Elsevier

Biosensors and Bioelectronics

Volume 74, 15 December 2015, Pages 601-611
Biosensors and Bioelectronics

Smartphone-based, sensitive µPAD detection of urinary tract infection and gonorrhea

https://doi.org/10.1016/j.bios.2015.07.014Get rights and content

Highlights

  • First demonstration of smartphone/μPAD assay for UTI/gonorrhea directly from undiluted human urine.

  • Filtration of urobilin by μPAD reduces false-positive green fluorescence signals.

  • Superior detection limit (10 CFU/mL) compared to lateral flow assay and nitrite strip (106 CFU/mL).

  • Wider dynamic range (up to 107 CFU/mL) through optimizing the particle amount on μPAD.

  • Reproducible assay results by improving the particles' stability through soft centrifugation.

Abstract

The presence of bacteria in urine can be used to monitor the onset or prognosis of urinary tract infection (UTI) and some sexually-transmitted diseases (STDs), such as gonorrhea. Typically, bacteria's presence in urine is confirmed by culturing samples overnight on agar plates, followed by a microscopic examination. Additionally, the presence of Escherichia coli in a urine sample can be indirectly confirmed through assaying for nitrite (generated by reducing nitrate in urine), however this is not sufficiently specific and sensitive. Species/strains identification of bacteria in a urine sample provides insight to appropriate antibiotic treatment options. In this work, a microfluidic paper analytical device (µPAD) was designed and fabricated for evaluating UTI (E. coli) and STD (Neisseria gonorrhoeae) from human urine samples. Anti-E. coli or anti-N. gonorrhoeae antibodies were conjugated to submicron particles then pre-loaded and dried in the center of each paper microfluidic channel. Human urine samples (undiluted) spiked with E. coli or N. gonorrhoeae were incubated for 5 min with 1% Tween 80. The bacteria-spiked urine samples were then introduced to the inlet of paper microfluidic channel, which flowed through the channel by capillary force. Data confirms that proteins were not filtered by μPAD, which is essential for this assay. Urobilin, the component responsible for the yellow appearance of urine and green fluorescence emission, was filtered by μPAD, resulting in significantly minimized false-positive signals. This filtration was simultaneously made during the μPAD assay and no pretreatment/purification step was necessary. Antibody-conjugated particles were immunoagglutinated at the center of the paper channel. The extent of immunoagglutination was quantified by angle-specific Mie scatter under ambient lighting conditions, utilizing a smartphone camera as a detector. The total μPAD assay time was less than 30 s. The detection limit was 10 CFU/mL for both E. coli and N. gonorrhoeae, while commercially available gonorrhea rapid kit showed a detection limit of 106 CFU/mL. A commercially available nitrite assay test strip also had a detection limit of 106 CFU/mL, but this method is not antibody-based and thus not sufficiently specific. By optimizing the particle concentration, we were also able to extend the linear range of the assay up to 107 CFU/mL. The proposed prototype will serve as a low-cost, point-of-care, sensitive urinalysis biosensor to monitor UTI and gonorrhea from human urine.

Introduction

Urinary tract infection (UTI) is a common problem among humans throughout the world. Although they are generally considered non-life-threatening, they may cause significant discomfort with potential for persistent reoccurrence, affecting the quality of a patient’s life. As high as 40–50% of UTIs are symptomatic and identified in primary care practices (Lin et al., 2011). Among those patients, 25% of them show retracted, reoccurring disease within 6 months (Lin et al., 2011).

Most UTIs are caused by Escherichia coli. In many cases, UTIs are associated with the E. coli found in the patient's own colon. Females show higher occurrence of UTI than males, due to the anatomical proximity of the urinary tract and anus in females. In some cases, UTI can spread from person to person, similar to sexually transmitted diseases (STDs). Some STD-causing bacteria can also reside in the urinary tract (similar to UTI), but are transmitted through sexual intercourse. Their symptoms are more severe and typically lead to higher medical costs (CDC, 2013). The most common STD that affects the urinary tract is gonorrhea, caused by Neisseria gonorrhoeae, which is the second most commonly reported STD in the United States (CDC, 2013). In 2013, there were 333,004 cases of gonorrhea reported by CDC (U.S. Centers for Disease Control and Prevention). Gonorrhea affects women as young as 15 (CDC, 2014).

Both bacterial infections, UTI and gonorrhea, are preventable by early detection, but this is arduous to do because most cases are often asymptomatic. CDC claims that a simplified screening method would provide a critical strategy for early identification and prevention (CDC, 2014). Currently available methods have such high detection limits that only cases showing symptoms are detected. At such stages, more treatments and costs are required due to further progression of the disease. Therefore, there is a critical need for early diagnostic methods with increased sensitivity and specificity for bacterial agents of UTIs and STDs. Such methods can greatly reduce the long-term consequences of diseases and reduce healthcare costs, which are of high demand by healthcare agencies around the world (CDC, 2013).

High concentrations of E. coli and N. gonorrhoeae induce the symptomatic UTI and gonorrhea, respectively. At high concentrations, both types of bacteria can cause serious health effects to the host. Most cases, however, go undetected with asymptomatic behavior, as mentioned previously. The gold standard of detecting E. coli and N. gonorrhoeae is by culturing the urine sample and quantifying the bacterial concentration, which takes half a day for E. coli and 48 h for N. gonorrhoeae. Lateral flow assays (LFAs; more precisely, lateral flow immunochromatographic assays), popularly known as rapid kits, are commercially available for detecting N. gonorrhoeae. However, they have high detection limits (typically 106 colony forming units or CFU per mL sample) and low sensitivities, making them unable to detect low concentrations of bacteria. This situation eventually causes complications with the development of an STD. There is currently no commercially available test for directly detecting E. coli (or other UTI-causing bacteria) from human urine. Colorimetric strips (not LFA) are available, detecting the presence of nitrites in urine. E. coli produces nitrate reductase that reduces nitrate in urine into nitrite (Hu et al., 2014, Lazzarini and Atkinson, 1961). These colorimetric strips are indirect, thus lacking specificity and sensitivity. In fact, these nitrite strips are unable to detect the low concentrations of bacteria necessary for the early detection of UTI or STD. Both direct gonorrhea LFAs and indirect nitrite strips typically require concentrated samples or overnight incubation to complement their insufficient detection limits, often combined with invasive sampling by inserting a swab into a patient's urinary tract. A newer platform is necessary, detecting multiple types of UTI- and STD-causing bacteria with high specificity and sensitivity, as well as very low limit of detection that enables early detection of such infections.

Lately, microfluidic paper analytical devices (µPADs) have gained much attention as an advantageous lab-on-a-chip (LOC) system. These LOC systems have previously been reported to include less sample and reagent consumption, lower power consumption, a lower per-unit cost, a reduced risk of contamination, higher reliability and functionality, automation and portability (Martinez et al., 2008, Park et al., 2013). µPADs are disposable, make disease diagnosis cost-effective, and can be operated without the assistance of trained medical personnel. These features make µPADs attractive for diagnostic applications in remote locations and the developing world (Sechi et al., 2013). Hence, numerous publications have recently appeared regarding µPADs, but these systems still face the challenging demand for better sensitivity, especially when relying on colorimetric detection for diagnosis (Yetisen et al., 2013).

While urinalysis can be performed on µPADs, detection of bacterial culture directly from human urine has been considered a laborious task. Urine is a complicated bio-fluid that usually requires sample pretreatment, such as purification and/or enrichment steps prior to analytical steps to determine specific components in the urine (Lin et al., 2011). Not to mention, the exposure of urine to benchtop-based protocols and devices increases risk of contamination and requires large sample volume. Literature reports that there is a necessity for automatic, miniaturized, inexpensive and easy-to-use microdevices for urinary sample pretreatment (Lin et al., 2011).

Previously, our group quantified the concentration of bacteria by evaluating the angle-specific Mie scatter signal from immunoagglutinated polystyrene particles (Fronczek et al., 2013, Park et al., 2013, You et al., 2011). By using paper microfluidic chips as our µPADs along with smartphone-based optical detection, E. coli was detected from waste water samples through quantifying the extent of immunoagglutination (Park and Yoon, 2015). However, our previous demonstration failed to detect E. coli concentrations higher than 105 CFU/mL (opposite problem to the conventional LFAs), due to the complex nature of E. coli colonies, i.e., they did not react with antibody-conjugated polystyrene particles sufficiently and reproducibly.

In Table 1, a summary of previously reported LOC-based assays for bacteria detection is provided. As reported, there are few LFA or µPAD methods that can efficiently detect bacteria. There are minimal amounts of methods (lines 1–4; Liao et al., 2006; Bercovici et al., 2011; Safavieh et al., 2012; Yang et al., 2011) that detect bacteria directly from a urine sample matrix using microfluidic platforms (not LFA or µPAD). Unfortunately, either pretreatment (filtration, centrifugation, and/or rinsing) was necessary or the detection limit was quite high (105–108 CFU/mL). There are a few methods that do report low detection limits within 100 CFU/mL (lines 3, 6, 12 and 13; Safavieh et al., 2012; Wang et al., 2012; Fang et al., 2012; Safavieh et al., 2014). Most of these “low detection limit” methods utilized loop-mediated isothermal amplification (LAMP), which amplified DNA targets to increase the concentration for their assays (lines 3, 12 and 13; Safavieh et al., 2012; Fang et al., 2012; Safavieh et al., 2014), as well as the use of relatively clean sample matrix (e.g. buffer; lines 12 and 13; Fang et al., 2012; Safavieh et al., 2014). In contrast to our near-real-time assay, none of the referenced methods are close to real-time and may take hours for results.

In this paper, we discuss our further development of µPADs for the detection of E. coli and N. gonorrhoeae with fast assay time, low detection limit, and broader and clinically relevant range of detection, all directly from the human urine sample. We aim to accomplish these goals without any purification/enrichment steps, towards near-real-time assays (<30 s). Additionally, we investigate how the paper fibers are beneficial in detecting bacteria from urine, specifically their filtration capability. E. coli concentration of 105 CFU/mL correlates with the onset of symptomatic UTI. As indicated previously, there is no commercially available LFA that specifically detects E. coli. Our interest is the low-level detection of 10–103 CFU/mL that may predict high risk of UTI development and allows early reaction to prevent UTIs. Specific and sensitive detection of other UTI-causing bacteria (most notably Staphylococcus saprophyticus) (Nicolle, 2008) should also be possible through switching the antibodies conjugated to the polystyrene particles. N. gonorrhoeae concentration of 105 CFU/mL correlates with the onset of symptomatic gonorrhea. Our interest is the detection of lower concentrations (10–103 CFU/mL) that may predict high risk of gonorrhea development, allowing for early reaction to prevent gonorrhea. For both cases, detection of higher bacterial concentrations (up to 107 CFU/mL; previously not demonstrated with this method) is demonstrated in this work, which correlates with substantially symptomatic UTI and gonorrhea. In addition, specificity experiments are also performed (again, previously not demonstrated with this method).

Overall, we aim to develop a point-of-care and near-real-time diagnostic tool with superior detection limit and sensitivity that can be used for wider ranges of bacterial concentrations and species, combined with the use of a smartphone. Previous studies show that the fast growing mobile phone market in the developing world has made camera phones a potential platform for the delivery of rapid diagnostics (Yetisen et al., 2013). Towards these goals, we attempt to evaluate and optimize the antibody conjugation procedure and particle concentration to make our assay fully adaptable to various disease conditions.

Section snippets

Urine samples with E. coli and N. gonorrhoeae

E. coli solutions were prepared from lypophilized E. coli K12 cell powders (Sigma-Aldrich; St. Louis, MO, USA) by culturing them in brain heart infusion broth (Remel; Lenexa, KS, USA) for 7.5 h at 37 °C. The culture reached its highest bacterial concentration of 108 CFU/mL. Serial dilutions were made using normal urine from pooled human donors (Lee Biosolutions; St. Louis, MO, USA). This serial urine dilution allowed us to prepare virtually undiluted, E. coli-spiked urine samples. E. coli culture

Optimization of particle loading to paper chips

Initially, the procedures and protocols described in our previous publications (Park et al., 2013, Park and Yoon, 2015) were used without any modifications in assaying E. coli and N. gonorrhoeae from undiluted human urine. The normalized scattering intensity values were quite small and not consistent from experiment to experiment. According to the microscopic images of the antibody-conjugated particles before and after the assays, we found non-specific aggregations before the assay, presumably

Discussion

Various available rapid tests for UTI and gonorrhea are nonspecific with high detection limits. The current reliant method for physicians to diagnose UTI and gonorrhea in patients is by extracting the tissue/bacteria sample from the urethra in men or from the cervix in women. The sample is sent to a lab where it is cultured for at least half a day for E. coli and 48 h for N. gonorrhoeae. Using this method, the patient receives their results days after visiting the physician. The available test

Acknowledgment

Funding for this research was provided by Seoul VioSys. Soohee Cho acknowledges the fellowship support from Richard A. Harvill graduate fellowship. The authors would also like to thank Dr. Daewoong “Dave” Suh and Dr. Kyujin Choi at Seoul VioSys for helpful discussions and guidance.

References (40)

  • W.-H. Chang et al.

    Biosens. Bioelectron.

    (2015)
  • C.E. Cornelius

    Liver function

  • C.F. Fronczek et al.

    Biosens. Bioelectron.

    (2013)
  • J.-H. Han et al.

    Biosens. Bioelectron.

    (2008)
  • J. Hu et al.

    Biosens. Bioelectron.

    (2014)
  • J. Jiang et al.

    Sens. Actuat. B

    (2014)
  • J.F. Kerkhoff et al.

    Clin. Chim. Acta

    (1968)
  • R.A. Lazzarini et al.

    J. Biol. Chem.

    (1961)
  • L.E. Nicolle

    Urol. Clin. N. Am.

    (2008)
  • M. Safavieh et al.

    Biosens. Bioelectron.

    (2012)
  • J.H. Yoo et al.

    Sens. Actuators B

    (2014)
  • D.J. You et al.

    Biosens. Bioelectron.

    (2011)
  • M. Bercovici et al.

    Anal. Chem.

    (2011)
  • Bio-Rad Laboratories, 1995. DC™ Protein Assay Instruction Manual....
  • J.N. Bixler et al.

    Proc. Natl. Acad. Sci. U.S.A.

    (2014)
  • CDC (U.S. Centers for Disease Control and Prevention), 2013. Incidence, Prevalence, and Cost of Sexually Transmitted...
  • CDC (U.S. Centers for Disease Control and Prevention), 2014. Sexually Transmitted Disease Surveillance 2013....
  • Craine, B.L., 2002. U.S. Patent Application Publication. U.S....
  • R. Elman et al.

    J. Exp. Med.

    (1925)
  • X. Fang et al.

    Lab Chip

    (2012)
  • Cited by (88)

    • Smartphone-based diagnostics for biosensing infectious human pathogens

      2023, Progress in Biophysics and Molecular Biology
    • Moving toward smart biomedical sensing

      2023, Biosensors and Bioelectronics
      Citation Excerpt :

      Among the different microfluidic-based POC diagnostic devices, paper-based biosensors coupled with smartphones have gained much attention owing to their portability, high throughput, reduced risk of contamination, rapidness, low per-unit cost, and flexibility (especially satisfactory for wearable sensors). There are many reports on the development of smartphone-based paper POC diagnostics for colorimetric sensing applications of a variety of targets including biomarkers of liver function (alkaline phosphatase (ALP) and aspartate aminotransferase (AST)) (Vella et al., 2012), nucleic acids (Cho et al., 2015), glucose (Choi et al., 2015; Jia et al., 2018; Mercan et al., 2021; Ngo et al., 2021; Soni and Jha, 2017; X. Wang et al., 2018; Liting Zhang et al., 2015), protein (Choi et al., 2015), biothiols (Kappi et al., 2018), hydrogen peroxide (H2O2) (Ngo et al., 2021), uric acid (UA) (Li et al., 2020; X. Wang et al., 2018), lactate (Liting Zhang et al., 2015), urea (Choi et al., 2021; Soni et al., 2018), cancer biomarkers (Aydindogan et al., 2020), and protease biomarkers (Liu et al., 2021) in divers biological fluids. The development of the ideal diagnostic devices/assays based on smartphone is also important particularly when the discourse around medical problems or seeking medical assistance may be deemed embarrassing due to embarrassment (social and cultural stigma) or fear (Culley et al., 2013; Tsao et al., 2021).

    View all citing articles on Scopus
    1

    These authors contributed equally to this work.

    View full text