2-APB inhibits volume-regulated anion channels independently from intracellular calcium signaling modulation.

It has previously been suggested that volume-regulated anion channels (VRACs) and store-operated channels (SOCs) interact with each other according to their expected colocalization in the plasma membrane of LNCaP cells. In order to study interactions between these two channels, we used 2-aminoethoxydiphenyl borate (2-APB) as a regular SOC inhibitor. Surprisingly 2-APB reduced VRAC activity in a dose-dependent manner (IC(50)=122.8 microM), but not 2,2-diphenyltetrahydrofuran (a structural analog of 2-APB). This effect was also present in keratinocytes. We conclude that 2-APB is an inhibitor of the VRAC family, and is also a potent tool to study the SOC-VRAC interaction in LNCaP cells.

The N-terminal domain I of human lactotransferrin binds specifically to phytohemagglutinin-stimulated peripheral blood human lymphocyte receptors.

Human lactotransferrin receptors have been recently characterized on mitogen-stimulated human lymphocytes [(1989) Eur. J. Biochem. 179, 481-487]. In order to define the lactotransferrin recognition site by these receptors, the binding to lymphocytes of several tryptic fragments, isolated from human lactotransferrin by mild tryptic hydrolysis [(1984) Biochim. Biophys. Acta 787, 90-96], has been investigated. The 30 kDa N-tryptic fragment (residues 4-281) and the re-associated N,C-tryptic complex bind to lactotansferrin lymphocyte receptor with a dissociation constant of 44 nM and 39 nM, respectively, similar to the value obtained for the native lactotransferrin (Kd = 46 nM). However, neither the N-terminal domain II (residues 91-257) nor the 50 kDa C-tryptic fragment (residues 282-703) are recognized. These results suggest that the binding site of human lactotransferrin by the lymphocyte receptor is located in the N-terminal lobe and more precisely in the N-terminal domain I (residues 4-90 and/or 258-281).

Structure and spatial conformation of the iron-binding sites of transferrins.

Transferrins are iron-binding glycoproteins involved in iron metabolism and antibacterial defense mechanisms. Since the discovery of transferrins, many studies have attempted to characterize the iron ligands and to establish the conformation of the iron-binding sites. From chemical and spectroscopic studies, it was generally accepted that iron was hexacoordinated to Tyr and His residues, to a water molecule and to a (bi)carbonate ion, electrostatically linked to an Arg residue. On the basis of these studies, on the one hand, and on the basis of the homologies between the amino acid sequences of transferrins, on the other hand, predicted data have been provided about the number and location of the iron ligands. Recent X-ray crystallography studies of human lactotransferrin have partially confirmed the above-mentioned predicted data and have brought invaluable information about the nature of the ligands and the conformation of the iron-binding site. On the basis of the obtained results, a scheme has been proposed in which the iron is coordinated to 2 Tyr, 1 His and 1 Asp residues, to a (bi)carbonate linked to an Arg residue and probably to a water molecule. The iron-binding site is located at the interface between the two domains which constitute each lobe of the transferrins.