(J-361) Low-Cost Electronics for Portable, Self-Contained, Triage Testing of Tuberculosis at the Point of Care

Annually, tuberculosis (TB) is responsible for 1.5 million deaths and is the leading infectious killer globally. Of the 10.6 million new cases of TB each year, 80% occur in low and middle-income countries (LMICs), and nearly 3 million go undiagnosed. The most common way to diagnose TB in LMICs is a century-old sputum smear microscopy (SSM) test that suffers from expensive equipment costs and poor sensitivity. Millions of cases are never traced as access to central laboratories that perform the SSM test is unreliable. As such, the WHO has outlined the need for a portable, accessible, and rapid alternative to the SSM test that can be powered without running electricity. Recent innovations in low-cost, current-to-voltage amplification circuits have demonstrated the feasibility of detecting faint fluorescence signals from samples using cheaply made, mass produced electronic components. Furthermore, a novel line of trehalose-conjugated solvatochromic dyes1 possess many favorable characteristics for a portable test such as eliminating several wash steps in the staining process, having high quantum yield, and being stable at room temperature for several months. This report describes the construction of a handheld device that amplifies the emission of said trehalose dyes to perform inference testing. The device was constructed for the low-cost of $65. Initial experiments with this device showed that it performed comparably to commercial plate readers in characterizing faint levels of fluorescence in surrogate trials that model the TB bacteria, making it highly promising for the detection of TB in resource-constrained settings.

The handheld device is powered by a circuit that uses a simple transimpedance amplification schematic to convert incident current from a photodiode into measurable voltage count values2. An orthogonal 405nm LED is coupled with a plano-convex lens to serve as an excitation source. An Arduino Mega2560 microcontroller is used to write the data into a spreadsheet. A 3D-printed box structure was built to encase the circuitry and hold the test tube in place, shielding the board and providing a dark environment for the sample. A commercial BioTek Synergy HTX plate reader was used to set the standard curves. Sodium fluorescein samples ranging in concentrations from 4-125µM  were measured in 200µL water background. Yellow-green carboxylate microspheres were prepared in 200µL of the same background in counts ranging from 400-20,000. Both the fluorescein and the microsphere readings were done in triplicates inside of polypropylene microcentrifuge tubes. The trehalose dye was modeled using the sodium fluorescein and microspheres because of their similar Stokes’ Shifts curves. Additionally, preliminary calculations using Stokes’ Drag Equation showed that the microspheres possessed a settling velocity in the same order of magnitude as the bacteria itself, meaning sample preparation of the TB bacteria can be roughly modeled by the microspheres. Microsphere pelleting was done via a 3D-printed centrifuge that achieves top speeds of ~7,500 RPM and 2,500 RCF3. The centrifuge weighed 10 grams and was made with parts costing under a dollar, making it highly compatible with LMIC settings.

Results: (238 words)
Performance of the device was measured by its R² coefficient across the linear working ranges of both the microspheres and sodium fluorescein and compared with results produced by a commercial plate reader. The regression generated by the device for sodium fluorescein showed excellent linearity at lower concentrations (4-125µM) with an R² value of 0.967. This value was comparable to the plate reader which had an R² value of 0.995. Both the device and the plate reader were capable of characterizing self-quenching effects at higher concentrations with the regression R² value of the device dropping to 0.847 when a 250µM datapoint was added. Next, fluorescent microspheres were diluted and added to 200µL samples at varying bead counts before being centrifuged down by hand. The device demonstrated similar results to the fluorescein trial, achieving an R² value of 0.931 with bead counts ranging from 400-20,000 and 0.989 when the count was reduced to 400-4400. 

Conclusions:

Performance of the device was cross-validated with the commercial plate reader to reliably plot fluorescence count values against the amount of fluorophore. For both the sodium fluorescein and microsphere trials, linearity was successfully characterized in low concentrations of both fluorophores and self-quenching was shown across larger concentration ranges. Future experiments will be done on H37Ra, an attenuated strain of TB, with the end goal of creating a complete procedure for rapidly testing TB in resource-constrained settings from raw sputum samples.

This work was generously supported by a grant from the Lemelson Foundation through VentureWell’s E-Team program. Lab equipment and supplies were supplied by both the Bowden lab and Haselton lab at Vanderbilt University. Seth Crawford, Prof. Locke, and Prof. Bowden assisted in device ideation, experimental setup, and grant logistics. Dalton Nelson and Megan Pask aided in day to day assistance with ordering and lab equipment and supplies.

[1] Kamariza, M., Keyser, S. G. L., Utz, A., Knapp, B. D., Ealand, C., Ahn, G., Cambier, C. J., Chen, T., Kana, B., Huang, K. C., & Bertozzi, C. R. (2021). Toward Point-of-Care Detection of Mycobacterium tuberculosis: A Brighter Solvatochromic Probe Detects Mycobacteria within Minutes. JACS Au, 1(9), 1368–1379. https://doi.org/10.1021/jacsau.1c00173

[2] Leeuw, T., Boss, E. S., & Wright, D. L. (2013). In situ measurements of phytoplankton fluorescence using low cost electronics. Sensors (Basel, Switzerland), 13(6), 7872–7883. https://doi.org/10.3390/s130607872

[3] Bhamla, M., Benson, B., Chai, C. et al. Hand-powered ultralow-cost paper centrifuge. Nat Biomed Eng 1, 0009 (2017). https://doi.org/10.1038/s41551-016-0009