This is the submitted version of the article: Peláez E.C., Estevez M.C., Mongui A., Menéndez M.-C., Toro C., Herrera-Sandoval O.L., Robledo J., García M.J., Portillo P.D., Lechuga L.M.. Detection and Quantification of HspX Antigen in Sputum Samples Using Plasmonic Biosensing: Toward a Real Point-of-Care (POC) for Tuberculosis Diagnosis. ACS infectious diseases, (2020). 6. : 1110 - . 10.1021/acsinfecdis.9b00502. Available at: https://dx.doi.org/10.1021/acsinfecdis.9b00502 ACS Infectious Diseases This document is confidential and is proprietary to the American Chemical Society and its authors. Do not copy or disclose without written permission. If you have received this item in error, notify the sender and delete all copies. Early diagnosis of Mycobacterium tuberculosis by detection and quantification of HspX antigen in sputum samples using plasmonic biosensing Journal: ACS Infectious Diseases Manuscript ID Draft Manuscript Type: Article Date Submitted by the n/a Author: Complete List of Authors: Pelaez, Enelia Cristina; ICN2-CSIC-CIBER BBN, Nanobiosensors and Bioanalytical Applications Estévez, Maria Carmen; ICN2-CSIC-CIBER BBN, Nanobiosensors and Bioanalytical Applications Mongui, Alvaro; Corporación CorpoGen, Departamento Biotecnología Molecular Menendez, M.-Carmen; Universidad Autonoma de Madrid, Facultad Medicina. Departamento de Medicina Preventiva, Salud Publica y Microbiología Toro, Carlos; La Paz University Hospital, Department of Microbiology Herrera-Sandoval, Oscar; Centro de Investigación y Desarrollo Tecnológico de la Industria Electro Electrónica y TIC Robledo, Jaime; Corporación para Investigaciones Biológicas, Laboratorio de Micobacterias Garcia, Maria J.; Universidad Autonoma de Madrid, Facultad Medicina. Departamento de Medicina Preventiva, Salud Publica y Microbiología del Portillo, Patricia; Corporación CorpoGen, Departamento Biotecnología Molecular Lechuga, Laura M. ; Institut Català de Nanociència i Nanotecnologia ACS Paragon Plus Environment Page 1 of 22 ACS Infectious Diseases 1 2 3 Early diagnosis of Mycobacterium tuberculosis by detection and quantification of 4 5 HspX antigen in sputum samples using plasmonic biosensing 6 7 8 E.- Cristina Peláeza,b,c, M.-Carmen Esteveza, Alvaro Monguic, M-Carmen Menéndezd, Carlos 9 Toroe, Oscar L. Herrera-Sandovalb,f, Jaime Robledog, Maria J. Garcíad, Patricia Del Portilloc, 10 and Laura M. Lechugaa* 11 12 aNanobiosensors and Bioanalytical Applications Group (NanoB2A), Catalan Institute of 13 Nanoscience and Nanotechnology (ICN2), CSIC, CIBER-BBN and BIST, Campus UAB 14 Bellaterra, 08193 Barcelona, Spain. 15 16 bCentro de Investigación y Desarrollo Tecnológico de la Industria Electro Electrónica y TIC 17 (CIDEI), Calle 45ª BIS # 19-09 Floor 2, Bogotá DC, Colombia. 18 19 cCorporación CorpoGen, Departamento Biotecnología Molecular. Carrera 4 # 20-41, Bogotá 20 DC, Colombia. 21 dDepartamento de Medicina Preventiva, Salud Publica y Microbiología, Facultad de Medicina, 22 Universidad Autonoma de Madrid, St Arzobispo Morcillo s/n, 28029 Madrid, Spain. 23 24 eDepartment of Microbiology. La Paz University Hospital, IdiPaz, St/Paseo de la Castellana 25 261. 28046 Madrid, Spain. 26 f ResearchCluster on Converging Sciences and Technology (NBIC), Universidad Central, Calle 27 21 #4-40, Bogotá DC, Colombia. 28 29 gCorporación para Investigaciones Biológicas (CIB), Laboratorio de Micobacterias, Carrera 30 72A 78B-141, 050034, Medellín, Colombia. 31 32 * Corresponding author: M.- Carmen Estevez: mcarmen.estevez@icn2.cat 33 34 Abstract 35 36 Advances occurred during the last years in the diagnosis of Mycobacterium tuberculosis (Mtb), 37 the causative agent of tuberculosis infection, have prompted increased survival rates of patients. 38 39 However, limitations related to the inefficiency of an early detection still remain: some 40 41 techniques and laboratory methods do not have enough specificity, most instruments are 42 expensive, and require handling by trained staff. In order to contribute to a prompt and effective 43 44 diagnosis of tuberculosis, we report the development of a portable, user-friendly and low-cost 45 biosensor device for its early detection. By using a label-free miniaturized Surface Plasmon 46 47 Resonance (SPR) biosensor, we have established a direct immunoassay for the direct detection 48 49 and quantification of the Heat shock protein X (HspX) of Mtb, a well-established biomarker of 50 this pathogen, directly in pre-treated sputum samples. The method relies on highly specific 51 52 monoclonal antibodies which are previously immobilized on the plasmonic sensor surface. This 53 technology allows the direct detection of the biomarker without amplification steps, showing a 54 55 Limit of Detection (LOD) of 0.63 ng mL-1 and a Limit of Quantification (LOQ) of 2.12 ng mL-1. 56 The direct analysis in pre-treated sputum shows significant differences in the HspX 57 58 concentration in patients with tuberculosis (with concentration levels in the order of 116-175 ng 59 60 mL-1) compared with non-tuberculosis infected patients (values below the LOQ of the assay). ACS Paragon Plus Environment ACS Infectious Diseases Page 2 of 22 1 2 3 4 5 Keywords: Tuberculosis diagnosis; HspX recombinant protein; SPR Immunosensor; Point-of- 6 Care device; human sputum samples 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 3 of 22 ACS Infectious Diseases 1 2 3 Tuberculosis (TB) is one of the most impactful infectious diseases with the highest morbidity 4 5 and mortality in human history.1,2 According to the World Health Organization (WHO) in 2017, 6 10.4 million new cases and 1.3 million deaths were attributed to this disease. Moreover, a third 7 8 of the world’s population is calculated to be latently infected, with an asymptomatic and non- 9 transmissible state.3–5 WHO also estimated that 13 % of tuberculosis patients have HIV5 and 10 11 480.000 cases per year are multidrug-resistant (MDR) owing to inadequate use of antibiotics, 12 13 poor adherence to anti-TB drugs and to insufficient monitoring of drug resistance.2,6,7 Around 3 14 million people who are not diagnosed each year contribute with the high incidence of TB, 15 16 becoming a source of dissemination of Mycobacterium tuberculosis (Mtb).3 There is a lack of 17 novel detection methods for the diagnosis of TB that can be used in low resource countries, 18 19 especially in remote populations and that can overcome for instance the low sensitivity of 20 21 microscopy and the delay in the diagnosis through bacteria culture (around two to six weeks).2,8 22 During the infection period, patients without treatment can transmit the bacillus to other people, 23 24 and therefore a prompt diagnosis could guarantees their survival.9,10 25 26 Currently, several TB diagnostic methods have been implemented and endorsed by the WHO 27 for the identification of Mtb, such as conventional bright field microscopy using Ziehl-Neelsen 28 29 (ZN) staining,11 LED-fluorescence microscopy12 and auramine O staining to identify the 30 31 presence of Acid-Fast Bacilli (AFB) in sputum smear.2,13 TB blood tests which evaluate the 32 immunological response to Mtb (i.e. Mantoux tuberculin skin test (TST), and Interferon-γ- 33 34 Release Assays (IGRA))14,15 are widely applied to diagnose latent TB infection (LTBI). 35 However, TST gives false-positive results in patients who have been vaccinated with Bacillus 36 37 Calmette-Guérin and both types of tests could give inaccurate responses in patients with 38 immunocompromised conditions as HIV/AIDS.16 In order to improve the efficiency, specificity 39 40 and accuracy of the diagnosis, WHO recommended nucleic acid amplification test (NAAT) 41 42 which uses PCR to amplify and detect mycobacterial rRNA or DNA directly in blood, sputum, 43 and other human biological fluids (i.e. Loop-mediated Isothermal Amplification PCR PURE- 44 45 TB-LAMP17,18 and GeneXpert MTB/RIF®.19 Furthermore, promising biomarkers of Mtb such as 46 lipoarabinomannan (LAM),20,21 ESAT-6,22 CFP-10,23 Ag85B,4 GlcB,24 and MPT-5125 found in 47 48 blood, sputum, cerebrospinal fluid and urine have been studied by ELISA.26–28 Nevertheless, 49 50 these techniques are expensive, and still require sophisticated equipment, trained personnel and 51 controlled environments, hindering the TB diagnosis in low-resource settings .18,29 52 53 For these reasons, WHO recognizes the need for new and more sensitive and specific strategies 54 55 to improve early TB diagnosis, such as the design of portable Point-of-Care (POC) devices, 56 which stand out as an alternative to the conventional techniques by offering significant 57 58 advantages as simplicity, portability and affordability. These features can eventually make them 59 accessible to the majority of population, thus helping reduce the global infection rate.3,30 Several 60 ACS Paragon Plus Environment ACS Infectious Diseases Page 4 of 22 1 2 3 POC devices based on the detection of Mtb biomarkers have been reported. For example, a 4 5 lateral flow immunochromatography for the detection of LAM in urine is currently 6 commercialized (i.e. Alere Determine™ TB LAM). This biomarker has also been detected in 7 8 urine using a nanophotonic POC biosensor in a label free configuration20 or in a planar 9 waveguide based biosensors using fluorescent labeled antibodies.21 Nevertheless, this 10 11 lipopolysaccharide is not only found in Mtb, but also in other mycobacteria.31 Other examples 12 13 are based on the use of Surface Plasmon Resonance (SPR) biosensing for the detection of the 14 CFP-10 antigen in urine,32,33 Ag85 in sputum34 and DNA fragments in plasma,24 the use of a 15 16 mobile-phone mediated plasmonic ELISA to detect ESAT-6 and CFP-10,35 or the use of 17 electrochemical devices for the detection of volatile organic compounds (VOC) of Mtb (i.e. 18 19 cyclohexane and benzene derivatives).36 The use of these biosensors can be inexpensive, simple 20 21 and fast, but also suffer drawbacks from variable and inconsistent sensitivity (14 – 86 %) and 22 low specificity (47 - 89 %).37–40 23 24 A critical point in developing a new POC test for TB diagnosis is the selection of the 25 26 appropriate biomarker to be detected. The heat shock protein HspX (known as α-crystalline 27 homologous or 16 KDa) is secreted along the growth curve of Mtb and is a promising Mtb 28 29 biomarker for the early diagnosis of TB24,26,41 or for the construction of new vaccines .42,43 Heat 30 31 shock proteins prevent the aggregation of denatured proteins and facilitate the reassembly by 32 other chaperone proteins. This function is particularly important in Mtb because it must 33 34 withstand extreme environments once it enters the alveolar macrophages, and in turn, forms the 35 granulomas in the lungs.44 Consequently, HspX stabilizes proteins and cellular structures to 36 37 survive in this hostile environment, providing a high degree of resistance to the bacteria, 38 escaping the host’s innate immune system. This protein is considered as one of the main 39 40 immunologically active mycobacterial antigen after infection and it is expressed at high levels 41 42 by bacterial pathogens during adaptation for intracellular survival,26 which makes it a promising 43 biomarker for the early, rapid and direct detection of Mtb.42 The detection of HspX for active 44 45 TB has been evaluated mainly by ELISA in human serum,25,45,46 pleural fluid,45,47 sputum45 and 46 cerebrospinal fluid (CSF).24,45 For latent TB detection has only tested by ELISA in serum.41,46,48 47 48 To evaluate the potential of the HspX as biomarker for the diagnosis of active tuberculosis, we 49 50 have implemented a quantitative and label-free SPR based immunoassay for its direct detection 51 52 in human samples. A direct immunoassay approach using specific monoclonal antibodies 53 previously immobilized on the sensor surface has been fully established, achieving excellent 54 55 sensitivity and reproducibility. Pre-treated sputum from human samples was directly evaluated, 56 exemplifying the achievement of a rapid and simple test for quantification of the HspX protein. 57 58 The SPR biosensor has been designed as a portable, user-friendly POC device that requires low 59 60 ACS Paragon Plus Environment Page 5 of 22 ACS Infectious Diseases 1 2 3 sample volume of pre-treated sputum from TB patients. Our approach could offer a valuable 4 5 alternative to current available strategies for the early detection and diagnosis of tuberculosis. 6 7 Results and Discussion 8 9 Cloning, expression and purification of rHspX protein 10 We have adopted a strategy to express the recombinant HspX protein because we need to obtain 11 12 a sufficient amount of this antigen in order to perform the optimization of the immunoassays 13 14 and the quantification of the pre-treated sputums. First, the PCR amplification of hspX gene 15 (436 bp) was confirmed by agarose gel electrophoresis as shown in Fig. S2 in SI. The amplified 16 17 gene was cloned into the vector pET100/D-TOPO and the ligation was evaluated by PCR 18 amplification with the T7 promoter primers of the vector (data not shown). PCR fragments of 19 20 the expected size (711 bp), corresponding to the hspX gene (436 bp) inserted in the TOPO 21 cloning site, were confirmed for correct orientation by a digestion with two restriction enzymes: 22 23 EcoRI and BgLII (Fig. S3 in SI). Sanger sequencing confirms the correct rhspX gene in two of 24 25 three selected clones (Fig. S4 in SI). Clone 10 had a mutation in the nucleotide 307 (an A for a 26 G) from start codon and was not used for further analysis. 27 28 Transformation of E. coli BL21 (DE3) cells were performed in order to express the rHspX by 29 30 induction with IPTG. According to Taylor et al. [43], HspX is expressed in the stationary phase. 31 Therefore, expression of rHspX (diluted 1/10 and 1/100) was analyzed at 4 and 16 h of growth 32 33 in a western blot using anti-HisG (1:10000) against the N-terminal 6xHis tag of the vector. Fig. 34 35 S5 in SI, shows the expression of a protein of 16 kDa according to the molecular weight size 36 marker. Expression at 4 h of induction was higher than at 16 h according with the intensity of 37 38 the bands (Fig. S5 in SI). Likewise, detection with mAb anti-HspX (1:5000) was assessed (data 39 not shown). Solubility tests were carried out to determine in which fractions the highest amount 40 41 of recombinant protein was present. The rHspX is obtained in both the soluble and the insoluble 42 43 fraction but a highest concentration is present in the insoluble fraction (Fig. S6 in SI). Protein 44 dimers and trimers are also observed due to bands with a double and triple molecular weight.49 45 46 The purification of the rHspX was performed by affinity chromatography (using the 6xHis- 47 48 tagged recombinant protein). The presence of the protein was visualized by SDS-PAGE and 49 western blot. The rHspX was eluted from the retention fraction R1 and R2, as can be seen in 50 51 Fig. S7 in SI. The protein was dialyzed, lyophilized and diluted in PBS buffer. A stock solution 52 53 with a final concentration of 84.63 µg mL-1 (as determined with the Bradford test) was used for 54 the assay development. 55 56 Direct immunoassay development 57 58 Ideally, a POC device should allow the detection of analytes in a single step and, if possible, 59 directly in the biological fluid. Thus, a direct assay, based on the immobilization of anti-HspX 60 ACS Paragon Plus Environment ACS Infectious Diseases Page 6 of 22 1 2 3 as the bioreceptor on the gold sensor chips was selected. A schematic representation of the 4 5 biofunctionalization strategy, based on the covalent attachment of the antibodies to carboxyl- 6 modified surfaces, and the subsequent direct detection are shown in Fig. 1. Parameters as the 7 8 composition of the carboxyl SAM and the buffer conditions were evaluated in order to provide 9 an optimal and efficient immobilization. Gold sensor chips were modified using thiol chemistry, 10 11 preparing a mixed SAM of organic thiols that incorporate a chain of ethylene glycol (EG) in 12 13 their structure to reduce non-specific adsorption of biological samples.50 A thiol with a 14 carboxylic end (EG-COOH) was used to attach the antibodies, and another one with a hydroxyl 15 16 end (EG-OH) to space out the antibodies. Two different ratios 1 mM (3:7 and 7:3) were 17 assessed and the effect of the pH in the immobilization was also evaluated. The immobilization 18 19 is more efficient at low pHs, mainly at pH 4.0-4.5 as it is summarized in the Fig. S9A in SI. 20 21 This can be due to the establishment of more favorable electrostatic interactions prior to the 22 coupling between the SAM and the antibody. In this way, the mAb is concentrated slightly on 23 24 the sensor surface, which increases the reaction yield. All steps of the immobilization of the 25 mAb (20 µg mL-1) for both ratios were monitored in-situ in real time by the plasmonic biosensor 26 27 (see Fig. 2A). 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 Fig. 1. Scheme of the direct assay employed for the detection of the HspX protein 49 50 According to the immobilization values, a pH = 4.0 was selected and the calibration curves 51 52 based on inter-assays (three calibration curves obtained with three different biofunctionalized 53 chips) for rHspX were obtained for both thiol ratios (see Fig. 2B). The detectability was similar 54 55 for both ratios (see Table 1), although the maximum signal of immobilization was higher for the 56 COOH:OH ratio of 7:3, which was finally selected. Representative real-time sensorgrams at 57 58 different rHspX concentrations in PBST are shown in Fig. 2C. In all cases, the average signal- 59 to-noise ratio was very similar (ΔλSPR= 0.0023 ± 0.0009). This low level of noise directly 60 ACS Paragon Plus Environment Page 7 of 22 ACS Infectious Diseases 1 2 3 correlated with a low LOD below 1 ng mL-1 (see Table 1). Furthermore, the coefficients of 4 5 variability (CV) of the inter-assays (especially for the selected combination) confirm the 6 excellent reproducibility and robustness of immunoassays. Finally, the exceptional robustness 7 8 of the antibody-coated surface robustness was demonstrated through successive regeneration 9 cycles, being possible to get reusable surfaces for at least 33 cycles (see Fig. S9B in SI). 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Fig. 2. (A) Real-time SPR sensorgram showing the immobilization steps over 1 mM COOH: 40 OH thiols ratios 3:7 and 7:3 (activation of carboxylic groups, mAb anti-HspX immobilization 41 (20 µg mL-1) in acetate buffer pH 4.0 and blocking with ethanolamine of remaining activated 42 groups). (B) Calibration curves of inter-assay in PBST over mAb anti-HspX biofunctionalized 43 44 (20 µg mL-1, pH 4.0) in COOH:OH ratio 3:7 (pink), 7:3 (green), and BSA as a negative control 45 (blue). (C) Real time sensorgrams showing the ΔλSPR for rHspX over mAb biofunctionalized 46 (20 µg mL-1, pH 4.0) in COOH:OH ratio 7:3 at different rHspX concentrations in PBST. 47 48 49 Sputum matrix effect on rHspX immunoassay 50 Sputum is a viscous mixture of cellular and biochemical composition of the peripheral airways 51 52 and the alveolar compartment that appears when there is a chronic inflammatory disease of the 53 respiratory tract such as asthma, chronic bronchitis, tuberculosis or cystic fibrosis. It consists of 54 55 95 % water and 5 % solids that vary according to the disease (i.e. carbohydrates, lipids, DNA, 56 filamentous actin, lipids and proteoglycans).51 This composition might affect the performance of 57 58 the biosensor (i.e. hindering antibody-antigen interaction or causing non-specific adsorptions). 59 60 Thus, in order to assess the diagnosis potential of the biosensor, we evaluated the feasibility of ACS Paragon Plus Environment ACS Infectious Diseases Page 8 of 22 1 2 3 directly measuring sputum samples. As can be observed in the Fig. 3A, untreated sputum 4 5 coming from non-tuberculous samples (UNTSS) directly injected resulted in a high signal 6 (ΔλSPR= 1.29 ± 0.023) (purple sensorgram), as a result of the non-specific adsorptions of the 7 8 viscous fluid onto the functionalized sensor chip. However, applying a standard 9 decontamination protocol commonly used with this kind of samples (i.e. PNTSS) resulted in a 10 11 much lower background (green sensorgram), but not enough to completely eliminate the 12 13 nonspecific binding. This was finally achieved by further diluting the pre-treated sample in 14 PBST 1:1 (red sensorgram). Under these conditions the interaction of the rHspX with the 15 16 immobilized antibody was unaffected (see orange and blue sensorgrams for rHspX sample 17 prepared in PBST and PNTSS:PBST 1:1, respectively). The Fig. S9C in the SI shows the real 18 19 time sensorgrams for different rHspX concentrations prepared in PNTSS (1:1) and the 20 21 calibration curve obtained with these conditions is graphed in the Fig. 3B and Table 1. Although 22 the signals are slightly lower in diluted pretreated sputum, the detectability parameters are 23 24 analogous within the linear range (from the LOQ around 2.1 – 2.4 ng mL-1 and LOD around 0.6 25 – 0.7 ng mL-1 to saturation signal). CV values, between 5 – 6 % for the inter-assays confirm the 26 27 excellent reproducibility to enable the direct detection of very low concentrations of HspX in 28 pre-treated sputum samples even after a dilution of 50 %. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Fig. 3. (A) Matrix effect of untreated non tuberculous sputum (UNTSS, purple line), pre-treated 45 non tuberculous sputum (PNTSS, green line), and PNTSS diluted in PBST 1:1 (red line) over an 46 antibody coated sensor chip (COOH:OH ratio 7:3). No rHspX has been added. Orange and blue 47 48 line show the rHspX binding (1.0 µg mL-1) in PBST and PTNSS:PBST 1:1 respectively. (B) 49 Calibration curves for rHspX protein in PBST and PNTSS:PBST 1:1. Each point represents the 50 mean ± SD of three replicates. 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 9 of 22 ACS Infectious Diseases 1 2 3 Table 1. Analytical features of direct immunoassays in buffer and pre-treated sputum 4 LOD (ng mL-1)a LOQ (ng mL-1)a 5 COOH:OH Buffer conditions Protein 6 Ratio Mean ± SD CV (%) Mean ± SD CV (%) 7 3:7 rHspX PBST 0.92 ± 0.14 15.2 3.13 ± 0.48 15.3 8 7:3 rHspX PBST 0.72 ± 0.04 5.55 2.41 ± 0.13 5.39 9 7:3 rHspX PNTSS:PBST (1:1) 0.63 ± 0.03 4.76 2.12 ± 0.11 5.19 a: LOD and CV obtained from inter-assays (three calibration curves obtained with three different chips) 10 11 12 Accuracy study with blind samples 13 14 The accuracy study was carried out evaluating seven blind samples (B1-B7) of rHspX protein 15 prepared in PNTSS:PBST 1:1 and measured by triplicate. Known concentrations were chosen 16 17 within and over the working range of the calibration curve. Samples B2, B4 and B6 required 18 additional dilutions (2, 3 and 5-fold respectively) to be within the working range. The results are 19 20 summarized in Table 2 and as can be observed, recoveries between 97 and 112 % were 21 22 obtained, which is within the commonly accepted values in conventional immunoassays (i.e. 80- 23 120 %), confirming good levels of accuracy in pre-treated sputum. 24 25 26 Table 2. Accuracy study with blind samples and the SPR immunosensor 27 Real concentration SPR immunosensor 28 Samples Recovery (%) CV (%) (ng mL-1) Mean ± SD (ng mL-1)a 29 B1 70 67.89 ± 1.97 96.99 2.90 30 B2 200 200.9 ± 3.35 100.4 1.67 31 B3 50 51.74 ± 2.31 103.5 4.46 32 B4 300 297.8 ± 7.00 99.27 2.35 33 B5 10 11.28 ± 0.67 112.8 5.94 34 B6 500 501.0 ± 10.2 100.2 2.03 35 B7 20 20.18 ± 1.45 100.9 7.18 a: average of three measurements. 36 37 38 Analysis of patient sputum samples 39 40 Table 3 shows the level of bacilli seen by microscopy (baciloscopy) as well as the time to 41 42 positivity and culture results of these samples. We have evaluated 12 pre-treated sputum 43 samples (S1-S12) from five patients, all diagnosed with TB (Table 3A). Up to 22 pre-treated 44 45 sputum samples from non-tuberculous patients (S13 – S34) were also included as negative 46 controls (Table 3B) as described in the experimental section. All control samples were negative 47 48 for Mtb. NTM (nontuberculous mycobacteria) species were identified in three of them (M. 49 50 intracellulare, M. avium and M. celatum respectively in samples S19, S22 and S26). 51 52 Each sample was diluted 1:1 with PBST and measured directly in real-time and replicated twice. 53 Signals were interpolated in the calibration curve with PNTSS (1:1) (see Fig 3B (red curve)). 54 55 The HspX concentrations for each sample are summarized in Table 3. The median, maximum 56 and minimum values are shown in Fig. 4 for TB and non-TB patients, where significant 57 58 statistical difference can be observed between them. We found a faint correlation between the 59 60 concentration of HspX and the level of baciloscopy. The presence of a metabolically variable ACS Paragon Plus Environment ACS Infectious Diseases Page 10 of 22 1 2 3 population of bacilli in each sample could explain that result. Remarkably, in tuberculous 4 5 samples previously confirmed by culture methods and Xpert MTB/Rif® but with negative 6 baciloscopy, the detection of the HspX protein was possible, which demonstrates a high 7 8 sensitivity of our biosensor device (Table 3). Moreover, detection of HspX by SPR biosensor 9 was possible at least one week before positivity of the culture (Table 3A) thus allowing an early 10 11 diagnosis of the disease. Finally, the SPR procedure developed showed to be Mtb specific, due 12 13 to the absence of cross reactivity with HspX protein of other slow grower mycobacteria. 14 15 From our knowledge this is the first work reporting HspX in sputum samples, whose 16 concentration ranges between 116 to 175 ng mL-1. These values are lower than those reported in 17 18 cerebrospinal fluid, where HspX has been detected in concentrations up to 18 μg mL-1.24 19 20 200 21 22 150 23 HspX (ng mL-1) 24 25 100 26 27 50 28 29 0 30 Non-TB patients TB patients 31 Real samples 32 33 34 Figure 4. Levels of HspX protein concentration of pre-treated sputum samples from tuberculous 35 (TB) patients (n=12) and non-TB patients (n=22). Median, maximum and minimum values are 36 shown. 37 38 39 Conclusions 40 41 We have designed and developed a complete strategy for the direct and label-free detection of 42 HspX TB antigen employing a SPR biosensor. Our approach provides a reliable, reproducible 43 44 and sensitive detection of the protein in sputum samples, it requires low volume of sputum 45 samples and has a short turnaround time (only 35 - 40 min considering pre-treatment of sputum 46 47 and analysis) which makes it especially attractive for an early diagnosis of TB. An effective 48 protocol for the expression and purification of HspX protein has been set up, obtaining a high 49 50 expression yield in just four hours. This has allowed obtaining a sufficient amount of protein to 51 52 perform the entire assay development and optimization. The biofunctionalization strategy based 53 on EG thiols provides a simple and robust sensing layer for biofunctionalization although the 54 55 pre-treatment of sputum is still needed to obtain a reliable analysis of the biomarker. Applying 56 this pre-treatment facilitates however its direct analysis, without the need of using any 57 58 additional amplification step, reaching a LOD of 0.63 ng mL-1 and a working range between 2.1 59 60 and approximately 125 ng mL-1. The biosensor based assay shows an excellent accuracy, as ACS Paragon Plus Environment Page 11 of 22 ACS Infectious Diseases 1 2 3 demonstrated with recovery values close to 100 % when blind samples were analyzed. Finally, 4 5 the analysis of real sputum samples from TB patients confirmed the diagnostic potential of the 6 biosensor tool, being possible to observe significant statistical differences between TB and non- 7 8 TB samples. Concentrations of HspX protein up to 175 ng mL-1 were detected. From our 9 knowledge, this is the first reported concentration values reported for this biomarker in sputum. 10 11 Although more data might be needed to confirm the sensitivity and specificity that we have 12 13 seeing of this biomarker for early detection, our results pave the way towards the 14 implementation of such analysis in a portable, miniaturized POC platform as a fast and user- 15 16 friendly tool for clinical diagnosis practice, which could eventually be used during the check-up 17 in primary healthcare units. 18 19 Table 3. Clinical sputum samples analysis 20 21 Sputum [HspX] (µg mL-1) TTP (A) Patient BLa CV (%) Culture 22 sample Mean ± SD (days) 23 S1 1 4 146.6 ± 34.1 23.26 7 + 24 S2 1 4 123.2 ± 10.1 8.20 10 + 25 S3 1 2 147.0 ± 4.4 2.99 ND ND Tuberculous patients 26 S4 1 4 162.9 ± 4.3 2.63 ND ND 27 S5 2 0 140.0 ± 15.3 10.91 18 + 28 29 S6 2 1 134.0 ± 29.1 21.74 10 + 30 S7 2 0 116.3 ± 1.8 1.61 16 + 31 S8 2 0 129.1 ± 24.4 18.92 12 + 32 S9 3 1 142.4 ± 0.4 0.30 14 + 33 S10 4 2 175.0 ± 4.0 2.30 14 + 34 S11 5 4 141.5 ± 1.7 1.21 7 + 35 36 S12 5 4 143.6 ± 1.3 0.90 7 + 37 Sputum [HspX] (µg mL-1) TTP (B) Patient BLa CV (%) Culture 38 sample Mean ± SD (days) 39 S13 7 0 2.13 ± 0.18 8.39 NA NA 40 S14 8 0 1.88 ± 0.18 9.50 NA NA 41 S15 9 0 1.62 ± 0.17 10.9 NA NA 42 S16 10 0 2.51 ± 0.36 14.2 NA NA 43 S17 11 0 2.38 ± 0.18 7.51 NA NA 44 S18 12 0 1.88 ± 0.17 9.50 NA NA Non-Tuberculous patients 45 46 S19 13 0 2.64 ± 0.18 6.80 NA NA 47 S20 14 0 2.50 ± 0.34 13.6 NA NA 48 S21 15 0 2.56 ± 0.07 2.80 NA NA 49 S22 16 0 2.69 ± 0.11 4.01 NA NA 50 S23 17 0 2.67 ± 0.09 3.35 NA NA 51 S24 18 0 2.70 ± 0.09 3.32 NA NA 52 S25 19 0 1.95 ± 0.29 14.6 NA NA 53 S26 20 0 2.42 ± 0.23 9.62 NA NA 54 S27 21 0 2.41 ± 0.14 5.95 NA NA 55 S28 22 0 2.67 ± 0.09 3.35 NA NA 56 S29 23 0 2.65 ± 0.20 7.45 NA NA 57 58 S30 24 0 2.67 ± 0.13 4.70 NA NA 59 S31 25 0 2.84 ± 0.25 8.86 NA NA 60 S32 26 0 1.86 ± 0.12 6.70 NA NA ACS Paragon Plus Environment ACS Infectious Diseases Page 12 of 22 1 2 3 S33 27 0 2.41 ± 0.21 8.90 NA NA 4 S34 28 0 2.55 ± 0.05 2.11 NA NA 5 a:BL: Baciloscopy levels: 0 = none; 1= paucibacilary to 4= multibacilary); TTP: time to positivity of the culture; ND: 6 no data; NA: not applicable 7 8 9 Experimental section 10 Study patients and sputum sampling 11 12 A total of 12 pre-treated tuberculous sputum samples (PTSS, S1-S12) from five patients were 13 14 included in the study, and 22 pre-treated non-tuberculous sputum samples (PNTSS) from 22 15 patients (S13-S34) were evaluated and employed during the optimization of the immunoassays. 16 17 In addition, two samples of untreated non-tuberculous sputum sample (UNTSS) were used as a 18 control of sputum matrix. All samples (pre-treated or untreated samples) were collected in 19 20 sterile containers for diagnostic purposes by the Service of Clinical Microbiology from La Paz 21 University Hospital Madrid (Spain). The local Ethics Committee Hospital La Paz approved the 22 23 study protocol. Standard diagnostic procedures were performed to test the presence of Mtb 24 25 bacilli in the samples, including fluorescent acid-fast staining and growth in liquid medium 26 BACTEC MGIT 960 (Becton Dickinson) as well as in solid medium Löwenstein-Jensen. The 27 28 GenoType Mycobacterium CM (Bruker) or Xpert MTB/Rif® (Cepheid) were used for Mtb 29 confirmation. 30 31 Pretreatment for decontamination of samples has followed standard procedures. In brief, a 32 33 mixture of N-acetyl-L-cysteine, sodium citrate and NaOH was added to the sample until 34 35 complete dilution and then it was neutralized with PBS buffer pH 6.8 (BBL™ MycoPrep™, 36 BD). Finally, samples were centrifuged at 4 ºC, 3000 g for 15 min. The supernatant was 37 38 discarded in a bottle with sodium hypochlorite (5 g L-1) leaving a sediment of 500 µL. These 39 pretreated samples were stored at -20 ºC until analysis. Aliquots of these pre-treated samples 40 41 (PTSS) were placed in Eppendorf tubes containing a guanidine chloride buffer (preservative 42 43 agent that halts the bacterial metabolism) and glass beads. The untreated non-tuberculous 44 sample (UNTSS) was placed only in the guanidine chloride buffer. In order to lyse the bacteria, 45 46 the aliquots were subjected to 7 pulses of lysis with a speed of 5.0 m s-1 during 60 s, in a 47 fastprep-24, and maintaining the samples in ice between pulses. The supernatant was collected 48 49 by centrifugation at 4 ºC, 12000 rpm during 5 min and stored at -70 ºC until analysis.52 The 50 obtained lysate supernatants were evaluated with the SPR biosensor. 51 52 53 54 Bacterial strains and growth conditions 55 Recombinant protein (rHspX) was produced in bacterial expression systems at Corporación 56 57 Corpogen (Bogotá, Colombia). Reference DNA H37Rv of Mtb was obtained from Corporación 58 para Investigaciones Biológicas (CIB) (Medellín, Colombia). Conservation strain E. coli TOP10 59 60 and competent strain E. coli BL21 (DE3) were acquired from Thermo Fisher (Massachusetts, ACS Paragon Plus Environment Page 13 of 22 ACS Infectious Diseases 1 2 3 USA). Strains were grown in Luria-Bertani broth (LB) and LB agar plates, ampicillin (100 µg 4 5 mL-1) was added to the media when required. 6 7 Cloning, expression and purification of rHspX 8 Amplification of the encoding hspX gene, 436 base pairs (pb) was carried out by PCR using as 9 10 template the genomic DNA of Mtb H37Rv. The following upstream and downstream primers 11 12 were used for the PCR: 5'-CAC CGC CAC CAC CCT TCC-3' and 5'-TCA GTT GGT GGA 13 CCG GAT CT-3'. The amplification product was visualized by electrophoresis in agarose gel 14 15 stained with SYBR Green. The PCR product was purified and ligated into the vector pET100/D- 16 TOPO (5760 bp) (Thermo Fisher Scientific, Massachusetts, USA) and cells of E. coli TOP10 17 18 were transformed. The plasmid map of pET100/D-TOPO vector with the gene hspX is depicted 19 in the Fig. S1 in SI. The clone containing the hspX gene was verified by sequence (Macrogen, 20 21 Inc, Korea). For the expression of the rHspX, the plasmid pET100-hspX was transformed in E. 22 23 coli BL21 (DE3) (Invitrogen, Carlsbad, CA, USA). The transformed E. coli was plated on LB 24 media containing ampicillin (100 µg mL-1). The recombinant clone was cultured in liquid media 25 26 (LB) supplemented with ampicillin and the expression of the protein was induced with 27 isopropyl-β-D-1-thiogalactopyranoside solution (IPTG, 0.5 mM), incubated at 37ºC, 200 rpm 28 29 for 4 and 16 h [27,43,45]. Cells were broken by FastPrep (Thermo Scientific, MA, USA), (3 30 31 cycles at a speed of 5.0 m s-1, for 40 s, leaving samples on ice for 5 min between each cycle). 32 33 The soluble and insoluble fractions of rHspX were analyzed by polyacrylamide gel 34 electrophoresis SDS-PAGE 14 % and by western blot. The recombinant protein was detected 35 36 using anti-HisG (anti-histidine) 1:10000 f or 1 h and with the monoclonal antibody anti-HspX 37 (Abcam, Cambridge, UK) 1:5000 for 2 h. Anti-mouse IgG 1:5000 was used as a secondary 38 39 antibody. The immunorecognition was detected by the peroxidase enzyme kit.53 Subsequently, 40 41 protein was purified by Ni2+ chromatography of 50 % Ni-NTA purification system following the 42 manufacturer instructions. Both, the non-retained fraction and the retained fraction were 43 44 evaluated by electrophoresis and western blot. Finally, the culture of E. coli BL21 (DE3) was 45 scaled to 1.5 L, purified and dialyzed by Float-A-lyzer G2 Dialysis device MWCO: 8-10 KDa 46 47 (Spectrumlabs, FL, USA) in Milli-Q water, then, lyophilized at 0.5 mbar and -51.7 ºC (Telstar 48 Cryodos, UK). The final concentration of HspX protein was quantified by Bradford’s reagent 49 50 (Bio-Rad, UK). 51 52 SPR biosensor device 53 54 The rHspX detection assays were developed with a homemade compact SPR device. A 55 schematic representation of the sensing principle and a picture of the plasmonic biosensor 56 57 platform are shown in Fig. S8A in the SI. The optical components of the SPR biosensor and the 58 fluidic channel to keep running fluid continuously over the gold sensor chip (glass surface 59 60 coated with Ti 1 nm/Au 49 nm) have been previously described.54 The resonance spectra of the ACS Paragon Plus Environment ACS Infectious Diseases Page 14 of 22 1 2 3 sensor chips shows a minimum in the reflection at λSPR (675 nm), which is monitored in real- 4 5 time via polynomial fit using custom designed readout software (see Fig. S8B in SI). Chemical 6 interactions occurring at the gold sensor surface (i.e. biointeraction events) generate variations 7 8 of the RI which correlates with wavelength displacements of the minimum (ΔSPR). Therefore, 9 10 SPR is directly related to mass changes resulting from binding events (leading to higher RI and 11 shifts to higher wavelengths) or desorptions (leading to lower RI and shifts to lower 12 13 wavelengths) occurring on the surface (Fig. S8C in SI). 14 15 Surface functionalization 16 17 Gold sensor chips were cleaned by heating and sonication for 1 min each step with acetone, 18 ethanol and Milli-Q water. Then, the chips were dried with N2 stream and dipped in a freshly 19 20 prepared piranha solution (H2SO4/H2O2, 7:3) for 1 min. Finally, they were rinsed with Milli-Q 21 water, ethanol and dried with N2 stream. A mixed self-assembled monolayer (SAM) was formed 22 23 on the gold sensor surface with a solution of thiols HS-C11-(EG)6-O-CH2-COOH (EG-COOH) 24 and HS-C11-(EG)4-OH (EG-OH) (EG-COOH:EG-OH,1 mM). The sensor chips were immersed 25 26 in the thiol solution and heated at 40 ºC for 10 min. Then, they were incubated overnight at 27 28 room temperature. The sensor chips were placed in the optical platform and closed with the 29 microfluidic cell for biofunctionalization. The antibody immobilization was performed under a 30 31 continuous flow of Milli-Q water. The carboxylic groups of the generated SAM were first 32 activated as carbodiimide esters by flowing a mixed solution of EDC/NHS (0.2 M /0.05 M) in 33 34 MES buffer for 30 min at 20 µL min-1. Subsequently, an anti-HspX monoclonal antibody (mAb) 35 solution (20 µg mL-1) was injected at 15 µL min-1. The remaining unreacted groups were then 36 37 deactivated with an aqueous solution of ethanolamine (1 M pH 8.5) for 2 min at 30 µL min-1. 38 39 Finally, antibody-biofunctionalized sensor chips were kept under a continuous flow of PBS or 40 PBST buffer at 30 µL min-1. 41 42 SPR direct immunoassay performance 43 44 A calibration curve of the detection of the rHspX was obtained after evaluating different protein 45 concentrations (between 7.8 and 1000 ng mL-1) from a stock solution of rHspX (84.6 µg mL-1). 46 47 Triplicate analysis was carried out for each concentration. Each standard sample was injected 48 49 randomly over the antibody coated sensor surface at a flow rate of 30 µL min-1. A regeneration 50 process to dissociate the antigen-antibody interaction was achieved by injecting 3 mM NaOH at 51 52 35 µL min-1 for 120 s. The average signal and standard deviation (ΔλSPR ± SD) was graphed 53 versus the injected analyte concentration. The data were fitted to a one-site specific binding 54 55 model curve: 56 𝑨𝒙 57 𝒚= 58 𝑩+ 𝒙 59 60 ACS Paragon Plus Environment Page 15 of 22 ACS Infectious Diseases 1 2 3 Where x is the concentration, y is the response signal, A is the extrapolated maximum signal 4 5 obtained by the specific analyte and B is related to the equilibrium binding constant. The LOD 6 (Limit of Detection) and the LOQ (Limit of Quantification) were calculated as the concentration 7 8 corresponding to the blank signal plus three and ten times its SD respectively. BSA was used to 9 confirm the specificity of the assays. Data analysis and fitting was performed using Origin Pro 10 11 and GraphPad-Prism software. 12 13 Real sputum sample evaluation 14 15 For the evaluation of the sputum samples, non-tuberculous samples (UNTSS and PNTSS) as 16 negative controls were tested. Undiluted UNTSS and PNTSS samples and PNTSS samples 17 18 diluted 1:1 with PBST were injected in the biosensor at a flow rate of 30 µL min-1. Each sample 19 was measured by triplicate. In order to evaluate the accuracy of the assay, seven blind samples 20 21 of known concentration of rHspX (B1-B6) were prepared in PNTSS using concentrations within 22 23 and over the working range and then diluted in PBST 1:1 prior analysis. Blind samples B2, B4 24 and B6 with concentrations over the working range, required an additional dilution factor of 2, 3 25 26 and 5-fold respectively. The % Recovery was calculated as: 27 [𝐁𝐥𝐢𝐧𝐝 𝐬𝐚𝐦𝐩𝐥𝐞 ]𝒄𝒂𝒍𝒄𝒖𝒍𝒂𝒕𝒆𝒅 28 Recovery %= × 𝟏𝟎𝟎 [𝐁𝐥𝐢𝐧𝐝 𝐬𝐚𝐦𝐩𝐥𝐞]𝒓𝒆𝒂𝒍 29 30 31 Finally, 15 tuberculous samples PTSS (S1-S15) and 22 non-tuberculous samples PNTSS (S16- 32 S37) were evaluated with the optimized protocol. Signals were interpolated in the calibration 33 34 curve performed with PNTSS (1:1). Real samples were also measured in triplicate and the 35 average, SD and the coefficient of variability (CV) were calculated. 36 37 38 39 Acknowledgements 40 41 The ICN2 is funded by the CERCA programme / Generalitat de Catalunya. The ICN2 is 42 supported by the Severo Ochoa Centres of Excellence programme, funded by the Spanish 43 44 Research Agency (AEI, grant no. SEV-2017-0706). We thank COLCIENCIAS for supporting 45 46 this work through the grant No. 0375-2013 with number project 6570577636375 and the 47 Universidad Central (Colombia) for supporting this work through the Research Cluster on 48 49 Converging Sciences and Technology (NBIC). We also thank the clinician of La Paz University 50 Hospital for helping in the recruitment of TB patients. 51 52 53 References 54 55 56 [1] Magana-Arachchi, D., Medagedara, D., Thevanesam, V. (2011) Molecular 57 characterization of Mycobacterium tuberculosis isolates from Kandy, Sri Lanka. Asian 58 59 Pacific J. Trop. Dis. 1, 181–186. doi:10.1016/S2222-1808(11)60024-8. 60 [2] Cheon, S.A., Cho, H.H., Kim, J., Lee, J., Kim, H.J., Park, T.J. (2016) Recent ACS Paragon Plus Environment ACS Infectious Diseases Page 16 of 22 1 2 3 tuberculosis diagnosis toward the end TB strategy. J. Microbiol. Methods. 123, 51–61. 4 5 doi:10.1016/j.mimet.2016.02.007. 6 [3] World Health Organisation, Global Health TB Report, 2018. 7 8 [4] Goletti, D., Petruccioli, E., Joosten, S.A., Ottenhoff, T.H.M. (2016) Tuberculosis 9 biomarkers: From diagnosis to protection. Infect. Dis. Rep. 8, 24–32. 10 11 doi:10.4081/idr.2016.6568. 12 13 [5] Dheda, K., Barry, C.E., Maartens, G. (2016) Tuberculosis. Lancet. 387, 1211–1226. 14 doi:10.1016/S0140-6736(15)00151-8. 15 16 [6] Wallis, R.S., Maeurer, M., Mwaba, P., Chakaya, J., Rustomjee, R.. Migliori, G.B., 17 Marais, B., Schito, M., Churchyard, G., Swaminathan, S., Hoelscher, M., Zumla, A. 18 19 (2016) Tuberculosis-advances in development of new drugs, treatment regimens, host- 20 21 directed therapies, and biomarkers. Lancet Infect. Dis. 16, e34–e46. doi:10.1016/S1473- 22 3099(16)00070-0. 23 24 [7] Marais, B.J. (2016) The global tuberculosis situation and the inexorable rise of drug- 25 resistant disease. Adv. Drug Deliv. Rev. 102, 3–9. doi:10.1016/j.addr.2016.01.021. 26 27 [8] Yon Ju Ryu, M.D. (2015) Diagnosis of pulmonary tuberculosis: Recent advances and 28 diagnostic algorithms. Tuberc. Respir. Dis. 78 (2), 64–71. doi:10.4046/trd.2015.78.2.64. 29 30 [9] Hong, S.C., Lee, J., Shin, H.C., Kim, C.M., Park, J.Y., Koh, K., Kim, H.J., Chang, C.L., 31 32 Lee, J. (2011) Clinical immunosensing of tuberculosis CFP-10 in patient urine by 33 surface plasmon resonance spectroscopy. Sens. Act. B Chem. 160, 1434–1438. 34 35 doi:10.1016/j.snb.2011.10.006. 36 [10] Viñuelas-Bayón, J., Vitoria, M.A., Samper, S. (2017) Rapid diagnosis of tuberculosis. 37 38 Detection of drug resistance mechanisms. Enferm. Infecc. Microbiol. Clin. 35, 520–528. 39 40 doi:10.1016/j.eimc.2017.01.015. 41 [11] Bansal, R., Sharma, P.K., Jaryal, S.C., Gupta, P.K., Kumar, D. (2017) Comparison of 42 43 Sensitivity and Specificity of ZN and Fluorescent Stain Microscopy with Culture as 44 Gold Standard. J. Tuberc. Res. 5, 118–128. doi:10.4236/jtr.2017.52013. 45 46 [12] Reza, L.W., Satyanarayna, S., Enarson, D.A., Kumar, A.M.V., Sagili, K., Kumar, S., 47 Prabhakar, L.A., Devendrappa, N.M., Pandey, A., Wilson, N., Chadha, S., Thapa, B., 48 49 Sachdeva, K.S., Kohli, M.P. (2013) LED-Fluorescence Microscopy for Diagnosis of 50 51 Pulmonary Tuberculosis under Programmatic Conditions in India. PLoS One. 8, e75566 52 doi:10.1371/journal.pone.0075566. 53 54 [13] Ismail, N.A., Omar, S. V., Lewis, J.J., Dowdy, D.W., Dreyer, A.W., Van Der Meulen, 55 H., Nconjana, G., Clark, D.A., Churchyard, G.J. (2015) Performance of a novel 56 57 algorithm using automated digital microscopy for diagnosing tuberculosis. Am. J. Respir. 58 Crit. Care Med. 191, 1443–1449. doi:10.1164/rccm.201502-0390OC. 59 60 [14] Ruhwald, M., Aggerbeck, H. , Gallardo, R.V., Hoff, S.T., Villate, J.I., Borregaard, B., ACS Paragon Plus Environment Page 17 of 22 ACS Infectious Diseases 1 2 3 Martinez, J.A., Kromann, I., Penas, A., Anibarro, L.L., de Souza-Galvão, M.L., Sánchez, 4 5 F., Rodrigo-Pendás, J.Á., Noguera-Julian, A., Martínez-Lacasa, X., Tuñez, M.V., 6 Fernández, V.L., Millet, J.P., Moreno, A., Cobos, N., Miró, J.M., Roldan, L., Orcau, A., 7 8 Andersen, P., Caylá, J.A. (2017) Safety and efficacy of the C-Tb skin test to diagnose 9 Mycobacterium tuberculosis infection, compared with an interferon γ release assay and 10 11 the tuberculin skin test: a phase 3, double-blind, randomised, controlled trial. Lancet 12 13 Respir. Med. 5, 259–268. doi:10.1016/S2213-2600(16)30436-2. 14 [15] Mazurek, G.H., Jereb, J., Vernon, A., LoBue, P., Goldberg, S., Castro, K. (2010) 15 16 Updated guidelines for using interferon gamma release assays to detect Mycobacterium 17 tuberculosis infection-United States. Morb. Mortal. Wkly Rep. 59, 1–25. 18 19 doi:10.1097/00043764-197203000-00020. 20 21 [16] Ling, D.I., Nicol, M.P., Pai, M., Pienaar, S., Dendukuri, N., Zar, H.J. (2013) Incremental 22 value of T-SPOT.TB for diagnosis of active pulmonary tuberculosis in children in a 23 24 high-burden setting: A multivariable analysis. Thorax. 68, 860–866. 25 doi:10.1136/thoraxjnl-2012-203086. 26 27 [17] Mori, Y., Kanda, H., Notomi, T. (2013) Loop-mediated isothermal amplification 28 (LAMP): Recent progress in research and development. J. Infect. Chemother. 19, 404– 29 30 411. doi:10.1007/s10156-013-0590-0. 31 32 [18] Nagai, K., Horita, N., Yamamoto, M., Tsukahara, T., Nagakura, H., Tashiro, K., Shibata, 33 Y., Watanabe, H., Nakashima, K., Ushio, R., Ikeda, M., Narita, A., Kanai, A., Sato, T., 34 35 Kaneko, T. (2016) Diagnostic test accuracy of loop-mediated isothermal amplification 36 assay for Mycobacterium tuberculosis: Systematic review and meta-analysis. Sci. Rep. 6, 37 38 3–9. doi:10.1038/srep39090. 39 40 [19] Chakravorty, S., Simmons, A.M., Rowneki, M., Parmar, H., Cao, Y., Ryan, J., Banada, 41 P.P., Deshpande, S., Shenai, S., Gall, A., Glass, J., Krieswirth, B., Schumacher, S.G., 42 43 Nabeta, P., Tukvadze, N., Rodrigues, C., Skrahina, A., Tagliani, E., Cirillo, D.M., 44 Davidow, A., Denkinger, C.M., Persing, D., Kwiatkowski, R., Jones, M., Alland, D. 45 46 (2017) The New Xpert MTB/RIF Ultra: Improving Detection of Mycobacterium 47 tuberculosis and Resistance to Rifampin in an Assay Suitable for Point-of-Care Testing. 48 49 MBio. 8, 1–12. doi:10.1128/mbio.00812-17. 50 51 [20] Ramirez-Priego, P., Martens, D., Elamin, A.A., Soetaert, P., Van Roy, W., Vos, R., 52 Anton, B., Bockstaele, R., Becker, H., Singh, M., Bienstman, P., Lechuga, L.M. (2018) 53 54 Label-Free and Real-Time Detection of Tuberculosis in Human Urine Samples Using a 55 Nanophotonic Point-of-Care Platform. ACS Sensors. 3, 2079–2086. 56 57 doi:10.1021/acssensors.8b00393. 58 [21] Mukundan, H., Kumar, S., Price, D.N., Ray, S.M., Lee, Y.J., Min, S., Eum, S., Kubicek- 59 60 Sutherland, J., Resnick, J.M., Grace, W.K., Anderson, A.S., Hwang, S.H., Cho, S.N., ACS Paragon Plus Environment ACS Infectious Diseases Page 18 of 22 1 2 3 Via, L.E., Barry, C., Sakamuri, R., Swanson, B.I. (2012) Rapid detection of 4 5 Mycobacterium tuberculosis biomarkers in a sandwich immunoassay format using a 6 waveguide-based optical biosensor. Tuberculosis. 92, 407–416. 7 8 doi:10.1016/j.tube.2012.05.009. 9 [22] Poulakis, N., Gritzapis, A.D., Ploussi, M., Leventopoulos, M., Papageorgiou, C. V., 10 11 Anastasopoulos, A., Constantoulakis, P., Karabela, S., Vogiatzakis, E., Tsilivakos, V. 12 13 (2016) Intracellular ESAT-6: A new biomarker for Mycobacterium tuberculosis 14 infection. Cytom. Part B - Clin. Cytom. 90, 312–314. doi:10.1002/cyto.b.21220. 15 16 [23] Ren, N., JinLi, J., Chen, Y., Zhou, X., Wang, J., Ge, P., Khan, F.A., Zhang, L., Hu, C., 17 Robertson, I.D., Chen, H., Guo, A. (2018) Identification of new diagnostic biomarkers 18 19 for Mycobacterium tuberculosis and the potential application in the serodiagnosis of 20 21 human tuberculosis. Microb. Biotechnol. 11, 893–904. doi:10.1111/1751-7915.13291. 22 [24] Haldar, S., Sankhyan, N., Sharma, N., Bansal, A., Jain, V., Gupta, V.K., Juneja, M., 23 24 Mishra, D., Kapil, A., Singh, U.B., Gulati, S., Kalra, V., Tyagi, J.S. (2012) Detection of 25 Mycobacterium tuberculosis GlcB or HspX Antigens or devR DNA Impacts the Rapid 26 27 Diagnosis of Tuberculous Meningitis in Children. PLoS One. 7, e0044630. 28 doi:10.1371/journal.pone.0044630. 29 30 [25] de Sousa, E.M., da Costa, A.C., Trentini, M.M., de Araújo Filho, J.A., Kipnis, A., 31 32 Junqueira-Kipnis, A.P. (2012) Immunogenicity of a Fusion Protein Containing 33 Immunodominant Epitopes of Ag85C, MPT51, and HspX from Mycobacterium 34 35 tuberculosis in Mice and Active TB Infection PLoS One. 7, e0047781 36 doi:10.1371/journal.pone.0047781. 37 38 [26] Rajpal, S.K., Snehal, S.W., Milind, S.P., Hemant, J.P., Girdhar, M.T., Hatim, F.D. 39 40 (2011) Mycobacterium Tuberculosis Heat Shock Protein 16 as a Potential Marker for 41 Latent TB: A Preliminary Findings. J. Clin. Cell. Immunol. 2, 115 doi:10.4172/2155- 42 43 9899.1000115. 44 [27] Shekhawat, S.D., Jain, R.K., Gaherwar, H.M., Purohit, H.J., Taori, G.M., Daginawala, 45 46 H.F., Kashyap, R.S. (2014) Heat shock proteins: Possible biomarkers in pulmonary and 47 extrapulmonary tuberculosis. Hum. Immunol. 75, 151–158. 48 49 doi:10.1016/j.humimm.2013.11.007. 50 51 [28] Wu, X., Yang, Y., Zhang, J., Li, B., Liang, Y., Zhang, C., Dong, M. (2010) Comparison 52 of antibody responses to seventeen antigens from Mycobacterium tuberculosis. Clin. 53 54 Chim. Acta. 411, 1520–1528. doi:10.1016/j.cca.2010.06.014. 55 [29] Pai, M., Behr, M., Dowdy, D., Dheda, K., Divangahi, M., Boehme, C., Ginsberg, A., 56 57 Swaminathan, S., Spigelman, M., Getahun, H., Menzies, D., Raviglione, M. (2016) 58 Tuberculosis. Lancet. 2, 1–23. doi:10.1016/S0140-6736(19)30308-3. 59 60 [30] Srivastava, S.K., van Rijn, C.J.M., Jongsma, M.A. (2016) Biosensor-based detection of ACS Paragon Plus Environment Page 19 of 22 ACS Infectious Diseases 1 2 3 tuberculosis. RSC Adv. 6, 17759–17771. doi:10.1039/C5RA15269K. 4 5 [31] Shah, M, Hanrahan, C, Wang, ZY, Dendukuri, N, Lawn, SD, Denkinger, CM, Steingart, 6 KR. (2016) Lateral flow urine lipoarabinomannan assay for detecting active tuberculosis 7 8 in HIV-positive adults. Cochrane Database of Systematic Reviews, 5, CD011420 9 10.1002/14651858.CD011420.pub2 10 11 [32] Zou, F., Wang, X., Qi, F., Koh, K., Lee, J., Zhou, H., Chen, H. (2017) Magneto- 12 13 plamonic nanoparticles enhanced surface plasmon resonance TB sensor based on 14 recombinant gold binding antibody. Sens. Act. B Chem. 250, 356–363. 15 16 doi:10.1016/j.snb.2017.04.162. 17 [33] Hong, S.C., Chen, H., Lee, JPark, ., H.-K., Kim, Y.S., Shin, H.-C., Kim, C.-M., Park, 18 19 T.J., Lee, S.J., Koh, K., Kim, H.-J., Chang, C.L., Lee, J. (2011) Ultrasensitive 20 21 immunosensing of tuberculosis CFP-10 based on SPR spectroscopy. Sens. Act. B Chem. 22 156, 271–275. doi:10.1016/j.snb.2011.04.032. 23 24 [34] Trzaskowski, M., Napiórkowska, A., Augustynowicz-Kopeć, E., Ciach, T. (2018) 25 Detection of tuberculosis in patients with the use of portable SPR device. Sens. Act. B 26 27 Chem. 260, 786–792. doi:10.1016/j.snb.2017.12.183. 28 [35] Shabut, A.M., Hoque Tania, M., Lwin, K.T., Evans, B.A., Yusof, N.A., Abu-Hassan, 29 30 K.J., Hossain, M.A. (2018) An intelligent mobile-enabled expert system for tuberculosis 31 32 disease diagnosis in real time. Expert Syst. Appl. 114, 65–77. 33 doi:10.1016/j.eswa.2018.07.014. 34 35 [36] Bhattacharyya, D., Smith, Y.R., Mohanty, S.K., Misra, M. (2016) Titania Nanotube 36 Array Sensor for Electrochemical Detection of Four Predominate Tuberculosis Volatile 37 38 Biomarkers. J. Electrochem. Soc. 163, B206–B214. doi:10.1149/2.0221606jes. 39 40 [37] Hsieh, S.C., Chang, C.C., Lu, C.C., Wei, C.F., Lin, C.S., Lai, H.C., Lin, C.W. (2012) 41 Rapid identification of Mycobacterium tuberculosis infection by a new array format- 42 43 based surface plasmon resonance method. Nanoscale Res. Lett. 7, 1–6. 44 doi:10.1186/1556-276X-7-180. 45 46 [38] Thanyani, S.T., Roberts, V., Siko, D.G.R., Vrey, P., Verschoor, J. A. (2008) A novel 47 application of affinity biosensor technology to detect antibodies to mycolic acid in 48 49 tuberculosis patients. J. Immunol. Methods. 332, 61–72. doi:10.1016/j.jim.2007.12.009. 50 51 [39] Steingart, K.R., Henry, M., Laal, S., Hopewell, P.C., Ramsay, A., Menzies, D., 52 Cunningham, J., Weldingh, K., Pai, M. (2007) Commercial serological antibody 53 54 detection tests for the diagnosis of pulmonary tuberculosis: A systematic review. PLoS 55 Med. 4, 1041–1060. doi:10.1371/journal.pmed.0040202. 56 57 [40] Wallis, R.S., Pai, M., Menzies, D., Doherty, T.M., Walzl, G., Perkins, M.D., Zumla, A. 58 (2010) Biomarkers and diagnostics for tuberculosis: progress, needs, and translation into 59 60 practice. Lancet. 375, 1920–1937. doi:10.1016/S0140-6736(10)60359-5. ACS Paragon Plus Environment ACS Infectious Diseases Page 20 of 22 1 2 3 [41] Castro-Garza, J., García-Jacobo, P., Rivera-Morales, L.G., Quinn, F.D., Barber, J., Karls, 4 5 R., Haas, D., Helms, S., Gupta, T., Blumberg, H., Tapia, J., Luna-Cruz, I., Rendon, A., 6 Vargas-Villarreal, J., Vera-Cabrera, L., Rodríguez-Padilla, C. (2017) Detection of anti- 7 8 HspX antibodies and HspX protein in patient sera for the identification of recent latent 9 infection by Mycobacterium tuberculosis. PLoS One. 12, e0181714. 10 11 doi:10.1371/journal.pone.0181714. 12 13 [42] Yousefi-Avarvand, A., Tafaghodi, M., Soleimanpour, S., Khademi, F. (2018) HspX 14 protein as a candidate vaccine against Mycobacterium tuberculosis: an overview. Front. 15 16 Biol. 13, 293–296. doi:10.1007/s11515-018-1494-2. 17 [43] Taylor, J.L., Wieczorek, A., Keyser, A.R., Grover, A., Flinkstrom, R., Karls, R.K., 18 19 Bielefeldt-Ohmann, H., Dobos, K.M., Izzo, A.A. (2012) HspX-mediated protection 20 21 against tuberculosis depends on its chaperoning of a mycobacterial molecule. Immunol. 22 Cell Biol. 90, 945-954. doi:10.1038/icb.2012.34. 23 24 [44] Kennaway, C.K., Benesch, J.L.P., Gohlke, U., Wang, L., Robinson, C. V., Orlova, E. V., 25 Saibi, H.R., Keep, N.H. (2005) Dodecameric structure of the small heat shock protein 26 27 Acr1 from Mycobacterium tuberculosis. J. Biol. Chem. 280, 33419–33425. 28 doi:10.1074/jbc.M504263200. 29 30 [45] Kaushik, A., Singh, U.B., Porwal, C., Venugopal, S.J., Mohan, A., Krishnan, A., Goyal, 31 32 V., Banavaliker, J.N. (2012) Diagnostic potential of 16 kDa (HspX, alpha-crystalline) 33 antigen for serodiagnosis of tuberculosis. Indian J. Med. Res. 135, 771–777. 34 35 [46] Zhang, C., Song, X., Zhao, Y., Zhang, H., Zhao, S., Mao, F., Bai, B., Wu, S., Shi, C. 36 (2015) Mycobacterium tuberculosis Secreted Proteins As Potential Biomarkers for the 37 38 Diagnosis of Active Tuberculosis and Latent Tuberculosis Infection. J. Clin. Lab. Anal. 39 40 29, 375–382. doi:10.1002/jcla.21782. 41 [47] Limongi, A., Olival, L., Conde, M.B., Junqueira-Kipnis, A.P. (2011) Determination of 42 43 levels of specific IgA to the HspX recombinant antigen of Mycobacterium tuberculosis 44 for the diagnosis of pleural tuberculosis. J Bras Pneumol. 37, 302–307. 45 46 [48] Rabahi, M.F., Junqueira-Kipnis, A.P., dos Reis, M.C.G., Oelemann, W., Conde, M.B. 47 (2007) Humoral response to HspX and GlcB to previous and recent infection by 48 49 Mycobacterium tuberculosis. BMC Infect. Dis. 7, 1–9. doi:10.1186/1471-2334-7-148. 50 51 [49] Fu, X., Zhang, H., Zhang, X., Cao, Y., Jiao, W., Liu, C., Song, Y., Abulimiti, A., Chang, 52 Z. (2005) A dual role for the N-terminal region of Mycobacterium tuberculosis Hsp16.3 53 54 in self-oligomerization and binding denaturing substrate proteins. J. Biol. Chem. 280, 55 6337–6348. doi:10.1074/jbc.M406319200. 56 57 [50] De Vos, K., Girones, J., Popelka, S., Schacht, E., Baets, R., Bienstman, P. (2009) SOI 58 optical microring resonator with poly(ethylene glycol) polymer brush for label-free 59 60 biosensor applications. Biosens. Bioelectron. 24, 2528–2533. ACS Paragon Plus Environment Page 21 of 22 ACS Infectious Diseases 1 2 3 doi:10.1016/j.bios.2009.01.009. 4 5 [51] Holz, O., Kips, J., Magnussen, H. (2000) Update on sputum methodology. Eur. Respir. 6 J. 16, 355–359. doi:10.1034/j.1399-3003.2000.16b26.x. 7 8 [52] Cubero, N., Esteban, J., Palenque, E., Rosell, A., Garcia, M.J. (2013) Evaluation of the 9 detection of Mycobacterium tuberculosis with metabolic activity in culture-negative 10 11 human clinical samples. Clin. Microbiol. Infect. 19, 273–278. doi:10.1111/j.1469- 12 13 0691.2012.03779.x. 14 [53] Bannantine, J.P., Stabel, J.R. (2000) HspX is present within Mycobacterium 15 16 paratuberculosis-infected macrophages and is recognized by sera from some infected 17 cattle. Vet. Microbiol. 76, 343–58. http://www.ncbi.nlm.nih.gov/pubmed/11000531. 18 19 [54] Soler, M., Estevez, M.C., Moreno, M.L., Cebolla, A., Lechuga, L.M. (2016) Label-free 20 21 SPR detection of gluten peptides in urine for non-invasive celiac disease follow-up. 22 Biosens. Bioelectron. 79, 158–164. doi:10.1016/j.bios.2015.11.097. 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment ACS Infectious Diseases Page 22 of 22 1 2 3 FOR TABLE OF CONTENTS ONLY 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment