This is the submitted version of the following article: Sena-Torralba A., Ngo D.B., Parolo C., Hu L., Álvarez-Diduk R., Bergua J.F., Rosati G., Surareungchai W., Merkoçi A.. Lateral flow assay modified with time-delay wax barriers as a sensitivity and signal enhancement strategy. Biosensors and Bioelectronics, (2020). 168. 112559: - . 10.1016/j.bios.2020.112559, which has been published in final form at https://dx.doi.org/10.1016/j.bios.2020.112559 © https://dx.doi.org/10.1016/j.bios.2020.112559. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ 1 Lateral flow assay modified with time-delay wax barriers as a 2 sensitivity and signal enhancement strategy. 3 4 Amadeo Sena-Torralba a†, Duy Ba Ngo b†, Claudio Parolo a, Liming Hu a, Ruslan Álvarez a, 5 Jose Francisco Bergua a, Giulio Rosati a, Werasak Surareungchai b d and Arben Merkoçi a c* 6 7 a. Nanobioelectronics & Biosensors Group, Institut Català de Nanociència i Nanotecnologia (ICN2), CSIC and 8 The Barcelona Institute of Science and Technology (BIST), Campus UAB, 08193, Bellaterra, Barcelona, Spain. 9 b. School of Bioresoures and Technology, King Mongkut’s University of Technology Thonburi (KMUTT), 10 Bangkok 10150, Thailand. 11 c. Catalan Institution for Research and Advanced Studies (ICREA), Pg. Lluís Companys 23, 08010 Barcelona, 12 Spain. 13 d. Nanoscience and Nanotechnology Graduate Research Program, Faculty of Science, KMUTT, Bangkok 10140, 14 Thailand 15 16 * arben.merkoci@icn2.cat 17 † Amadeo Sena-Torralba and Duy Ba Ngo contributed equally to this work. 18 19 Keywords: Lateral flow, Sensitivity enhancement, Wax barrier, Point-of-Care, Optical biosensor 20 21 22 23 Abstract 24 25 The ease of use, low cost and quick operation of lateral flow assays (LFA) have made them 26 some of the most common point of care biosensors in a variety of field. However, their 27 generally low sensitivity has limited their use for more challenging applications, where 28 the detection of low analytic concentrations is required. Here we propose the use of 29 soluble wax barriers to selectively and temporarily accumulate the target and label 30 nanoparticles on top of the test line (TL). This extended internal incubation step 31 promotes the formation of the immune-complex, generating a 2.8-fold sensitivity 32 enhancement and up to 96% signal enhancement compared to the conventional LFA for 33 Human IgG (H-IgG) detection. 34 35 1. Introduction 36 During the last decade we have been observing an ever-growing need for fast and reliable 37 sensing devices for healthcare, environmental and safety applications.(Turner, 2013, 38 Malhotra et al., 2005) Both developed and developing countries are increasingly relying 39 on the use of point of care tests (POCT) to keep up with the saturation of their health-care 40 systems, to monitor the quality of their environmental resources and to prevent possible 41 threats. The POCTs represent not only convenient analytical tools for rural regions (they 42 are cheap, portable and easy to use), but they are also establishing themselves as key 43 monitoring devices in the busy centralized areas.(Price, 2001, Land et al., 2019, Chin et 44 al., 2012) For example a faster diagnosis means faster therapy initiation and higher 45 chances to control the spread of a disease; and it also means a quicker response time for 46 starting a remediation campaign in a contaminated river. In this context LFAs are 47 probably the most used POCT covering the widest variety of applications.(Quesada- 48 González and Merkoçi, 2015, Posthuma-Trumpie et al., 2009) They serve as an excellent 49 tool for healthcare,(Chen et al., 2016, Brangel et al., 2018) safety(Quesada-González et al., 50 2018, Mirasoli et al., 2012) and environmental applications(Hassan et al., 2019, 51 Schubert-Ullrich et al., 2009, Raeisossadati et al., 2016). 52 Being fast and easy to use are key features for the POCTs, but they also come with the 53 drawback that the sensitivity of the POCTs has no match with laboratory-based, slower, 54 multi-steps sensing techniques. For example in the LFA the binding event between bio- 55 receptors (antibodies, aptamers) and their targets happens in the order of seconds (the 56 time required for the flow to pass the TL)(Gasperino et al., 2018, Miller et al., 2018), while 57 in Enzyme-Linked Immunosorbent Assays (ELISA)(Van Weemen and Schuurs, 1971) 58 incubation for several hours is common. Although for many applications the sensitivity 59 of LFAs is sufficient, finding a way to improve it without affecting their being the POC 60 would open a plethora of new applications. Here we are describing just that, a method to 61 increase the time of the bio-recognition event, without affecting the ease of use and the 62 overall sensing time of the LFAs. 63 The concept at the base of LFAs, is the use of capillary movement to guide the sample 64 through different functional membranes.(Quesada-González and Merkoçi, 2015, 65 Muhammad Sajid et al., 2015) Taking as example an immune-sandwich (non- 66 competitive) assay, first the sample encounters the sample pad, where pre-stored 67 biochemical reagents stabilizes its pH and ionic strength at their optimal values to 68 maximize the signal to noise ratio.(Millipore, 2013) Then it flows to the conjugate pad, 69 where the detection antibodies (generally labelled with coloured nanoparticles) 70 recognize the target in the first bio-recognition event.(Quesada-González et al., 2019) 71 Then the solution flows along the detection pad, where the complex target/labelled- 72 antibody binds to the capture antibodies of the TL, during the second bio-recognition 73 event. Finally the sample after passing through the control line (CL), reaches the 74 absorbent pad at the end of the LFA. Surprisingly most of the reported works trying to 75 improve the sensitivity of LFA have focused on increasing the time for the first bio- 76 recognition event, but depending on the flow-rate of the pads it can already take minutes. 77 (Zhang et al., 2019, Tsai et al., 2018) Instead of the second bio-recognition event is always 78 in the order of seconds, representing the real bottle neck for the sensitivity of the test. 79 The possibility to slow down or even temporarily stop the flow above the test line would 80 increase the time for both bio-recognition events, boosting the sensitivity of the LFAs. 81 Here we present how we achieved it by placing a soluble wax barrier 1 mm after the TL. 82 Looking at previous works we find that wax structures were previously employed in LFAs 83 just to control the flow (Giokas et al., 2014, Lai et al., 2019, Phillips et al., 2016, Rivas et 84 al., 2014), but never to temporarily stop it. While some paper-based microfluidic devices 85 did incorporate real barrier to achieve incubation steps, but they were introduced before 86 the test line and were based on salt (He et al., 2019) or sugar(Lutz B, Liang T, Fu E, 87 Ramachandran S, Kauffman P, 2013), which once dissolved can affect the ionic strength 88 of the working buffer and thus the binding reaction. Instead, here, the wax barrier is 89 printed on the surface of the nitrocellulose membrane and melted in order to assure its 90 penetration through the membrane pores creating a hydrophobic barrier just after the 91 TL, which is the place where the second bio-recognition event is taking place. The 92 conjugate solution together with the sample solution flows along the nitrocellulose 93 membrane towards the test line, giving time for the first binding event to occur. Once 94 arrived to wax barrier, the solution stops temporarily on top of the test line until the 95 barrier breaks due to the surfactant present on the strip (figure 1). In this sense, the 96 thickness of the wax barrier and the concentration of surfactant can be carefully selected 97 in order to modulate the incubation time and determine the ones that provide the highest 98 sensitivity. 99 100 101 102 Figure 1. Schematic representation of the proposed strategy for the detection of H-IgG. 103 (A) In the standard LFA the flow is constantly moving towards the absorbent pad and the 104 bio-recognition event occurs within seconds. Few labelled antibodies are captured in TL, 105 thus the signal intensity is weak. (B) In the LFA modified with a wax barrier, the flow is 106 temporarily stopped on the TL. This increases the time for the bio-recognition event and 107 boosts the assay’s sensitivity. (C) Fluorescent microscope pictures (40X) of the wax 108 barrier on the LFA strip. The wax barrier temporarily retains the solution. Once broken 109 by the Tween-20, the solution goes through the barrier. 110 111 2. Experimental section 112 2.1. Materials and reagents 113 Goat anti-human IgG antibody, human IgG, bovine serum albumin (BSA), 114 Tetrachloroauric acid (HAuCl4), trisodium citrate, phosphate buffer saline (PBS) tablets, 115 sodium phosphate basic and dibasic, sodium tetraborate, boric acid sucrose and Tween 116 20 were purchased from Sigma Aldrich. Chicken anti-goat antibody was purchased from 117 Abcam. Nitrocellulose membranes (CN95 and CN150) were purchased from Sartorius 118 Stediem and CNPF200 from mdi. Cellulose membrane (CFSP001700) was purchased in 119 Merck Millipore, glass fiber (Standard14) from GE Healthcare and supporting adhesive 120 cards were purchased from Kenosha. Decol food colorant red powder was purchased in 121 Torten Deko and wax ink (Xerox 108R00935 black) was purchased in Xerox. 122 2.2. Instruments 123 Wax printer Xerox ColorQube 8580, Bioreagent dispenser (Imagene Technology), 124 Microscope Olympus cellSense, Centrifuge Allegra 64R, Lateral flow strips cutter 125 (Shanghai Kinbio Tech), Spectrophotometer SpectraMax ID3, SkanMulti (Skannex), TEM 126 Technai F20. 127 2.3. Synthesis and characterization of gold nanoparticles 128 The gold nanoparticles (AuNPs) were synthesized following the Turkevich 129 method(Turkevich et al., 1951) and were characterized with TEM (~ 15 nm) and 130 spectrophotometer (absorbance peak at 515 nm) (figure S1 and S2), respectively. The 131 AuNPs were adjusted to pH 9.2 using borate buffer (10 mM pH 9.2) and centrifuged at 132 14000 rpm for 30 minutes. Finally, they were re-suspended in 250 μL of PBS (10 mM, pH 133 7.4, 5% sucrose, 1% BSA and 0.5% tween-20) and dried in the glass fiber overnight. 134 Moreover, the sample pad was blocked with PBS (10 mM, pH 7.4, 0.5% BSA and 0.05% to 135 0.1% tween-20) and dried overnight. 136 2.4. Conjugation of the AuNPs with anti H-IgG 137 The AuNPs were conjugated to antibodies against H-IgG following the procedure 138 previously reported by our group.(Ambrosi et al., 2007) In order to check if the AuNPs 139 were successfully conjugated with the antibodies, the absorbance spectrum was 140 evaluated. As observed in figure S2, the maximum absorbance peak of the conjugated the 141 AuNPs has a little red shift (about 5 nm), indicating the presence of the antibodies on the 142 surface of the AuNPs. 143 144 3. Results and discussion 145 3.1. Evaluation of the retention time 146 When developing the POCT, reproducibility is one of the most important parameters to 147 optimize. For this reason, we fixed the lateral flow strip dimensions to 6 x 0.3 cm and 148 verified that the bed volume corresponds to 70 µL (figure S3). We then proceeded to 149 optimize the barrier fabrication by printing black lines with different widths (0.01, 0.03, 150 0.05 y 0.1 mm) onto CN95 nitrocellulose membrane, and melting homogeneously the wax 151 at 95 ° C for 5 minutes in the oven (figure S4). Next, we evaluated the time required for 152 breaking the barriers at different concentrations of Tween-20, following the former 153 optimized parameters. In particular, we fabricated lateral flow strips containing different 154 concentrations of Tween-20 (0.05%, 0.1%, 0.5% and 0.1%) in the sample pad and the 155 AuNPs in the conjugate pad. The retention time was measured from the moment at which 156 the AuNPs reached the wax barrier until they crossed the barrier. 157 158 As a result, we generated a matrix providing the wax barrier width and Tween-20 159 concentration that provide different retention times (Table 1 and figure S5). As expected, 160 the retention time increases increasing barrier width and decreasing Tween-20 161 concentration. In particular, we achieved the longest retention time (12 minutes) by 162 employing barriers of 0.05 mm and a concentration of Tween-20 of 0.1%; while the 163 shortest retention time (1 minute) required a barrier width of 0.01 mm and 0.5% Tween- 164 20. Width of 0.1 mm and Tween-20 concentration of 0.05%, showed to be not suitable for 165 the assay as the barriers broke once the AuNPs solution was already dried out. 166 167 Moreover, we printed the barriers in a slower flow rate nitrocellulose membrane 168 (CNPF200 mdi), in order to evaluate how the use of smaller pore sizes can affect the 169 retention time. In this case, the pressure applied by the wax printing process compacted 170 the pore of the nitrocellulose, thereby the flow could not even reach the TL. 171 172 Table 1. Retention times on CN95 with different wax barrier width and %Tween-20 Wax width (mm) 0.5% Tween-20 0.1% Tween-20 0.05% Tween-20 0.01 1.40 ± 0.42 min 3.93 ± 1.55 min NB 0.03 2.25 ± 0.13 min 7.64 ± 1.19 min NB 0.05 7.23 ± 0.86 min 11.45 ± 1.33 min NB 0.10 NB NB NB 173 * NB: No Breaking, which means that the solution dried out before breaking the wax barrier. 174 175 3.2. Evaluation of the sensitivity enhancement in LFA 176 Finally, after determining the optimum conditions to achieve the longest flow retention, 177 we performed the lateral flow immunoassay for H-IgG detection. In order to check if there 178 is a correlation between the retention time on the TL and the sensitivity enhancement, 179 we fabricated the strips without barriers and with barriers width of 0.01, 0.03 and 0.05 180 mm. In this regard, we used CN95 and 0.1% Tween-20 as surfactant to break the barriers. 181 We were expecting to see a progressive sensitivity enhancement when using higher 182 width barriers, obtaining the highest sensitivity by using 0.05 mm width barrier, since 183 this is the one that provides the longest retention time. Next, we prepared the strips by 184 fixing anti H-IgG antibodies (1 mg mL-1) in the TL and anti-goat antibodies (1 mg mL-1) in 185 the CL, both in 10 mM PB buffer pH 7.4. Moreover, we conjugated the AuNPs with anti H- 186 IgG by following the procedure previously reported by our group,(Ambrosi et al., 2007) 187 explained in the experimental section. Furthermore, we performed calibration curves 188 using serial dilutions of H-IgG (0 ng mL-1 to 103 ng mL-1). Finally, we evaluated the signal 189 intensity in both the TL and CL the by taking a picture of the strips using a LFA scanner 190 and analyzing them with Image J software.(Schneider et al., 2012) We took the pictures 191 of the strips once the AuNPs solution had completely reached the absorbent pad. 192 Eventually, we used the signal values of the TL and the CL after background signal 193 subtraction, and normalized the signal value by dividing TL over CL. 194 195 Figure S6 shows the calibration curves obtained when using CN95 and CN150 without 196 the barriers and CN95 with 0.01, 0.03 and 0.05 mm width barriers. As observed the 197 normalized optical density increases upon the detection of higher concentrations of H- 198 IgG. Figure S7 shows the pictures of the lateral flow strips after performing the assays. 199 In order to evaluate the sensitivities obtained with each condition we performed a linear 200 fitting from 10 to 1000 ng mL-1 and we considered the slope values (figure 2A). As we 201 were expecting, a higher sensitivity is achieved when using wider barriers, as this assures 202 a longer interaction time between the antibodies in TL, AuNPs and the analyte. The 203 highest sensitivity is achieved when using a 0.05 mm width barrier, which provides a 2.8- 204 fold enhancement compared to the same nitrocellulose without barriers. It is noteworthy 205 that, conversely to the reported, there isn’t an outstanding sensitivity improvement when 206 reducing the flow rate by using a smaller pore nitrocellulose membrane, as in the case of 207 CN150. (Millipore, 2013, NanoComposix, 2016) Besides, we have proved that the 208 outstanding sensitivity enhancement is obtained only when temporarily stopping the 209 flow above the TL. 210 211 Moreover, we calculated the limit of detection (LoD) and quantification (LoQ) achieved 212 for every tested condition. The LoD was calculated as 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑑𝑒𝑛𝑠𝑖𝑡𝑦𝐿𝑂𝐷 = 𝑏𝑙𝑎𝑛𝑘 + 213 3 𝜎𝑏𝑙𝑎𝑛𝑘 (i.e. the corresponding value of blank sample plus 3 times its standard deviation). 214 (Armbruster and Pry, 2008) The LoD for CN95, CN150 and CN95 with 0.01, 0.03 and 0.05 215 mm width barriers were 35.01, 73.23, 21.67, 18.71, 14.47 ng·mL-1, respectively. 216 Furthermore, the LoQ, calculated as 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑑𝑒𝑛𝑠𝑖𝑡𝑦𝐿𝑂𝐷 = 𝑏𝑙𝑎𝑛𝑘 + 10 𝜎𝑏𝑙𝑎𝑛𝑘 , were 217 2129.09, 1909.76, 1164.38, 87.49, 41.19 ng·mL-1, respectively. We obtained the lowest 218 LoD and LoQ values by setting the barriers at the condition that permitted the highest 219 retention time of the conjugate solution on the TL. 220 221 In figure 2B, we show pictures of the LFA strips after performing the calibration curve 222 assay from 10 to 1000 ng mL-1 using CN95, CN150 non-modified with the wax barriers 223 and CN95 modified with 0.01, 0.03 and 0.05 mm width barriers. As observed, there is a 224 signal enhancement in TL when introducing the wax barriers in to the system. The 225 percentage signal enhancement was calculated taking as reference the normalized optical 226 density of the strips with CN95 and without the wax barriers (Table S2). As observed, the 227 percentage signal increase is higher for increasing concentrations of H-IgG. The highest 228 percentage signal increase is 96%, and was achieved for 1000 ng mL-1 when using CN95 229 with a 0.05 mm width barrier. 230 231 Finally, we proved that a longer incubation time of the conjugate solution over the TL 232 didn’t favor the generation of non-specific signal in TL. As shown in figure S7, the signal 233 in TL for blank samples (0 ng mL-1 of H-IgG) is not higher when using wider wax barriers, 234 as we obtained the highest TL signal for the blank sample when using the CN150 235 nitrocellulose membrane. 236 237 238 239 240 Figure 2. The LFA for H-IgG detection using time-delay barriers. (A) Calibration curve 241 from 10 to 1000 ng mL-1 showing the optical density linear range when using CN95 242 (Optical density = 0.0631 ln [H-IgG (ng mL-1)] + 0.0433 r2=0.98) and CN150 non-modified 243 with the wax barriers (Optical density = 0.0695 ln [H-IgG (ng mL-1)] + 0.0627 r2=0.99) 244 and CN95 modified with 0.01 (Optical density = 0.0925 ln [H-IgG (ng mL-1)] + 0.0124 245 r2=0.99), 0.03 (Optical density = 0.1359 ln [H-IgG (ng mL-1)] - 0.1165 r2=0.96) and 0.05 246 mm width barriers (Optical density = 0.1767 ln [H-IgG (ng mL-1)] - 0.2056 r2=0.98). (B) 247 Pictures of the LFA strips after performing the calibration curve assay from 10 to 1000 248 ng mL-1 using CN95, CN150 non-modified with the wax barriers and CN95 modified with 249 0.01, 0.03 and 0.05 mm width barriers. 250 251 4. Conclusion 252 In conclusion, we have developed a LFA modified with time-delay barriers as a strategy 253 for sensitivity and signal enhancement. We have achieved an inner incubation time of 12 254 minutes, which produced a 2.8-fold enhancement in sensitivity and up to 96% signal 255 enhancement. We believe the proposed method represents an outstanding solution to 256 improve the analytical performance of LFAs, without sacrificing their ease-of-use and low 257 cost. 258 259 Acknowledgments 260 We acknowledge financial support from NACANCELL project PCIN-2016-066 261 (program Euronanomed 2). This work is also funded by the CERCA 262 Program/Generalitat de Catalunya. The ICN2 is funded by the CERCA programme 263 / Generalitat de Catalunya. ICN2 acknowledges the support of the Spanish 264 MINECO for the Project MAT2017-87202-P and through the Severo Ochoa Centers 265 of Excellence Program under Grant SEV2201320295. Amadeo Sena-Torralba 266 acknowledge Autonomous University of Barcelona (UAB) for the possibility of 267 performing this work inside the framework of Biotechnology Ph.D. Programme. 268 Duy Ba Ngo gratefully acknowledges Petchra Pra Jom Klao Doctoral, King 269 Mongkut’s University of Technology Thonburi (KMUTT) in Thailand and Sensor 270 Technology, School of Bioresources and Technology, KMUTT, Thailand. 271 272 References 273 Ambrosi, A., Castañeda, M.T., Killard, A.J., Smyth, M.R., Alegret, S., Merkoçi, A., 2007. 274 Double-codified gold nanolabels for enhanced immunoanalysis. Anal. Chem. 79, 275 5232–5240. 276 Armbruster, D.A., Pry, T., 2008. Limit of blank, limit of detection and limit of 277 quantitation. Clin. Biochem. Rev. 29 Suppl 1, S49–52. 278 Brangel, P., Sobarzo, A., Parolo, C., Miller, B.S., Howes, P.D., Gelkop, S., Lutwama, J.J., Dye, 279 J.M., McKendry, R.A., Lobel, L., Stevens, M.M., 2018. A Serological Point-of-Care Test 280 for the Detection of IgG Antibodies against Ebola Virus in Human Survivors. ACS 281 Nano 12, 63–73. https://doi.org/10.1021/acsnano.7b07021 282 Chen, Y., Sun, J., Xianyu, Y., Yin, B., Niu, Y., Wang, S., Cao, F., Zhang, X., Wang, Y., Jiang, X., 283 2016. A dual-readout chemiluminescent-gold lateral flow test for multiplex and 284 ultrasensitive detection of disease biomarkers in real samples. Nanoscale 8, 285 15205–15212. https://doi.org/10.1039/c6nr04017a 286 Chin, C.D., Linder, V., Sia, S.K., 2012. Commercialization of microfluidic point-of-care 287 diagnostic devices. Lab Chip 12, 2118–2134. https://doi.org/10.1039/c2lc21204h 288 Gasperino, D., Baughman, T., Hsieh, H. V., Bell, D., Weigl, B.H., 2018. Improving Lateral 289 Flow Assay Performance Using Computational Modeling. Annu. Rev. Anal. Chem. 290 11, 219–244. https://doi.org/10.1146/annurev-anchem-061417-125737 291 Giokas, D.L., Tsogas, G.Z., Vlessidis, A.G., 2014. Programming fluid transport in paper- 292 based microfluidic devices using razor-crafted open channels. Anal. Chem. 86, 293 6202–6207. https://doi.org/10.1021/ac501273v 294 Hassan, A.H.A., Bergua, J.F., Morales-Narváez, E., Mekoçi, A., 2019. Validity of a single 295 antibody-based lateral flow immunoassay depending on graphene oxide for highly 296 sensitive determination of E. coli O157:H7 in minced beef and river water. Food 297 Chem. 297, 124965. https://doi.org/10.1016/j.foodchem.2019.124965 298 He, X., Liu, Z., Yang, Y., Li, L., Wang, L., Li, A., Qu, Z., Xu, F., 2019. Sensitivity Enhancement 299 of Nucleic Acid Lateral Flow Assays through a Physical-Chemical Coupling Method: 300 Dissoluble Saline Barriers. ACS Sensors 4, 1691–1700. 301 https://doi.org/10.1021/acssensors.9b00594 302 Lai, Y., Tsai, C., Hsu, J., Lu, Y., 2019. Microfluidic Time-Delay Valve Mechanism on Paper- 303 Based Devices for Automated Competitive ELISA. Micromachines 10, 837. 304 Land, K.J., Boeras, D.I., Chen, X.S., Ramsay, A.R., Peeling, R.W., 2019. REASSURED 305 diagnostics to inform disease control strategies, strengthen health systems and 306 improve patient outcomes. Nat. Microbiol. 4, 46–54. 307 https://doi.org/10.1038/s41564-018-0295-3 308 Lutz B, Liang T, Fu E, Ramachandran S, Kauffman P, Y.P., 2013. Dissolvable fluidic time 309 delays for programming multi-step assays in instrument-free paper diagnostics. 310 Lab Chip 13, 2840–7. https://doi.org/10.1038/jid.2014.371 311 Miller, B.S., Parolo, C., Turbé, V., Keane, C.E., Gray, E.R., McKendry, R.A., 2018. 312 Quantifying Biomolecular Binding Constants using Video Paper Analytical Devices. 313 Chem. - A Eur. J. 24, 9783–9787. https://doi.org/10.1002/chem.201802394 314 Millipore, M., 2013. Rapid Lateral Flow Test Strips Considerations for Product 315 Development. 316 Mirasoli, M., Buragina, A., Dolci, L.S., Guardigli, M., Simoni, P., Montoya, A., Maiolini, E., 317 Girotti, S., Roda, A., 2012. Development of a chemiluminescence-based quantitative 318 lateral flow immunoassay for on-field detection of 2,4,6-trinitrotoluene. Anal. Chim. 319 Acta 721, 167–172. https://doi.org/10.1016/j.aca.2012.01.036 320 Muhammad Sajid, Abdel-Nasser Kawde, Muhammad Daud, 2015. Designs, formats and 321 applications of lateral flow assay: A literature review. J. Saudi Chem. Soc. 19, 689– 322 705. https://doi.org/10.1016/J.JSCS.2014.09.001 323 NanoComposix, 2016. Lateral flow assay development guide. nanoComposix Lateral 324 Flow Handb. v.1.4. 325 Phillips, E.A., Shen, R., Zhao, S., Linnes, J.C., 2016. Thermally actuated wax valves for 326 paper-fluidic diagnostics. Lab Chip 16, 4230–4236. 327 https://doi.org/10.1039/c6lc00945j 328 Posthuma-Trumpie, G.A., Korf, J., Van Amerongen, A., 2009. Lateral flow (immuno)assay: 329 Its strengths, weaknesses, opportunities and threats. A literature survey. Anal. 330 Bioanal. Chem. 393, 569–582. https://doi.org/10.1007/s00216-008-2287-2 331 Price, C.P., 2001. Regular review: Point of care testing. Br. Med. J. 322, 1285–1288. 332 https://doi.org/10.1136/bmj.322.7297.1285 333 Quesada-González, D., Jairo, G.A., Blake, R.C., Blake, D.A., Merkoçi, A., 2018. Uranium (VI) 334 detection in groundwater using a gold nanoparticle/paper-based lateral flow 335 device. Sci. Rep. 8, 8–15. https://doi.org/10.1038/s41598-018-34610-5 336 Quesada-González, D., Merkoçi, A., 2015. Nanoparticle-based lateral flow biosensors. 337 Biosens. Bioelectron. 73, 47–63. https://doi.org/10.1016/j.bios.2015.05.050 338 Quesada-González, D., Sena-Torralba, A., Wicaksono, W.P., de la Escosura-Muñiz, A., 339 Ivandini, T.A., Merkoçi, A., 2019. Iridium oxide (IV) nanoparticle-based lateral flow 340 immunoassay. Biosens. Bioelectron. 132, 132–135. 341 https://doi.org/10.1016/j.bios.2019.02.049 342 Raeisossadati, M.J., Danesh, N.M., Borna, F., Gholamzad, M., Ramezani, M., Abnous, K., 343 Taghdisi, S.M., 2016. Lateral flow based immunobiosensors for detection of food 344 contaminants. Biosens. Bioelectron. 86, 235–246. 345 https://doi.org/10.1016/j.bios.2016.06.061 346 Rivas, L., Medina-Sánchez, M., de la Escosura-Muñiz, A., Merkoçi, A., 2014. Improving 347 sensitivity of gold nanoparticle-based lateral flow assays by using wax-printed 348 pillars as delay barriers of microfluidics. Lab Chip 14, 4406–4414. 349 https://doi.org/10.1039/c4lc00972j 350 Schneider, C.A., Rasband, W.S., Eliceiri, K.W., 2012. NIH Image to ImageJ: 25 years of 351 image analysis. Nat. Methods 9, 671–675. https://doi.org/10.1038/nmeth.2089 352 Schubert-Ullrich, P., Rudolf, J., Ansari, P., Galler, B., Führer, M., Molinelli, A., 353 Baumgartner, S., 2009. Commercialized rapid immunoanalytical tests for 354 determination of allergenic food proteins: An overview. Anal. Bioanal. Chem. 395, 355 69–81. https://doi.org/10.1007/s00216-009-2715-y 356 Tsai, T.T., Huang, T.H., Chen, C.A., Ho, N.Y.J., Chou, Y.J., Chen, C.F., 2018. Development a 357 stacking pad design for enhancing the sensitivity of lateral flow immunoassay. Sci. 358 Rep. 8, 1–10. https://doi.org/10.1038/s41598-018-35694-9 359 Turkevich, J., Stevenson, P.C., Hillier, J., 1951. A study of the nucleation and growth 360 processes in the synthesis of colloidal gold. Discuss. Faraday Soc. 11, 55–75. 361 Turner, A.P.F., 2013. Biosensors: Sense and sensibility. Chem. Soc. Rev. 42, 3184–3196. 362 https://doi.org/10.1039/c3cs35528d 363 Van Weemen, B.K., Schuurs, A.H.W.M., 1971. Immunoassay using antigen-enzyme 364 conjugates. FEBS Lett. 15, 232–236. https://doi.org/10.1016/0014- 365 5793(71)80319-8 366 Zhang, S.F., Liu, L.N., Tang, R.H., Liu, Z., He, X.C., Qu, Z.G., Li, F., 2019. Sensitivity 367 enhancement of lateral flow assay by embedding cotton threads in paper. Cellulose 368 26, 8087–8099. https://doi.org/10.1007/s10570-019-02677-6 369 370 Supplementary Material 371 Supplementary material can be found at: 372 https://www.journals.elsevier.com/biosensors-and-bioelectronics/XXXX 373 374