Contents lists available at ScienceDirect Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios Real time observation and automated measurement of red blood cells agglutination inside a passive microfluidic biochip containing embedded reagents Maxime Huet a , Myriam Cubizolles a, ⁎ , Arnaud Buhot b,c,d a Univ. Grenoble Alpes, F-38000 Grenoble, France. CEA LETI MlNATEC Campus, F-38054 Grenoble, France b Univ. Grenoble Alpes, INAC-SPRAM, F-38000 Grenoble, France c CEA, INAC-SPRAM, F-38000 Grenoble, France d CNRS, INAC-SPRAM, F-38000 Grenoble, France ARTICLE INFO Keywords: Agglutination assay Passive microfluidic Embedded reagents Automated image processing Real time detection ABO blood typing ABSTRACT The process of agglutination is commonly used for the detection of biomarkers like proteins or viruses. The multiple bindings between micrometer sized particles, either latex beads or red blood cells (RBCs), create aggregates that are easily detectable and give qualitative information about the presence of the biomarkers. In most cases, the detection is made by simple naked-eye observation of agglutinates without any access to the kinetics of agglutination. In this study, we address the development of a real-time time observation of RBCs agglutination. Using ABO blood typing as a proof-of-concept, we developed i) an integrated biological protocol suitable for further use as point-of-care (POC) analysis and ii) two dedicated image processing algorithms for the real-time and quantitative measurement of agglutination. Anti-A or anti-B typing reagents were dried inside the microchannel of a passive microfluidic chip designed to enhance capillary flow. A blood drop deposit at the tip of the biochip established a simple biological protocol. In situ agglutination of autologous RBCs was achieved by means of embedded reagents and real time agglutination process was monitored by video recording. Using a training set of 24 experiments, two real-time indicators based on correlation and variance of gray levels were optimized and then further confirmed on a validation set. 100% correct discrimination between positive and negative agglutinations was performed within less than 2 min by measuring real-time evolution of both correlation and variance indicators. 1. Introduction The process of agglutination is commonly used for the detection of biomarkers like proteins or viruses present in solution or in biological samples. The multiple bindings to micrometer sized particles, either latex beads or red blood cells (RBCs), create aggregates that are easily detectable and lead to qualitative information about the presence or not of the biomarkers (Golchin et al., 2012; Jemima et al., 2014). Hemagglutination assays are widely employed to characterize viruses and bacteria that naturally agglutinate RBCs, especially for influenza and veterinary diagnosis (Fan et al., 2012; Gopinath and Kumar, 2013; Horie et al., 2009; Idelevich et al., 2014). In addition, hemagglutination assays can also be used to detect the presence of antigens on RBCs by specific probes. In this case agglutinins such as IgM antibodies may be used as reagents. An example of such assays is forward blood typing (Li et al., 2015; Noiphung et al., 2015; Voak, 1990). Moreover, some studies focused on the detection of biomarkers via hemagglutination, using a bispecific reagent. The latter is able on one hand to interact with red blood cells and on the other hand to recognize the targeted biomarker (Chen et al., 2007; Gupta and Chaudhary, 2006; John et al., 1990). Hemagglutination Inhibition Assays (HIA) are also considered to determine the affinity of inhibitors of agglutinins (Cecioni et al., 2015). Agglutinates can be generated using various protocols, based on slide (Pandya and Kirby, 1981), tube (Li et al., 2005), microplate (Spindler et al., 2001), gel column techniques (Seth et al., 2012) or on paper based devices (Noiphung et al., 2015). Usually, these protocols are performed manually. In most cases, the detection of the agglutina- tion is made by simple naked-eye observation (Chattopadhyay et al., 2014; Gupta and Chaudhary, 2003). Sometimes, automation of the analysis is included to overcome bias and variability induced by manual protocols (Charrière et al., 2015; Nguyen et al., 2016). Nevertheless, in http://dx.doi.org/10.1016/j.bios.2016.09.068 Received 17 June 2016; Received in revised form 14 September 2016; Accepted 19 September 2016 ⁎ Correspondence to: CEA Grenoble, 17 rue des Martyrs, 38054 Grenoble Cedex 9, France. E-mail addresses: maxime.huet@cea.fr (M. Huet), myriam.cubizolles@cea.fr (M. Cubizolles), arnaud.buhot@cea.fr (A. Buhot). Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx 0956-5663/ © 2016 Elsevier B.V. All rights reserved. Available online xxxx Please cite this article as: Huet, M., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.09.068