ABSTRACT
Urban sanitation infrastructure is inadequate in many low-income countries, leading to the presence of highly concentrated, uncontained fecal waste streams in densely populated areas. Combined with mechanisms of aerosolization, airborne transport of enteric microbes and their genetic material is possible in such settings but remains poorly characterized. We detected and quantified enteric pathogen-associated gene targets in aerosol samples near open wastewater canals (OWCs) or impacted surface waters and control sites in La Paz, Bolivia; Kanpur, India; and Atlanta, USA via multiplex reverse-transcription qPCR (37 targets) and ddPCR (13 targets). We detected a wide range of enteric targets, some not previously reported in extramural urban aerosols, with more frequent detections of all enteric targets at higher densities in La Paz and Kanpur near OWCs. We report density estimates ranging up to 4.7 × 102 gc per m3air across all targets including heat stabile enterotoxigenic E. coli, C. jejuni, enteroinvasive E. coli/Shigella spp., Salmonella spp., norovirus, and Cryptosporidium spp. An estimated 25%, 76%, and 0% of samples containing positive pathogen detects were accompanied by culturable E. coli in La Paz, Kanpur, and Atlanta, respectively, suggesting potential for viability of enteric microbes at the point of sampling. Airborne transmission of enteric pathogens merits further investigation in cities with poor sanitation.
INTRODUCTION
With few exceptions, large cities in low- and middle-income countries (LMICs) have inadequate sanitation infrastructure1–3. Unsafe water and sanitation enable the transmission of enteric pathogens from infected individuals to susceptible hosts via direct contact or through the environment in multiple interconnected pathways4,5. While a rich and rapidly growing body of literature describes microbial risks associated with direct or indirect exposure to fecal contamination in a wide variety of settings, relatively few studies have examined the potential for transmission of enteric pathogens via the aeromicrobiological pathway in cities in LMICs. In these settings, the transport of enteric pathogens in aerosols may be possible due to a confluence of factors: inadequate sanitation infrastructure resulting in concentrated flows of fecal wastes, a high disease burden resulting in high-risk waste containing human enteric pathogens, high population density, and environmental conditions that may be conducive to the aerosolization of concentrated fecal wastes. The aerosolization, transport, and deposition of microbial pathogens in cities lacking good sanitation could lead to exposure either through inhalation or through ingestion via other pathways (e.g., food, water, direct contact)6.
Aerosolization of biological material is known to be possible via several mechanisms including bubble bursting7–9, evaporation, raindrop impaction10,11, and others12–15. The creation and persistence of bioaerosols can be associated with a range of variables related to environmental conditions and the built environment including rain events16–19, meteorological conditions20–22, urban surface waters and water features6,23,24, wastewater treatment unit processes that include mechanical mechanisms25,26, and other infrastructure. The mechanisms behind aerosolization and transport of microorganisms from liquid surfaces and the microbial effects on droplet lifetime have been well-characterized under controlled conditions7,8. Laboratory studies have revealed that bubbles in contaminated water surfaces may experience conditions manipulated by microorganisms, allowing for smaller, more numerous, and higher velocity droplets to transition from water to air9. It has been shown that pathogens can be released via raindrop impaction, allowing for aerosol transport10,11. Despite an increasingly nuanced understanding of the physics of microbial transmission via aerosols, less is known about viability and persistence of infectious pathogens in this medium.
Among important human pathogens transmitted in aerosols, respiratory viruses including SARS-CoV-2 and influenza are the best characterized, primarily in indoor settings, given the central role of inhalation in exposures27–31. In the case of sanitation-related pathogens – where primary exposure is generally ingestion rather than inhalation – studies in high-risk, extramural (outdoor) settings in the USA and in other high-income countries have revealed that bioaerosols containing enteric microbes are common near point sources of concentrated fecal waste. Studies have primarily focused on ambient air surrounding wastewater treatment plants32–43 and in the context of land application of biosolids44–54; several studies have examined bioaerosols surrounding composting facilities55,56, meat markets57, impacted urban sites58–63, and concentrated animal feeding operations64–66.
Of the studies examining bioaerosols at extramural urban sites, a majority have included a limited range of enteric microbes, generally reporting fecal indicator bacteria including members of the coliform group if using culture methods60,67,68, or reporting wastewater-associated or fecal genera typically from 16S sequencing data69–73 with rare population classification to the species or strain level. Targeted measurement of specific enteric pathogens is rare, partly because the presence of important enteric pathogens is unexpected outside settings where these infections are common, and few studies have been conducted where the burden of sanitation-related infections is high. While sequencing provides key information about airborne microbial communities by revealing relative abundance, quantitative estimates for specific pathogens are needed in exposure assessment and transport modeling74 and to further assess the potential public health relevance of this poorly understood pathway of transmission. Among previously reported studies of ambient urban aerosols, few have been located in LMICs despite the fact that cities in these countries are generally characterized by inadequate sanitation infrastructure and high population density. These conditions lead to widespread fecal contamination in close proximity to people, contributing to a high burden of disease. Diarrheal diseases caused an estimated 1.8 million deaths worldwide in 201775, with LMICs bearing a disproportionate burden of morbidity and mortality.
Based on previous literature on the presence of enteric microbes in aerosols from well-studied settings in wealthy countries, we hypothesized that aerosolized enteric pathogens could be present and quantifiable where urban sanitation is lacking. We assessed this hypothesis in two cities with poor sanitation and in one city with established and maintained wastewater infrastructure as a reference site.
METHODS
Sampling locations
We conducted sampling in Kanpur, India (May – July 2017); La Paz, Bolivia (March 2018, June 2018, March 2019, June and July 2019); and Atlanta, Georgia, USA (March 2018-January 2019). Kanpur has distinct dry (October to June) and rainy (July to September) seasons; we sampled from May to August to capture both periods. Similarly, we intentionally sampled in La Paz during both rainy (December to March) and dry (May to August) seasons.
Kanpur is densely populated (Nagar district: 4.6 million people, population density of 1500 persons/km2)76 with a majority of untreated industrial, agricultural, and sewage waste conveyed via a system of uncovered canals (open wastewater canals, OWCs) discharging to the Ganges River77,78. In La Paz, a network of rivers receive untreated sewage discharge, industry effluent, and stormwater runoff; most of the waterway flows in a series of engineered channels79,80, also characterized as OWCs. The largest of these is the highly impacted Choqueyapu River, flowing through central La Paz (population: 900,000, 900 persons/km2)81,82 where it is joined by tangential tributaries including the Orkojahuira, Irpavi, and Achumani rivers. In past studies, this river system and its basin – eventually flowing into the Amazon – has been shown to contain a diverse and rich array of enteric microbes indicating high levels of fecal contamination79,82,83. As a reference site, Atlanta is characterized by having an established and maintained subsurface wastewater infrastructure, although urban surface waters in Atlanta’s watershed experience elevated levels of fecal-associated pathogens84–86 due to nonpoint source pollution and combined sewer overflows87,88. The city of Atlanta’s population density is an estimated 1500 persons/km2 89, though sampling locations near impacted streams were in suburban locations at lower than mean population density.
We identified 18 sites in Kanpur, 37 sites in La Paz, and 8 sites in Atlanta meeting the following criteria: (1) proximity to sources of bioaerosols (<1 km) containing enteric microbes, OWCs in the cases of India and Bolivia and impacted surface waters in Atlanta; (2) public and ground level accessibility; and (3) unintrusive to members of the community during multi-hour sampling events. In Kanpur, we selected a control site greater than 1 km away from known OWCs and located on the Indian Institute of Technology IIT-Kanpur’s campus. The campus is a controlled private area with limited access to non-students and non-faculty, is less densely populated, has underground piped sewerage, and has a much lower animal presence. In La Paz, we identified two control sites >1 km from known concentrated wastewaters or other contaminated sources: (1) Chacaltaya, a weather station and environmental observatory located at 5380 m in elevation and far from human habitation and (2) Pampalarama, an unimpacted site near the Choqueyapu headwaters in a protected natural area. In Atlanta, we sampled at eight sites adjacent to impacted streams and rivers in Atlanta’s watershed: the Chattahoochee River, Proctor Creek, Foe Killer Creek, and South Fork Peachtree Creek. Additionally, we sampled on the roof of our laboratory and at ground level on Georgia Tech’s campus (located in Midtown Atlanta), >1 km from surface waters.
Bioaerosol sampling, extraction and analysis
We used a combination of high-volume filtration and aerosol impaction in sampling across sites. We used the ACD-200 BobCat Dry Filter Continuous Air Sampler (InnovaPrep, Drexel, MO, USA) with sterile and single-use 52 mm electret filters and a flow rate of 200 L/min for downstream molecular analysis post extraction. We applied a sterile single-use wet foam carbon compressed elution kit (InnovaPrep, Drexel, MO, USA) to flush the filter following the manufacturer’s instructions, yielding approximately 6 mL of liquid eluate90. The pre-sterile single use filter and elution kit is intended to prevent cross-contamination between sequential samples; all filter handling and elution was conducted aseptically via manufacturer instructions and the high volume filter was wiped down with 10% bleach and 70% ethanol before and after sampling. As a control assessment for any residual background signal in the elution buffer and filters, we applied the eluate to the filters in the laboratory, extracted the blank eluate, and used this as a template for limit of detection calculations as well as sample blank controls in molecular analyses.
We treated the eluate with guanidine thiocyanate-based universal extraction (UNEX; Microbiologics, St. Cloud, MN, USA) lysis buffer in a 1:1 ratio, storing the mix in cryovials for sample transport to the laboratory. Samples were stored in a -80°C freezer immediately after transport to the laboratory, avoiding freeze-thaw by distributing the sample and buffer mixture into multiple aliquots. As a process control prior to extraction, we spiked the mix with 5uL of Inforce 3 Bovine Vaccine (Zoetis, Parsippany, NJ) containing bovine respiratory syncytial virus (BRSV) and bovine herpes virus (BoHV). Briefly, we conducted manual DNA and RNA extraction with the following steps: 1) added 300 uL of the mixture and 300 uL of 70% ethanol to a HiBind mini column (Omega BioTek, Norcross, GA), 2) following manufacturer advice to increase DNA yield, repeated step (1) three times, using a total of 450 uL of sample eluate, 3) washed the filter column with 100% ethanol, 4) and finally, eluted the nucleic acids and stored them in 50-75 µL of 10 mM Tris-1 mM EDTA (pH 8) at -80°C until further analysis91. Previous experiments have demonstrated an optimized recovery efficiency for Cryptosporidium parvum of around 50%92 and comparable crossing threshold (CT) values to other commercially available kits when assessed using real-time PCR91. In total, we collected 75 high-volume air samples from La Paz (71 collected near OWCs and 4 collected from control sites >1 km from OWCs), 53 high-volume air samples from Kanpur (45 collected near OWCs and 8 collected from one control site >1 km from OWCs), and 15 high-volume air samples in Atlanta (10 collected near impacted surface waters and 5 collected from the control site >1 km from impacted sites). We detail the time of day of sampling grouped by morning samples (AM, within the time period of 7am-12pm) and afternoon samples (PM, within the time period of 12pm-7pm) in addition to relevant information as described in the EMMI MIQE guidelines in Supporting Information93,94. The average volume sampled with the high-volume sampler was 47.5 m3 in La Paz (95% confidence interval = 42.3, 52.8), 36.3 m3 in Kanpur (95% confidence interval = 35.8, 36.7), and 28.3 in Atlanta (95% confidence interval = 30.4, 26.2).
For all high-volume air samples in Kanpur (n=53), we applied 1 mL of BobCat eluate to Compact Dry-EC (CD-EC) plates (Hardy Diagnostics, Santa Maria, CA, USA)95 for culture of total coliform and E. coli. Concurrent with high-volume air sampling in La Paz (n=31) and in Atlanta (n=15), we simultaneously used the Six-Stage Viable Andersen Cascade Impactor (ACI) with plates in six partitioned chambers at a flow rate of approximately 28.5 L/min for 1 hour to collect size-resolved bioaerosols in the size range of 0.65 to >7 µm (ACI, Thermo Scientific™, USA)96. We used AquaTest medium (Sisco Research Laboratories PVT. LTD., India) in the ACI to detect E. coli97–99. All culture samples were incubated at 37 °C and counted per the manufacturer’s instructions after 18-24 hours for colony forming units (CFUs). Due to poor potential recovery of culturable E. coli in high-volume samples from Kanpur, we subsequently used the more sensitive ACI in later sampling in La Paz and Atlanta. The average volume sampled in La Paz and Atlanta with the ACI was 1.30 m3 (95% confidence interval = 1.23, 1.37) and 0.970 m3 (95% confidence interval = 0.930, 1.01) respectively.
Meteorological data
In La Paz, we collected solar UV irradiance (UVB, 280–320 nm), temperature (°C), and relative humidity (%) data from a stationary radiometer (Yankee Environmental Systems, Turners Falls, MA, USA) located approximately 2-5 km from sampling sites depending on location. In Kanpur, we collected on site data for each sample. We collected temperature (°C), and relative humidity (%) using the Vaisala HUMICAP Humidity and Temperature Transmitter Series HMT330 (Helsinki, Finland) and windspeed and direction with an anemometer (Ambient Weather, Arizona, USA). We conducted linear regression analyses to determine meteorological effects on enteric pathogens in aerosols. All analysis were conducted in R version 4.0.2100. We assessed the samples on a daily level as well as disaggregated the data by whether samples were taken in the morning (8am-12pm) or in the afternoon (12 pm and later).
Enteric pathogen screening: multiplex qPCR
As a first step in screening enteric targets, we analyzed high-volume aerosol samples using a custom multiplex qPCR-based TaqMan Array Card (TAC) assembled and optimized by Thermo Fisher Scientific (Waltham, MA, USA). Select targets included globally important sanitation-related viruses (pan-adenovirus, pan-astrovirus, pan-enterovirus, norovirus GI/II, rotavirus A-C, and sapovirus I/II/IV/V), bacteria (Aeromonas spp., Campylobacter coli, Clostridium difficile, 9 Escherichia coli virulence genes (SI Table 1), Enterococcus faecalis, Enterococcus faecium, Mycobacterium tuberculosis, Salmonella spp., Vibrio cholerae, Yersinia spp.), protozoa (Cryptosporidium parvum, Entamoeba histolytica, Giardia duodenalis), and helminths (Trichuris trichiura, Ascaris lumbricoides) as well as multiple internal controls Though many of the targets are pathogenic, the card also includes some that may be commensal or of uncertain health relevance101, particularly in settings where asymptomatic carriage of common enteric pathogens is common102. We detail methods, descriptive statistics, targets, specific classifications of strains and types included in these assays, and their pathogenic relevance in SI Text and SI Table 1, also previously described by Capone et al 103 and in accordance with the MIQE guidelines104.
Quantitative molecular assays
ddPCR
For density estimation, we conducted absolute quantification of 12 enteric pathogen targets in high-volume aerosol samples via Droplet Digital PCR (ddPCR; QX200 Droplet Digital PCR System, Bio-Rad, Hercules, CA, USA). Targets included nucleic acids associated with selected viruses (adenovirus A-F, pan-enterovirus, norovirus GI, and norovirus GII), bacteria (Campylobacter jejuni, enteroinvasive E. coli (EIEC)/Shigella/spp., heat stabile enterotoxigenic E. coli (ST-ETEC), and two targets for Salmonella spp.), and protozoa (Cryptosporidium spp. and Giardia duodenalis assemblage B) and are detailed in SI Table 2. Targets represent a subset from TAC chosen based on public health relevance in the context of the global enteric disease burden102. Although by necessity a sub-set of globally important diarrheagenic pathogens, it includes those responsible for the highest burdens of morbidity and mortality resulting from acute diarrhea105 and that have been implicated in large-scale studies of diarrheal etiology102,106–109. We experimentally determined LODs for each assay using a probit analysis outlined by Stokdyk et. al110. A complete and detailed description of methods, pathogen targets, assay conditions, and all digital MIQE111 requirements are included in the SI. All data and metadata are accessible in a publicly accessible data repository (https://osf.io/NP5M9/)93
In addition to stratifying by sampling city and season (for La Paz and Kanpur), we further disaggregated data by distance from nearby putative sources. We collected GPS coordinates for each sample site and estimated the linear distance from the nearest OWC or impacted surface water. We use a priori-defined categories of 0-10 m and more than 10 m from OWCs (or impacted surface waters in Atlanta) and assessed the number of unique target detections per total assays we ran on each sample at each distance category. We used a Wilcoxon rank sum test to evaluate whether molecular detections decreased as a function of increasing distance from OWCs. All analyses were completed in R version 4.0.2100 and significance was based on 95% confidence (α=0.05).
RESULTS
Culturable fecal indicator bacteria (FIB) in aerosols
Because we could not measure pathogen viability directly across all targets at the point of sampling, we used culturable FIB as a proxy measure to indicate potential for viability of enteric microbes. Of the 45 air samples we collected in Kanpur near OWCs (<1 km) and analyzed by culture, 61% had detectable E. coli with an average concentration and 95% CI of 1.5 ± 1.3 CFU/m3air across positive detections. Across all 45 samples including non-detects, the average concentration of E. coli was 0.92 ± 0.41 CFU/m3air. All control samples (n=7) taken >1 km away from observed fecal contamination were negative for culturable E. coli. In La Paz, adjacent to the Choqueyapu River and its adjoining tributaries, of the 28 air samples in close proximity to uncontained waste (<1 km) and analyzed for viable coliform, 52% were positive for E. coli with an average concentration and 95% CI of 11 ± 3.8 CFU/m3air. Across all 28 samples including non-detects, the average concentration of E. coli was 5.3 ± 2.1 CFU/m3air. We did not detect any E. coli in the samples taken >1 km away from uncontained waste (n=4). The size distribution capabilities of the ACI revealed that 27% of culturable E. coli were under 2.1 μm, the size cutoff for fine aerosol particles112. No samples in Atlanta were positive for E. coli though we detected total coliform (Figure 2).
Screening of enteric microbes in aerosols
We analyzed a subset of 40 high-volume samples from Kanpur, 23 high-volume samples from La Paz, and 13 high-volume samples from Atlanta for the presence and absence of 42 molecular targets including those specific to an a priori-defined list of globally important enteric viruses, bacteria, and protozoa. To compare across cities where we collected different numbers of samples, we assessed detections in the context of total volume collected at each site, normalizing to 2000 m3 for direct comparison. In La Paz, we detected genes associated with astrovirus and Aeromonas spp. in one of the three control samples. In Kanpur, we detected one each of the following molecular targets in control samples (n=8): adenovirus, non-typable rotavirus, Aeromonas spp., EIEC/Shigella spp., and Yersinia spp. At the control site in Atlanta (n=6), we detected Aeromonas spp. in one sample.
Positive detections for a priori-defined enteric pathogen-associated gene targets were enriched in samples taken less than 1 km from known fecal waste flows across all cities. In Kanpur, 53% of samples (n=13) were positive for at least one target, 28% (n=10) were positive for at least two targets, and 3% (n=4) were positive for at least five targets. Among these positive detections were genes associated with two protozoan parasites (Cryptosporidium parvum and Giardia duodenalis), four viruses (pan-astrovirus, pan-enterovirus, norovirus GII, and rotavirus), and five bacteria (Aeromonas spp., Campylobacter coli, pathogenic enteroaggregative E. coli (EAEC), Enterococcus faecium, and Vibrio cholerae). In Bolivia, 76% of samples (n=16) were positive for at least one target, 62% (n=13) were positive for at least two targets, and 19% (n=4) were positive for at least five targets. Among these positive detections were five viral targets (adenovirus 40/41, pan-adenovirus, pan-astrovirus, pan-enterovirus, and norovirus GII), and nine bacterial targets (Aeromonas spp., EAEC, ST-ETEC, heat labile (LT)-ETEC, EIEC/Shigella spp., Enterococcus faecium, Salmonella spp., and Yersinia spp.). In Atlanta, 6 of 13 samples were positive for one target (46%). Targets detected included one virus (pan-adenovirus) and two bacteria (Aeromonas spp. and Campylobacter coli). There were two Atlanta samples adjacent to the Chattahoochee River positive for adenovirus nucleic acid and one adjacent to Proctor Creek with Campylobacter coli nucleic acid, both above their respective 95% LODs (SI Table 2). The Chattahoochee commonly experiences levels of fecal indicator bacteria beyond EPA-recommended limits for recreational use, above 235 CFU per 100 mL86,113,114. To compare across cities, we normalized the collection volume to 2000 m3 and calculated the expected number of positives based on measured values. We also highlight the fraction of samples with positive detects that also tested positive for viable E. coli through culture (Figure 3). An estimated 25%, 76%, and 0% of samples containing positive pathogen detects were accompanied by culturable E. coli in La Paz, Kanpur, and Atlanta, respectively. We detected the highest pathogen diversity in La Paz, with genes specific to two enteric protozoa, ten bacteria, and six viruses among pre-defined targets. In La Paz, we detected multiple genes associated with a diverse array of pathogenic E. coli including EIEC/Shigella, LT and ST-ETEC, and two gene targets for EAEC. In Kanpur, we detected genes associated with two protozoa, five bacteria, and four viruses among pre-defined targets. In Atlanta, we detected gene targets associated with two bacteria and one virus with no co-detection of culturable E. coli. A linear regression model revealed a positive correlation between positive qPCR detects and viable E. coli detects in La Paz (p=0.02).
Quantitative estimation of enteric microbes in aerosols
We analyzed high-volume samples from the three sites for quantitative estimation of select, pre-defined gene targets associated with globally important enteric pathogens. We censored raw data using the 95% LOD as a conservative threshold for positivity. In Atlanta, we quantified Shigella spp./EIEC (3 detections). In La Paz, we quantified ST-ETEC (2 detections), Shigella spp./EIEC (16 detections), Campylobacter jejuni (2 detections), Salmonella spp. ttr gene (3 detections), pan-enterovirus (3 detections), adenovirus A-F (1 detection), norovirus GI (3 detections), norovirus GII (3 detections), and Cryptosporidium spp. (3 detections). In Kanpur, we quantified Campylobacter jejuni (1 detection in 53), norovirus GI (33 detections), norovirus GII (1 detection), MS2 (1 detection), and Cryptosporidium spp. (3 detections). Among the detections above the 95% LOD threshold, target densities ranged from 1.5×101 gc per m3air to 4.7×102 gc per m3air (Figure 4).
We observed some differences in detection of enteric targets (gene copy density estimates) across meteorological variables, though interpretation of these associations is constrained by limited sample size. In La Paz, temperature was the only variable we measured that had a statistically meaningful effect on target density. In a stratified analysis from morning sample times only, our linear regression model showed that as temperature increased, pan-adenovirus, EIEC/Shigella spp., and norovirus GII density decreased (p=0.02, p=0.04, and p=0.02 respectively). We detected no significant effect of any variables on target densities later in the day. Among all samples and with analysis restricted to average daily meteorological measurements, we detected no statistically meaningful effects of any variables we examined on pathogen target densities. When grouped by season (rainy or dry, for La Paz), we found that E. coli density increased in the rainy season (p=0.05), on average. Mean temperature, relative humidity, and UV were higher in the rainy season. In Kanpur, we observed higher norovirus GI and GII densities in the dry season, before the monsoon (p=0.003 and p=0.05 respectively). We detected norovirus GI at greater densities in the morning compared to samples taken in the afternoon (p=0.03). In assessing the effects of meteorological measurements on target densities, we found that as relative humidity increased, norovirus GI density decreased (p=0.012). Similarly, as relative humidity increased, norovirus GI density significantly decreased (p=0.006).
We stratified data by distance from OWCs (for Kanpur and La Paz) or impacted urban water sources (Atlanta) (Figure 5). The number of unique targets detected decreased as distance from OWCs increased in La Paz and Kanpur. We observed a downward trend in density for one target in La Paz (EIEC/Shigella spp.) and Kanpur (norovirus GI). Overall, we observed a clear decrease (p=0.005) in the probability of detecting any positive molecular pathogen target (including ddPCR and qPCR detections above LOD) between samples collected within 10 m and samples collected greater than 10 m from OWCs in La Paz and Kanpur and impacted urban surface waters in Atlanta. We also observed a decrease in the probability of detecting bacterial, viral, and protozoan molecular pathogen targets (p=0.006, p=0.02, and p=0.009 respectively) between samples collected within 10 m and samples collected greater than 10 m from potential sources. We observed no differences in probability of detection across all targets by seasonality (rainy/dry) in La Paz and Kanpur.
DISCUSSION
Open sewers conveying domestic, institutional, commercial, and industrial effluent are common in cities in LMICs. They may also serve important drainage functions including flooding control. Sewers may be open to the atmosphere because solid waste can clog closed drains, and a lack of adequate solid waste management in some cities makes OWCs a rational approach to removing concentrated waste away from human habitation. They may pose risks, however, both to downstream communities and people in close proximity to open urban wastewater flows.
Our results suggest that in cities of LMICs with poor sanitation infrastructure and the presence of concentrated, uncontained fecal waste streams in open sewers, genes specific to enteric microbes, many pathogenic, are present in aerosols and may disseminate in the local environment. Overall, detection of fecal microbes in aerosols was higher than we expected at each of our study sites, suggesting this phenomenon may be common in similar settings. The diversity and density of enteric microbes was enriched in La Paz and Kanpur compared with reference sampling in Atlanta and was greater near open wastewater canals and impacted surface waters compared to sites farther from known potential sources. The co-detection of culturable E. coli in a high percentage of samples in La Paz and in Kanpur suggests, indirectly, that some pathogens detected could have been viable at the point of sampling. We observed no culturable E. coli in aerosols sampling in Atlanta. The health risk implications of the presence of aerosolized enteric microbes in these settings are unknown but merit further study.
Our study included a range of pathogen targets of global public health relevance, many of which have not been previously detected in urban outdoor aerosols where infection risk is a clear possibility due to the proximity of concentrated waste and high population density. In La Paz, we quantified ST-ETEC in two aerosol samples at densities of 28 gc/m3air and 150 gc/m3air. ETEC was responsible for 50,000 deaths worldwide in 2016115 but has not been previously quantified in extramural aerosols in cities where transmission is endemic. Also in La Paz, we report the first quantitative estimate of EIEC/Shigella (ipaH gene, n=16) in a similar setting at densities ranging from 1.8 gc/m3air to 53 gc/m3air. We detected and quantified other enteric bacteria not previously observed in extramural urban aerosols such as Campylobacter coli and Salmonella spp., though they have been previously quantified in air near concentrated animal feeding operations116,117.
We observed comparable prevalence of Aeromonas spp. associated nucleic acids in Atlanta, La Paz, and Kanpur with 8, 9, and 7 positive detections per 2000 m3air at each site respectively. Aeromonas spp. have been consistently detected in environmental media in a variety of settings118. Although some aeromonads are important human pathogens, in 2016 Aeromonas spp. were responsible for only 1% of total diarrheal deaths and only 19 of 36 subtypes are known to be pathogenic.115,119
Among viral detects in Kanpur and in La Paz, norovirus GI and GII may be most risk-relevant. We detected norovirus GI and GII at the highest average density across all targets in Kanpur aerosol samples (320 gc/m3air and 150 gc/m3air respectively). In La Paz, we detected norovirus GII at the highest average density across all targets (13 gc/m3air) and norovirus GI at a mean of 2.4 gc/m3air. Norovirus is relatively resistant to inactivation in environmental media120, may persist on environmental surfaces for up to two weeks121, has been shown to survive in aerosols122–126, and has an estimated relatively low median infectious dose between 18 and 103 virus particles127. It is among the most widely prevalent viral enteric pathogens, globally, with an estimated 33,000 cases and 20,000 deaths per year115,128,129. Norovirus transport through bioaerosols and subsequent deposition and indirect exposure through ingestion remains poorly characterized.
We are aware of only one previous study reporting detection of enteric protozoan parasites in air samples, from rural Mexico, by microscopy; the study reported 8 of 12 samples positive for Cryptosporidium and 10 of 12 samples positive for Giardia, possibly via aerosolization of soil130. By comparison we detected Giardia duodenalis via qPCR in 22% of samples in La Paz and 18% of samples in Kanpur, with 3 and 5 positive detections per 2000 m3air at each site respectively. Cryptosporidium was present in 9% and 3% of samples in La Paz and Kanpur respectively with 1 positive detection per 2000 m3air at both sites. We quantified Cryptosporidium spp. in aerosol samples via ddPCR in La Paz (n=3) and in India (n=2) at average densities ranging from 9.3 gc/m3air to 560 gc/m3air, the second time Cryptosporidium has been reported in an aerosol and the first quantitative estimate. Cryptosporidium is among the most prevalent etiologies of moderate to severe diarrhea globally102,131.
A small number of previous studies have identified the likely presence of aerosolized fecal material and potential for pathogen transmission in bioaerosols in similar settings. Our previous studies in La Paz and Kanpur reported detection of culturable Enterobacteriaceae and E. coli in ambient urban air that was hypothesized to also originate from OWCs132–134. A study in Malawi measured the presence of enteric microbes in ambient air (enterotoxigenic E. coli) before, during, and after pit latrine emptying, confirming that these microbes increased in nearby aerosols during pit latrine emptying events67. In 22 samples of outdoor aerosols collected in Mumbai, 28 species of culturable bacteria were identified including several opportunistic pathogens: Staphylococcus spp., Serratia plymuthica, Serratia haemolyticus, and Enterobacter aerogenes60. Bacterial bioaerosols including opportunistic pathogens have similarly been identified using 16S rDNA sequencing near composting facilities in India56. Staphylococcus aureus and other opportunistic pathogens have been identified in urban environments in the Philippines135. A larger study from Beijing, China identified many of the same genera in addition to 16 species of Pseudomonas (some potential opportunistic pathogens) and the possibly fecal-associated genera Enterococcus, Escherichia, Vibrio, and Yersinia59.
The concurrent detection of culturable E. coli in many samples from La Paz and Kanpur suggests that some of these important pathogenic bacteria (including pathogenic E. coli), viruses, and protozoa we detected may have been viable at the point of sampling136. As a commonly used fecal indicator bacterium, E. coli suggests – though does not conclusively demonstrate – the presence of aerosolized fecal material in samples137. Culturable E. coli also may indicate recently aerosolized material, since vegetative bacteria are not persistent in the aerosolized state and may be quickly inactivated if ideal conditions are not met.102 Other sanitation-related pathogens, such as those showing greater persistence in the environment (e.g., Cryptosporidium oocysts) may persist for longer periods in environmental media138 than co-occurring vegetative bacteria139, suggesting that culturable E. coli may conservatively represent the potential viability of other enteric pathogens. We further note that we have used culturable E. coli as an indicator of viability only, rather than an indicator of exposure risk: no relevant standards for E. coli or other FIB in aerosols exist that would indicate a threshold of exposure risk in these settings. Also, differences in methods for E. coli capture and culture were different across sites preventing a direct quantitative comparison of the data to inform risk on the basis of E. coli counts. More work on the comparative survival of pathogens in extramural aerosols is needed. However, such studies are challenging to conduct, because the methods for capturing aerosolized enteric pathogens at low densities – requiring high flow-rates – are not ideal for preserving viability140–142.
Size-resolved densities of culturable E. coli indicate the presence of enteric bacteria in aerosols ranging from 0.6 μm to 7 μm. Culturable E. coli that were captured under 2.1 μm112 indicate possible suspension of bioaerosols over a period of hours, with a settling velocity in still air of 0.5 m/hour for a typical particle with a 2 μm diameter143, indicating high transport potential in air near OWCs. Larger diameters indicate that the fecal bacteria may deposit on surfaces more rapidly. The size-resolved density estimates of FIB may or may not indicate the size ranges associated with other enteric microbes, including viral and protozoan pathogens, but may be useful as an initial proxy for use in transport modeling for enteric microbes. Though more work is needed to characterize exposure potential based on aerosol size, and although some enteric pathogens have been known to cause respiratory infections144–146, the most likely pathways of exposure relevant to bioaerosols in this setting are associated with surface deposition of pathogens and subsequent transport via other well understood pathways represented in the F-diagram5.
There is some epidemiological evidence that proximity to concentrated fecal waste streams in urban areas can be related to enteric infection risk. Contreras et al.147, assessing the spatial relationship between household proximity to fecal contaminated surface water canals used for crop irrigation and diarrheal disease in children in multiple municipalities in the Mezquital Valley in Mexico, determined that compared to children under 5 living within 10 m from a canal, children living 100 m from a canal had 45% lower odds of diarrhea and children living 1000 m from a canal had 70% lower odds of diarrhea. They further estimated that 24% of all diarrheal cases in the study and 50% of all cases within 100 m from a canal were attributable to canal exposure. The authors posited aerosolization of pathogens from canals as a potential pathway of exposure in this setting. Our results confirm the presence of enteric pathogens in aerosols at these distances and we report a statistically meaningful decrease in bacterial, viral, and protozoan molecular detects as distance increases from a potential source. Though the density range of gene copies per m3 that we quantified in aerosols is lower than what we might find in concentrated sources such as wastewaters, microbial infection risks attributable to aerosols in these settings cannot be ruled out given co-detection of culturable FIB and close proximity to population centers. Important next steps include further epidemiological studies examining risks attributable to this pathway as well as mechanistic studies including quantitative microbial risk assessment. Such studies would be useful in estimating both direct and indirect infection risks that account for microbial transport and viability in aerosols and following deposition via other pathways of relevance in fecal-oral transmission, including comparative risk assessment that examines the relative contributions of multiple pathways of exposure. Our quantitative estimates of specific pathogens are a necessary first step towards further work in understanding the implications of the presence of these microbes in air, including fate and transport modeling and risk assessment.
This study had a number of important limitations that deserve consideration. First, the viability of enteric pathogens detected by molecular methods in aerosols cannot be assumed, even with co-culture of fecal indicator bacteria. Although E. coli viability in samples may indirectly indicate potential viability of other microbes present in bioaerosols, and vegetative bacteria may represent a conservative proxy since they have been shown to survive relatively poorly in aerosols138,148, we did not measure viability of other microbes directly. The methods we used for high-volume samples present high pressure and desiccating conditions that may reduce viability of microbes captured on the filter149, and may also have resulted in an underestimation of culturable total coliforms and E. coli in Kanpur; in contrast, the ACI used in Bolivia and Atlanta is suitable for coliform survival and growth since bioaerosols are impacted at a lower flow rate onto nutrient-rich, semi-solid agar media. High-volume aerosol sampling generally presents conditions that are known to limit recovery of viable microbes150. Capture methods that preserve viability, such as impingement65 and water-vapor condensation28 typically operate at relatively low flow rates (8 - 13 liters per minute), requiring extended periods to capture targets present at low densities. We further acknowledge that pathogen-specific nucleic acids in aerosols could be attributable to either viable or inactivated microbes, or may exist as extracellular genetic material in the environment. Second, while our data suggest concentrated fecal waste streams as potential sources of aerosolized sanitation-related microbes in nearby air samples, and we observed a trend of decreasing density with increasing distance from fecal waste streams, we cannot definitively conclude that enteric microbes in aerosols derive from these sources. Cities with poor sanitation infrastructure typically have many contaminated sites, including OWCs and drains but also open latrine pits67,151, composting sites56, or uncontained solid waste152,153. Animals and animal waste may be common and could be aerosolized. Further work is needed on methods for source-tracking of bioaerosols, including via sequencing approaches. Third, our molecular results are likely conservative representations of target densities in ambient air of the sampling locations based on laboratory experiments with the high-volume sampler that reveal recovery efficiencies of 1 μm particles ranging from 73% in conditions most similar to those of our study150 to 101% under controlled laboratory conditions154. Finally, this study included a limited sample size across the wide range of fecal-related targets we sought, constraining statistical power for assessment of risk factors for pathogen detection across sites. Further studies of specific pathogen transport under specific controlled conditions are needed to fully describe mechanisms of aerosolization, transport, deposition, viability and persistence in aerosols, and risk of exposure to humans.
We highlight aerosol transport as playing a potentially important and understudied role in the spread of microbes originating in fecal wastes in outdoor environments with uncertain implications for human exposure, infection, and disease transmission. Fecal-oral transmission of enteric pathogens is often summarized in the so-called F-diagram, describing key media that serve as direct and indirect sources of enteric pathogen spread: water, soil, hands, fomites and surfaces, food, and flies5,155,156. Aerosols should be added to this conceptual framework. Aerosols may allow for transport of enteric pathogens between and among media, contributing to the spread of fecal contamination and associated microbes, resulting in potential for greater exposure via contact, inhalation, or ingestion either directly or indirectly following deposition on a surface, food, water or other subsequent exposure pathway6,147. In many settings, multiple relevant pathways of exposure may exist157, and aerosols may be one more whose risk relevance remains uncharacterized but cannot be excluded from consideration. There have been no previous attempts to estimate the global, national, or local burdens of morbidity and mortality attributable to airborne transmission of enteric pathogens in urban outdoor environments, partly because this phenomenon is poorly understood and has not been studied in either epidemiological research or risk assessment at scale. Indeed, our paper is the first to quantitatively describe the potential for transmission in settings where enteric pathogens are common and therefore represent potentially important exposure risks. We hope this work stimulates further work in epidemiology and infectious disease research to advance further understanding of this pathway and to estimate attributable burdens of disease and death, where appropriate. A complete accounting of enteric pathogen transport is required to design intervention strategies with the potential to control exposures. In high-burden settings such as rapidly densifying cities in LMICs with poor sanitation, enteric pathogen transport via aerosols near concentrated fecal waste flows merits further investigation. Direct measures of pathogen viability and persistence in aerosols, exposure modeling, quantitative microbial risk assessment, and epidemiological studies would be useful next steps in further characterizing the public health relevance of this phenomenon.
Data Availability
All study data are available at Open Science Framework: https://osf.io/NP5M9/.
TOC ART: FOR TABLE OF CONTENTS ONLY
ACKNOWLEDGEMENTS
This material is based upon work supported by the National Science Foundation under grant number 1653226. Many thanks to colleagues at IIT-Kanpur including Hari Shankar, Harish Vishwakarma, Shivshankar Mishraa and numerous others who helped with instrumentation and shared their laboratory space. We further acknowledge the kind people who shared their homes and their tea with us as we sampled in the community. Thanks to all colleagues at Universidad Católica Boliviana and Universidad Mayor de San Andrés, including undergraduate assistants.
Footnotes
SYNOPSIS We detected and quantified molecular targets associated with important enteric pathogens in outdoor aerosols in cities with poor sanitation to assess the potential role of the aeromicrobiological pathway in enteric infection transmission in such settings.
This version includes a substantial revision following peer-review, with added and amended material in the Abstract, Introduction, Methods, Results, and Discussion.
ABBREVIATIONS
- (LMIC)
- Low-and middle-income country
- (OWC)
- open wastewater canal
- (BRSV)
- Bovine respiratory syncytial virus
- (BoHV)
- Bovine herpes virus
- (TAC)
- Taqman Array Card
- (ddPCR)
- Droplet digital PCR
- (EIEC)
- Enteroinvasive E. coli
- (ST-ETEC)
- Heat stabile enterotoxigenic E. coli
- (EAEC)
- Enteroaggregative E. coli
- (LT)
- Heat labile
- (LOD)
- Limit of detection
REFERENCES
- (1).↵
- (2).
- (3).↵
- (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).↵
- (61).
- (62).
- (63).↵
- (64).↵
- (65).↵
- (66).↵
- (67).↵
- (68).↵
- (69).↵
- (70).
- (71).
- (72).
- (73).↵
- (74).↵
- (75).↵
- (76).↵
- (77).↵
- (78).↵
- (79).↵
- (80).↵
- (81).↵
- (82).↵
- (83).↵
- (84).↵
- (85).
- (86).↵
- (87).↵
- (88).↵
- (89).↵
- (90).↵
- (91).↵
- (92).↵
- (93).↵
- (94).↵
- (95).↵
- (96).↵
- (97).↵
- (98).
- (99).↵
- (100).↵
- (101).↵
- (102).↵
- (103).↵
- (104).↵
- (105).↵
- (106).↵
- (107).
- (108).
- (109).↵
- (110).↵
- (111).↵
- (112).↵
- (113).↵
- (114).↵
- (115).↵
- (116).↵
- (117).↵
- (118).↵
- (119).↵
- (120).↵
- (121).↵
- (122).↵
- (123).
- (124).
- (125).
- (126).↵
- (127).↵
- (128).↵
- (129).↵
- (130).↵
- (131).↵
- (132).↵
- (133).
- (134).↵
- (135).↵
- (136).↵
- (137).↵
- (138).↵
- (139).↵
- (140).↵
- (141).
- (142).↵
- (143).↵
- (144).↵
- (145).
- (146).↵
- (147).↵
- (148).↵
- (149).↵
- (150).↵
- (151).↵
- (152).↵
- (153).↵
- (154).↵
- (155).↵
- (156).↵
- (157).↵