Biosensors and Bioelectronics 70 (2015) 254–260
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Development of immunochromatographic strip test using ﬂuorescent, micellar silica nanosensors for rapid detection of B. abortus antibodies in milk samples Swati S. Vyas a, Sushma V. Jadhav b, Sharmila B. Majee b, Jayanthi S. Shastri c, Vandana B. Patravale a,n a b c
Department of Pharmaceutical Sciences and Technology, Institute of Chemical Technology, Matunga, Mumbai 400019, India Department of Microbiology, Bombay Veterinary College, Parel, Mumbai 400012, India Topiwala National Medical College & B. Y. L. Nair Charitable Hospital, Dr. A. L. Nair Road, Mumbai 400008, India
art ic l e i nf o
a b s t r a c t
Article history: Received 14 January 2015 Received in revised form 17 March 2015 Accepted 19 March 2015 Available online 20 March 2015
Presence of bacteria such as Brucella spp. in dairy products is an immense risk to public health. Point of care immunoassays are rapid in that they can quickly screen various samples in a relatively short amount of time, are sensitive, speciﬁc and offer a great advantage in accurate and fast diagnosis of infectious diseases. We have fabricated a point of care rapid diagnostic assay that employs ﬂuorescent, micellar silica nanosensors capable of speciﬁcally detecting Brucella IgG antibodies in milk samples of afﬂicted animals. Currently, point of care detection assays are not commercially available for ﬁeld testing of farm animals using milk samples. The nanosensing allows precise detection of antibodies with low sample volumes (50 μl). We demonstrate recognition of B. abortus antibodies through capture by ﬂuorescent silica nanosensors using spiked and raw milk samples validated by ELISA and PCR. The test results are accurate and repeatable with high sensitivity and speciﬁcity, and a short assay time of 10 min for antigenic recognition and do not require any sample processing procedures such as isolation and separation. Additionally, well deﬁned antigenic components and surface biomarkers of various disease causing microbes can be broadly incorporated within the purview of this technology for accurate and rapid detection of suspected bovine pathological conditions, and can largely enable rapid ﬁeld testing that can be implemented in farms and food industry. & 2015 Elsevier B.V. All rights reserved.
Keywords: Point-of-care Nanosensors Detection Silica Immunochromatography Fluorescence
1. Introduction Brucellosis is a neglected infectious disease that has a high prevalence with reports of fresh 50,0000 cases emerging annually (Seleem et al. 2010). According to WHO reports, brucellosis is a global zoonotic infection and endemic in Mediterranean basin, Turkey, Anatolia regions, India, Mexico, Central and South America and Arabian Peninsula. The disease is transmitted to humans through consumption of unpasteurized dairy products and direct contact with afﬂicted animals (Atluri et al. 2011; Escobar et al. 2013; Monath 2013). Medical diagnosis of suspected cases in farm animals like buffaloes and cattle is often delayed as detection majorly relies upon previous history of abortion and time consuming culture techniques, which continue to persist as the gold standard for n
Corresponding author. E-mail address: [email protected]
http://dx.doi.org/10.1016/j.bios.2015.03.045 0956-5663/& 2015 Elsevier B.V. All rights reserved.
brucellosis detection in addition to lack of accessibility to diagnostic testing facilities (Andriopoulos et al. 2007; Praud et al. 2012). Apart from these, agglutination and complement ﬁxation based serological immunoassays have been developed and include Coombs test, standard tube agglutination test (STAT), and Rose Bengal test, but these suffer from cross reactions, false positives and in some tests non-speciﬁcity as well (Fernando Padilla Poester 2010; Yohannes et al. 2012). A major disadvantage associated with these techniques is that they require blood samples and therefore subject animals to painful and invasive sample collection procedures. Identiﬁcation of brucellosis from non-serological samples such as milk has been achieved using newer techniques such as polymerase chain reaction (PCR) and primary binding assays such as ELISAs and give sensitive, accurate detection but nevertheless rely on the availability of skilled personnel, costly immunoreagents, complex instrumentation and sophisticated testing facilities (Bruning et al. 1999; Hamdy and Amin 2002; Padmavathy et al. 2012; Vanzini et al. 2001). Fluorescence polarization assay
S.S. Vyas et al. / Biosensors and Bioelectronics 70 (2015) 254–260
(FPA) has by far proven to be the most effective diagnostic method for brucellosis detection and has been employed in ﬁeld tests due to its simplicity, rapidity, high sensitivity, no cross reactions and accuracy (Konstantinidis et al. 2007). However, a FPA is not recognized as a point-of-care diagnostic as it too requires ﬂuorescence polarization analyzers which are available at variable costs. Point-of-care screening devices for brucellosis are not available, and urgently required for rapid screening of samples with at least with a qualitative yes/no result. We report in this paper the ﬁrst rapid, point-of-care nanosensing immunodetection technology for brucellosis that can be performed using nitrocellulose membrane as a strip based test and analyzes non-serological milk samples. The antigenic components of Brucella spp. are largely composed of lipopolysaccharides, found in the outer membrane region of the microorganisms, and anti-lipopolysaccharide antibodies have been identiﬁed in clinical samples (Goldbaum et al. 1992). These include secretory antibodies in endogenous ﬂuids such as expressed milk, saliva and serum (Ariza et al. 1992; Kraus and Konno 1963; Nielsen et al. 1996; Wheatcroft 1957).The antigenic lipopolysaccharides component in Brucella spp. is primarily of two types: smooth and rough (Barrio et al. 2009). The strains that affect humans have outer membranes composed of smooth lipopolysaccharides bearing the O-antigen, whereas animals are afﬂicted with both smooth and rough strains (Cardoso et al. 2006). Nevertheless, the lipopolysaccharides from both the types of strains can bind speciﬁc antibodies and can work as surface biomarkers on nanosensors enabling capture of Brucella antibodies for diagnosing brucellosis. Moreover, lipopolysaccharides from different Brucella spp. subtypes can be readily obtained using the same extraction and isolation methods with reasonable yields. In contrast to most of the agglutination based methods for detection that use whole cell microbes, isolated lipopolysaccharides as antigens, even though pyrogenic, offer a greater advantage than whole cell bacteria in that they are avirulent, non-infectious, speciﬁc, comparatively benign, and hassle free in handling. We have therefore incorporated lipopolysaccharides from Brucella spp. as antigenic components in our immunoassay coupled with a biosensing nanosystem for quick detection. We demonstrate in our work that ﬂuorescent silica nanosensors (FSNs) developed using lipopolysaccharides isolated from Brucella spp. subtypes can aid accurate, rapid and sensitive detection of brucellosis. Silica nanosystems form rigid, robust structures, are chemically inert and easy to fabricate (Liberman et al. 2014; Matsushita et al. 2014; Slowing et al. 2007; Whitaker and Furst 2014; Zhu et al. 2014). Each silica nanoparticle can contain a large number of photochemically active species allowing bright luminescence and a high quantum yield (Ambrogio et al. 2011; Folling et al. 2008; Huang et al. 2013; Kim et al. 2013). Additionally, silica nanoparticles with ﬂuorophores as cargo warrant a stable, robust, sensitive nanosensing system, as ﬂuorescence is a detection methodology with a high signal to noise ratio, and easily distinguishes compounds based on a separate excitation and emission spectra. Silica nanosystems are additionally amenable to incorporating colored dye molecules, and readily adopt the color of the dye molecules contained within, serving as a visual detection aid. The nanosensing technology offers ﬂexibility to include most dyes as a means of detection depending upon the extent of structural association and retention within the nano template. FSNs have been explored in this paper for the ﬁrst time as immunosensors for detection of milk samples in a lateral assay format. We report an immunochromatographic diagnostic test (ICDT) based on FSNs, our core biosensing technology, as a promising nanodiagnostic approach for detection of intact Brucella antibodies in unprocessed non-serological milk samples which eliminates painful blood collection procedures in animals. The developed
ICDT provides a reliable, accurate and reproducible result in accordance with PCR and milk ELISA results and with high sensitivity and speciﬁcity for antibodies against Brucella abortus.
2. Materials and methods 2.1. Materials, spectral measurements and data analysis Tetraethylorthosilicate, triethoxyvinylsilane, Tween 20, bovine serum albumin, potassium dihydrogen phosphate, sodium chloride, anti-bovine IgG antibodies, Rhodamine B and ﬂuorescein were purchased from Sigma-Aldrich. Monoclonal Brucella abortus antibodies were procured from Abcams. Poloxamer 407 (Lutrols F127) was kindly gifted by BASF, Mumbai. Deionized water (Millipore Milli-Q grade) with a resistivity of 18.2 MΩ-cm was used throughout this study. Dynamic light scattering (DLS) in triplicates and zeta potential (under an electric ﬁeld as an average of 30 measurements) were performed using Zeta Sizer Nano ZS, Worcestershire, UK (Malvern Zetasizer 3000HS and He/Ne laser at scattering angles of 173° at 25 °C). UV–visible spectra and absorbance (OD values) were obtained using Epoch microplate spectrophotometer (Bio Tek Instruments Inc., USA). Fluorescence estimation was conducted using Synergy 2 Multi mode microplate reader (Bio Tek Instruments Inc., USA) with an excitation wavelength set at 540 nm and at slit widths (both excitation and emission) of 5 nm. TEM imaging was performed by adding a drop of FSNs to the surface of a carbon ﬁlm on a copper grid and then observed by a FEI Tecnai 12BT TEM (Peabody, MA, USA) at an accelerating voltage of 100 kV. For 3D imaging using confocal microscopy, FSNs were applied to a glass slide surface and covered by a glass cover slip sealed to the slide. The slide was placed inverted under the microscope and observed for FSNs using LSM Confocal ﬂuorescence laser scanning microscope at 100 magniﬁcation with appropriate ﬁlters excited using helium-neon laser. 2.2. Study population and clinical sampling of sera and milk Blood and milk samples were collected from one hundred and three buffaloes from three buffalo farms at Pune, India where cases of abortion had occurred. Sera was obtained from 5 ml whole blood by centrifugation, collected and stored at 20 °C until further testing. Milk samples were obtained during their routine milking time, collected in 20 ml tubes and stored at 20 °C and analyzed within 48 h. 10 samples were used as negative control obtained from a buffalo herd at Shirwal, Pune which was clinically, microbiologically and serologically free from Brucella infection. 2.3. Extraction of lipopolysaccharides from B. abortus B. abortus S19 strain was procured from Indian Veterinary Research Institute (IVRI), Izatnagar, Uttar Pradesh and validated by PCR. Lipopolysaccharides were extracted from B. abortus by hot SDS extraction and proteinase K digestion with few modiﬁcations (Garin-Bastuji, 1980). Extraction method of Garin-Bastuji and coworkers for isolation of O-lipopolysaccharides from B. abortus S19 strain was adopted with slight modiﬁcations (Garin-Bastuji et al. 1990). Brieﬂy, B. abortus was grown on nutrient agar base prepared with 5% equine serum. Brucella cells were harvested in distilled water, following which the bacteria were inactivated by heating at 60 °C for 1 h. The contents were centrifuged at 12000 rpm for 30 min at 4 °C and the pellet obtained was treated separately to extract and isolate lipopolysaccharides antigen. 0.5 g of bacterial cells thus treated were taken in tube, to which 10 ml Tris buffer (0.0625 M Tris buffer of pH 6.8) containing 2% SDS was added. The contents were heated at 100 °C for 10 min and
S.S. Vyas et al. / Biosensors and Bioelectronics 70 (2015) 254–260
subsequently cooled at 55 °C. About 7.5 μl of Proteinase K was added to this and the tube was incubated at 55 °C for 3 h, following which it was stored at 20 °C overnight. 0.2 g of SDS was then added and the tube was heated at 100 °C for 10 min followed by centrifugation at 12,000 rpm for 30 min at 20 °C. 3 ml of supernatant was then taken in a fresh tube and 9 ml ( 3) isopropanol was added for the precipitation of lipopolysaccharides at 4 °C. The tube was re-centrifuged at 12,000g for 30 min at 4 °C, the protein-lipopolysaccharides lysate (pellet) was dissolved in distilled water and the supernatant discarded. A second extraction was performed with isopropanol by the same process and the lipopolysaccharides-protein lysate was re-dissolved in 3 ml water. 20 μl of RNAse and DNAse I each were introduced at concentration of 0.01 mg/ml to the tube and allowed to stand at 37 °C for 30 min. Subsequently, 5 μl of Proteinase K (0.001 mg/ml) was added, kept at 55 °C for 3 h and left undisturbed at 20 °C overnight. A third extraction with isopropanol was conducted, and the ﬁnal sediment containing lipopolysaccharides was dissolved in water and further lyophilized. 0.5 g cells yielded 15 mg of lipopolysaccharides. All the lyophilized fractions were collected together and stored at 4 °C. 2.4. Nanosensor (FSNs) construction FSNs were prepared by a modiﬁed micellar template method of synthesis involving the hydrolysis and condensation of tetraethylorthosilicate (TEOS) (Huo et al. 2006). Brieﬂy, Poloxamer 407 was solubilized in 10 ml of 1 N hydrochloric acid above its critical micelle concentration to mediate formation of polymeric micelles of size approximately 45 nm. The appropriate amount of Rhodamine B (10 mg) (or ﬂuorescein) was then incorporated in this micellar solution using high speed magnetic stirring for 2hrs which was ﬁltered using 0.22 m pore sized ﬁlter membrane and then re-subjected to magnetic stirring. Following this, 1 g TEOS and 0.3 g triethoxyvinylsilane were added to the micellar solution and stirring was continued further for 2 h. The ﬁnal reaction mixture was then dialyzed against deionized water for 48 h and then with PBS pH 7.4 for 6hrs with continuous replacement of dialysis medium after every 2 h. The puriﬁed FSNs thus obtained were then incubated with lipopolysaccharides (1 mg/ml) from B. abortus in a shaker incubator at 37 °C for 1hr. Excess lipopolysaccharides were removed by centrifugation (20000 rpm, 10 min). Bovine serum albumin 1% was then added, incubated in a shaker incubator for 30 min and puriﬁed by centrifugation (13500 rpm, 15 min) with PBS washing for 3 runs at 4 °C. The pellet comprising of antigenic FSNs was resuspended in 10 mM PBS (pH 7.4) containing 1% sucrose, 1% bovine serum albumin and used for further experimentation. Fluorescence spectra of the particles were obtained using Synergy 2 multi mode microplate reader. 2.5. ELISA, PCR and milk ring test (MRT) Milk and serum samples were validated for brucellosis using ELISA and PCR. Milk ELISA kits for brucellosis detection were purchased from IDEXX Laboratories Inc., USA and were used for qualitative testing of samples. Competitive ELISA kits (Svanovirs Brucella-Ab C-ELISA) were purchased from Boehringer Ingelheim Svanova, Sweden for the detection of Brucella antibodies from serum specimens. Sample to positive percentages (%S/P) were calculated for each sample using instructions mentioned in the kit. Samples with %S/P o45% were negative and those with %S/P 445% were positive for brucellosis. For conducting validation of samples using PCR (Mastercycler nexus, Eppendorf AG, Hamburg, Germany), DNA was ﬁrst isolated and quantiﬁed spectrophotometrically at A260 (NanoDrop, Thermo Fischer Scientiﬁc Inc., USA) for each sample (both milk and sera). Template primer
BCSP31 speciﬁc to B. abortus biovars [Primers B4 (5′-TGG CTC GGT TGC CAA TAT CAA-3′) and B5 (5′-CGC GCT TGC CTT TCA GGT CTG3′)], was employed to target and amplify a fragment of the 31- kDa OMP (outer membrane protein) of size 223 bp (Baily et al., 1992). The DNA extraction procedures and PCR protocols are described in the electronic supplementary material (ESM). MRTs were conducted on milk samples using B. abortus MRT antigen (Institute of animal health and veterinary biological, Bangalore) for comparison of sensitivity and speciﬁcity. Ring formations with deeper color in the cream layer than in the skim region were seen for positive milk samples. 2.6. Milk ICDT The milk ICDT strip was assembled as previously described for serum rapid immunochromatographic test strip (Wang et al., 2010). Fiberglass strips (0.5 cm 30 cm, Ambala, India) were immersed into a solution of antigenic FSNs for 15 min, dried under vacuum and stored in sealed plastic bags with desiccant at 4 °C. Nitrocellulose membranes (2.5 cm 30 cm, Ambala, India) were imprinted with two 1 μl bands as test (1000 μg/ml LPS, B. abortus) and control (isolated Brucella IgGs from bovine brucella positive sera) and dried at room temperature. Sample pads (ﬁberglass, 1.6 cm 30 cm, Ambala, India) were pretreated with 10 mM PBS containing 0.1% Tween 20 (pH 7.4) and dried under vacuum. The ﬁnal test strip was assembled on adhesive coated polyester backing (6 cm 30 cm) by overlaying sample pads, conjugate pads, nitrocellulose membranes and cellulose ﬁber absorbent pads, cut into 4 mm strips and housed in plastic cassettes. For performing the test, 50 μl milk sample was added to the sample pad followed by addition of 50 μl of running ﬂuid (10 mM PBS, pH 7.4). Upon contacting the conjugate release matrix, the sample reacted with the test and control spots and in case of positive samples, ﬂuorescent spots were seen at both test and control regions whereas in case of negative samples, only the control spot ﬂuoresced upon excitation. The overall test takes 15 min to complete. 2.7. Statistical analysis To assess each modiﬁcation of FSNs, we treated the experimental data with non-parametric Kruskal–Wallis One way ANOVA with the Bonferroni correction for multiple comparisons with statistical signiﬁcance upper limit ﬁxed at 0.05 (npo 0.05, nn p o0.01, nnnp o0.001, nnnnp o0.0001). Error bars represent standard error of the mean. The evaluation of sensitivity, speciﬁcity and overall comparison of ICDT with other testing methodologies was performed by applying Two way ANOVA analysis with statistical signiﬁcance threshold ﬁxed at 0.05. A clinically acceptable range was taken as the interval within which the variation between the experimental data obtained from each method would lie within a 95% conﬁdence interval. Statistical analyses were performed using GraphPad Prism (Release 5, GraphPad Software Inc., La Jolla, CA, USA).
3. Results and discussion This study describes a new approach for determining Brucella antibodies in non-serological liquid media (milk) by ICDT. Brucella antibodies in milk samples scavenged by FSNs are captured on the surface of nitrocellulose membrane, which is modiﬁed by lipopolysaccharide (LPS) adsorption for antibody-FSN immobilization (Fig. 1). Antigenic FSNs were prepared by conjugating LPS to the FSN surface to mediate antibody recognition and binding. To evaluate antigenic FSNs assisted capture of antibodies on the nitrocellulose membrane surface, a range of LPS concentrations were
S.S. Vyas et al. / Biosensors and Bioelectronics 70 (2015) 254–260
Fig. 1. Representation of the ﬂuorescence-activatable immunochromatographic diagnostic test (ICDT) for detection of brucellosis.
tested and concentration dependant ﬂuorescence intensity was observed (Fig. S1A). The maximal ﬂuorescence emission of bare FSNs (not conjugated to LPS) was observed at 514 nm depicted in Fig. S1B. Fluorescein in its native free acid structural form shows maximal emission at 514 nm and no shift in wavelength of maximum emission of ﬂuorescein was observed after its incorporation into FSNs. This could be attributed to the transparent optical properties of silica that permit molecular light absorption and emission properties of ﬂuorescein to remain intact post entrapment. Fig. S1C illustrates the formation of FSNs. Self assembled micelles of the triblock copolymer, poly(ethylene glycol)-blockpoly(propylene glycol)-block-poly(ethylene glycol) formed by association of polymeric alkyl chains accommodate entrapped molecules of ﬂuorescein due to hydrophobic forces of interaction between aryl groups of ﬂuorescein, and propylene and ethylene residues of the triblock copolymer. DLS data showed that these micelles are of approximately 50 nm in size and upon silica condensation around the micellar structure more compact, robust structures (FSNs) of approximately 30 nm with stable ﬂuorescence signals over 6 months are obtained. The narrow size distribution of FSNs was conﬁrmed by the polydispersity index value (0.171). The chemical and particulate stability of any nanosized ﬂuorescent detection probe is essential in obtaining in quantitative and sensitive analytical output, as the optical properties of such probes are highly susceptible to the surrounding milieu and variations in ﬂuorescence signals might produce erroneous data. The aqueous dispersions of FSNs prepared were tested for batch to batch variation for their consistency in particle size and ﬂuorescence signal output (Fig. S2A). The data revealed high chemical and particulate stability of FSNs as evidenced by DLS and ﬂuorescence measurements which could be accounted for by the presence of the long chain hydrocarbons of the triblock copolymer, and the robust silica shell that prevented the active ﬂuorescent dye units from leaching out via diffusion processes. The photostable and transparent nature of the silica shell permitted high quantum yield with optimal detectable ﬂuorescence, and the abundance of surface hydroxyl groups in conjunction with a negative surface charge rendered high aqueous stability to FSNs. The inﬂuence of time on the stability of FSNs was assessed at room temperature is shown in Fig. S2B. The particles showed less variation in size and remained in the 25–30 nm range for up to 6 months. The ﬂuorescence signal strength was consistent up to 6 months as depicted. This stability could be achieved as FSNs are homogenously distributed in the aqueous environment and the ﬂuorophore molecules are maintained intact within the silica architecture.
We also assessed the inﬂuence of varied pH conditions on the particulate stability of FSNs. The particle size of FSNs remained stable over a wide range of pH conditions (Fig. S2C), however dissociating at higher pH over 12 and showing 50 nm particle size of native micellar structure conﬁrming silica disintegration at higher alkaline pH. The surface charges of FSNs determined from zeta potential analysis were close to neutral with a slight negative potential ( 3.12 mV) at pH 7. Confocal microscopy revealed nearly spherical morphologies of FSNs which existed exclusively as ﬂuorescing particulate entities showing emissions at 514 nm under dark ﬁeld (Fig. 2A). Background ﬂuorescence was subtracted for clear visibility of the particles. TEM images revealed compact structures (stained with 2% urinyl acetate) of FSNs (Fig. 2B) which were closely rotund, and concurred in particle size with DLS data. The silica casing around each FSN particle, ranging from 25 to 30 nm, could be recognized with a gray contrast, due to a suitable degree of staining of the sample. The hirsute appearance of the FSNs packed closely into a spherical silica based construct, was also observed and could be caused by the self-assembled molecular orientation of the triblock copolymer micelles revealing polymer brushes (hairs) across the interstitial space interspersed across the surface of FSNs. Native architecture of silica nanoparticles not loaded with ﬂuorescein also revealed similar morphology and particle dimensions as that of FSNs. After obtaining FSNs we next computed the amount of LPS bound to the FSN surface using a previously reported carbocyanine dye based assay developed for LPS from E. coli and T. palladium (Zey and Jackson 1973). Free LPS molecules exhibit concentration dependant absorbance quenching properties for a ﬁxed amount of carbocyanine dye (1,1′-Diethyl-4,4′-carbocyanine iodide) recorded at 640 nm which can be ﬁtted into a linear regression equation and used to extrapolate LPS amounts in unknown samples. We performed the assay procedure for testing the absorbance quenching ability of LPS extracted from B. abortus in which LPS quantitatively reduced the absorbance of carbocyanine (Figure S3 in the ESM). As carbocyanine dye is subject to ﬂuctuations in wavelength as reported by Zey and Jackson, we analyzed the dye prior to performing the assay and found that carbocyanine in PBS showed an absorption maxima at wavelength of 640 nm. We demonstrated that the carbocyanine quenching effect of LPS remained intact post binding to FSN surface (Figure S4A). Bare FSNs showed minimal quenching of absorbance of carbocyanine dye at 640 nm (Figure S5 in the ESM). The self absorbing power of FSNs (A0) was nulliﬁed by computing the difference from the obtained OD values for each test sample (AT) to get the actual absorbance
S.S. Vyas et al. / Biosensors and Bioelectronics 70 (2015) 254–260
Fig. 2. Confocal imaging revealed nearly spherical structures of FSNs (A) and TEM image of FSNs (B) showed agreement with the particle size data.
values (ΔA), ΔA¼ A0-AT for assessment of dye quenching effect by LPS bound FSNs at 640 nm. As shown by the trend in ΔAmax values, LPS quantitatively lowered absorbance of carbocyanine dye and reached a plateau for incubations of FSNs with LPS concentrations higher than 1 mg/ml (Figure S4B). This concentration (1 mg/ml) was the cut off value to achieve maximum lipopolysaccharide binding to FSNs and was employed in preparation of antigenic FSNs to obtain the best detectable signal. The maximum amount of LPS that could be bound to total FSN surface was estimated from the standard curve equation, ΔA ¼ 0.0169Cþ2.266 as 65 ppm, where C is the concentration of LPS and this value remained nearly steady despite incubations with high amounts of LPS. The FSN particles were puriﬁed by centrifugation to remove unbound LPS prior to recording the readings. We determined the optimal time for adsorption of LPS onto surface of FSNs. Corresponding absorbance data for the cut off LPS concentration and incubation time with FSNs are presented in Fig. S4C. Time resolved quenching effect was seen at varied incubation times, and reached a plateau after 1hr incubation time, indicating that maximum binding of LPS onto the FSN surface was achieved within this time duration. We assessed the inﬂuence of LPS surface binding on the overall ﬂuorescence of FSNs to evaluate variations in signal strength arising from surface modiﬁcations brought about by physical adsorption forces. There was no wavelength shift from 514 nm seen due to LPS binding to FSNs (Fig. S4D). The ﬂuorescence intensity remained nearly unchanged after LPS adsorption. Subsequent to each modiﬁcation, particulate surfaces were washed with PBS to lessen deviations in wavelength and ﬂuorescence intensity measurements. The detection sensitivity of the ICDT was determined by analyzing spiked milk samples (standardized against a reference positive control serum considered positive for unvaccinated bovine samples) at varied dilutions compared against reference signal (antigenic FSNs bound to Brucella IgGs stationed at the control line of the ICDT) for determination of detectable signal; milk clinically, serologically and microbiologically free from bacteria was used as negative control. The working principle of ICDT is based on sandwich assay; if Brucella antibodies are present in the milk sample, both the test and control line will glow. For a negative result, there will be no analyte and only the control line (antigenic
FSNs bound to IgGs stationed at the control line) will glow. The test results were interpreted by visual observation under UV lamp excitation and quantitatively recorded on nitrocellulose membrane using a ﬂuorescence plate reader in precise area scanning mode. LPS concentration of 1000 μg/ml showed highest signal output on the nitrocellulose membrane as previously shown (Figure S1A) and was therefore applied to the test line for antibody capture. Signal to noise ratios were calculated and deﬁned as the ratios of ﬂuorescence intensities of test samples to that of blank (intensity at the control line). Fig. 3 reveals that the intensities of the test zone (Fig. 3A), and the S/N ratios (Fig. 3B) were increased with increasing concentrations of serum antibodies. LPS binding through particulate surface adsorption yielded stable antigenic FSNs with preservation of antibody detection capabilities on the ICDT. Therefore, based on their quality attributes, FSNs can be universally applied as excellent detection tools using ICDT as a platform. We compared test results of ICDT with commercially available tests for brucellosis in real milk samples from buffaloes (Fig. 4). With PCR test used as a reference test, ICDT had high speciﬁcity (100%) and sensitivity (100%) for milk samples. There was an excellent agreement between the results obtained by PCR and ICDT for clinical samples. Also, there was complete agreement between ICDT and milk ELISA kits. Milk ELISA kits use antigens expressed from stable cell lines as the coating antigen and showed high speciﬁcity than that of milk ring test (MRT) (90%). ICDT developed for milk samples had high sensitivity (100%), much higher than serum tube agglutination test (STAT) (31%) and comparable to Rose Bengal testing (RBT). RBT for brucellosis is the most extensively used and inexpensive test and showed only one false positive result from serum samples tested. In fact, milk ICDT gave 100% sensitivity and speciﬁcity for the detection of B. abortus for the 103 samples tested and showed complete agreement with PCR results. Milk ELISA tests also showed 100% sensitivity with the milk samples. The sensitivity of ICDT was comparable or better than commercially available MRT (98.89%). Milk ICDT was very speciﬁc for brucella and showed 100% agreement with milk ELISA and PCR results for all of the 103 milk samples tested. Test results of PCR (milk), milk ELISA and serum RBT are further depicted in supporting information (Fig. S7 in the ESM). The corresponding serum samples from the same animals tested by PCR
S.S. Vyas et al. / Biosensors and Bioelectronics 70 (2015) 254–260
Fig. 3. Determination of detection limit of ICDT in spiked milk samples. (A) Milk samples spiked with varied dilutions of sera containing Brucella antibodies (Ab) tested on the ICDT and compared against blank (FSNs bound to Brucella IgGs stationed at the control line of the ICDT) for determination of detectable signal. Strengths with varied Ab titres (o 100) showed detectable ﬂuorescence emission. Statistical treatment was performed using Kruskal–Wallis One way ANOVA with Bonferroni correction for multiple comparisons. Error bars represent SEM with n¼ 6. (np o 0.05, nnpo 0.01, nnnpo 0.001, nnnnpo 0.0001). (B) S/N ratios computed for series of [Ab] levels tested with antigenic FSNs on the ICDT platform upon application of 50 μl of milk samples. Error bars denote SEM, n¼ 6.
Fig. 4. Sensitivity and speciﬁcity analyses of various tests using real samples from buffaloes. ICDT concurred with commercially available milk ELISA tests and PCR tests showing 100% sensitivity and speciﬁcity. MRT and STAT results showed signiﬁcant variation from the reference test (PCR) and also from the ICDT. Two way ANOVA analysis was applied to the results for comparison of data sets (p o0.05).
demonstrated similar results. Nitrocellulose membrane of a speciﬁc grade (pore size 5–8 μ) was incorporated in ICDT developed for milk samples as compared to conventional serum immunochromatographic strips (15 μ), as non-serological samples are less complex than serum or whole blood samples and have low peripheral content. The test results concur that antibody testing in milk samples for bovine brucellosis is reasonably equivalent in terms of reproducibility and accuracy in comparison to serum samples. In comparison to ELISA and PCR, the ICDT has the advantage of requiring no expensive equipment, cultured cells, live bacteria, or skilled operators and the test is very easy to read. In the ICDT a few microlitres of the sample is applied, two apple green ﬂuorescent spots on the membrane indicate the test is positive for brucellosis and appearance of one spot on the control line is indicative of a negative sample. The stronger the ﬂuorescence appears the more Brucella speciﬁc antibodies are present in the sample. The intensity of the test could however reach a plateau because the amount of test reagent (LPS) applied and immobilized on the test line is constant, and even though a very high amount of antibodies are present in the sample, the amount of IgGs present at the test line is limited and therefore the ﬂuorescence cannot become stronger. However, even though the test shows excellent disparity between Brucella positive and negative samples and is robust, it requires UV excitation source to read the result. The ICDT nevertheless uses inexpensive components and is highly sensitive and speciﬁc for
testing non-serological samples. The ICDT shows excellent results for strongly positive samples with antibody titers over 25 IU/ml. DNA isolation prior to PCR testing, used as the gold standard in this paper for comparison, is crucial for identiﬁcation and works best in strongly positive samples. Milk ELISA according to manufacturer's instructions can detect strongly positive samples with a reasonable antibody titer cut off displaying signiﬁcantly high OD values for such samples. The ICDT has been validated by comparing the test results of over a hundred samples with the results obtained from PCR. PCR was used as the standard test for comparison due a high level of accuracy, but it involves extensive DNA isolation procedures, is time consuming and routine use is expensive and requires transportation to specialized lab testing facilities. The ICDT uses isolated lipopolysaccharides from B. abortus as antigens to detect Brucella antibodies in collected clinical specimens. The lipopolysaccharides from Y. enterocolitica have some structural resemblance with those from Brucella spp. and therefore the ICDT could show some cross reactivity with Yersinia antibodies (Munoz et al. 2005). The test nevertheless, maintains its speciﬁcity and sensitivity to a large extent. The developed test is a useful method to identify animal patients with brucellosis using unprocessed milk samples and could be employed in routine ﬁeld tests.
4. Conclusions In this study, we have presented antigen tagged ﬂuorescent silica nanoprobes to facilitate detection of Brucella antibodies in bovine milk with only a few microlitres of sample using immunochromatographic testing module. The limit of detection of this functional nanoprobe is many orders of magnitude lower than the currently employed tube agglutination and milk ring tests and comparable to conventional immunoassays. The high sensitivity of this probe is based on the high loading efﬁciency of ﬂuorescein within FSN core structures and high binding properties of FSNs to bacterial lipopolysaccharides with retention of antibody capturing abilities on the ICDT. Such improved detection systems would enable the screening of biomarkers in less complex non-serological samples such as milk, thus supplementing quicker therapeutic intervention and outcomes. Although ICDT based on silica nanosensors was taken only as an example to demonstrate the proof-of-concept for brucellosis detection in non-serological media, we expect that this technology can be extended to rapid detection of other biomarkers clinically required using easily obtainable samples collected by more compliant procedures by
S.S. Vyas et al. / Biosensors and Bioelectronics 70 (2015) 254–260
avoiding painful jugular vein blood withdrawals from bovine animals. Lastly, ICDT detection was performed in nitrocellulose strips housed in plastic cassettes, the most accepted user friendly diagnostic system that can be made readily available, particularly in developing countries, making this technique low-cost and accessible.
Acknowledgements The work was supported by the Indian Council of Medical Research (ICMR) (No. 30/3/15/2008-ECD-II), New Delhi. Authors would like to thank Dr. Naﬁsa Balasinor, National Institute for Research in Reproductive Health, Parel, Mumbai for confocal imaging of nanosensors.
Appendix A. Suplementary information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.bios.2015.03.045.
References Ambrogio, M.W., Thomas, C.R., Zhao, Y.L., Zink, J.I., Stoddart, J.F., 2011. Acc. Chem. Res. 44 (10), 903–913. Andriopoulos, P., Tsironi, M., Deftereos, S., Aessopos, A., Assimakopoulos, G., 2007. Int. J. Infect. Dis.: IJID: Off. Publ. Int. Soc. Infect. Dis. 11 (1), 52–57. Ariza, J., Pellicer, T., Pallares, R., Foz, A., Gudiol, F., 1992. Clin. Infect. Dis.:Off. Publ. Infect. Dis. Soc. Am. 14 (1), 131–140. Atluri, V.L., Xavier, M.N., de Jong, M.F., den Hartigh, A.B., Tsolis, R.M., 2011. Annu. Rev. Microbiol. 65, 523–541. Barrio, M.B., Grillo, M.J., Munoz, P.M., Jacques, I., Gonzalez, D., de Miguel, M.J., Marin, C.M., Barberan, M., Letesson, J.J., Gorvel, J.P., Moriyon, I., Blasco, J.M., Zygmunt, M.S., 2009. Vaccine 27 (11), 1741–1749. Bruning, A., Bellamy, K., Talbot, D., Anderson, J., 1999. J. Virol. Methods 81 (1-2), 143–154. Cardoso, P.G., Macedo, G.C., Azevedo, V., Oliveira, S.C., 2006. Microb. Cell Factories 5, 13. Escobar, G.I., Jacob, N.R., Lopez, G., Ayala, S.M., Whatmore, A.M., Lucero, N.E., 2013.
Comp. Immunol. Microbiol. Infect. Dis. 36 (6), 575–580. Padilla Poester, Fernando, K.N, Ernesto Samartino, Luis, Ling Yu, Wei, 2010. Diagn. Brucell. Open Vet. Sci. J. 4, 46–60. Folling, J., Polyakova, S., Belov, V., van Blaaderen, A., Bossi, M.L., Hell, S.W., 2008. Small 4 (1), 134–142. Garin-Bastuji, B., Bowden, R.A., Dubray, G., Limet, J.N., 1990. J. Clin. Microbiol. 28 (10), 2169–2174. Goldbaum, F.A., Rubbi, C.P., Wallach, J.C., Miguel, S.E., Baldi, P.C., Fossati, C.A., 1992. J. Clin. Microbiol. 30 (3), 604–607. Hamdy, M.E., Amin, A.S., 2002. Vet. J. 163 (3), 299–305, London, England: 1997. Huang, X., Aguilar, Z.P., Li, H., Lai, W., Wei, H., Xu, H., Xiong, Y., 2013. Anal. Chem. 85 (10), 5120–5128. Huo, Q., Liu, J., Wang, L.Q., Jiang, Y., Lambert, T.N., Fang, E., 2006. J. Am. Chem. Soc. 128 (19), 6447–6453. Kim, H.C., Hong, H.G., Yoon, C., Choi, H., Ahn, I.S., Lee, D.C., Kim, Y.J., Lee, K., 2013. J. Colloid Interface Sci. 393, 74–79. Konstantinidis, A., Minas, A., Pournaras, S., Kansouzidou, A., Papastergiou, P., Maniatis, A., Stathakis, N., Hadjichristodoulou, C., 2007. Eur. J. Clin. Microbiol. Infect. Dis.: Off. Publ. Eur. Soc. Clin. Microbiol. 26 (10), 715–721. Kraus, F.W., Konno, J., 1963. Ann. NY Acad. Sci. 106 (2), 311–329. Liberman, A., Mendez, N., Trogler, W.C., Kummel, A.C., 2014. Surf. Sci. Rep. 69 (2-3), 132–158. Matsushita, H., Mizukami, S., Sugihara, F., Nakanishi, Y., Yoshioka, Y., Kikuchi, K., 2014. Angew. Chem. Int. Ed. Engl. 53 (4), 1008–1011. Monath, T.P., 2013. Vaccine 31 (46), 5321–5338. Munoz, P.M., Marin, C.M., Monreal, D., Gonzalez, D., Garin-Bastuji, B., Diaz, R., Mainar-Jaime, R.C., Moriyon, I., Blasco, J.M., 2005. Clin. Diagn. Lab. Immunol. 12 (1), 141–151. Nielsen, K., Smith, P., Gall, D., Perez, B., Cosma, C., Mueller, P., Trottier, J., Cote, G., Boag, L., Bosse, J., 1996. Vet. Microbiol. 52 (1–2), 165–173. Padmavathy, B., Vinoth Kumar, R., Jaffar Ali, B.M., 2012. J. Nanobiotechnol. 10, 8. Praud, A., Champion, J.L., Corde, Y., Drapeau, A., Meyer, L., Garin-Bastuji, B., 2012. BMC Vet. Res. 8, 68. Seleem, M.N., Boyle, S.M., Sriranganathan, N., 2010. Vet. Microbiol. 140 (3-4), 392–398. Slowing, II, Trewyn, B.G., Giri, S., Chen, H.T., Lin, V.S., 2007. Acc. Chem. Res. 40 (9), 846–853. Vanzini, V.R., Aguirre, N.P., Valentini, B.S., Torioni de Echaide, S., Lugaresi, C.I., Marchesino, M.D., Nielsen, K., 2001. Vet. Microbiol. 82 (1), 55–60. Wheatcroft, M.G., 1957. J. Dental Res. 36 (1), 112–117. Whitaker, K.A., Furst, E.M., 2014. Langmuir: ACS J. Surf. Colloids 30 (2), 584–591. Yohannes, M., Gill, J.P., Ghatak, S., Singh, D.K., Tolosa, T., 2012. Rev. Sci. Tech. 31 (3), 979–984. Zey, P., Jackson, S., 1973. Appl. Microbiol. 26 (2), 129–133. Zhu, Y., Sundaram, H.S., Liu, S., Zhang, L., Xu, X., Yu, Q., Xu, J., Jiang, S., 2014. Biomacromolecules 15 (5), 1845–1851.