Surveillance of emerging infectious diseases is essential for implementing preventive and control strategies during outbreaks; this surveillance is often performed using clinical testing of individuals. 1 However, large-scale clinical testing to track infectious diseases within populations is resource-intensive and expensive. [2][3][4][5] Wastewater-based epidemiology (WBE) tracks the prevalence of pathogens in a community via regular wastewater testing. 6 WBE is a powerful public health tool that can act as an early warning system and inform health officials about the spread of infectious diseases in a community [7][8][9] and allocation of resources in an outbreak. 10 During the COVID-19 pandemic, WBE has provided valuable information on SARS-CoV-2 and infection rates in rural and urban settings such as university dormitories, nursing homes, wastewater treatment plants, and others. [11][12][13] Beyond SARS-CoV-2, WBE can measure the emergence/presence/abundance of different viruses (e.g., influenza, mpox), bacteria (e.g., E. coli, V. cholerae, antibiotic-resistant strains), fungi (e.g., C. auris), and other biological markers indicative of infectious diseases. [14][15][16][17][18] Unfortunately, WBE implementation has been inequitable, with disproportionate utilization at large cities' municipal wastewater treatment plants and limited application in low-resource and rural locations. WBE can be particularly useful in low-resource settings where access to health care is limited, and individual clinical testing faces logistical challenges. 19,20 WBE minimizes the cost of disease monitoring at a population level compared to traditional clinical testing. However, widespread adoption of WBE in low-resource settings is hindered by the lack of infrastructure, limited technical capacity, and the need for trained personnel. Those limitations are seen at almost every step of the WBE workflow (Figure 1). Wastewater samples are collected using sampling methods from either a sewage manhole, pit latrine, lagoon, or wastewater treatment plant and are stored and transported to a central laboratory in sterile containers at refrigerated temperatures. 21 However, the maintenance of the cold chain storage, especially for long travel distances in remote areas, is often compromised, facilitating degradation of the nucleic acids in the wastewater sample. 22 Stabilization techniques have been shown to mitigate this degradation, and our group has previously demonstrated that nucleic acid extraction provides significant stabilization, even at room temperature. 23 Moreover, wastewater samples are usually highly diluted due to rainfall, agricultural runoff, and industrial and nonexcrement wastewater. This results in low and even undetectable concentrations of nucleic acid in wastewater samples. Therefore, a concentration step is also recommended to achieve reliable signals. 24 Common concentration methods such as polyethylene glycol precipitation, ultrafiltration, and ultracentrifugation are widely used in (and limited to) centralized laboratories. Those methods require long processing times (>90 min), and expensive equipment such as ultracentrifugation systems and vacuum pumps and are followed by a separate extraction step that adds to the time and cost per sample. 25 Solid-phase extraction and magnetic bead-based extraction are commonly used to isolate nucleic acid from wastewater samples. However, the cost and complexity of those methods and the need for trained staff remain a huge barrier to the adoption in lowresource settings. 26 The final step is to identify and quantify the isolated nucleic acids using an amplification method or sequencing techniques. The commonly used polymerase chain reaction (PCR) method needs a thermal cycler and typically takes more than 1 h to complete. Therefore, there is a need for an easy-to-use and inexpensive concentration, extraction, and identification workflow that is suitable for WBE applications in low-resource settings. 27,28 A point-of-use wastewater testing device can revolutionize WBE, in the same way point-of-care (POC) testing has revolutionized clinical testing. 29,30 Integrated POC testing devices typically test clinical samples with relatively high biomarker concentrations, enabling them to test relatively small sample volumes. [31][32][33][34] However, targets are often very diluted in wastewater samples, which requires concentration and extraction from a higher volume of sample. Also, many POC devices utilize microfluidic components to manipulate small volumes of samples and reagents, 35 which often also limits sample input volume.
In this work, we have utilized immiscible filtration assisted by surface tension (IFAST) to concentrate and extract nucleic acids from a high volume of wastewater to an elution buffer. IFAST uses the immiscibility of oil in water as a barrier between the sample and other buffers, allowing the nucleic acid boundmagnetic beads to move easily between buffers using a simple magnet. 36 Based on the IFAST technique, an integrated, easy-touse, and inexpensive concentration-extraction-identification device (CEID) has been developed to detect low-prevalence pathogens such as SARS-CoV-2 in environmental samples (Figure 2). In the CEID, nucleic acids are captured by magnetic beads and are moved through a series of oil and wash buffer wells for concentration and extraction purposes. Coupled with loopmediated isothermal amplification (LAMP) identification, CEID enables fast semiquantitative identification of the target pathogen. IFAST integrated with LAMP has been previously investigated as a simple method for nucleic acid extraction and detection; 37,38 however, its application in wastewater surveillance remains unexplored. This technology is most impactful in rural and low-resource settings, where fast and low-cost environmental surveillance can lead to the democratization of public health tools and offer more equitable health care. Also, with several (re)emerging pathogens on the horizon (e.g., H5N1 influenza, cholera), such CEIDs can enable early detection of an emerging pathogen, facilitating a timely public health response. Overall, CEID aims to enhance WBE capacity in rural and lowresource setting through simplified and integrated concentration, extraction, and identification of pathogen.
Chip Fabrication and Preparation. The chip was fabricated out of 1/8 in (3.175 mm) thick Poly(methyl methacrylate) (PMMA) sheets (McMaster-Carr, 8589K41) using a CO 2 laser cutter (Universal Laser Systems, VersaLaser VLS3.50). The adhesive films (ThermoFisher Scientific, AB-1170) for the enclosing of the chambers were also cut using the laser cutter. The chips and adhesive films were designed in CorelDraw (version 15.2), where cutting and engraving lines were specified. Refer to Figure S1 for a dimensioned drawing of the chip and enclosing films. Following the fabrication, the CEIDs were sprayed with 70% ethanol, wiped with Kimwipes and left to dry. Then, the chips and pressure sensitive adhesive films were aligned and bonded together using a film sealing paddle (Figure 2A). The bonded CEIDs were kept inside a clean container for future use. The whole process for fabrication and preparation of each CEID took about 7 min.
The overall size of the CEID is 50 × 127 mm, where the wash well is 5 × 10 mm (equal to 160 μL), and the oil and elution wells are 5 × 5 mm (equal to 2 drops or 80 μL). The sample well uses a 30 × 45 mm rectangle and a 30 × 20 mm triangle (total of 5.12 mL). Wells are connected by channels that converge from a width of 5 mm to 1.5 mm, with a height of 0.44 mm. While the wells were cut out of the PMMA sheet, the connection channels were fabricated using the engraving option of the laser cutter.
On-Chip Nucleic Acid Concentration and Extraction. In the first step, the chip was loaded with the appropriate reagents. Briefly, 2 drops of silicone oil with a viscosity of 1000 cSt (1:2 mixing ratio of 50 cSt and 10000 cSt silicone oils (Sigma-Aldrich, 378356 and 378402)) were added to each of the oil chambers, and 160 and 80 μL of nuclease-free water were added to the wash and elution wells, respectively. Next, 2 mL of wastewater sample was transferred to a tube and mixed with 3 mL of lysis buffer containing 4 M Guanidine Thiocyanate (GTC) (ThermoFisher, AM9422), and 10 mM 4-morpholinethanesulfonic acid (MES) sodium salt (Sigma-Aldrich, M3671) dissolved in 1:1 v/v absolute ethanol/DI water and 0.1% v/v Tween 20 (Sigma-Aldrich, P9416). Then, 60 μL of each of two different sizes of paramagnetic particles (PMPs) (Cytiva, Serasil-Mag #29357369 and #29357374), were added to the sample.
The tube was then inverted a few times to ensure the mixing of the reagents and PMPs. The mixture of sample, lysis buffer, and PMPs (total of 5.12 mL) was then loaded to the sample well. An external magnet (K&J Magnetics, D4X0DIA-N52) collected all the PMPs inside the sample well. The PMPs were then pulled through the series of wells by a simple linear manual movement of the magnet. While PMPs can move through oil, the interface between oil and aqueous buffers acted as a barrier to prevent flowing of the buffers, resulting in IFAST (Figure 2B). Once in the elution well, the PMPs were mixed by pipetting up and down. The CEID was then placed on a hot plate at 65 °C for 5 min to ensure elution of nucleic acids. In the final step, the PMPs were separated using a magnet and the extracted nucleic acid was collected.
ESP Nucleic Acid Extraction. Exclusion-based Sample Preparation (ESP) has been previously used to extract nucleic acids from environmental samples, as well as clinical samples. [39][40][41][42] For more information about the ESP protocol, please refer to Supporting Information or previously published work. 43 RT-qPCR Assay. For the purposes of quality control and comparison, SARS-CoV-2 and CrAssphage nucleic acids were measured in all extracted samples from the CEID and ESP using RT-qPCR. CrAssphage is a bacteriophage that is abundantly found in wastewater samples and is used as an intrinsic positive control for this study. 44 The N1 gene was used for the SARS-CoV-2 assay with a primer and probe listed in Table S1, as recommended by the CDC. 45 The primer and probe sequences for CrAssphage are also listed in Table S1. For more details on RT-qPCR, please refer to Supporting Information.
RT-LAMP Assay. We used colorimetric RT-LAMP as a less resource-intensive end point, as it only requires a constant temperature and visual detection. 46 The SARS-CoV-2 RT-LAMP assay targets the nucleocapsid gene, with six primers, including forward, backward, forward inner, backward inner, loop forward, and loop backward primers. The sequences for all the LAMP primers are listed in Table S2. The primers are designed and evaluated by the New England Biolabs LAMP Primer Design Tool (lamp.neb.com) and ordered through custom oligos from ThermoFisher. Primers were received lyophilized and rehydrated to 100 mM using nuclease-free water. For a final 1× concentration, the concentration of each primer included: 200 nM for forward and backward, 400 nM for loop forward and loop backward, and 1600 nM for forward and backward inner primers. All assays were performed in a total of 25 μL reaction volume, consisting of 10 μL of sample, 12.5 μL of WarmStart Colorimetric LAMP 2× Master Mix with UDG (NEB, M1804L), and 2.5 μL of 10× primer mix. SARS-CoV-2 genomic RNA (NR-52508, Isolate USA-CA4/2020, BEI Resources) was added to one of the reaction wells as the positive control. For the no template control (NTC), nucleasefree water was added to the reaction well. In the case of a positive sample, the reaction color changed from pink to yellow, while a negative sample stayed pink throughout the reaction. For more details on on-chip and off-chip RT-LAMP, please refer to Supporting Information.
Real-Time Analysis of On-Chip RT-LAMP. A smartphone was used to take time-lapse images of the on-chip RT-LAMP reaction in 1 min intervals. For consistency of the light in images, a cardboard box was placed on top of the hot plate and an LED was used to light up the box. The resulting images were then analyzed to measure crossing time (Ct) using ImageJ. Briefly, 30 images are stacked together, and after splitting the RGB channels, the mean gray value was measured in the regions of interest in the green channel. A crossing time (Ct) was determined when the green channel increased more than 10% of the rolling average calculated from the previous 10 time points.
Contamination Carry-Over and PMPs Recovery Measurement. The contamination carry-over from one well to the other was measured using the fluorescent dye acridine orange (ThermoFisher Scientific, A1301) as the model contaminant, due to its stable fluorescence across the various buffers utilized in this study. In brief, acridine orange was dissolved in the combination of lysis, sample, and PMPs at a concentration of 10 mg/mL. PMPs were then extracted and transferred to the next well. PMPs were then mixed with a pipet, captured on a magnet and the elution buffer was collected. The elution buffer fluorescent intensity was measured using the plate reader with excitation at 490 nm and emission at 520 nm. A standard curve was obtained by measuring the fluorescent intensity of serial dilutions with known concentrations of acridine orange. The amount of carry-over was determined by fitting the fluorescent intensity of the elution buffer to the standard curve. Moreover, the effect of carry-over contamination is studied on both RT-qPCR and RT-LAMP, to ensure that CEID has minimal carryover that does not inhibit the nucleic acid amplification step. PMPs' recovery was also measured using their autofluorescence at 635 nm. A standard curve was formed to measure the concentration of PMPs in solution.
Wastewater Sample Collection. Effluent wastewater samples from a local hospital were collected from a manhole connected to the building's sewage system. A passive sampler, also known as a Moore swab, was used, in which five pieces of gauze pads (ThermoFisher Scientific, Catalog No. 13-761-52) were tied together and suspended in the wastewater flow using a string. The Moore swab was collected after 24 h and transferred to the lab inside a sealed plastic bag. The swab was then squeezed to extract the entrapped liquid and particulate matter and transferred into a 50 mL conical tube for further analysis. Moore swab is a simple, yet effective sample collection method and has been shown to perform equivalently to composite sampling and to outperform grab sampling. 47 Additionally, influent composite samples were collected at a local wastewater treatment plant (WWTP) using an autosampler over 24 h. While the Moore swab represents a simple and instrument-free approach to sample collection, the autosampler is recognized as the gold standard in wastewater surveillance. In this study, we employed both techniques to demonstrate the compatibility of our CEID with these diverse sampling methods.
We designed the assay to work with open-source, nonproprietary reagents and fabricated the CEID from readily available, affordable materials. The reagents used in this chip build upon our previous published works, in which we used GTC-based lysis buffer and silica coated PMPs to capture nucleic acids. Silica coated PMPs from different vendors were tested previously, and SeraSil-Mag PMPs from Cytiva were chosen due to their high magnetic responsiveness and binding capacity. 43 Optimal performance was defined as the point where the chosen PMPs, exhibited both ease of movement through channels and buffers, as well as maximized nucleic acid capture. To optimize the amount of PMPs for optimal performance, wastewater was extracted on chip with five different PMPs volumes (5, 10, 20, 60, 120 μL). The purified nucleic acids were measured by qPCR for RNA (SARS-CoV-2) and DNA (CrAssphage) targets, as well as Qubit for RNA and DNA. None of the PMP volumes occluded the channels (height of 440 μm) during the magnetic transfer in the channels and there was no sign of residual PMPs in each well. As shown in Figure 3A-D, increasing the PMPs volume increased the yield of both DNA and RNA. However, both qPCR and Qubit measurements showed there was not a significant difference between 60 and 120 μL of PMPs. Compared with 40 μL of PMPs, the extracted nucleic acids were significantly higher when 60 μL of PMPs were used. As a result, 60 μL of PMPs was chosen for further experiments.
The IFAST technology relies on the dominance of surface tension forces over other forces (e.g., gravity force) to establish and retain a barrier between different buffers. The Bond number (Bo = ΔρgL 2 /γ) can be used to explain the relative effect of gravitational forces to surface tension forces, where ρ is the density of the liquid, g is the acceleration of gravity, L is the characteristic length scale, and γ is the surface tension. For a system where surface tension forces are strong enough to minimize the influence of gravity, the Bond number need to be much smaller than 1. As Bond number is proportional to L 2 , decreasing the dimensions quickly shifts Bo into a regime dominated by surface tension. As a result, in CEID, the height of the connection channels between wells are minimized to stabilize the interfaces, while allowing the PMPs to pass. As connection channels are fabricated through laser engraving, the number of engraving passes determines the channel heights. We tested different repetitions of engravings ranging from 1 to 5 passes and measured the resulting channel height (Table S3). The experiments have shown that the resulting channel heights are small enough to have a stable interface between the two immiscible phases. Moreover, the PMPs can be moved throughout this broad range of channel heights. However, for channel heights around or smaller than 300 μm, some of the PMPs may be left behind and the magnet require manual correction to successfully transfer all the PMPs across each oil barrier. As a result, to prevent PMPs loss and occlusion, we selected an optimized channel height of 440 μm, equivalent to 4 engraving passes.
On the other hand, the purification of the IFAST relies on a stable immiscible phase barrier that permits the passage of PMPbound nucleic acids in a magnetic field while minimizing lysate carryover. As a result, the interfacial energy between the lysate and oil plays a key role in determining a stable barrier between the two wells after the passage of the PMPs. The ability of the lysate to pinch and recede after passage of the PMPs is related to the energy required to separate the lysate from the surface. According to the Dupréequation, 48 the work of adhesion per unit area w ( ) adh is measured by the decrease in the surface/lysate interface and increase in the surface/oil and lysate/oil interface:
where γ LO is the interfacial energy between the lysate and oil, γ SO is the oil surface energy, and γ SL is the lysate surface energy. If this work is smaller than the lysate/oil interfacial energy, the high interfacial energy will pinch off the lysate. On the other hand, if this work is larger than the lysate/oil interfacial energy, the PMPs will drag the lysate into the oil phase and eventually connect the lysate well to the wash well. The balance of these energies can be further explained by the Young's Equation for a droplet of lysate on a surface covered with oil:
where θ is contact angle of the lysate on a surface covered with oil. Combining eqs 1and 2 results in w cos 1
For a stable regime, where the lysate pinches off and PMPs move to the oil well with minimal carry-over, w adh must be less than γ LO , requiring the cos θ to be negative. This concept enables the prediction of the stable interface with a simple contact angle measurement where θ > 90°. In this work, we have measured the contact angle of the lysate on the pressure sensitive film covered with silicone oil. As shown in Figure 3E, the contact angle is about 135°, which ensures the minimal carry-over of the lysate into the following wells. Top and side images of the movement of the PMPs through the interface are shown in Figure 3F and Video S1. As expected, the lysate pinches off and recedes back to the lysate well after PMPs move to the oil well.
Moreover, we have also tested a variety of silicone oils with viscosities ranging from 50 to 10000 cSt. Since all these silicone oils results in a contact angle of more than 90°, the lysate/oil interface is stable (Figures S3 and S4). However, when filling the wells, lower viscosity oils fill up the channels faster and quickly move to the neighboring wells, resulting in a depletion of oil in the oil well. As a result, there is a need to refill the oil well after addition of sample and reagents, which is not ideal. On the other hand, higher viscosity oils result in a very slow movement of the PMPs and might require manual correction to move the PMPs (Table S4). This result can also be explained through the Stokes law for a moving particle in a fluid; u =F mag /6πrη, where η is the viscosity of the fluid. Based on these results, we chose silicone oil with a viscosity of 1000 cSt as our immiscible phase.
A key determinant of CEID performance is the loss of PMPs during transfer. To quantify this, three independent operators executed the CEID protocol, and the residual PMPs volume was measured. On average, 1.3 μL of beads were lost during transfer, corresponding to a PMP recovery of 97.8% (Figure S5).
Since wastewater is a complex matrix containing numerous inhibitory molecules, it is crucial to evaluate CEID's effectiveness in depleting these inhibitors. After the optimization of chip geometry and oil properties, we have measured the amount of carry-over from one well to the other using a fluorescent molecule (acridine orange) as the model contaminant (refer to Figure S6 for acridine orange standard curve). Our results show that on average, 0.86% of the acridine orange was carried over from one well to the next well during the PMP extraction (Figure 3G). By implementing a wash step, only 0.0074% of the lysate contamination (wastewater and lysis buffer) was present in the elution well. To understand whether this amount of contamination inhibited the amplification step, we measured the effect of different concentrations of sample contamination on RT-qPCR and RT-LAMP using three different wastewater samples. These wastewater samples come from various sources with variable turbidity and levels of inhibitory molecules. As shown in Figure S7, RT-qPCR is completely inhibited for lysate contamination concentrations of higher than 1% in the elution buffer. However, RT-LAMP showed some positive results for contamination concentrations of around 5%, although the Ct dramatically increased, leading to a reduction in sensitivity. As mentioned by others, 49,50 RT-LAMP is more resistant to inhibition compared to RT-qPCR. Our results showed that the CEID contamination carry-over did not inhibit RT-qPCR and RT-LAMP. While we have tested 3 different wastewater samples with varying levels of turbidity, it is important to note that each wastewater sample is unique, and some might have higher concentrations of inhibitory molecules that can affect the amplification step.
On-Chip Nucleic Acid Extraction. We further evaluated the performance of the CEID for the extraction of SARS-CoV-2 from wastewater. Wastewater samples were spiked with heatinactivated SARS-CoV-2 at different concentrations and extracted using three different methods; CEID, ESP (refer to the experimental section), and Qiagen AllPrep DNA/RNA Micro Kit, which is commonly used for extraction of DNA/RNA from wastewater (a "gold standard" in this case). The spiked concentrations ranged from 10000 to 100 Copies per mL (Cp/ mL) of wastewater. After extraction using these methods, SAR-CoV-2 was quantified using RT-qPCR on the elution buffer from each of the methods. As shown in Figure 4, the on-chip extraction resulted in higher SARS-CoV-2 copies per elution compared to ESP and Qiagen at different spiked concentrations. On-chip and ESP extractions successfully detected SARS-CoV-2 at concentrations as low as 100 Cp/mL, whereas the Qiagen extraction method failed to extract the virus at this concentration. Moreover, the variations between replicates were generally smaller for the on-chip extraction, leading to less errors in measurement. As shown in Table S5, while the ESP had a higher average recovery rate (82.4%) than the on-chip method (35.9%), the on-chip protocol processes a substantially larger input volume. As a result, the effective concentration factors were 2.1X for ESP and 9X for the on-chip method.
Colorimetric RT-LAMP Limit of Detection for SARS-CoV-2. In the first step, the limit of detection (LoD) of colorimetric RT-LAMP for SARS-CoV-2 was measured using serial dilutions of spiked samples. The serial dilutions were made by spiking known amounts of SARS-CoV-2 genomic RNA into a background of extracted nucleic acids from a negative wastewater sample (i.e., the original wastewater did not show any nonspecific amplification, confirmed by RT-qPCR and RT-LAMP). Interestingly, our data showed that, unlike RT-qPCR, the presence of background nucleic acids improved the RT-LAMP assay by decreasing the Ct (Figure S8). Other studies reported the same effect, where carrier RNA (background nucleic acids) was used to improve the RT-LAMP. 51,52 In our case, wastewater contained a high concentration of RNA, which can work as a carrier RNA to improve amplification efficiency. As a result, spiked dilutions were prepared in extracted wastewater, rather than nuclease-free water. These experiments were performed off-chip, on a 96-well plate and inside a plate reader, to accurately measure the RT-LAMP Ct.
The amplification curves are shown in Figure 5A using the ratio of absorbance at 560 and 480 nm. This ratio can quantify the amount of color change from pink to yellow. RT-LAMP was able to amplify SARS-CoV-2 down to 5 copies per reaction. However, due to the presence of 6 primers in the RT-LAMP reaction, primer dimers can form, and nonspecific amplification can happen at later times in the reaction. In our case, one of the replicates of our negative control (0 copies per reaction) started to show amplification at around 45 min. New England Biolabs suggests running the RT-LAMP for 30 min. By choosing a cutoff time of 30 min, the off-chip assay limit of detection was measured to be somewhere between 10 and 100 copies per reaction, which agrees with prior works. 46,[53][54][55] As shown in Figure 5B, at high concentrations of the target, the RT-LAMP Ct correlated with the number of copies per reaction (R 2 = 0.96), where higher concentrations lead to a faster color change. However, at concentrations lower than 100 copies per reaction, a similar correlation did not hold between the Ct time and the target concentration, and the variations between replicates were high, leading to a loss of accuracy. This suggested that RT-LAMP can be used as a qualitative method for concentrations lower than 100 copies per reaction.
In the next step, we performed RT-LAMP on our chip wells (on-chip RT-LAMP) and analyzed the color change in real-time using a phone camera (iPhone 13, Apple). Our previous experiences with SARS-CoV-2 detection in wastewater showed that the copies per reaction are usually in the range of 0-500. 11,27,41,43 Therefore, for on-chip RT-LAMP, we explored the LoD using lower concentrations of SARS-CoV-2 (0-400 copies per reaction). While the RT-LAMP Ct correlated with the SARS-CoV-2 concentration (R 2 = 0.78), the high variation between replicates made it difficult to use on-chip RT-LAMP for quantitative measurements at low concentrations (Figure 5C). Based on replicate positivity shown in Figure 5D, at 50 or more copies per reaction 3 out 3 replicates were positive. At 25 and 12 copies per reaction 2 and 1 replicates were positive, respectively. This agreed with the off-chip RT-LAMP, which showed an LoD between 10 and 100. The negative control (0 copies per reaction) did not show a sign of positivity at 30 min, supporting the use of a 30 min cut off for our assay. It should be noted that the greater variability seen at the lowest concentration is not an inherent difference between the on-chip and off-chip assays; it simply arises from the different negative Ct we applied (60 min for the off-chip assay and 30 min for the on-chip assay).
Analysis of Spiked Wastewater Samples with Integrated CEID. As shown in Figure 6A and Video S2, wastewater samples mixed with lysis buffer and PMPs were loaded into the chip and a magnet dragged PMPs through oil and buffers leading to the elution of nucleic acids in nuclease-free water. In the last step, the extracted nucleic acids along with positive and negative controls were mixed with colorimetric RT-LAMP Mastermix in the assigned wells. The CEID was placed on the hot plate at 65 °C. As shown in Figure 6B, the negative and positive wastewater samples were visibly distinguished by the color change. The whole process took 40 min, including 5 min for extraction, 5 min for elution, and 30 min for detection (hands-off time, Table S14).
To study the performance of the integrated CEID, wastewater samples from different locations were spiked with varying amounts of heat-inactivated SARS-CoV-2. Nucleic acids were then extracted, and SARS-CoV-2 was detected using on-chip colorimetric RT-LAMP. An example of the final results for positive or negative wastewater is shown in Figure 6B. Additionally , the concentration of SARS-CoV-2 was measured using ESP (our standard lab method) and RT-qPCR. As the concentration of SARS-CoV-2 increased, more of the replicates became positive. As shown in Figure S10A, there was not a significant difference in the concentration of SARS-CoV-2 between 0 and 1 RT-LAMP positive replicates. However, for wastewater samples with 2 or 3 RT-LAMP positive replicates, there was a significant difference between the RT-qPCR measured concentrations. As a result, wastewater samples with 0 or 1 positive replicates are considered negative or very low concentration samples. Moreover, the RT-LAMP Ct was plotted against SARS-CoV-2 concentration in wastewater, which showed a relatively low R2 value (0.48), due to the low copies per reaction concentration (Figure S10B). Therefore, we used the RT-LAMP positivity to determine the limit of detection for the integrated CEID which was 113 Cp/mL of wastewater (Figure 6C).
Wastewater Surveillance of SARS-CoV-2 Using the Integrated CEID. The CEID was used to monitor the presence of SARS-CoV-2 over 6 weeks in the wastewater samples from two different locations, a wastewater treatment plant, and a hospital sewage manhole. The samples were processed by both ESP and the integrated CEID. Our standard lab method included ESP sample processing and RT-qPCR to measure the concentration of SARS-CoV-2 in wastewater. The results were then reported in copies per mL of wastewater on the left Y-axis in Figure 7A-D. For the integrated CEID both the RT-LAMP replicate positivity (out of 3 and RT-LAMP Ct (in minutes) were reported. As shown in Figure 7A,B, the trend of SARS-CoV-2 concentration was also seen in the replicate positivity (Kendall's Tau value of 0.64 and 0.79, which are considered "high correlation"). As an example, for location A, RT-qPCR results from samples on 04/30 and 05/07 indicated an increase in the concentration of SARS-CoV-2 which agreed with higher RT-LAMP replicate positivity, as well as lower RT-LAMP Ct. The RT-LAMP Ct was also highly correlated with the SARS-CoV-2 concentrations as measured by RT-qPCR (shown by Kendall's Tau values of -0.64 for both locations). Based on these results, CEID measured SARS-CoV-2 in wastewater in a semiquantitative manner and indicated SARS-CoV-2 concentration trends with replicate positivity, as well as RT-LAMP time to positivity.
■ CONCLUSION WBE offers a powerful tool for monitoring infectious diseases and identifying emerging hotspots by analyzing wastewater for genetic material from pathogens like viruses or bacteria. WBE provides a cost-effective and noninvasive way to assess the health of entire communities without relying on individual testing. This is particularly valuable in regions with limited healthcare infrastructure or resources, where access to medical facilities may be restricted, and traditional disease surveillance is challenging. 56,57 Moreover, these low resource settings can serve as the breeding ground for emerging pathogens, which emphasizes the importance of adapting WBE in these regions. In LMICs and rural settings, WBE can uncover previously undetected disease hotspots, enabling targeted interventions in areas that might otherwise be overlooked. However, WBE adaption in LMICs and rural areas is hindered due to the need for a centralized lab with infrastructure, expensive instruments, and skilled personnel. As a result, there is a need for simplified and easy-to-use technologies to analyze wastewater samples for pathogens. A point-of-use device, that can process the wastewater sample at the point-of-sample collection, can greatly increase the capacity of WBE in LMICs and rural areas. By leveraging local wastewater systems, this approach offers a scalable and practical solution for global health monitoring, particularly in underserved communities.
To date, there are only a few studies that aim to develop a point-of-use device for wastewater testing of infectious diseases. Boza et al. evaluated a field deployable RT-LAMP device, called MINI, to measure SARS-CoV-2 in wastewater. 58 While the device can perform RT-LAMP in the field, the concentration and extraction steps were performed in the lab using an automated liquid handler, which is not portable, limiting the ability of the workflow to be performed in the field. Similarly, Bivins et al., used RT-LAMP for SARS-CoV-2 detection in wastewater, but the RNA extraction step is performed in the lab using a Qiagen kit, which requires a centrifuge. 53 Other studies have also tried RT-LAMP to simplify WBE, but the concentration and extraction steps still need to be performed in the lab. [59][60][61] The only studies that integrate all the steps (i.e., concentration, extraction, and identification) into the same platform use a paper microfluidic device. 62,63 Both of these studies use a syringe and wax-printed paper to concentrate and extract the nucleic acids, and then RT-LAMP for target amplification. While, all the steps are integrated into a field ready device, there are still multiple liquid handling steps during the paper-based extraction, including addition of wash buffers and elution buffer and folding the paper multiple times, with a sample-to-answer time of around 1.5 h.
In this work, we have successfully developed a device (CEID) with the potential for rapid, and cost-effective point-of-use nucleic acid concentration, extraction, and identification from wastewater. The CEID uses nonproprietary, open-source reagents and affordable materials to perform the concentration and extraction of nucleic acid from wastewater samples. The target nucleic acids are then identified using a colorimetric RT-LAMP assay, where the final results can be visualized through a color change. The total per sample cost is about $11 for the CEID, with the RT-LAMP Mastermix accounting for more than 60% of this cost (Table S6). With further developments, the cost for the Mastermix can be lowered by using bulk order of the Mastermix from NEB or developing an in-house Mastermix recipe. Moreover, the CEID only requires a hot plate at 65 °C, which can be reached using a battery-powered PTC heating element. As a comparison, the ESP and Qiagen methods cost about $34 and $61 per sample respectively, and require high speed centrifuge and PCR machine, which are expensive, bulky, and not available in many parts of the world (Tables S7-S11). In terms of time, the CEID takes about 40 min, of which only 5 min is hands-on time. As a comparison, ESP and Qiagen requires around 3 h to complete, with 35 and 55 min hands-on time, respectively (Tables S12-S14). The CEID effectively measured the presence of SARS-CoV-2 in wastewater samples from a hospital sewage stream and a wastewater treatment plant over the course of 6 weeks. The results showed good correlation with our standard lab approach (i.e., ESP and RT-qPCR) and CEID. While CEID is not able to quantify the amount of SARS-CoV-2, it can reveal the trends of pathogen in wastewater in a semiquantitative manner. CEID has the potential to expand to other infectious diseases by implementing RT-LAMP assays for other pathogens such as influenza or RSV.
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.analchem.5c00831.
https://doi.org/10.1021/acs.analchem.5c00831Anal. Chem. 2025, 97, 13140-13150