¿Podrían ayudarnos los algoritmos de machine learning en la predicción de hemorragia masiva a nivel prehospitalario? / Could machine learning algorithms help us in predicting massive hemorrhage at the prehospital level?

Objetivo: Comparación de la capacidad predictiva de diferentes algoritmos de machine learning (AML) respecto a escalas tradicionales de predicción de hemorragia masiva en pacientes con enfermedad traumática grave (ETG). Diseño: Sobre una base de datos de una cohorte retrospectiva con variables clínicas prehospitalarias y de resultado de hemorragia masiva se realizó un tratamiento de la base de datos para poder aplicar los AML, obteniéndose un conjunto total de 473 pacientes (80% entrenamiento, 20% validación). Para la modelización se realizó imputación proporcional y validación cruzada. El poder predictivo se evaluó con la métrica ROC y la importancia de las variables mediante los valores Shapley. Ámbito: Atención extrahospitalaria del paciente con ETG. Pacientes: Pacientes con ETG atendidos en el medio extrahospitalario por un servicio médico extrahospitalario desde enero de 2010 hasta diciembre de 2015 y trasladados a un centro de trauma en Madrid. Intervenciones: Ninguna. Variables de interés principales: Obtención y comparación de la métrica ROC de 4 AML: random forest, support vector machine, gradient boosting machine y neural network con los resultados obtenidos con escalas tradicionales de predicción. Resultados: Los diferentes AML alcanzaron valores ROC superiores al 0,85, teniendo medianas cercanas a 0,98. No encontramos diferencias significativas entre los AML. Cada AML ofrece un conjunto de variables diferentes, pero con predominancia de las variables hemodinámicas, de reanimación y de deterioro neurológico. Conclusiones: Los AML podrían superar a las escalas tradicionales de predicción en la predicción de hemorragia masiva. (AU)

Objective: Comparison of the predictive ability of various machine learning algorithms (MLA) versus traditional prediction scales for massive hemorrhage in patients with severe traumatic injury (ETG). Design: On a database of a retrospective cohort with prehospital clinical variables and massive hemorrhage outcome, a treatment of the database was performed to be able to apply the different MLA, obtaining a total set of 473 patients (80% training and 20% validation). For modeling, proportional imputation and cross validation were performed. The predictive power was evaluated with the ROC metric and the importance of the variables using the Shapley values. Setting: Out-of-hospital care of patients with ETG. Participants: Patients with ETG treated out-of-hospital by a prehospital medical service from January 2010 to December 2015 and transferred to a trauma center in Madrid. Interventions: None. Main variables of interest: Obtaining and comparing the ROC curve metric of 4 MLAs: random forest, support vector machine, gradient boosting machine and neural network with the results obtained with traditional prediction scales. Results: The different MLA reached ROC values higher than 0.85, having medians close to 0.98. We found no significant differences between MLAs. Each MLA offers a different set of more important variables with a predominance of hemodynamic, resuscitation variables and neurological impairment. Conclusions: MLA may be helpful in patients with massive hemorrhage by outperforming traditional prediction scales. (AU)

Could machine learning algorithms help us predict massive bleeding at prehospital level?

OBJECTIVE: Comparison of the predictive ability of various machine learning algorithms (MLA) versus traditional prediction scales (TPS) for massive hemorrhage (MH) in patients with severe traumatic injury (STI). DESIGN: On a database of a retrospective cohort with prehospital clinical variables and MH outcome, a treatment of the database was performed to be able to apply the different AML, obtaining a total set of 473 patients (80% training, 20% validation). For modeling, proportional imputation and cross validation were performed. The predictive power was evaluated with the ROC metric and the importance of the variables using the Shapley values. SETTING: Out-of-hospital care of patients with STI. PARTICIPANTS: Patients with STI treated out-of-hospital by a out-of-hospital medical service from January 2010 to December 2015 and transferred to a trauma center in Madrid. INTERVENTIONS: None. MAIN VARIABLES OF INTEREST: Obtaining and comparing the "Receiver Operating Characteristic curve" (ROC curve) metric of four MLAs: "random forest" (RF), "vector support machine" (SVM), "gradient boosting machine" (GBM) and "neural network" (NN) with the results obtained with TPS. RESULTS: The different AML reached ROC values higher than 0.85, having medians close to 0.98. We found no significant differences between AMLs. Each AML offers a different set of more important variables with a predominance of hemodynamic, resuscitation variables and neurological impairment. CONCLUSIONS: MLA may be helpful in patients with HM by outperforming TPS.

Alternative AKT2 splicing produces protein lacking the hydrophobic motif regulatory region.

Three AKT serine/threonine kinase isoforms (AKT1/AKT2/AKT3) mediate proliferation, metabolism, differentiation and anti-apoptotic signals. AKT isoforms are activated downstream of PI3-kinase and also by PI3-kinase independent mechanisms. Mutations in the lipid phosphatase PTEN and PI3-kinase that increase PIP3 levels increase AKT signaling in a large proportion of human cancers. AKT and other AGC kinases possess a regulatory mechanism that relies on a conserved hydrophobic motif (HM) C-terminal to the catalytic core. In AKT, the HM is contiguous to the serine 473 and two other newly discovered (serine 477 and tyrosine 479) regulatory phosphorylation sites. In AKT genes, this regulatory HM region is encoded in the final exon. We identified a splice variant of AKT2 (AKT2-13a), which contains an alternative final exon and lacks the HM regulatory site. We validated the presence of mRNA for this AKT2-13a splice variant in different tissues, and the presence of AKT2-13a protein in extracts from HEK293 cells. When overexpressed in HEK293 cells, AKT2-13a is phosphorylated at the activation loop and at the zipper/turn motif phosphorylation sites but has reduced specific activity. Analysis of the human transcriptome corresponding to other AGC kinases revealed that all three AKT isoforms express alternative transcripts lacking the HM regulatory motif, which was not the case for SGK1-3, S6K1-2, and classical, novel and atypical PKC isoforms. The transcripts of splice variants of Akt1-3 excluding the HM regulatory region could lead to expression of deregulated forms of AKT.

USP19 deubiquitinates EWS-FLI1 to regulate Ewing sarcoma growth.

Ewing sarcoma is the second most common pediatric bone and soft tissue tumor presenting with an aggressive behavior and prevalence to metastasize. The diagnostic translocation t(22;11)(q24;12) leads to expression of the chimeric oncoprotein EWS-FLI1 which is uniquely expressed in all tumor cells and maintains their survival. Constant EWS-FLI1 protein turnover is regulated by the ubiquitin proteasome system. Here, we now identified ubiquitin specific protease 19 (USP19) as a regulator of EWS-FLI1 stability using an siRNA based screening approach. Depletion of USP19 resulted in diminished EWS-FLI1 protein levels and, vice versa, upregulation of active USP19 stabilized the fusion protein. Importantly, stabilization appears to be specific for the fusion protein as it could not be observed neither for EWSR1 nor for FLI1 wild type proteins even though USP19 binds to the N-terminal EWS region to regulate deubiquitination of both EWS-FLI1 and EWSR1. Further, stable shUSP19 depletion resulted in decreased cell growth and diminished colony forming capacity in vitro, and significantly delayed tumor growth in vivo. Our findings not only provide novel insights into the importance of the N-terminal EWSR1 domain for regulation of fusion protein stability, but also indicate that inhibition of deubiquitinating enzyme(s) might constitute a novel therapeutic strategy in treatment of Ewing sarcoma.

Validation of extracellular ligand-receptor interactions by Flow-TriCEPS.

OBJECTIVE: The advent of ligand-based receptor capture methodologies, allows the identification of unknown receptor candidates for orphan extracellular ligands. However, further target validation can be tedious, laborious and time-consuming. Here, we present a methodology that provides a fast and cost-efficient alternative for candidate target verification on living cells. RESULTS: In the described methodology a ligand of interest (e.g. transferrin, epidermal growth factor or insulin) was conjugated to a linker (TriCEPS) that carries a biotin. To confirm ligand/receptor interactions, the ligand-TriCEPS conjugates were first added onto living cells and cells were subsequently labeled with a streptavidin-fluorophore and analyzed by flow cytometry (thus referred as Flow-TriCEPS). Flow-TriCEPS was also used to validate identified receptor candidates when combined with a siRNA knock down approach (i.e. reduction of expression levels). This approach is versatile as it can be applied for different classes of ligands (proteins, peptides, antibodies) and different cell lines. Moreover, the method is time-efficient since it takes advantage of the large variety of commercially available (and certified) siRNAs.

Proteasomal Degradation of the EWS-FLI1 Fusion Protein Is Regulated by a Single Lysine Residue.

E-26 transformation-specific (ETS) proteins are transcription factors directing gene expression through their conserved DNA binding domain. They are implicated as truncated forms or interchromosomal rearrangements in a variety of tumors including Ewing sarcoma, a pediatric tumor of the bone. Tumor cells express the chimeric oncoprotein EWS-FLI1 from a specific t(22;11)(q24;12) translocation. EWS-FLI1 harbors a strong transactivation domain from EWSR1 and the DNA-binding ETS domain of FLI1 in the C-terminal part of the protein. Although Ewing cells are crucially dependent on continuous expression of EWS-FLI1, its regulation of turnover has not been characterized in detail. Here, we identify the EWS-FLI1 protein as a substrate of the ubiquitin-proteasome system with a characteristic polyubiquitination pattern. Using a global protein stability approach, we determined the half-life of EWS-FLI1 to lie between 2 and 4 h, whereas full-length EWSR1 and FLI1 were more stable. By mass spectrometry, we identified two ubiquitin acceptor lysine residues of which only mutation of Lys-380 in the ETS domain of the FLI1 part abolished EWS-FLI1 ubiquitination and stabilized the protein posttranslationally. Expression of this highly stable mutant protein in Ewing cells while simultaneously depleting the endogenous wild type protein differentially modulates two subgroups of target genes to be either EWS-FLI1 protein-dependent or turnover-dependent. The majority of target genes are in an unaltered state and cannot be further activated. Our study provides novel insights into EWS-FLI1 turnover, a critical pathway in Ewing sarcoma pathogenesis, and lays new ground to develop novel therapeutic strategies in Ewing sarcoma.

PI3K/AKT signaling modulates transcriptional expression of EWS/FLI1 through specificity protein 1.

Ewing sarcoma (ES) is the second most frequent bone cancer in childhood and is characterized by the presence of the balanced translocation t(11;22)(q24;q12) in more than 85% of cases, generating a dysregulated transcription factor EWS/FLI1. This fusion protein is an essential oncogenic component of ES development which is necessary for tumor cell maintenance and represents an attractive therapeutic target. To search for modulators of EWS/FLI1 activity we screened a library of 153 targeted compounds and identified inhibitors of the PI3K pathway to directly modulate EWS/FLI1 transcription. Surprisingly, treatment of four different ES cell lines with BEZ235 resulted in down regulation of EWS/FLI1 mRNA and protein by ~50% with subsequent modulation of target gene expression. Analysis of the EWS/FLI1 promoter region (-2239/+67) using various deletion constructs identified two 14 bp minimal elements as being important for EWS/FLI1 transcription. We identified SP1 as modulator of EWS/FLI1 gene expression and demonstrated direct binding to one of these regions in the EWS/FLI1 promoter by EMSA and ChIP experiments. These results provide the first insights on the transcriptional regulation of EWS/FLI1, an area that has not been investigated so far, and offer an additional molecular explanation for the known sensitivity of ES cell lines to PI3K inhibition.

PLK1 phosphorylates PAX3-FOXO1, the inhibition of which triggers regression of alveolar Rhabdomyosarcoma.

Pediatric tumors harbor very low numbers of somatic mutations and therefore offer few targets to improve therapeutic management with targeted drugs. In particular, outcomes remain dismal for patients with metastatic alveolar rhabdomyosarcoma (aRMS), where the chimeric transcription factor PAX3/7-FOXO1 has been implicated but problematic to target. In this report, we addressed this challenge by developing a two-armed screen for druggable upstream regulatory kinases in the PAX3/7-FOXO1 pathway. Screening libraries of kinome siRNA and small molecules, we defined PLK1 as an upstream-acting regulator. Mechanistically, PLK1 interacted with and phosphorylated PAX3-FOXO1 at the novel site S503, leading to protein stabilization. Notably, PLK1 inhibition led to elevated ubiquitination and rapid proteasomal degradation of the PAX3-FOXO1 chimeric oncoprotein. On this basis, we embarked on a preclinical validation of PLK1 as a target in a xenograft mouse model of aRMS, where the PLK1 inhibitor BI 2536 reduced PAX3-FOXO1-mediated gene expression and elicited tumor regression. Clinically, analysis of human aRMS tumor biopsies documented high PLK1 expression to offer prognostic significance for both event-free survival and overall survival. Taken together, these preclinical studies validate the PLK1-PAX3-FOXO1 axis as a rational target to treat aRMS.

Molecular mechanism of regulation of the atypical protein kinase C by N-terminal domains and an allosteric small compound.

Protein kinases play important regulatory roles in cells and organisms. Therefore, they are subject to specific and tight mechanisms of regulation that ultimately converge on the catalytic domain and allow the kinases to be activated or inhibited only upon the appropriate stimuli. AGC protein kinases have a pocket in the catalytic domain, the PDK1-interacting fragment (PIF)-pocket, which is a key mediator of the activation. We show here that helix αC within the PIF-pocket of atypical protein kinase C (aPKC) is the target of the interaction with its inhibitory N-terminal domains. We also provide structural evidence that the small compound PS315 is an allosteric inhibitor that binds to the PIF-pocket of aPKC. PS315 exploits the physiological dynamics of helix αC for its binding and allosteric inhibition. The results will support research on allosteric mechanisms and selective drug development efforts against PKC isoforms.

2-(3-Oxo-1,3-diphenylpropyl)malonic acids as potent allosteric ligands of the PIF pocket of phosphoinositide-dependent kinase-1: development and prodrug concept.

The protein kinase C-related kinase 2 (PRK2)-interacting fragment (PIF) pocket of phosphoinositide-dependent kinase-1 (PDK1) was proposed as a novel target site for allosteric modulators. In the present work, we describe the design, synthesis, and structure-activity relationship of a series of 2-(3-oxo-1,3-diphenylpropyl)malonic acids as potent allosteric activators binding to the PIF pocket. Some congeners displayed AC(50) values for PDK1 activation in the submicromolar range. The potency of the best compounds to stabilize PDK1 in a thermal stability shift assay was in the same order of magnitude as that of the PIF pocket binding peptide PIFtide, suggesting comparable binding affinities to the PIF pocket. The crystal structure of PDK1 in complex with compound 4h revealed that additional ionic interactions are mainly responsible for the increased potency compared to the monocarboxylate analogues. Notably, several compounds displayed high selectivity for PDK1. Employing a prodrug strategy, we were able to corroborate the novel mechanism of action in cells.

Substrate-selective inhibition of protein kinase PDK1 by small compounds that bind to the PIF-pocket allosteric docking site.

The PIF-pocket of AGC protein kinases participates in the physiologic mechanism of regulation by acting as a docking site for substrates and as a switch for the transduction of the conformational changes needed for activation or inhibition. We describe the effects of compounds that bind to the PIF-pocket of PDK1. In vitro, PS210 is a potent activator of PDK1, and the crystal structure of the PDK1-ATP-PS210 complex shows that PS210 stimulates the closure of the kinase domain. However, in cells, the prodrug of PS210 (PS423) acts as a substrate-selective inhibitor of PDK1, inhibiting the phosphorylation and activation of S6K, which requires docking to the PIF-pocket, but not affecting PKB/Akt. This work describes a tool to study the dynamics of PDK1 activity and a potential approach for drug discovery.

Use of a fluorescent ATP analog to probe the allosteric conformational change in the active site of the protein kinase PDK1.

There is growing interest in exploring allosteric sites on proteins for drug discovery. At the center of the regulation of many protein kinases from the AGC family there is an allosteric site termed "PIF-pocket." The regulated binding of a C-terminal region of the kinase to the PIF-pocket, within the small lobe of the catalytic core, modulates the activity of AGC kinases. Small compounds that bind to the PIF-pocket can mimic its physiological mechanism of regulation and modulate the kinase activity in vitro, e.g., small compounds can activate the phosphoinositide-dependent protein kinase 1 (PDK1). Compounds binding to an allosteric site on a protein kinase may produce conformational changes at the ATP-binding site within the active site of the kinase domain. We here describe a fluorescent method using the ATP analog TNP-ATP that allows evaluating the allosteric conformational changes at the ATP-binding site of PDK1 triggered by small compounds binding to the PIF-pocket.

Allosteric regulation of protein kinase PKCζ by the N-terminal C1 domain and small compounds to the PIF-pocket.

Protein kinases are key mediators of cellular signaling, and therefore, their activities are tightly controlled. AGC kinases are regulated by phosphorylation and by N- and C-terminal regions. Here, we studied the molecular mechanism of inhibition of atypical PKCζ and found that the inhibition by the N-terminal region cannot be explained by a simple pseudosubstrate inhibitory mechanism. Notably, we found that the C1 domain allosterically inhibits PKCζ activity and verified an allosteric communication between the PIF-pocket of atypical PKCs and the binding site of the C1 domain. Finally, we developed low-molecular-weight compounds that bind to the PIF-pocket and allosterically inhibit PKCζ activity. This work establishes a central role for the PIF-pocket on the regulation of PKCζ and allows us to envisage development of drugs targeting the PIF-pocket that can either activate or inhibit AGC kinases.

4-benzimidazolyl-3-phenylbutanoic acids as novel PIF-pocket-targeting allosteric inhibitors of protein kinase PKCζ.

Protein kinase inhibitors with an allosteric mode of action are expected to reach, in many cases, higher selectivity for the target enzyme than ATP-competitive compounds. Therefore, basic research is aiming at identifying and establishing novel sites on the catalytic domain of protein kinases which might be targeted by allosteric inhibitors. We previously published the first structure-activity relationships (SARs) for allosteric activators of protein kinase PDK1. Here, we present the design, synthesis, and SAR data on a series of novel compounds, 4-benzimidazolyl-3-phenylbutanoic acids, that inhibit the atypical protein kinace C (PKC) ζ via binding to the PIF-pocket. Key positions were identified in the compounds that can be modified to increase potency and selectivity. Some congeners showed a high selectivity toward PKCζ, lacking inhibition of the most closely related isoform, PKCι, and of further AGC kinases. Furthermore, evidence is provided that these compounds are also active toward cellular PKCζ without loss of potency compared to the cell-free assay.

Regulation of the interaction between protein kinase C-related protein kinase 2 (PRK2) and its upstream kinase, 3-phosphoinositide-dependent protein kinase 1 (PDK1).

The members of the AGC kinase family frequently exhibit three conserved phosphorylation sites: the activation loop, the hydrophobic motif (HM), and the zipper (Z)/turn-motif (TM) phosphorylation site. 3-Phosphoinositide-dependent protein kinase 1 (PDK1) phosphorylates the activation loop of numerous AGC kinases, including the protein kinase C-related protein kinases (PRKs). Here we studied the docking interaction between PDK1 and PRK2 and analyzed the mechanisms that regulate this interaction. In vivo labeling of recombinant PRK2 by (32)P(i) revealed phosphorylation at two sites, the activation loop and the Z/TM in the C-terminal extension. We provide evidence that phosphorylation of the Z/TM site of PRK2 inhibits its interaction with PDK1. Our studies further provide a mechanistic model to explain different steps in the docking interaction and regulation. Interestingly, we found that the mechanism that negatively regulates the docking interaction of PRK2 to the upstream kinase PDK1 is directly linked to the activation mechanism of PRK2 itself. Finally, our results indicate that the mechanisms underlying the regulation of the interaction between PRK2 and PDK1 are specific for PRK2 and do not apply for other AGC kinases.

Structure and allosteric effects of low-molecular-weight activators on the protein kinase PDK1.

Protein phosphorylation transduces a large set of intracellular signals. One mechanism by which phosphorylation mediates signal transduction is by prompting conformational changes in the target protein or interacting proteins. Previous work described an allosteric site mediating phosphorylation-dependent activation of AGC kinases. The AGC kinase PDK1 is activated by the docking of a phosphorylated motif from substrates. Here we present the crystallography of PDK1 bound to a rationally developed low-molecular-weight activator and describe the conformational changes induced by small compounds in the crystal and in solution using a fluorescence-based assay and deuterium exchange experiments. Our results indicate that the binding of the compound produces local changes at the target site, the PIF binding pocket, and also allosteric changes at the ATP binding site and the activation loop. Altogether, we present molecular details of the allosteric changes induced by small compounds that trigger the activation of PDK1 through mimicry of phosphorylation-dependent conformational changes.

3,5-Diphenylpent-2-enoic acids as allosteric activators of the protein kinase PDK1: structure-activity relationships and thermodynamic characterization of binding as paradigms for PIF-binding pocket-targeting compounds.

The modulation of protein kinase activities by low molecular weight compounds is a major goal of current pharmaceutical developments. In this line, important efforts are directed to the development of drugs targeting the conserved ATP binding site. However, there is very little experience on targeting allosteric, regulatory sites, different from the ATP binding site, in protein kinases. Here we describe the synthesis, cell-free activation potency, and calorimetric binding analysis of 3,5-diphenylpent-2-enoic acids and derivatives as allosteric modulators of the phosphoinositide-dependent kinase-1 (PDK1) catalytic activity. Our SAR results combined with thermodynamic binding analyses revealed both favorable binding enthalpy and entropy and confirmed the PIF-binding pocket of PDK1 as a druggable site. In conclusion, we defined the minimal structural requirements for compounds to bind to the PIF-binding pocket and to act as allosteric modulators and identified two new lead structures (12Z and 13Z) with predominating binding enthalpy.

Allosteric activation of the protein kinase PDK1 with low molecular weight compounds.

Organisms rely heavily on protein phosphorylation to transduce intracellular signals. The phosphorylation of a protein often induces conformational changes, which are responsible for triggering downstream cellular events. Protein kinases are themselves frequently regulated by phosphorylation. Recently, we and others proposed the molecular mechanism by which phosphorylation at a hydrophobic motif (HM) regulates the conformation and activity of many members of the AGC group of protein kinases. Here we have developed specific, low molecular weight compounds, which target the HM/PIF-pocket and have the ability to allosterically activate phosphoinositide-dependent protein kinase 1 (PDK1) by modulating the phosphorylation-dependent conformational transition. The mechanism of action of these compounds was characterized by mutagenesis of PDK1, synthesis of compound analogs, interaction-displacement studies and isothermal titration calorimetry experiments. Our results raise the possibility of developing drugs that target the AGC kinases via a novel mode of action and may inspire future rational development of compounds with the ability to modulate phosphorylation-dependent conformational transitions in other proteins.