INTRODUCTION
The fungus
Aspergillus fumigatus plays an important role in the environment as a decomposer, recycling nutrients from decaying plant matter into the soil. This highly sporulating mold is commonly found in woodchip piles and compost from household waste, sewage, sludge, and moldy hay (
1), where its thermotolerance enables it to proliferate during the thermogenic phase of composting when temperatures reach 40 to 60°C (
2). The small size of
A. fumigatus spores (2 to 3 μm), and their hydrophobicity means they are easily aerosolized and transported on air currents, making
A. fumigatus a globally ubiquitous fungus (
3). Exposure to this mold is medically important, and it is estimated that humans inhale several hundred
A. fumigatus spores per day (
4), which can trigger an immunoinflammatory response resulting in severe asthma with fungal sensitization (SAFS) or allergic bronchopulmonary aspergillosis (ABPA) (
5). The size of the spores allows them to bypass mucociliary clearance in the lung (
6), whereupon they must then evade clearance by the host innate and adaptive immune responses (
7). If they survive, germinated spores establish in lung cavities, where they can eventually cause chronic pulmonary aspergillosis (CPA). CPA affects apparently immunocompetent individuals with an existing lung condition, such as ABPA, chronic obstructive pulmonary disease, tuberculosis, or lung cancer, or an underlying immune dysfunction due to diabetes, rheumatoid arthritis, or alcoholism (
8). If the host immune system is unable to prevent spores from entering the bloodstream, then invasive aspergillosis (IA) develops, which is a life-threatening infection associated with ∼58% survival (
9). Individuals who are immunocompromised due to treatment with immunosuppressants, chemotherapy, or HIV/AIDS infection are at greatest risk of IA (
9). Furthermore, individuals admitted to intensive care units with severe influenza infection are at risk of developing influenza-associated pulmonary aspergillosis, which is associated with increased mortality (
10). A similar disease is now being observed for COVID-19-associated pulmonary aspergillosis in individuals with severe COVID-19 infection (
11). It was estimated that in the UK in 2011 there were ∼178,000 individuals living with ABPA, 3,600 with CPA, and 2,900 with IA, plus an additional 377 to 1,345 cases of IA in critical care patients (
12). The number of patients in the UK presenting with infections that are resistant to one or more of itraconazole (ICZ), voriconazole (VCZ), and posaconazole (PCZ)—the frontline triazole drugs for treating aspergillosis—increased from 3 to 7% between 1999 and 2001 to 14 to 20% between 2007 and 2009 (
13). Triazole-resistant infections are associated with treatment failure, salvage therapy with more toxic antifungals and increased case fatality rates (CFRs), with CFRs up to 88% reported for triazole-resistant IA (
14).
Triazole resistance is most commonly caused by polymorphisms in the
cyp51A gene, which results in increased production of, or configurational changes in, lanosterol-14a-demethylase, an enzyme involved in ergosterol biosynthesis and the binding target of triazole drugs. An environmental route for the acquisition of triazole-resistant infections has been proposed due to the increase of infections caused by
A. fumigatus isolates with a tandem repeat (TR) in the promoter region of
cyp51A coupled with single-nucleotide polymorphisms in the coding region leading to amino acid substitutions in the protein, which are frequently recovered from air and soil samples globally (
15). This is likely due to the use of the fungicides epoxiconazole, tebuconazole, propiconazole, difenoconazole, and bromuconazole, which have similar molecular structures to the medical triazoles and show cross-resistance (
16). In 2008, these were the second, third, sixth, ninth, and seventeenth most sprayed triazoles in agriculture in the UK, respectively (
17). In agriculture, triazoles are applied to wheat, beans, carrots, oilseed rape, soft fruits and vines; in horticulture, they are used to sterilize bulbs and to control fungal diseases in lawns and ornamentals; and in industry they are used as wood preservatives and antifouling agents in leather, paper, textiles, paints, and adhesives (
17).
The UK government is committed to reducing carbon dioxide emissions by diverting waste from landfill and incineration to composting (
18), and compost features in the government’s Food 2030 strategy for improving the productive capacity of soil (
19). Compost producers accept input material from agriculture, horticulture, forestry, wood, and paper processing, leather and textile industries, and household and garden waste, which are highly likely to contain triazole residues. In 2007, 90% of composting facilities in the UK produced compost in open windrows (
20); where organic waste is shredded, mixed, and placed in uncovered rows that are turned regularly during the composting process to improve oxygenation of the waste and to distribute heat and moisture. Composting facilities are known to produce large numbers of
A. fumigatus spores (
20–27), with resulting negative health impacts on compost handlers (
28–36), and there is evidence from the Netherlands that composting material also produces large numbers of triazole-resistant spores (
37,
38). In 2017, UK households spent approximately £450 million on compost (
39) and apply it more liberally to their gardens at 300 tonnes per hectare (t/ha) than the 50 t/ha applied to agricultural land (
40). Furthermore, more than a third of households with access to a garden report composting their garden and/or kitchen waste (
41). This means that a substantial proportion of the UK population is handling compost on a regular basis, with potential exposure to high levels of
A. fumigatus spores that may have developed triazole resistance from composts that contain triazole residues. Indeed, there have been reports of hypersensitivity pneumonitis (
42) and IA (
43–47) in apparently immunocompetent individuals following gardening activities; however, no clinical links following exposure to triazole resistant spores have been documented.
The aims of this study were to (i) determine the numbers of triazole-susceptible and -resistant A. fumigatus spores in soil samples collected from residential gardens in the UK, (ii) characterize the cyp51A polymorphisms responsible for resistance, and (iii) find environmental variables associated with the presence/numbers of A. fumigatus spores in soil samples. In order to simultaneously sample a wide range of UK gardens, we were assisted by a network of citizen-scientists trained in the collection of samples that may contain A. fumigatus. Our aim was to ascertain whether gardening activities may lead to exposure to triazole-resistant genotypes of this mold that could present a risk to susceptible individuals. Based on our findings, we present thoughts on how these exposure risks in susceptible individuals might be mitigated.
DISCUSSION
In this study, 5,174
A. fumigatus isolates were cultured from 509 soil samples collected by 249 citizen scientists from their gardens across the UK (
48). Of these soil samples, 327 (64%) grew
A. fumigatus isolates, and 101 (20%) grew isolates that were resistant to tebuconazole at a concentration of 6 mg/L. The percentage of soils that grew
A. fumigatus in this study was lower than the 78% of soils collected by Sewell et al. from several sites across South West England, including parks, cemeteries, public gardens, flower beds outside hospitals, a lavender farm, a forest, and farmland (
49). However, the percentage of soils in this study that grew tebuconazole-resistant
A. fumigatus isolates was greater than the 6% of soils in Sewell et al. that grew
A. fumigatus with increased MICs to ICZ, VCZ, and/or PCZ (
49). Of the 5,174
A. fumigatus isolates cultured in this study, 736 (14%) were resistant to tebuconazole, which is greater than the 6% prevalence of triazole-resistant
A. fumigatus reported by Tsitsopoulou et al. from urban and rural soils in South Wales (
50) and the absence of triazole-resistance detected by van der Torre et al. in isolates cultured from soils adhered to vegetables grown in the UK (
51). This prevalence of 14% is also greater than the 9% in experimental cropland and 12% in commercial wheat fields in the UK reported by Fraaije et al. (
52); however, it is less than the 37% prevalence in isolates cultured from flower bulbs bought from a garden center in Dublin reported by Dunne et al. (
53). In this study, the average concentration of
A. fumigatus from positive soil samples was 316 CFU/g, which is higher than the 43.5 CFU/g in agricultural soils and 106 CFU/g in urban soils from Greater Manchester reported by Bromley et al. (
54) and considerably higher than the 0 to 10 CFU/g reported from woodlands, grass verges, experimental cropland, and commercial wheat fields across the UK reported by Fraaije et al. (
52). Given that
A. fumigatus is often considered to be ubiquitous in the environment, it is intriguing that 36% of the soil samples collected in this study did not grow this mold. We speculate that
A. fumigatus spores and mycelial fragments in garden soils are killed by triazole residues from dipped bulbs (
53), for example, if they have not developed triazole resistance. It is also possible that
A. fumigatus is outcompeted by other microbes, especially in soils that have not experienced the high temperatures that are associated with composting.
Of the 736
A. fumigatus isolates that grew on tebuconazole at 6 mg/L, 93 (13%) did not regrow from short-term storage in the fridge, which left 643 (87%) isolates for sequencing of the
cyp51A promoter and gene coding regions. Similar to existing UK studies (
49,
50,
54), the predominant mutation identified in this study was TR
34/L98H (
n =
535; 73%). Of these isolates, 22 had amino acid substitutions in cyp51A in addition to L98H. Six isolates had T289A, I364V, and G448S amino acid substitutions, in addition to TR
34/L98H, which has been previously detected in Korea in a patient with IA (
55) and in Japan on tulip bulbs imported from The Netherlands (
56). TR
68/L98H was detected in one isolate, which was found to be two repeats of the 34-bp insert, and in three isolates TR
34 was detected without any accompanying amino acid substitutions, which was first detected in an environmental isolate collected from Scotland (
57). TR
46/Y121F/T289A was detected in 16 (2%) isolates and was accompanied by S363P, I364V, and G448S in four additional isolates, a combination reported from The Netherlands in 2018 (
52). Additional polymorphisms detected in this study included TR
53, which has been previously reported from flower fields in Colombia (
58) and from a patient with multiple-azole-resistant
A. fumigatus osteomyelitis in The Netherlands (
59), and TR
92/Y121F/M172I/T289A/G448S, which has been previously detected in flower bulb waste in The Netherlands (
38) and is two repeats of the 46-bp insert. There were 33 (4%) isolates in this study that did not contain any TRs: five contained I242V, one contained C270R, and 27 had no amino acid substitutions in
cyp51A. I242V is the only single
cyp51A amino acid substitution detected in this study to have been reported in studies summarizing
cyp51A polymorphisms (
60–63), which may suggest these polymorphisms occurred
in situ. The 28 isolates that did not contain any
cyp51A polymorphisms may well be using non-
cyp51A mechanisms for triazole resistance, such as the overexpression of efflux pumps,
cyp51B overexpression, cholesterol import, or
hapE mutation, which were not explored in this study (
64).
The only environmental variable measured in this study that was found to have a significant effect on whether a sample grew A. fumigatus or on the numbers of A. fumigatus grown was the garden location from which the sample was collected. The greatest concentration of A. fumigatus was cultured from a bag of manure at 600 CFU/g, followed by homemade compost heap samples at 505 CFU/g, commercial compost bag samples at 451 CFU/g, and pot/planters containing commercial compost at 321 CFU/g. Soil samples that did not contain compost grew fewer A. fumigatus isolates: 254 CFU/g from pot/planters and 204 CFU/g from borders. Similar observations were made for tebuconazole-resistant A. fumigatus, with concentrations of 128 to 289 CFU/g recorded for samples containing compost and of 98 to 127 CFU/g for samples without compost. As citizen scientists were only asked to indicate one garden location from which the soil sample was collected, it is possible that the concentrations of A. fumigatus spores from borders and pot/planters were inflated by the recent addition of compost that was not indicated on the questionnaire. In the regression models, soils collected from commercial compost bags, homemade compost heaps, pot/planters, and pot/planters plus commercial compost had significantly greater odds of growing A. fumigatus and tebuconazole-resistant A. fumigatus (P < 0.01 for all associations) compared to soil samples collected from borders. Furthermore, samples collected from commercial compost bags, homemade compost heaps, and pot/planter plus commercial compost grew significantly more A. fumigatus colonies compared to samples collected from borders. No association was found for garden locations sampled from and numbers of tebuconazole-resistant A. fumigatus.
Several existing studies have looked for triazole-resistant
A. fumigatus specifically in compost in the UK and globally, and the findings have been highly variable. Tsitsopoulou et al. collected 11 compost samples from agricultural fields, a horticultural nursery and public areas across South Wales that grew 10
A. fumigatus isolates in all—none of which were triazole resistant (
50). Dunne et al. do not report how many samples they collected from commercial compost bought from a garden center in Dublin or how many
A. fumigatus were cultured from these samples; only that one isolate was triazole-resistant (
53). Sewell et al. collected two samples from a compost heap in London that, combined with three samples collected from a flower bed ∼500 m away, gave a 60% prevalence of triazole-resistant
A. fumigatus (
49). Pugliese et al. sampled from composting orange peel in Italy and found
A. fumigatus concentrations of 8.8 × 10
3 CFU/g at the start of the process rising to 605.7 × 10
3 CFU/g by the end, and yet none of the 30 isolates selected for susceptibility testing were triazole resistant (
65). Santoro et al. sampled from 11 green and brown composts across Spain, Hungary, and Italy and found concentrations of
A. fumigatus ranging from 100 to 10.6 × 10
3 CFU/g; none of the 30 isolates selected for susceptibility testing were triazole resistant (
66). Ahangarkani et al. screened isolates cultured from 300 compost samples collected in Iran and detected 57 isolates with elevated MICs to ICZ and VCZ (
67). Zhang et al. collected 114 samples from a plant waste stockpile over 16 months in the Netherlands and detected >10
3 A. fumigatus CFU/g in 74% of samples, with the prevalence of triazole-resistant
A. fumigatus averaging 50% across all samples (
37). Also in The Netherlands, Schoustra et al. found concentrations of triazole-resistant
A. fumigatus of 200 CFU/g in household green waste, (1.5 to 1.8) × 10
3 CFU/g in compost heaps in residential gardens, up to 2.3 × 10
5 CFU/g in flower bulb waste, and up to 8.4 × 10
4 CFU/g in organic waste from landscaping (
38).
The key findings of this study are that 64% of soil samples collected from residential gardens in the UK grew
A. fumigatus and that 20% of samples grew tebuconazole-resistant
A. fumigatus. This means that individuals are very likely to be exposed to both
A. fumigatus and triazole-resistant
A. fumigatus spores that are aerosolized from soil when they are undertaking gardening activities (
43–47). Although this study has not undertaken susceptibility testing for the tebuconazole-resistant
A. fumigatus isolates against medical triazoles, the most commonly detected
cyp51A polymorphisms TR
34/L98H and TR
46/Y121F/T289A are associated with elevated MICs to ICZ, PCZ, and VCZ (
68). Furthermore, Hodiamont et al. reported a clinical isolate containing TR
53 as being resistant to ICZ and VCZ, with reduced susceptibility to PCZ (
59). This study also reports that the likelihood of being exposed to
A. fumigatus and triazole-resistant
A. fumigatus spores is significantly greater when handling commercial or homemade compost compared to soils in borders or pots/planters. The 14% prevalence of triazole resistance detected in garden soil samples in this study is higher than most existing studies that have sampled from rural and urban locations in the UK, which is likely being driven by the concentrated application of compost in residential settings. The National Aspergillosis Centre advises that people take care when opening bags of compost and recommends wearing a facemask while doing so to avoid dust inhalation. Currently, the only health warning on commercial compost bags is for women to not handle compost without gloves if they are pregnant, presumably to avoid toxoplasmosis infection (
69). The evidence presented here supports the recommendation for users to wear a mask while handling compost and the introduction of health warnings on bags of compost with regard to inhaling
A. fumigatus. Measures could also be taken by compost producers to sterilize the composting before packaging, thereby killing viable
A. fumigatus spores and eliminating the immediate hazard it poses to the user.