A Novel Framework for Biomedical Imaging Integrating 3D Printed Device, Confocal Microscopy and Large Scale Data Analytics

Swain, Sarpras and Giri, Lopamudra (2019) A Novel Framework for Biomedical Imaging Integrating 3D Printed Device, Confocal Microscopy and Large Scale Data Analytics. PhD thesis, Indian institute of technology Hyderabad.

Full text not available from this repository. (Request a copy)


Current methods of drug screening and disease diagnosis are mainly based on biomarker information obtained at a particular time point. In contrast, functionality and physiological response of a living system are mostly guided by the dynamic characteristics. However, a major challenge lies in obtaining high-resolution data and development of data analytics that are useful for pharmaceutical companies and in hospital settings for developing useful methods based on dynamical features. Data acquisition and analysis become particularly challenging in the measurement and interpretation of network activity in eye and brain. In this regard, recent trends in neuronal activity measurement focus on microelectrode based techniques. Such microelectrode array (MEA), fabricated from biocompatible materials and inserted in animal brain, makes measurements with high time resolution but low spatial resolution. Yet, in order to investigate the synchronicity and activity in single neurons for various network topologies, it is essential to obtain information on the micro-organization of the neuronal network and spike patterns. As a method of obtaining information at subcellular spatial resolution, molecular imaging has been suggested. In particular, modulation of calcium oscillations in human neuronal system is an indication of underlying physiological changes that includes neuro-glia communication to perform information transfer across cells or neuronal synchronization that plays a role in learning and memory formation. Specifically, dysregulation of calcium signaling results in neurodegenerative diseases such as diabetic retinopathy and Alzheimer’s disease. Although current research on disease diagnosis are mostly carried out by conventional method such as immunological and biochemical assays, we propose calcium imaging for measurement of the activity level for studying drug-cell interactions and pathological conditions. Specially, we focus on studying the cell-drug interaction through measurement of neuronal activity using confocal microscopy. Quantification of Ca2+ spike trains in drug-induced Ca2+ oscillation in neurons remains challenging due to the inherent heterogeneity in primary culture obtained from retina and hippocampus. Presently, there is no systematic data analysis framework available for statistical modeling of neuronal activity after treatment with drugs and identification of various activity levels. To address this, we used quantitative confocal imaging and clustering analysis, and present the characterization of Ca2+ spiking in GPCR targeting drug-treated primary culture of hippocampal neurons. A systematic framework, instead of an intuition-based method, was used for selection of the cluster number and distance metric. The results discern neurons with diverse Ca2+ response patterns, including active cells and silent cells in a neuron population in absence and presence of GPCRtargeted drugs. We also show that the clustering patterns of Ca2+ spiking can be controlled using appropriate selection of GABAB and mGluR targeting drugs, baclofen and DHPG. Next, we implement k-means clustering method to evaluate the stress level of in vitro model of human primary mixed retinal culture (MRC) obtained from cadaveric retina. The results indicate that the clustering pattern undergoes significant modification under hypoxia. To the best of our knowledge, this is the first report of automated analysis of large number of time-lapse videos and clustering of calcium oscillation in human MRC. Further, measurement of synchronicity between two neurons and correlating with particular network topology remains challenging. Hence, investigation on understanding the role of network in tuning the synchronicity remains sparse. In this context, we have implemented the clustering method to detect the level of synchronicity in a neuron population for various network topologies. In order to identify the synchronous neurons, we performed k-means clustering based on frequency and coefficient of variation of interspike interval. Further, the synchronicity was measured through evaluation of the correlation coefficient and cross-correlation function between Ca2+ spiking in two neurons. Based on results, it can be speculated that neuronal connectivity and branching density play an essential role in tuning the synchronicity of the network. Live cell imaging and 3D imaging of larger sections is one of the key aspects of characterization of 3D tissue models for drug screening, regenerative medicines and other biomedical applications. However, such high-resolution imaging in time and space is greatly hindered because of expensive glass bottom device and imaging chambers having suitable optical properties that are required for imaging using confocal microscopy. In this context, we present a simple and cost-effective scheme for biochip fabrication using 3D printing. To the best of our knowledge, this is the first instance of fused deposition modeling (FDM) based fabrication of a glass bottom PDMS device having an assembly viii with 0.17 mm glass coverslip that is compatible for imaging using laser scanning confocal microscopy. First, we evaluated the FDM based 3D printer to print the template having various geometries using acrylonitrile butadiene styrene (ABS) and show that these templates can be used to fabricate spiral channel in PDMS with high fidelity. Secondly, we demonstrate that the proposed device is capable of multiple operations within the same chip including cell culture, mixing reagents, morphology monitoring as well as live-dead assay using LSCM. Thirdly, we demonstrate the imaging performance of the 3D printed chip through time-lapse imaging of cytosolic calcium in microglia and carcinoma cells cultured within the chip. Additionally, we show the feasibility of performing immunohistochemistry assay and confocal Z-stack imaging in the same chip. The fabrication protocol offers several advantages over existing methods in terms of reduction in the cost of printing material, transparency of the device material and compatibility to LSCM. The proposed framework integrating imaging, statistical tools, and 3D printing yields an on-chip data acquisition and data analytics that has the potential to be used in drug testing and evaluation of disease model. The same framework can also be used for toxicity testing, selection of drug dose, disease diagnosis and other applications where high-resolution imaging and real-time monitoring is crucial.

[error in script]
IITH Creators:
IITH CreatorsORCiD
Giri, Lopamudrahttp://orcid.org/0000-0002-2352-7919
Item Type: Thesis (PhD)
Uncontrolled Keywords: Functional imaging, Large scale data analysis, 3D printing, k-means clustering, Drug screening, Biochip
Subjects: Chemical Engineering
Divisions: Department of Chemical Engineering
Depositing User: Team Library
Date Deposited: 12 Jul 2019 10:12
Last Modified: 21 Sep 2019 09:24
URI: http://raiith.iith.ac.in/id/eprint/5709
Publisher URL:
Related URLs:

    Actions (login required)

    View Item View Item
    Statistics for RAIITH ePrint 5709 Statistics for this ePrint Item