ABSTRACT
Tuberculosis (TB) remains a leading cause of infectious mortality globally, yet most cases cannot be epidemiologically linked even with extensive contact investigations and whole genome sequencing. Consequently, there remain major gaps in our understanding of where and when M. tuberculosis (Mtb) exposures occur. We aimed to investigate whether Mtb can be detected in environments where TB patients were recently present, which could serve as a tool for characterizing exposure risk. We collected 389 environment surface (ES) swabs from two high TB burden prisons in Brazil, sampling 41 (n=340) cells occupied by individuals with active TB and 7 (n=49) cells from individuals without TB. In a subset of pooled swabs (n=6) and a swab from a cigarette lighter from the cell with active TB patients, we enriched Mtb DNA using RNA-bait hybrid capture assays and performed whole genome sequencing. In prison cells, Mtb DNA was detected in 55/340 (16%) of ES swabs from cells occupied by active TB patients and none (0/49) from cells in which no active TB patients were present. Mtb was detected in 13/16 (81%) prison cells occupied by the individuals with high/medium sputum Xpert Mtb load and 8/25 (32%) with low/very low sputum Mtb load (p=0.003). Seven hybrid capture samples had a median genomic coverage of 140X. rpoB mutations conferring high-level rifampin resistance were detected in 3/7 ES swabs. Mtb was frequently detectable in environments recently occupied by individuals with active TB. This approach could be applied in congregate environments to identify and characterize high-risk settings for Mtb exposure.
1. Introduction
Despite the availability of highly effective treatment regimens and reliable rapid diagnostics, more than 10 million new cases of TB are estimated to occur every year 1. A major obstacle to reducing this burden is ongoing transmission of infection. Classical TB transmission studies predominantly rely on detecting Mtb either from patient-derived samples such as sputum or bioaerosols generated by an active TB patient at a single time-point2-4. However, a major limitation of this approach is that infectious individuals remain undiagnosed for several weeks to months, making it challenging to identify locations of transmission5. Using molecular data combined with household and social contacts, studies have reported that only a minority of Mtb infections in adults are attributable to these identifiable contacts6-8. A study in KwaZulu-Natal in South Africa found that only 12% of the transmission was molecularly linked to close contacts9. In another HIV-prevalence community in South Africa, only 19% of clustered TB cases could be attributed to household transmission10. These challenges have also been independently observed in rural Malawi where only 9% of TB cases were attributable to known contacts11. Taken together, these findings suggest that most Mtb infections are acquired in congregate settings outside the household; however, there is little direct evidence about where such transmission occurs.
Several emerging approaches have been undertaken in attempts to identify high risk locations for Mtb exposure in community settings. These include social network analysis among individuals diagnosed with TB12, use of wearable GPS loggers to identify durations individuals spend in various environments13, and mapping of ‘rebreathed air’ across communal environments using portable carbon dioxide monitors14. These approaches provide insights into potential exposure risks, though such evidence is indirect in the absence of Mtb detection in these environments. We hypothesized that Mtb DNA could be detectable on surfaces in environments where active TB patients were recently present, providing direct evidence about exposure risk in these environments. To test this hypothesis, we developed a scalable environmental sampling and molecular testing approach for Mtb DNA detection and validated it in a very high-risk environment--prisons. We further demonstrated that Mtb whole genome sequences can be recovered from environmental samples using an RNA-bait hybrid capture approach.
2. Methods
2.1. Ethics statement
The clinical study was approved by the institutional review boards (IRB) of the Federal University of Grande Dourados, Brazil (IRB##3.483.377) and Stanford University (IRB# 40285). All participants were over 18 years of age and provided written informed consent to participate.
2.2. Tuberculosis screening in prisons
As a part of an active case finding program, our team performed mass screening for TB in two prisons in Brazil (Penitenciária Estadual de Dourados and Estabelecimento Penal Jair Ferreira de Carvalho) by chest radiography and sputum testing by GeneXpert MTB/RIF Ultra assay and culture15. All incarcerated individuals were offered screening and >90% agreed to participate. Sputum testing was performed for all individuals who were able to produce sputum, irrespective of reported symptoms. Based on these data, we identified 41 cells occupied by active TB patients, in which we undertook environmental surface sampling.
2.3. Assay development and validation
We first performed a pilot study at Stanford Hospital, CA, USA, to develop and validate the environment sampling assays (Supplemental methods and results). Sterile Hydraflock swabs (Longhorn Primeswab LH-11-10) were used to collect ES swabs. Surfaces covering an area roughly the size of an adult palm (approx. 70 cm2) were swabbed with a hyraflock swab that was pre-wet with buffer (1X TE or 1X PBS). After collection, the ES swabs were immediately dispensed in 500ul buffer and transported to the molecular lab. We tested the Mtb DNA detection methods on two different platforms: GeneXpert MTB/RIF Ultra (Cepheid, Sunnyvale, USA) and StepOnePlus real time PCR machine (Applied Biosystems) using insertion sequence (IS) 6110 and the region of difference 9 (RD-9) custom TaqMan assays described previsouly16. IS6110 is typically present in multiple, but variable numbers of copies in the Mtb genome, making it a sensitive target for detection but precluding quantification of genome copies, whereas RD-9 is present as a single copy, enabling genome copy quantification.
2.3.1 Custom TaqMan assays
For custom TaqMan assays, we collected ES swabs in 500ul 1X TE buffer (pH=7.4) (Millipore Sigma) for DNA extraction. DNA was extracted using QIAamp DNA mini kit (Cat # 51304). All qPCR reactions were performed in 20ul volume using TaqMan™ Environmental Master Mix 2.0 (Cat# 4396838) with 500nM of each forward and reverse primers and 250nM of probes. The samples were analyzed in technical duplicates on a StepOnePlus real time PCR using the following cycling conditions; initial denaturation (95°C;10 minutes), 40 cycles (denaturation 95 °C 15 seconds and 60°C 30 seconds annealing and extension). The samples were considered positive only when the amplification was observed in both duplicates of IS6110 and RD-9 assays. DNA copies were quantified using a RD-9 standard curve derived from Mtb H37Rv ATCC strain (Catalogue#25618DQ) (Supplementary Figure 1).
2.3.2 Analysis on GeneXpert MTB/RIF Ultra
For GeneXpert analysis, we collected ES swabs in 500ul sterile 1X PBS (pH=7.4) (Cytiva HyClone™) and used directly without manual DNA extraction. Briefly, the swabs were vortexed for 30 seconds and the samples were mixed with Xpert MTB/RIF buffer in 1:1 (sample to buffer ratio). The samples were vortexed for 5 seconds followed by 5 minutes incubation at room temperature. The entire 1ml sample was loaded into GeneXpert MTB/RIF Ultra cartridge for Mtb DNA detection. GeneXpert Mtb load was graded semiquantitatively as high (Ct <16), medium (Ct 16<22), low (Ct 22–28), very low (Ct >28) and trace).
2.4. Environmental surface sampling in prison cells
We collected a total 389 ES swabs from 48 cells. Forty-one cells were occupied by at least one individual with TB diagnosed within the past 14 days. Seven prison cells were occupied by individuals without known active TB (Table 1). We collected 340 swabs from cells with active TB and 49 swabs from cells without active TB patients. For security reasons, the ES swab samples were collected by the incarcerated individuals who were provided brief instructions by the study staff. Prior swab collection, study staff briefly explained the purpose and importance of collecting ES swabs from the prison cells followed by a demonstration on the swabbing technique and dispensing the swab in the collection tube. The study staff waited outside the cell instructing the incarcerated individuals about which surfaces they should swab (Supplementary Figure 2). The swabs were collected in five categories-clothing, bedding, bathroom, personal utensils, and other surfaces in the cell which included floors, windows, walls, ceilings, and doors. As additional low-risk control sites, we collected 70 additional ES swabs from apartments (n=7) and a non-TB lab (n=1) at the Federal University of Grande Dourados, Brazil.
2.5. Sample collection for whole genome sequencing
Based on the GeneXpert results on ES swabs, we identified seven cells with highest rates of Mtb DNA positivity and collected additional 85 ES swabs for DNA enrichment with RNA baits-based hybrid capture assays and whole genome sequencing. The swabs were collected in 500ul 1X TE buffer. We first processed 15 swabs randomly chosen from a batch of 85 samples and individually analyzed those on IS6110 TaqMan assay. Only one swab collected from a cigarette lighter which was earlier positive on GeneXpert MTB/RIF Ultra was positive on the custom TaqMan assay. Therefore, to maximize the chances of detecting Mtb DNA, instead of analyzing the swabs individually, we divided the swabs from each cell into three categories-utensils, bed & clothing and other surfaces. We pooled the samples within each category before DNA elution to get total three swabs (one from each category) per cell. DNA was extracted as described above and eluted in 30ul nuclease-free water.
2.6. Mtb culture from paired sputum samples
To compare the enriched genomes of ES swab samples with the genomes of Mtb present in the sputum of individuals occupied those cells at the time of swab sampling, we further cultured Mtb from three sputum samples using Ogawa-Kudoh Method described previously17. Briefly, a sterile cotton swab was dipped and swirled in 2ml sputum and submerged into a 4% solution of sodium hydroxide (NaOH) for 2 minutes to decontaminate the sample. The excess of 4% NaOH was removed by compression of the swab against the flask wall and the sample was inoculated in Ogawa-Kudoh medium (pH 6.4). The cultures were incubated at 37°C and observed for growth twice a week for 60 days. Mtb DNA was extracted using Cetyltrimethyl ammonium bromide (CTAB) method described previously16.
2.7. Hybrid capture assays and whole genome sequencing
We performed IS6110 TaqMan qPCR on DNA extracted from 21 pooled swabs corresponding to seven cells. Six pooled samples from three cells and one swab from a cigarette lighter, all positive on IS6110 assay, were considered for hybrid capture assays. DNA from ATCC H37Rv was used as a positive control. All sequencing libraries were prepared using Illumina DNA Prep library preparation kit (Cat # 20018704) following manufacturer instructions. Following library preparation, we pooled eight libraries consisting of seven samples from prison cells and one with H37Rv DNA (ATCC; 104 copies). To achieve maximum genome coverage in hybrid capture reactions, we split the libraries into 4 parts (7ul each) and enriched each library using biotinylated RNA-baits (myBaits®; Arbor Biosciences) based on the inferred ancestral genome of the Mtb complex, which is genetically equidistant to all Mtb lineages. We performed the baits protocol for four replicate reactions following the standard protocol published by the manufacturer using a hybridization and wash temperature of 60°C. Briefly, the concentrated library pools were first split into four reactions, each undergoing hybridization incubation (60°C, 24 hours). These reactions were then bound to streptavidin-coated magnetic beads and washed (60°C). Finally, each reaction was split into two 15uL aliquots and performed a post-capture PCR amplification of both aliquots (8 total PCR reactions) using 2x KAPA HiFi HotStart Ready Mix (Roche) and 10x KAPA Library Quantification Primer Mix (Roche). The post-capture amplification reaction consisted of an initial denaturation step at 98°C (2 minutes), then 12 cycles of 98°C (20 seconds), 60°C (30 seconds), and 72°C (45 seconds). Final extension was at 72°C (5 minutes) and held at 8°C until 0.8x Agencourt Ampure SPRI beads cleanup. A total 20uL of cleaned enrichments were analyzed on the Illumina Miseq in paired-end runs of 2×292bp length.
2.8. Whole genome sequencing and bioinformatic methods
We trimmed low-quality bases (Phred-scaled base quality < 20) and removed adapters with Trim Galore (stringency=3)18. We used CutAdapt to further filter reads (--nextseq-trim=20 -- minimum-length=20 --pair-filter=any) 19.To exclude potential contamination, we used Kraken2 to taxonomically classify reads and removed reads that were not assigned to the Mycobacterium genus or that were assigned to a Mycobacterium species other than Mtb 20. We mapped reads with bwa v. 0.7.15 (bwa mem)21 to the H37Rv reference genome (NCBI Accession: NC_000962.3) and removed duplicates with sambamba 22. We called variants with GATK 4.1 HaplotypeCaller23, setting sample ploidy to 1, and GenotypeGVCFs, including non-variant sites in output VCF files. We included variant sites with a minimum depth of 10X and a minimum variant quality score 40 and constructed consensus sequences with bcftools consensus24, excluding indels, and used SNP-sites to extract a multiple alignment of SNPs between sampled genomes25. We excluded SNPs in previously defined repetitive regions (PPE and PE-PGRS genes, phages, insertion sequences and repeats longer than 50 bp)26. We measured drug resistance associated mutations with MykrobePredictor v0.8.0, using the ‘201901’ database of genomic predictors of resistance27. We identified lineage and drug resistance associated mutations with TBProfiler v.2.8.6 28. Our bioinformatic pipeline is available at https://github.com/ksw9/mtb_pipeline.
2.9. Phylogenetic analysis
We used the R package ape to measure the number of pairwise SNP differences between samples21. Missing sites, “Ns” in the alignment, were not counted as pairwise differences. To explore if heterozygous calls in samples with mixed infections were contributing to observed pairwise distances between samples, we excluded heterozygous sites with a minor allele frequency >5% using bcftools and recalculated pairwise SNP distances.
We fit maximum likelihood trees with RAxML-ng 1.0.1 and used a GTR substitution model and a Stamatakis ascertainment bias correction for invariant sites in our alignment. We divided the number of invariant sites by 1000 to avoid issues created by small branch lengths. We defined nucleotide stationary frequencies as frequencies in the reference genome.
2.10. Data analysis
We compared the detection of at least one swab in a cell by GeneXpert semi-quantitative grade using Fisher’s exact test. We used mixed effects logistic regression with random effect for cell to assess the relationship between Xpert semiquantitative grade and swab positivity. All statistical analyses were performed using R29.
3. Results
3.1. Mtb DNA detection in prison cells
We sampled 48 prison cells, 41 of which were occupied by 45 active TB patients. At any time point during swab collection, each cell was occupied with multiple individuals without TB sharing a cell with active TB patients (median= 14; IQR, 11-17). Thirty-seven cells housed one active TB patient each, whereas four cells accommodated two active patients. Approximately half (22/45; 49%) of participants with diagnosed TB reported cough at the time of swab collection. At the time of diagnosis, sputum GeneXpert semi-quantitative grade was high in 14/45 (31%) participants, medium in 6/45 (13%), low in 14/45 (31%), and very low in 11/45 (24%). TB treatment had already been initiated for 56% (25/45) for a median of 3 days (IQR, 1-10) at the time of sample collection (Table 1).
Of the 340 swabs collected from 41 prison cells with active TB patients, Mtb DNA was detected in 55/340 (16%) of ES swabs. We did not detect any Mtb DNA (0/49) in swabs collected from non-TB lab and residential apartments which served as negative controls. Mtb DNA positivity on custom TaqMan assays was 6/70 (9%) which was relatively higher on GeneXpert MTB/RIF Ultra 49/270 (18%). Among those positive on GeneXpert, the majority were trace (37/49; 76%) and remainder were low (5/49; 10%) or very low (7/49; 14%). Mtb DNA was detected on various surfaces, and positivity did not significantly differ across surface location or type: clothing, 25% (2/8); bedding, 23% (16/71); bathroom surfaces, 17% (8/48); personal utensils, 14% (6/42); other surfaces in the cell, 13% (23/171) (Figure 1A).
Of the 41 cells occupied by active TB patients, Mtb was detected in at least one sample in 13/16 (81%) of the cells housing TB patients with high or medium Xpert sputum bacillary load and in 8/25 (32%) of the cells occupied by individuals with low or very low sputum Mtb load (p=0.003; Figure 1B). On a per-sample basis, 27% of samples (44/161) were positive from cells with medium or high sputum Xpert cases and 6% of samples (11/179) were positive from cells with low or very low sputum Xpert. Mtb detection on ES swabs was positively correlated with sputum Xpert bacillary load (p<0.001)
3.2. Hybrid capture assays and whole genome sequencing
Because of the relatively low quantity of Mtb DNA in ES swab samples compared to routinely sequenced cultured sputum, we took an RNA hybrid capture approach, allowing us to selectively enrich for Mtb DNA from a mixed template. Six pooled swabs including three from utensils, two from bed and clothing, and 1 from other surfaces in addition to one swab from a cigarette lighter corresponding to three cells were positive on IS6110 qPCR and were selected for hybrid capture and whole genome sequencing (Table 3). To compare captured ES swab samples to sequenced cultured sputum, we additionally sequenced three cultured sputum samples from the Mtb positive individual in the same cells.
The seven captured ES swab samples had a median genomic coverage of 140X (range 23 to 209) (Supplementary Figure 2). A mean of 98% of the Mtb genome had a depth of at least 10X; 65% had a depth of at least 100X. The median efficiency of our hybrid capture assay, which we defined as the percentage of total sequenced reads that taxonomically assigned to the Mycobacterium genus and were not assigned to a nontuberculous Mycobacterium species was 92% (range 65% to 98%).
3.3. Phylogenetic analysis
Mtb sequence data from captured ES swabs allowed us to assign taxonomy and investigate phylogenetic relatedness between samples. All seven captured ES swab samples were assigned to lineage 4, the predominant lineage in the state, and nested within previously sampled genomic diversity from the state30 (Figure 3).
The three ES swabs from cell 1 were closely related (pairwise distance: 0-2 SNPs). In all three, we found evidence that both sub-lineages 4.8 and 4.1.2.1 were present; while the culture from that cell was assigned to lineage 4.3.4.2 (Table 3). We predicted that consensus sequences generated from environmental samples with two distinct sub-lineages could be chimeric, thus overestimating the genetic distance between these consensus sequences and the matched culture. To test this, we excluded heterozygous sites (Methods) from the environmental sample consensus sequences. Revised consensus sequences from the three ES swab samples were identical to one another and were 23 SNPs distant from the matched cultured isolate in cell 1 (85 SNPs distant before excluding heterozygous sites).
The three captured ES swab samples and one cultured sample from Cell 2 were all assigned to sub-lineage 4.1.2.1 and were closely related, with consensus sequences 0-3 SNPs distant (Table 3). When excluding heterozygous sites (Methods), consensus sequences were 0-2 SNPs distant.
Both sub-lineages 4.1.2.1 and 4.1.1.3 were present in the single captured ES swab sample from cell 3 and the matched cultured sample was assigned to sub-lineage 4.1.2.1 (Table 3). Again, we predicted that the presence of multiple strains in the could lead to an over-estimate of genetic distances between samples. We found that the environmental samples were genetically distant from the cultured isolate from the same cell (pairwise distance: 55 SNPs; when excluding heterozygous sites: 23 SNPs).
3.4. Antibiotic resistance prediction
We predicted antimicrobial resistance profiles using a database of confirmed mutations associated with resistance phenotypes31. All three of the captured ES swab samples from cell 1 carried the same rifampicin resistance conferring mutation (rpoB Ser450Leu). All three of these environmental samples were heteroresistant: both the susceptible and resistant alleles were present at moderate coverage depth, possibly on the distinct sub-lineages identified. 71% (50/70) of reads corresponded to the resistance-conferring mutation for the bed & clothing sample, 63% (85/134) for the utensils sample, and 60% (47/78) from the other surfaces sample. Additionally, a 70 base pairs deletion (799-869) in the katG gene associated with isoniazid resistance, was present in 18% of reads in the cell 1 utensils sample. It is likely that the rpoB and katG mutations, supported by 63% and 18% of reads respectively, were on different genomic backgrounds and do not reflect the presence of multi-drug resistant Mtb. The corresponding culture sample from a patient in cell 1 was drug sensitive. The remaining captured ES swab samples and corresponding cultures from the two other cells contained no known drug-resistance conferring mutations.
4. Discussion
In this study, we report the development and testing of a novel sampling approach based on detecting Mtb DNA on environment surfaces where active TB patients were present. We found that Mtb DNA was detectable in most environments where individuals with active TB had recently resided, and that detection was positively correlated with sputum bacillary load. By adapting our sampling method for GeneXpert, which is widely available in tuberculosis-endemic countries, this approach could be utilized in resource-constrained settings with minimal training or equipment. We further found that we could sequence Mtb genomes directly from environmental samples by using an RNA-bait hybrid capture approach, achieving sufficient coverage for lineage identification, phylogenetic analyses, and variant calling for resistance-associated mutations. Taken together, these findings suggest that environmental sampling could be a feasible approach for generating new dimensions of data for studying tuberculosis transmission.
Several historical studies in late 19th and early 20th century demonstrated the detection of Mtb in natural and built environments. In 1888, Cornet collected dust from TB medical wards, residential apartments, and asylums with active TB patients. The dust was subcutaneously inoculated into guinea pigs. Among those inoculated with dust from TB sites, 16.4% developed TB, while none of the guinea pigs in negative control group developed TB32. Later, in 1920, Rogers conducted similar experiments by treating dust with 2% sodium hydroxide solution subsequently inoculating the guinea pigs. In a series of experiments, seven out of ten guinea pigs were diagnosed with disseminated TB33. While these findings provide compelling evidence that Mtb may be detectable in the environment, most of these studies were performed in an era when identification of Mtb at molecular level was not available. This is particularly important in the context of environment samples, where slow growing Mtb can be outgrown by other mycobacteria, which may also share similar clinical or phenotypic features. In 2015, Velayati et al, attempted to isolate Mtb from 1,500 randomly collected soil and water samples from three counties in Tehran34. The samples were inoculated on Löwenstein–Jensen culture media and monitored for growth for several weeks. Mtb growth was observed in 1% soil and 10% water samples. DNA isolated from positive cultures was used to perform spoligotyping and mycobacterial interspersed repetitive units-variable number of tandem repeats (MIRU-VNTR) analysis. Typing of environmental and clinical isolates revealed partial overlap in Mtb diversity present in the area34. These findings provided the first molecular evidence of detecting Mtb in environment. However, the screening strategy employed was expensive, time consuming, labor intensive, and required culturing hundreds of samples. Additionally, spoligotyping and MIRU-VNTR do not provide information on drug resistance and may lack sufficient data required to study transmission networks. Our novel sampling method allowed detection of Mtb on surfaces by direct swab sampling. By pooling samples, we further sequenced Mtb from environment samples for lineage identification, phylogenetic analysis, and variant calling. To the best of our knowledge, this is the first study to demonstrate the molecular detection of Mtb in environment using targeted qPCR assays and whole genome sequencing.
Previous studies have demonstrated that it is possible to detect Mtb from air samples in high-risk settings such as hospitals, clinics, and schools in TB endemic settings35-37. The utility of air sampling is that it can be used to directly characterize airborne exposure risk. However, it requires that the infectious individual have been present in the sampling location within minutes of the sampling, which limits its sensitivity in communal settings. Additionally, air sampling requires more sophisticated equipment and sampling that is conducted over a period of hours. Surface sampling, by contrast, can be performed in seconds using low-cost disposable swabs. It may be more sensitive, by detecting cumulative environmental contamination, but does not provide the temporal resolution for recent presence of an infectious individual as does air sampling. Surface sampling may therefore be more useful for identifying high risk locations than for identifying the presence of infectious individuals.
We detected Mtb DNA on multiple types of surfaces and locations, including vertical (walls) and horizontal ones (e.g. floors, bedding). While the study was not designed to provide comparative information about locations within rooms, we did not observe any systematic trends in positivity by location or material type. Given that Mtb travels in aerosolized droplet nuclei and can disseminate throughout rooms, it is likely to deposit on surfaces throughout rooms. Based on our experience collecting and analyzing samples, we speculate that surfaces that are less frequently cleaned (e.g. elevated or hard-to-reach surfaces) might be more likely to contain Mtb DNA. Additionally, personal utensils which frequently come in close proximity with the mouth of an infectious individual may also have higher chances of Mtb DNA detection if not cleaned thoroughly. Interestingly, we detected Mtb DNA on cigarette lighters from two cells analyzed on GeneXpert MTB/RIF Ultra. These lighters were re-swabbed after one week for hybrid capture assays, and we detected Mtb DNA on one of the lighters which matched with the Mtb genome from paired sputum culture. Additionally, a key variable which likely affects the recovery of Mtb DNA from surfaces is how frequently they are cleaned. Cells from the study prisons were regularly cleaned; by contrast, tuberculosis isolation rooms at the hospital only undergo light cleaning while occupied, followed by a terminal decontamination when the patient leaves. It’s possible that this (along with the high sputum smear grade) partially explains the higher recovery rate of Mtb from TB isolation rooms (supplementary material).
Using RNA hybrid capture assays, we were able to enrich and fully sequence Mtb genomes from ES swabs, which generated sufficient coverage for lineage identification and high-resolution variant calling. This is the first use of hybrid capture for enriching Mtb from non-sputum samples and demonstrates the potential for environmental sampling to inform our understanding of circulating Mtb diversity, including that of drug-resistance associated genotypes. ES swabs from two cells showed evidence of mixed infection and did not match with the paired sputum culture suggesting detection of an undiagnosed case or DNA from individual occupied the cell previously. We also detected rpoB (S450L) and 70 base pairs deletion in katG mutations conferring rifampin and isoniazid resistance respectively. However, no patient-derived clinical isolate containing rifampicin resistance was identified in that prison at the time of swab collection, despite a concomitant mass screening campaign in which >90% of prison residents participated. One possible explanation to these unexpected findings is frequent movement of incarcerated individuals between prisons cells. This observation is particularly important in case of drug resistant TB, where the infected individuals who usually share a cell with several healthy individuals continue to amplify the transmission as they move from one cell to another. This concerning finding suggests that ES sampling might provide information on Mtb strains which were missed through clinical screening.
The results of this study should be interpreted with the context of several limitations. First, we studied environments (hospital rooms, prisons) in which individuals with TB are present for prolonged periods of time, and positivity in environments where TB patients are present for shorter periods requires further investigation. Future studies should evaluate communal environments such as public transit, churches, schools, shops, bars and other places where individuals congregate. Second, we only sequenced ES swabs from a small subset of prison cells which had high ES swab positivity. Therefore, our study lacks data to interpret how frequently the Mtb DNA detected on the surface was indeed from the active patient present in that environment or we detected Mtb DNA accumulated over the period in the same prison cells. Lastly, our study lacks data on Mtb culture from environment samples. We attempted to culture Mtb from 75 swabs collected from positive prison cells (data not shown). However, none of the swab cultures were positive for Mtb after several weeks of incubation. This could be due to differentially culturable organisms or in vitro competition with other environmental organisms.
5. Conclusions
In summary, we found that Mtb DNA is frequently detectable in environments recently occupied by individuals with active TB. We demonstrated that environmental sampling for Mtb can be performed using a simple swabbing procedure and assayed on an automated molecular platform widely used in low- and middle-income countries. We further found that complete Mtb genome sequences can be recovered from environmental samples using an RNA-bait hybrid capture approach, with sufficient depth to allow variant calling and phylogenetic analysis. Our findings highlight the potential for harnessing information derived from environmental sampling in studying TB transmission.
Data Availability
Data supporting the findings of this manuscript are available in the Supplementary Information files or from the corresponding author upon request. Access to the supplementary figure 2 is available from the corresponding author upon request Our bioinformatic pipeline is available at https://github.com/ksw9/mtb_pipeline.
Data availability
Data supporting the findings of this manuscript are available in the Supplementary Information files or from the corresponding author upon request. Access to the supplementary figure 2 is available from the corresponding author upon request Our bioinformatic pipeline is available at https://github.com/ksw9/mtb_pipeline.
CRediT authorship contribution statement
Renu Verma: Study design; method development, validation, and data collection; Formal analysis, Visualization; Writing – original draft & editing. Flora Martinez Figueira Moreira: Sample collection; data collection; Writing – review & editing. Agne Oliveira do Prado Morais: Sample collection; Writing – review & editing. Katharine S. Walter: Writing – original draft & editing, data analysis and visualization. Paulo César Pereira dos Santos: Sample collection; data collection; Writing – review & editing. Eugene Kim: Data collection; Writing – review & editing. Thiego Ramon Soares: Sample collection; Writing – review & editing. Rafaele Carla Pivetta de Araujo: Sample collection; Writing – review & editing. Bruna Oliveira da Silva: Sample collection; Writing – review & editing. Andrea da Silva Santos: Sample collection; Writing – review & editing. Julio Croda: Study design; Supervision; Formal analysis; Writing – review & editing Jason R. Andrews: Conceptualization; Supervision; Study design; Writing – original draft & editing; Funding acquisition; Formal analysis; sample collection.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Funding
This study was supported by NIH DP2 AI131082 and NIH AI130058.
Acknowledgements
We thank the prison staff for their support and incarcerated individuals from prisons Penitenciária Estadual de Dourados (PED) and Estabelecimento Penal Jair Ferreira de Carvalho (EPJFC), Brazil for collecting ES swabs from their cells. We also thank Leonardo Martinez from Boston University for his support.