Electrochemical Sensing Platforms for Tuberculosis-Associated VOCs from Patient Breath

Abstract

In this presentation two point-of-care electrochemical sensing platforms (gas and liquid phase) to detect volatile organic compounds (VOCs) present in the breath of tuberculosis (TB) patients will be discussed. Rapid screening using patient breath is significantly faster, less expensive and more convenient than traditional sputum-based tests, and up to one third of adults and the majority of young children with TB are unable to produce sufficient sputum for traditional culture testing, making the development of non-sputum-based tests a priority in TB screening and diagnostics. Methyl nicotinate (MN) and methyl p-anisate (MPA) have been identified as specific biomarkers for TB in patient breath and culture by GC-MS [1], and are not found in readily in ambient air or the breath of healthy patients, making them good candidates for TB screening. Previous modelling studies [2] suggest that these biomarkers complex with certain transition metals (such as cobalt and copper) given a particular bias voltage and oxidation state of the metal. This known interaction was used as the basis for the sensing platforms described.A solid-state titanium dioxide nanotube array-based sensor is used to directly detect TB biomarkers in the gas phase, utilizing a divalent cobalt-functionalized surface (Co-TNA) to specifically bind MN and MPA [3]. The original sensor design consisted of a simple two-electrode setup and chronoamperometry at binding voltages specific to the VOCs of interest was used to selectively sense these biomarkers. However, the lack of a separate reference electrode in this system can cause a high level of variability in the sensors, limiting sensitivity and specificity. In order to improve reproducibility and specificity, a solid-state, electron-conductive polymer “electrolyte” of known resistance is used to create a gas-phase three-electrode system, utilizing silver screen-printed references and counter electrodes. This new configuration ensures more consistent interactions with the analyte of interest between sensors in the amperometric mode, and allows the sensor to be operated using more complex scanning techniques traditionally used in liquid systems. Scanning potential methods such as cyclic voltammetry (CV) have shown promise in detecting these VOCs more specifically than chronoamperometry alone. For example, the CV of a Co-TNA sensor shows a reduction peak at about -0.5 V vs the silver quasi-reference electrode when exposed to methyl nicotinate vapor which is not present when exposed to simple humid air (figure 1). In preliminary field studies, CVs of the Co-TNA sensor in response to the breath of sick patients show artifacts not present in healthy controls. Results comparing suspected TB patients (Xpert positive and negative) as well as healthy controls will be presented.Because the VOCs associated with Tuberculosis are only semi-volatile and readily dissolve in water at significantly high concentrations (mM range), collection of these compounds from breath and condensate into an aqueous solution can provide a simple concentration method when very low levels of VOC are present. In addition, breath condensate can be frozen and saved for later analysis when immediate testing is not available. Therefore, we have developed a liquid-phase sensing platform utilizing engineered electroactive solutions (EASs) containing a functional metal ion which can be customized to the analyte of interest without the need for a more expensive, highly specific working electrode as used in the gas-phase platform. Rather than observing interactions of the biomarker at the specific electrode surface, a functional metal solution of known electrochemical activity is used to detect the VOC of interest by observing the change electrochemically when it is added to the EAS. In this application, copper is preferred over cobalt because of its highly reversible redox reactions and the ability to observe multiple redox peaks within a relatively small potential window. Testing of this platform with a divalent copper EAS has shown that the cyclic voltammogram of a copper redox pair changes significantly when as little as 1 mM methyl nicotinate (MN) is added to the solution. The variation of peak oxidation voltage with MN concentration can be seen in figure 2. In order to further understand the reactions taking place and specifically identify the biomarkers, square wave voltammetry (SWV) is employed to show which oxidation states of the functionalized metal are complexing with the biomarker of interest. The SWV for the copper EAS alone shows three distinct peaks for each of the redox reaction couples present in the voltage window (less than ±1 V). When a small amount of MN is added, the size, shape and in some cases peak-voltage of each distinct peak is affected differently, implying a difference in reaction to specific valence states of copper (figure 3). When methyl p-anisate (MPA) was added at the same concentration, different interactions were observed. In this way, a “fingerprint” method can be used to identify biomarkers once their known interaction is established.