Search

Estimating hospital mortality of patients is important in assisting clinicians to make decisions and hospital providers to allocate resources. This paper proposed a multi-task recurrent neural network with attention mechanisms to predict patients' hospital mortality, using reconstruction of patients' physiological time series as an auxiliary task. Experiments were conducted on a large public electronic health record database, i.e., MIMIC-III. Fifteen physiological measurements during the first 24 h of critical care were used to predict death before hospital discharge. Compared with the conventional simplified acute physiology score (SAPS-II), the proposed multi-task learning model achieved better sensitivity (0.503 ± 0.020 versus 0.365 ± 0.021), when predictions were made based on the same 24-h observation period. The multi-task learning model is recommended to be updated daily with at least a 6-h observation period, in order for it to perform similarly or better than the SAPS-II. In the future, the need for intervention can be considered as another task to further optimize the performance of the multi-task learning model.

Due to the poor self-regenerative capacity of native cartilage, the treatment of articular cartilage defects is a significant clinical challenge. The cartilage tissue engineered scaffolds with high chondrogenic capacity might provide an efficient strategy for the repair of cartilage defects. Herein, we report an aligned porous poly ( l -lactic acid) (PLLA) electrospun fibrous membrane with a biomimetic surface for the rapid regeneration of cartilage tissue. The PLLA electrospun fibers aligned in a single direction and generated ellipse-shaped nanopores in situ onto the surface of the fibers. Subsequently, chondroitin sulfate (CS) was coated on the surface of the fibers using polydopamine (PDA) as an adhesive polymeric bridge. The in vitro results revealed that the aligned porous composite membranes with a biomimetic surface (PL/PDA/CS) could stimulate the attachment and proliferation of chondrocytes and rabbit bone marrow mesenchymal stem cells (rBMSCs), as well as inducing chondrogenic gene expression in rBMSCs. Furthermore, the PL/PDA/CS membranes significantly facilitated the filling of defects and the regeneration of hyaline cartilage-like tissue in vivo. The results suggested that the hierarchical aligned porous fibrous electrospun membrane with a biomimetic surface modification is a potential therapeutic strategy for highly efficient cartilage defect repair.

The diabetic non-healing wound is one of the most common complications of diabetics. The long-term stimulus of oxidative stress, inflammation and infection caused by the hyperglycemic microenvironment in the wound site always leads to a delayed healing process of the diabetic wound. To address this issue, in this study, we prepared an asiatic acid (AA)-embedded aligned porous poly(l-lactic acid) (PLLA) electrospun fibrous scaffold (AA-PL) for accelerating diabetic wound healing. The results showed that the electrospun fibers with nanopores on the surfaces were aligned in a single direction, while the AA was well embedded in the fibers and can be continuously released from them. The in vitro results revealed that the AA-PL scaffolds can effectively alleviate the H2O2-induced oxidative stress damage to HaCat cells and downregulate the LPS-induced pro-inflammatory cytokine (IL-1β, TNF-α, IL6) gene expression in RAW 264.7 macrophage cells. Moreover, the growth of E. coli and S. aureus could be inhibited by the AA-PL scaffolds. The in vivo study further demonstrated that the AA-PL scaffolds can accelerate the re-epithelization, angiogenesis and extracellular matrix formation of a wound by relieving the high oxidative stress, inflammation and infection in the diabetic wound site. This study suggests that the combination of hierarchical structures (nanopores on the aligned fibers) with the controllable release of AA from the scaffolds is an efficient and innovative strategy for the treatment of diabetic non-healing wounds.

Here, we present a new series of hydrazide-bearing class I selective HDAC inhibitors designed based on panobinostat. The cap, linker, and zinc-binding group were derivatized to improve HDAC affinity and antileukemia efficacy. Lead inhibitor 13a shows picomolar or low nanomolar IC50 values against HDAC1 and HDAC3 and exhibits differential toxicity profiles toward multiple cancer cells with different FLT3 and p53 statuses. 13a indirectly inhibits the FLT3 signaling pathway and down-regulates master antiapoptotic proteins, resulting in the activation of pro-caspase3 in wt-p53 FLT3-ITD MV4-11 cells. While in the wt-FLT3 and p53-null cells, 13a is incapable of causing apoptosis at a therapeutic concentration. The MDM2 antagonist and the proteasome inhibitor promote 13a-triggered apoptosis by preventing p53 degradation. Furthermore, we demonstrate that apoptosis rather than autophagy is the key contributing factor for 13a-triggered cell death. When compared to panobinostat, 13a is not mutagenic and displays superior in vivo bioavailability and a higher AUC0-inf value.

NLRP3 inflammasome plays critical roles in a variety of human diseases and represents a promising drug target. In this study, we established the in vivo functional activities of JC124, a previously identified NLRP3 inflamma-some inhibitor from our group, in mouse models of Alzheimer’s disease and acute myocardial infarction. To understand the chemical space of this lead structure, a series of analogues were designed, synthesized, and biologically characterized. The results revealed the critical roles of the two substituents on the benzamide moiety of JC124. On the other hand, modifications on the sulfonamide moiety of JC124 are well tolerated. Two new lead compounds, 14 and 17, were identified with improved inhibitory potency (IC(50) values of 0.55 ± 0.091 and 0.42 ± 0.080 μM, respectively). Further characterization confirmed their selectivity and in vivo target engagement. Collectively, the results strongly encourage further development of more potent analogues based on this chemical scaffold.

In our continuing efforts to develop novel neuroprotectants for Alzheimer’s disease (AD), a series of analogs based on a lead compound that was recently shown to target the mitochondrial complex I were designed, synthesized and biologically characterized to understand the structure features that are important for neuroprotective activities. The results from a cellular AD model highlighted the important roles of the 4-OH on the phenyl ring and the 5-OCH(3) on the indole ring of the lead compound. The results also demonstrated that the β-keto moiety can be modified to retain or improve the neuroprotective activity. Docking studies of selected analogs to the FMN site of mitochondrial complex I also supported the observed neuroprotective activities. Collectively, the results provide further information to guide optimization and development of analogs based on this chemical scaffold as neuroprotectants with a novel mechanism of action for AD.

In our efforts to develop novel small-molecule inhibitors for the NOD-like receptor family pyrin-domain-containing 3 (NLRP3) inflammasome as potential disease-modifying agents to treat neurological disorders including multiple sclerosis (MS), a hydroxyl sulfonamide analogue JC-171 has been rationally designed and biologically characterized both in vitro and in vivo. Our studies established that JC-171 dose dependently inhibited LPS/ATP-induced interleukin-1β (IL-1β) release from J774A.1 macrophages with an IC50 of 8.45 ± 1.56 μM. Selective inhibition of the NLRP3 inflammasome induced IL-1β release by this compound was also confirmed using mouse bone-marrow-derived macrophages and LPS-challenged mice in vivo. Furthermore, immunoprecipitation study revealed that JC-171 interfered with NLRP3/ASC interaction induced by LPS/ATP stimulation. More importantly, JC-171 treatment delayed the progression and reduced the severity of experimental autoimmune encephalomyelitis (EAE), a mouse model of MS, in both prophylactic and therapeutic settings. This coincided with blocking of IL-1β production and a pathogenic Th17 response. Collectively, these results suggest that JC-171 is a selective NLRP3 inflammasome inhibitor with biological activity in vivo, thus strongly encouraging further development of this lead compound as a potential therapeutic agent for human MS.

We have designed a deep-learning model, an “Artificial Intelligent Endoscopist (a.k.a. AI-doscopist)”, to localise colonic neoplasia during colonoscopy. This study aims to evaluate the agreement between endoscopists and AI-doscopist for colorectal neoplasm localisation. AI-doscopist was pre-trained by 1.2 million non-medical images and fine-tuned by 291,090 colonoscopy and non-medical images. The colonoscopy images were obtained from six databases, where the colonoscopy images were classified into 13 categories and the polyps’ locations were marked image-by-image by the smallest bounding boxes. Seven categories of non-medical images, which were believed to share some common features with colorectal polyps, were downloaded from an online search engine. Written informed consent were obtained from 144 patients who underwent colonoscopy and their full colonoscopy videos were prospectively recorded for evaluation. A total of 128 suspicious lesions were resected or biopsied for histological confirmation. When evaluated image-by-image on the 144 full colonoscopies, the specificity of AI-doscopist was 93.3%. AI-doscopist were able to localise 124 out of 128 polyps (polyp-based sensitivity = 96.9%). Furthermore, after reviewing the suspected regions highlighted by AI-doscopist in a 102-patient cohort, an endoscopist has high confidence in recognizing four missed polyps in three patients who were not diagnosed with any lesion during their original colonoscopies. In summary, AI-doscopist can localise 96.9% of the polyps resected by the endoscopists. If AI-doscopist were to be used in real-time, it can potentially assist endoscopists in detecting one more patient with polyp in every 20–33 colonoscopies.

NLRP3 inflammasomes have recently emerged as an attractive drug target for neurodegenerative disorders. In our continuing studies, a new chemical scaffold was designed as selective inhibitors of NLRP3 inflammasomes. Initial characterization of the lead HL16 demonstrated improved, however, nonselective inhibition on the NLRP3 inflammasome. Structure-activity relationship studies of HL16 identified a new lead, 17 (YQ128), with an IC50 of 0.30 ± 0.01 μM. Further studies from in vitro and in vivo models confirmed its selective inhibition on the NLRP3 inflammasome and its brain penetration. Furthermore, pharmacokinetic studies in rats at 20 mg/kg indicated extensive systemic clearance and tissue distribution, leading to a half-life of 6.6 h. However, the oral bioavailability is estimated to be only 10%, which may reflect limited GI permeability and possibly high first-pass effects. Collectively, these findings strongly encourage development of more potent analogues with improved pharmacokinetic properties from this new chemical scaffold.

Background Neuroinflammation is an essential player in many neurological diseases including traumatic brain injury (TBI). Recent studies have identified that inflammasome complexes are responsible for inflammatory responses in many pathological conditions. Inflammasomes are intracellular multiprotein complexes which regulate the innate immune response, activation of caspase-1, production of pro-inflammatory cytokines IL-1β and IL-18, and induction of cell death (pyroptosis). Among inflammasome family members, the nucleotide-binding domain leucine-rich repeats family protein 3 (NLRP3) is the most extensively studied and its activation is induced following TBI. As a novel target, drug development targeting the formation and activation of NLRP3 inflammasome is a prospective therapy for TBI. We have recently developed a small molecule JC124 with specificity on NLRP3 inflammasome. In this study, we explored the therapeutic value of JC124 for TBI treatment. Methods Adult male Sprague-Dawley rats were subjected to a moderate cortical impact injury. Following TBI, animals received 4 doses of JC124 treatment with the first dose starting at 30 min, the second dose at 6 h after TBI, the third and fourth doses at 24 or 30 h following TBI, respectively. Animals were sacrificed at 2 days post-injury. Brain tissues were processed either for ELISA and western blotting analysis for inflammatory response, or for histological examination to assess degenerative neurons, acute inflammatory cell response and lesion volume. Results We found that post-injury treatment with JC124 significantly decreased the number of injury-induced degenerating neurons, inflammatory cell response in the injured brain, and cortical lesion volume. Injured animals treated with JC124 also had significantly reduced protein expression levels of NLRP3, ASC, IL-1 beta, TNFα, iNOS, and caspase-1. Conclusion Our data suggest that our novel NLRP3 inhibitor has a specific anti-inflammatory effect to protect the injured brain following TBI.

Curcumin and two bivalent compounds, namely 17MD and 21MO, both obtained by conjugation of curcumin with a steroid molecule that acts as a membrane anchor, were comparatively studied. When incorporated into 1,2-dipalmitoyl-sn-glycero-3-phosphocholine the compounds showed a very limited solubility in the model membranes. Curcumin and the two bivalent compounds were also incorporated in membranes of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine and quenching the fluorescence of pure curcumin or of the curcumin moiety in the bivalent compounds by acrylamide it was seen that curcumin was accessible to this water soluble quencher but the molecule was somehow located in a hydrophobic environment. This was confirmed by quenching with doxyl-phosphatidylcholines, indicating that the curcumin moieties of 17MD and 21MO were in a more polar environment than pure curcumin itself. NOESY-MAS-(1)HNMR analysis supports this notion by showing that the orientation of curcumin was parallel to the plane of the membrane surface close to C2 and C3 of the fatty acyl chains, while the curcumin moiety of 17MD and 21MO positioned close to the polar part of the membrane with the steroid moiety in the centre of the membrane. Molecular dynamics studies were in close agreement with the experimental results with respect to the likely proximity of the protons studied by NMR and show that 17MD and 21MO have a clear tendency to aggregate in a fluid membrane. The anchorage of the bivalent compounds to the membrane leaving the curcumin moiety near the polar part may be very important to facilitate the bioactivity of the curcumin moiety when used as anti-Alzheimer drugs.

Algorithms for localising colorectal polyps have been studied extensively; however, they were often trained and tested using the same database. In this study, we present a new application of a unified and real-time object detector based on You-Only-Look-Once (YOLO) convolutional neural network (CNN) for localizing polyps with bounding boxes in endoscopic images. The model was first pre-trained with non-medical images and then fine-tuned with colonoscopic images from three different databases, including an image set we collected from 106 patients using narrow-band (NB) imaging endoscopy. YOLO was tested on 196 white light (WL) images of an independent public database. YOLO achieved a precision of 79.3% and sensitivity of 68.3% with time efficiency of 0.06 sec/frame in the localization task when trained by augmented images from multiple WL databases. In conclusion, YOLO has great potential to be used to assist endoscopists in localising colorectal polyps during endoscopy. CNN features of WL and NB endoscopic images are different and should be considered separately. A large-scale database that covers different scenarios, imaging modalities and scales is lacking but crucial in order to bring this research into reality.

A chronic wound in diabetic patients is usually characterized by poor angiogenesis and delayed wound closure. The exploration of efficient strategy to significantly improve angiogenesis in the diabetic wound bed and thereby accelerate wound healing is still a significant challenge. Herein, we reported a kind of aligned porous poly ( l -lactic acid) (P l LA) electrospun fibrous membranes containing dimethyloxalylglycine (DMOG)-loaded mesoporous silica nanoparticles (DS) for diabetic wound healing. The P l LA electrospun fibers aligned in a single direction and there were ellipse-shaped nano-pores in situ generated onto the surface of fibers, while the DS were well distributed in the fibers and the DMOG as well as Si ion could be controlled released from the nanopores on the fibers. The in vitro results revealed that the aligned porous composite membranes (DS-PL) could stimulate the proliferation, migration and angiogenesis-related gene expression of human umbilical vein endothelial cells (HUVECs) compared with the pure P l LA membranes. The in vivo study further demonstrated that the prepared DS-PL membranes significantly improved neo-vascularization, re-epithelialization and collagen formation as well as inhibited inflammatory reaction in the diabetic wound bed, which eventually stimulated the healing of the diabetic wound. Collectively, these results suggest that the combination of hierarchical structures (nanopores on the aligned fibers) with the controllable released DMOG drugs as well as Si ions from the membranes, which could create a synergetic effect on the rapid stimulation of angiogenesis in the diabetic wound bed, is a potential novel therapeutic strategy for highly efficient diabetic wound healing. Statement of Significance A chronic wound in diabetic patients is usually characterized by the poor angiogenesis and the delayed wound closure. The main innovation of this study is to design a new kind of skin tissue engineered scaffold, aligned porous poly ( l -lactic acid) (P l LA) electrospun membranes containing dimethyloxalylglycine (DMOG)-loaded mesoporous silica nanoparticles (DS), which could significantly improve angiogenesis in the diabetic wound bed and thereby accelerate diabetic wound healing. The results revealed that the electrospun fibers with ellipse-shaped nano-pores on the surface were aligned in a single direction, while there were DS particles distributed in the fibers and the DMOG as well as Si ions could be controllably released from the nanopores on the fibers. The in vitro studies demonstrated that the hierarchical nanostructures (nanopores on the aligned fibers) and the controllable released chemical active agents (DMOG drugs and Si ions) from the DS-PL membranes could exert a synergistic effect on inducing the endothelial cell proliferation, migration and differentiation. Above all, the scaffolds distinctly induced the angiogenesis, collagen deposition and re-epithelialization as well as inhibited inflammation reaction in the wound sites, which eventually stimulated the healing of diabetic wounds in vivo. The significance of the current study is that the combination of the hierarchical aligned porous nanofibrous structure with DMOG-loaded MSNs incorporated in electrospun fibers may suggest a high-efficiency strategy for chronic wound healing.

Previously a novel ubenimex-fluorouracil (5-FU) conjugate, BC-01 was identified and validated as a potent CD13 inhibitor with marked in vitro and in vivo antitumor potency. Herein, further structural modifications of the linker part of BC-01 was carried out to get more potent and stable ubenimex–fluorouracil conjugates. It was striking that most of these conjugates showed even more potent CD13 inhibitory activities than BC-01 and the approved CD13 inhibitor ubenimex. One representative compound 12a displayed significant in vitro anti-proliferation, pro-apoptosis, anti-metastasis, anti-angiogenesis and CD13+ cell elimination effects. In vitro stability and in vivo pharmacokinetic study revealed that compound 12a could release ubenimex and 5-FU slowly, which could act as a mutual prodrug of ubenimex and 5-FU. Compared with 5-FU or 5-FU plus ubenimex, 12a exhibited superior in vivo antitumor growth efficiency, even in our mice model of 5-FU-resistant liver cancer. Moreover, 12a exhibited more potent in vivo anti-metastasis and lifespan extension effects compared to the approved 5-FU prodrug capecitabine. Collectively, these results suggest that further optimization and evaluation of 12a as a promising anticancer candidate are warranted to develop effective therapeutic agents for human liver cancer.

Pharmacological inhibition of histone deacetylases (HDACs) has been successfully applied in the treatment of a wide range of disorders, including Parkinson's disease, infection, cardiac diseases, inflammation, and especially cancer. HDAC inhibitors (HDACIs) have been proved to be effective antitumor agents by various stages of investigation. At present, there are two opposite focuses of HDACI design in the cancer therapy, highly selective inhibitor strategy and dual- or multitargeted inhibitors. The former method, which is supposed to elucidate the function of individual HDAC and provide candidate inhibitors with fewer side effects, has been widely accepted by the inhibitor developer. The latter approach, though less practiced, has promising potential for the antitumor therapy based on HDACIs. Effective HDACIs, some of which are in clinic anticancer research, have been developed by both methods. In order to gain insight into HDACI design, the strategies and achievements of the two diverse methods are reviewed.

Histone deacetylases are a class of enzymes that play important roles in post translational modifications of histones by deacetylating the lysine residues as well as interacting with various non-histone proteins. This type of enzymes is closely related to oncogenesis and has been proved to be attractive targets for designing novel anti-cancer agents. Over the last 10 years, a large number of HDACs have entered pre-clinical and/or clinical trials. Among these drug candidates, the pan-HDAC inhibitor, panobinostat demonstrated high therapeutic potential as monotherapy and combined therapy in both preclinical models and clinical cancer patients. In this review, we have mainly focused on the recent progress of the clinical studies about panobinostat, and discussed its anti-cancer effects and molecular rationale for the treatment strategies.

Herein, a novel mutual prodrug BC-A1 was discovered by integrating ubenimex and gemcitabine into one molecule. Biological characterization revealed that compound BC-A1 could maintain both the anti-CD13 activity of ubenimex and the cytotoxic activity of gemcitabine in vitro. Further characterization also demonstrated that compound BC-A1 exhibited significant anti-invasion and anti-angiogenesis effects in vitro. The preliminary stability test of BC-A1 revealed that it could release gemcitabine in vitro. The in vivo anti-tumor results in liver cancer showed that at the same dosage, oral administration of BC-A1 was as potent as intraperitoneal administration of gemcitabine. This warranted the further research and development of the orally active prodrug BC-A1 because gemcitabine can not be orally administrated in clinic.

We designed and synthesized a novel mutual prodrug, named BC-01 (3), by integrating ubenimex and Fluorouracil (5-FU) into one molecule based on prior research results that showed that a combination of the aminopeptidase N (CD13) inhibitor, ubenimex, and the cytotoxic antitumor agent, 5-FU, exhibited improved in vitro and in vivo antitumor efficiency. 3 showed potent inhibitory activity against CD13 enzymatic activity. Compared with ubenimex, 3 exhibited more potent anti-angiogenesis effects, and compared with the approved 5-FU prodrug, capecitabine, 3 exhibited more potent tumor growth inhibitory and anti-metastasis effects. Additionally, compared with 5-FU or 5-FU plus ubenimex, 3 also exhibited a superior antitumor efficiency even in our 5-FU-resistant mice model. Other antitumor agents could be conjugated with ubenimex using this strategy to obtain novel mutual prodrugs with promising antitumor potency.

In our efforts of developing novel compounds as potential anticancer agents, a series of tamibarotene analogs containing Zn(2+) -binding moieties were designed and developed. Biological characterization identified compound 7b as the most potent one with improved antiproliferative activities against multiple cancer cell lines, compared to parent compound tamibarotene. Further characterization also demonstrated that compound 7b exhibited moderate activities as a histone deacetylase inhibitor with IC50 of 1.8 ± 0.1 μm, thus suggesting that this could contribute to the improved antiproliferative activities of 7b. Pharmacokinetic studies revealed that compound 7b could release tamibarotene after administration and prolong the circulation time of tamibarotene, and this may also potentially contribute to the improved antiproliferative activities. Collectively, the results demonstrated that compound 7b could serve as a new lead for further development of more potent analogs as potential anticancer agents.

Histone deacetylases (HDACs) inhibitors have multiple effects targeting the cancer cells and have become one of the promising cancer therapeutics with possibly broad applicability. Combination of HDAC inhibitors with the cytotoxic fluorouracil (5-FU) showed additive and synergistic effects both in vitro and in vivo. To explore the possibility in cancer therapy of a bivalent agent that combines two bioactive groups within a single molecular architecture, we designed and synthesized new dual-acting compounds by combining the bioactive fragment of MS-275, a clinical HDACs inhibitor, with cytotoxic agent 5-FU. The target compounds 9a and 9b showed comparable HDACs inhibition with MS-275 and moderate antiproliferative acitivities against six cancer cells lines.

Tamibarotene (Am80), a poorly water-soluble drug for the treatment of acute promyelocytic leukemia (APL), loaded nanostructured lipid carrier (Am80-NLC) was developed and characterized previously. The purpose of the present work was to develop PEGylated nanostructured lipid carrier (PEG-NLC) for intravenous delivery of Am80, with the aim to further extend the circulation in blood and decrease the adverse events. Am80-loaded PEG-NLC (Am80-PEG-NLC) modified with PEG-40 stearate (PEG40-SA, molecular weight 2000 Da) was formulated by the method of melt-emulsification and low temperature-solidification technique. Am80-NLC was developed as well as control. Based on the optimized results of single-factor screening experiment, the average drug entrapment efficiency, the mean particle size, and zeta potential of Am80-NLC and Am80-PEG-NLC were found to be 89.8-94.3%, 178.9-201.6 nm, and -37.74 to -20.1 mV, respectively. In vitro drug release of Am80-NLC and Am80-PEG-NLC possessed a sustained release characteristic and their release behavior was in accordance with the Ritger-Peppas equation. In vivo, after intravenous (i.v.) injection to rats, the mean residence time (MRT) of Am80-PEG-NLC group was significantly prolonged and the AUC value was improved as well compared with the Am80-NLC group. Furthermore, the biodistribution in mice showed that Am80-PEG-NLC preferentially decreased the accumulation of Am80 in kidney and increased the drug concentration in brain after i.v. injection. In conclusion, Am80-PEG-NLC may be a potential delivery system for Am80 in the treatment of APL.