Graves’ orbitopathy (GO) is a disfiguring and sometimes blinding disease, characterised by inflammation and swelling of orbital tissues, with fibrosis and adipogenesis being predominant features. Little is known about the disease aetiology and the molecular mechanisms driving the phenotypic changes in orbital fibroblasts are unknown. Using fibroblasts isolated from the orbital fat of undiseased individuals or GO patients, we have established a novel in vitro model to evaluate the dual profile of GO cells in a three-dimensional collagen matrix; this pseudo-physiological 3D environment allows measurement of their contractile and adipogenic properties. GO cells contracted collagen matrices more efficiently than control cells following serum or TGFβ1 stimulation, and showed a slightly increased ability to proliferate in the 3D matrix, in accordance with a fibro-proliferative phenotype. GO cells, unlike controls, also spontaneously differentiated into adipocytes in 3D cultures - confirming an intrinsic adipogenic profile. However, both control and GO cells underwent adipogenesis when cultured under pathological pressure levels. We further demonstrate that a Thy-1-low population of GO cells underlies the adipogenic - but not the contractile - phenotype and, using inhibitors, confirm that the contractile and adipogenic phenotypes are regulated by separate pathways. In view of the current lack of suitable treatment for GO, we propose that this new model testing the duality of the GO phenotype could be useful as a preclinical evaluation for the efficacy of potential treatments.
Our findings suggest that orbital fibroblasts contain a population of cells that fulfil the criteria defining MSC. This subpopulation may be increased in GO, possibly underlying the complex differentiation phenotype of the disease.
Glaucoma filtration surgery is one of the most effective methods for lowering intraocular pressure in glaucoma. The surgery efficiently reduces intra-ocular pressure but the most common cause of failure is scarring at the incision site. This occurs in the conjunctiva/Tenon’s capsule layer overlying the scleral coat of the eye. Currently used antimetabolite treatments to prevent post-surgical scarring are non-selective and are associated with potentially blinding side effects. Developing new treatments to target scarring requires both a better understanding of wound healing and scarring in the conjunctiva, and new means of delivering anti-scarring drugs locally and sustainably. By combining plastic compression of collagen gels with a soft collagen-based layer, we have developed a physiologically relevant model of the sub-epithelial bulbar conjunctiva/Tenon’s capsule interface, which allows a more holistic approach to the understanding of subconjunctival tissue behaviour and local drug delivery. The biomimetic tissue hosts both primary human conjunctival fibroblasts and an immune component in the form of macrophages, morphologically and structurally mimicking the mechanical proprieties and contraction kinetics of ex vivo porcine conjunctiva. We show that our model is suitable for the screening of drugs targeting scarring and/or inflammation, and amenable to the study of local drug delivery devices that can be inserted in between the two layers of the biomimetic. We propose that this multicellular-bilayer engineered tissue will be useful to study complex biological aspects of scarring and fibrosis, including the role of inflammation, with potentially significant implications for the management of scarring following glaucoma filtration surgery and other anterior ocular segment scarring conditions. Crucially, it uniquely allows the evaluation of new means of local drug delivery within a physiologically relevant tissue mimetic, mimicking intraoperative drug delivery in vivo .
Spiking neural networks, thanks to their sensitivity to the timing of the inputs, are a promising tool for unsupervised processing of spatio-temporal data. However, they do not perform as well as the traditional machine learning approaches and their real-world applications are still limited. Various supervised and reinforcement learning methods for optimising spiking neural networks have been proposed, but more recently the evolutionary approach regained attention as a tool for training neural networks. Here, we describe a simple evolutionary approach for optimising spiking neural networks. This is the first published use of evolutionary algorithm to develop hyperparameters for fully unsupervised spike-timingdependent learning for pattern clustering using spiking neural networks. Our results show that combining evolution and unsupervised learning leads to faster convergence on the optimal solutions, better stability of fit solutions and higher fitness of the whole population than using each approach separately.
Abstract-Spiking neural networks have been previously used to perform tasks such as object recognition without supervision. One of the concerns relating to the spiking neural networks is their speed of operation and the number of iterations necessary to train and use the network. Here, we propose a biologically plausible model of a spiking neural network which is used in multiple, separately trained copies to process subsets of data in parallel. This ensemble of networks is tested by applying it to the task of unsupervised classification of spatio-temporal patterns. Results show that despite different starting weights and independent training, the networks produce highly similar spiking patterns in response to the same class of inputs, enabling classification with fast training time.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.