Over the past decade, deep neural networks (DNNs) have demonstrated remarkable performance in a variety of applications. As we try to solve more advanced problems, increasing demands for computing and power resources has become inevitable. Spiking neural networks (SNNs) have attracted widespread interest as the third-generation of neural networks due to their event-driven and low-powered nature. SNNs, however, are difficult to train, mainly owing to their complex dynamics of neurons and non-differentiable spike operations. Furthermore, their applications have been limited to relatively simple tasks such as image classification. In this study, we investigate the performance degradation of SNNs in a more challenging regression problem (i.e., object detection). Through our in-depth analysis, we introduce two novel methods: channel-wise normalization and signed neuron with imbalanced threshold, both of which provide fast and accurate information transmission for deep SNNs. Consequently, we present a first spiked-based object detection model, called Spiking-YOLO. Our experiments show that Spiking-YOLO achieves remarkable results that are comparable (up to 98%) to those of Tiny YOLO on non-trivial datasets, PASCAL VOC and MS COCO. Furthermore, Spiking-YOLO on a neuromorphic chip consumes approximately 280 times less energy than Tiny YOLO and converges 2.3 to 4 times faster than previous SNN conversion methods.

BackgroundBovine pericardium is one of the most widely used materials in bioprosthetic heart valves. Immunologic responses have been implicated as potential causes of limited durability of xenogenic valves. This study aimed to determine the effectiveness of decellularization and α-galactosidase (α-gal) to remove major xenoreactive antigens from xenogenic tissues.Materials and MethodsRecombinant Bacteroides thetaiotaomicron (B. thetaiotaomicron) α-gal or decellularization, or both were used to remove α-gal from bovine pericardium. It was confirmed by α-gal-bovine serum albumin-based enzyme-linked immunosorbent assay (ELISA), high-performance anion exchange chromatography, flow cytometry, 3,3'-diaminobenzidine-staining, and lectin-based ELISA. The mechanical properties of bovine pericardium after decellularization or α-gal treatment were investigated by tests of tensile-strength, permeability, and compliance. Collagen fiber rearrangement was also evaluated by a 20,000× transmission electron microscope (TEM).ResultsRecombinant B. thetaiotaomicron α-gal could effectively remove α-gal from bovine pericardium B. thetaiotaomicron (0.1 U/mL, pH 7.2) while recombinant human α-gal removed it recombinant human α-gal (10 U/mL, pH 5.0). There was no difference in the mechanical properties of fresh and recombinant α-gal-treated bovine pericardium. Furthermore, the TEM findings demonstrated that recombinant α-gal made no difference in the arrangement of collagen fiber bundles with decellularization.ConclusionRecombinant B. thetaiotaomicron α-gal effectively removed α-gal from bovine pericardium with a small amount under physiological conditions compared to human recombinant α-gal, which may alleviate the harmful xenoreactive immunologic responses of α-gal. Recombinant α-gal treatment had no adverse effects on the mechanical properties of bovine pericardium.

Extracellular vesicles (EVs) carrying tumor cell-derived programmed death-ligand 1 (PD-L1) interact with programmed death 1 (PD-1)-producing T cells, thus significantly lowering a patient's response to immune checkpoint blockade drugs. No drug that reinvigorates CD8+ T cells by suppressing EV PD-L1 has been approved for clinical usage. Here we have identified macitentan (MAC), an FDA-approved oral drug, as a robust booster of antitumor responses in CD8+ T cells by suppressing tumor cell-derived EV PD-L1. Methods: EV was analyzed by the data from nanoparticle tracking, immunoblotting analyses, and nano-flow cytometry. Antitumor immunity was evaluated by luciferase assay and immune phenotyping using flow cytometry. Clinical relevance was analyzed using the cancer genome atlas database. Results: MAC inhibited secretion of tumor-derived EV PD-L1 by targeting the endothelin receptor A (ETA) in breast cancer cells and xenograft models. MAC enhanced CD8+ T cell-mediated tumor killing by decreasing the binding of PD-1 to the EV PD-L1 and thus synergizing the effects of the anti-PD-L1 antibody. MAC also showed an anticancer effect in triple-negative breast cancer (TNBC)-bearing immunocompetent mice but not in nude mice. The combination therapy of MAC and anti-PD-L1 antibody significantly improved antitumor efficacy by increasing CD8+ T cell number and activity with decreasing Treg number in the tumors and draining lymph nodes in TNBC, colon, and lung syngeneic tumor models. The antitumor effect of MAC was reversed by injecting exogenous EV PD-L1. Notably, ETA level was strongly associated with the innate anti-PD-1 resistance gene signature and the low response to the PD-1/PD-L1 blockade. Conclusion: These findings strongly demonstrate that MAC, already approved for clinical applications, can be used to improve and/or overcome the inadequate response to PD-1/PD-L1 blockade therapy.

It has been reported that the immune response due to α-Gal epitopes is an important factor in tissue valve failure. The elimination of the interaction between the natural anti-Gal antibodies and α-gal epitopes on the xenografts is a prerequisite to the success of xenografts in humans. Previously, we reported that the green coffee bean α-galactosidase could remove all α-Gal epitopes from cell surface of porcine aortic valve and pericardial tissue, but it has limitations on cost effectiveness. In this study we wanted to know whether the recently produced recombinant human α-galactosidase A has the same effective enzymatic activity as green coffee bean α-galactosidase in removing α-Gal epitopes from the same tissues. After treating fresh porcine aortic valve and pericardial tissue with recombinant α-galactosidase A, each sample was stained with Griffonia simplicifolia type I isolectin B4 indirect immunoperoxidase avidin-biotin technique. We then examined whether the α-Gal epitopes were reduced or abolished in each consecutive concentration of recombinant α-galactosidase A by comparing the degree of the Griffonia simplicifolia isolectin B4 staining. As a result, the recombinant α-galactosidase A could remove cell surface α-Gals on porcine aortic valve and pericardial tissue as effectively as green coffee bean α-galactosidase.