Six Feet under Microbiota: Microbiologic Contamination and Toxicity Profile in Three Urban Cemeteries from Lisbon, Portugal

Cemeteries are potential environmental reservoirs of pathogenic microorganisms from organic matter decomposition. This study aimed to characterize the microbial contamination in three cemeteries, and more specifically in grave diggers’ facilities. One active sampling method (impingement method) and several passive sampling methods (swabs, settled dust, settled dust filters and electrostatic dust cloths—EDC) were employed. The molecular detection of Aspergillus sections and SARS-CoV-2, as well as mycotoxin analysis, screening of azole resistance, and cytotoxicity measurement were also conducted. Total bacteria contamination was 80 CFU·m−2 in settled dust samples, reached 849 CFU·m−2 in EDC and 20,000 CFU·m−2 in swabs, and ranged from 5000 to 10,000 CFU·m−2 in filters. Gram-negative bacteria (VRBA) were only observed in in settled dust samples (2.00 × 105 CFU·m−2). Regarding Aspergillus sp., the highest counts were obtained in DG18 (18.38%) and it was not observed in azole-supplemented SDA media. SARS-CoV-2 and the targeted Aspergillus sections were not detected. Mycophenolic acid was detected in one settled dust sample. Cytotoxic effects were observed for 94.4% filters and 5.6% EDC in A549 lung epithelial cells, and for 50.0% filters and 5.6% EDC in HepG2 cells. Future studies are needed in this occupational setting to implement more focused risk management measures.


Introduction
In Portugal, cemeteries are perceived as historical and religious monuments where people usually go to remember their lost loved ones [1]. The burial of corps and human
Regarding Aspergillus sp., the highest value of the genera was obtained in DG18 (18.38%) comparatively with MEA (4.11%). In DG18, the most contaminated matrixes with Aspergillus sp. were filter samples (27.71%). The genus was also present in surface swab samples (23.08%), EDC (11.25%), and settled dust (25.00%). On MEA, the matrices with the highest values of the genera were surface swab samples (4.55%), followed by EDC (2.79%) and filters (2.50%). The genus was not identified in settled dust samples ( Figure 2).

Detection of SARS-CoV-2 and the Targeted Fungal Sections
Considering all the environmental samples collected, SARS-CoV-2 was not detected as well as the four Aspergillus sections investigated.

Mycotoxins Results
From the total of 64 samples analyzed only one settled dust sample showed contamination by a single mycotoxin. The mycophenolic acid was the mycotoxin detected with a valued below the limit of quantification (20 µg/kg).

Correlation and Comparison Analysis
From the correlation analysis (Table 3), it can be concluded that: (i) in the EDC, higher fungal counts in DG18 is related to higher counts in SDA, ITZ, and VCZ and that higher counts in SDA and ITZ are related to higher values in VCZ media; (ii) in the settled dust filters, higher fungal counts in MEA is related with higher counts in DG18, higher counts in SDA and in VCZ and that higher counts in SDA and ITZ are related to higher counts in VCZ; (iii) in swabs, higher counts in DG18 is related with counts in SDA (Table 3). Table 3. Heatmap of the study of the relationship between fungal contamination (MEA and DG18) and fungal resistance (SDA ITZ, VCZ, and PSZ) in each sampling method. Results of the Spearman correlation coefficient.

Method
Media As for fungal resistance, statistically significant differences were detected in SDA (with higher values in swabs), ITZ (with higher values in EDC) and VCZ (with values in EDC) media (Table 4).
From the comparison of the fungal contamination between sampling methods, statistically significant differences were detected on MEA, and it was verified that the filters and swabs were the ones with the highest counts (Supplementary material- Table S1).

Correlation and Comparison Analysis
Regarding the diversity of species found considering MEA, it was found that the settled dust (H = 1.33; D = 3.65) was the sampling method in which greater diversity was detected, as it was the one with the highest values of Shannon and Simpson indices. Of note, when the species richness (number of species) exceeds 10, Simpson's index values are mainly influenced by the uniformity between the detected amounts of each species (Supplementary material- Table S1).

Discussion
To our knowledge this is the first attempt to assess cemeteries' occupational environment focusing on microbial contamination and mycotoxins exposure. In addition, it was the first time that a comprehensive sampling campaign, as well as the assays employed, were used in this specific setting. The sampling approach was similar to our previous studies in different occupational environments in Portugal. The use of passive sampling methods to assess exposure, in what concerns the microbial contamination, allows to overcome the expected fluctuation due to a wide range of factors such as humidity levels, ventilation, human occupancy and their activities, environmental characteristics, water infiltrations, and outdoor air [19,20]. Additionally, in this specific setting, soil can be brought inside the workers' facilities increasing the microbial contamination indoors [13][14][15]. The obtained results also followed the trend obtained in different studies where different sampling methods and culture media were employed. Indeed, different results were obtained in what concerns sampling methods, since swabs presented the higher bacterial contamination, while settled dust presented the highest fungal counts and species diversity. Aspergillus sections were more frequently observed in DG18, since this medium restricts the growth of other fungi with higher growth rates, such as Mucorales order [21,22]. Fumigati section was only observed in MEA corroborating also the results found in previous developed studies [23,24].
Fungal contamination, and more specifically Aspergillus sections distribution observed, was different in the different sampling methods as expected, since besides the media and the indoor environment assessed, also the sampling method influence the sections distribution [25]. The Aspergillus sections reported (Fumigati, Circumdati, Nidulantes, Aspergilli, and Nigri) have toxigenic potential [26], and some present clinical significance (Fumigati, Nigri, and Aspergilli) [23,27]. These Aspergillus sections are considered indicators of harmful fungal contamination [27,28], and a rigorous monitoring should be implemented to avoid their presence indoors.
The screening of azole resistance revealed the absence of Aspergillus sp. in either tested azole concentration (using the EUCAST values for susceptibility testing of Aspergillus fumigatus). Interestingly, the presence of fungi in azole media was only confirmed by EDC sampling. The species recovered (Cladosporium sp. and C. sitophila) have no clinical relevance, and no conclusions on their antifungal resistance profile can be drawn, as the used azole breakpoint values are not defined for the species.
The correlations between sampling devices (EDC, settled dust, swabs) in different media revealed that fungal counts recovered from EDC and swabs were better correlated regarding DG18 and SDA, whereas from settled dust, MEA was best correlated with SDA. This might be related with fungal total contamination, which was higher in the azole screening for EDC (9.56 × 10 4 CFU·m −2 ·day −1 ) and swabs (2.90 × 10 5 CFU·m −2 ·day −1 ), compared to settled dust (5.40 × 10 1 CFU·g −1 ·day −1 ).
The culture dependent methods allowed the observation of several Aspergillus sections in an extensive number of samples, with molecular tools failing the same sections detection. Despite this divergence, it is of relevance to use both assays, as they provide different information. In fact, molecular tools allow fast, specific, precise, and sensitive detection of the target microorganisms. Notably, they also can detect dormant or dead microorganisms and can differentiate toxigenic strains from regular fungal strains [24,29]. Though culturebased methods underestimate the total counts of microorganisms, these methods are vital since the microorganisms viability is of critical importance to predict health risks, since it affects inflammatory and cytotoxic responses [30]. This strengthens the importance of joining both culture dependent and independent methods in occupational exposure assessments [20].
As reported only one settled dust sample showed contamination by mycophenolic acid, produced mainly by Penicillium sp., with a value below the limit of quantification. Other occupational settings have shown higher mycotoxins contamination that might be related with many factors, such as the occupational environment characteristics (e.g., humidity, temperature, and availability of fungal nutrients) and the raw materials being used and handled [18,31]. However, this mycotoxin has the potential for causing immune dysregulation that in the long run may be related to increased oncologic morbidity and susceptibility to infections [32].
In this study, lung epithelial cells were used as a model for exposure by inhalation, and HepG2 cells as a model for hepatotoxicity [33]. Cells were incubated at controlled conditions with dilutions of wash extracts of filters and EDC. Settled dust filters revealed to be much more cytotoxic than EDC, both in A549 lung epithelial cells and hepatic cells.
Although not determined statistically, this might be related with the observed differences in higher maximum total bacterial contamination in filters (1.00 × 10 4 CFU·m −2 ) compared to EDC (8.49 × 10 2 CFU·m −2 ). Gram-negative bacteria contamination and fungal contamination (MEA and DG18) were similar among filters and EDC, thus, probably not responsible for the observed differences in cytotoxicity. Fungal diversity, however, might explain differences in cytotoxicity. For example, Aspergillus sp. is highly cytotoxic [34], with differences among Aspergillus sections [35][36][37]. In this study, augmented Aspergillus sp. dis-tribution in filters (compared to EDC) consisted of sections Nidulantes (only in filters, MEA), Circumdati (higher contamination in filters, DG18) and Aspergilli (only in filters, DG18). Of these, Nidulantes is among the most pathogenic to humans Aspergillus species, being mycotoxigenic (able to produce sterigmatocystin, penicillin, cotanin, and nidulotoxin) [38]. Most fungal metabolites, such as mycotoxins, are cytotoxic to different cellular structures. The best described mycotoxins produced by Aspergillus species, aflatoxins and ochratoxins, can act towards target cells, cellular structures, and their internal processes [39].
The IC50 levels were 0.016 g/mL in one EDC only, both in A549 cells and HepG2 cells. Regarding filters, IC50 levels ranged from 0.25 to 1 cm 2 /mL in A549 cells, and from 0.125 to 1 cm 2 /mL in HepG2 cells. Previous studies reported a cytotoxicity of Aspergillus metabolites in A549 cells as ranging from 44 to 61 µM [40]. Moreover, in other attributes such as particle size, MVOCs (not assessed in this study) may have cytotoxic effects [41].

Conclusions
Overall, this study provided a useful contribute to unveil the microbial contamination in cemeteries and estimate workers exposure to the microbial contamination characterized, since the sampling locations were defined based on the tasks developed by the workers and places where they spent more time. This is a very specific context with no available data published on microbiological contamination and exposure of the workers involved in the activities developed. Additionally, the data produce allowed to characterize exposure and identify some measures to prevent exposure. It was also possible to reinforce the positive features of the sampling approaches followed (combining active with the passive sampling methods), as well as the assays applied (use of different culture media).
Future studies are needed in this occupational setting regarding mycotoxins and cytotoxicity measurement to identify potential fungal producers and or triggers for mycotoxin production allowing to implement more focused risk management measures.

Graveyards Assessed
This study sampling campaign was conducted between May and June 2021 in three cemeteries located in Lisbon (Figure 3). There were 41 workers in Cemetery 1 (G1), 22 in Cemetery 2 (G2), and 13 in Cemetery 3 (G3), and all worked in regular 8-h shifts. In G1 there were 1 cemetery coordinator, 1 foreman, 7 administrative workers, 18 gravediggers, 8 gravediggers from the crematorium, 4 drivers of heavy machinery and special vehicles, 1 paver, and 1 locksmith. In G2 there were 1 cemetery coordinator, 1 foreman, 5 administrative workers, 5 gravediggers, 8 gravediggers from the crematorium, and 2 auxiliary workers. In G3 there were 1 cemetery coordinator, 1 foreman, 3 administrative workers, 5 gravediggers, 2 auxiliary workers and 1 concierge, accounting for a total of 76 workers in all the assessed cemeteries. The sampling sites were the Administrative service, Repository room, Crematory, Bar, Kitchen, Canteen (tables and Self-Service area), male and female locker room, and drivers' and gravediggers' locker room.

Sampling Approach and Characterization through Culture Dependent-Methods
A multi-sampling approach protocol was performed by using active and passive sampling methods and through normal working days (Figure 2). An impinger device-Coriolis µ air sampler (Bertin Technologies, Montigny-le-Bretonneux, France)-was employed for SARS-CoV-2 detection. The passive sampling methods used in the implemented sampling campaign were settled dust, surfaces swabs and electrostatic dust cloths (EDC) (Figure 4).
all the assessed cemeteries. The sampling sites were the Administrative service, Repository room, Crematory, Bar, Kitchen, Canteen (tables and Self-Service area), male and female locker room, and drivers' and gravediggers' locker room.

Sampling Approach and Characterization through Culture Dependent-Methods
A multi-sampling approach protocol was performed by using active and passive sampling methods and through normal working days ( Figure 2). An impinger device-Coriolis μ air sampler (Bertin Technologies, Montigny-le-Bretonneux, France)-was employed for SARS-CoV-2 detection. The passive sampling methods used in the implemented sampling campaign were settled dust, surfaces swabs and electrostatic dust cloths (EDC) (Figure 4).  Settled dust was collected through a vacuum cleaner (HOOVER Brave BV71_BV10 A2, Solon, OH, USA) with 1 × 4 collector filter attached (also used for further analyses) and a composite sample of the settled dust filters was obtained by vacuuming all the identified sampling sites [8,42]. Floor surfaces from the same sampling sites were swabbed following the same procedures already reported  to allow microbial quantification. The EDC were placed above 1.5 m high on a shelf in each sampling site for 30 days. All Settled dust was collected through a vacuum cleaner (HOOVER Brave BV71_BV10 A2, Solon, OH, USA) with 1 × 4 collector filter attached (also used for further analyses) and a composite sample of the settled dust filters was obtained by vacuuming all the identified sampling sites [8,42]. Floor surfaces from the same sampling sites were swabbed following the same procedures already reported  to allow microbial quantification. The EDC were placed above 1.5 m high on a shelf in each sampling site for 30 days. All samples obtained (settled dust, floor surface swabs, filters, EDC) were preserved in sterilized bags or in transport tubes (swabs) and transported under refrigeration (0-4 • C) to the laboratory for further analyses [42].
Swabs obtained from swabbing the floor were washed with 1 mL of 0.1% Tween 80 saline (0.9% NaCl) for 30 min on the orbital shaker (250 rpm, 30 min). The same procedure was applied on a piece (2 cm 2 ) of each filter used from vacuuming. A composite sample with the settled dust was washed in a ratio of 1 g per 9.1 mL of 0.1% Tween 80 saline (0.9% NaCl) for 30 min at 250 rpm [42]. EDC were weighted and processed with 20 mL of the same washing solution.
After incubation at 27 • C for 5-7 days for fungi (MEA and DG18) and at 30 • C (TSA) and 35 • C (VRBA) for 7 days for mesophilic bacteria and coliforms (Gram negative bacteria), respectively, microbial quantification was achieved (colony-forming units, CFU·g −1 , CFU·m −2 , CFU·m −2 ·day −1 ) as previously published [8]. Fungal species were identified microscopically following instructions reported by De [44]. Negative controls were used to ensure the inexistence of background contamination, namely: culture media (all samples) and control samples extracts (vacuuming filters, swabs, EDC) without prior use were analyzed to the same assays.

Azole Resistance Screening
The extracts from the passive sampling methods from cemeteries were screened for azole resistance using an adaptation of the EUCAST method and breakpoints for Aspergillus fumigatus [45,46]. For that purpose, Sabouraud dextrose agar (SDA) (Frilabo, Maia, Portugal) was used either alone (as control) or supplemented with the following medical azoles: 4 µg/mL itraconazole (ITZ), 2 µg/mL voriconazole (VCZ), and 0.5 µg/mL posaconazole (PSZ). The washed extracts of the collected passive samples (prepared as described in Section 2.2) were inoculated in the SDA supplemented media, and the media plates were incubated at 27 • C, to allow the growth of all fungi present in sample. As negative control, the reference strain A. fumigatus ATCC 204305 (provided by National Health Institute Doutor Ricardo Jorge) was used, and a pan-azole-resistant A. fumigatus (also provided by National Health Institute Doutor Ricardo Jorge, IP) was used as positive control. After three of incubation, fungal colonies were counted and prepared for microscopically identification [47].

Sampling and Molecular Detection of SARS-CoV-2 and Targeted Aspergillus Sections
Concerning SARS-CoV-2, composite surface samples were obtained by swabbing defined areas from each sampling site in each cemetery (Administrative service, Repository room, Crematory, Bar, Kitchen, Canteen (Tables and Self-Service area), Male and female locker room, drivers' and gravediggers' locker room) (Figure 2), using sterile cotton swabs moistened in Buffer NVL (specific for SARS-COV-2 assessment) and following the same steps applied on swabs applied to assess other microorganisms (fungi and bacteria).
Air samples of 600 L were obtained in all sampling locations employing the impinger Coriolis µ air sampler (Bertin Technologies, Montigny-le-Bretonneux, France) with a flow rate of 300 L/min collected into a conical vial containing 5 mL Buffer NVL (NZY Viral RNA Isolation kit (MB40701) component).
Samples were kept refrigerated (until 4 • C) for a maximum of 24 h before RNA was extracted from the obtained sample (1.5 mL in surface samples and 5 mL in air samples) with the NZY Viral RNA Isolation kit, from Nzytech, following to manufacturer's instructions. One step-RT qPCR was performed using NZYSpeedy One-step RT-qPCR probe Master Mix and specific procedures were followed (Supplementary Material) with primers and probes published by CDC (available on: https://www.cdc.gov/coronavirus/2019-ncov/ lab/rt-pcr-panel-primer-probes.html, accessed on 23 September 2020), (Supplementary material- Table S2). qPCR was performed on BioRad CFX96 PCR machine. To detect possible PCR inhibitors an internal control was added to each PCR.

Mycotoxins Analysis
Thirty-nine samples were screened for mycotoxins presence: in 18 samples collected from the filters, in 18 EDC and in 3 settled dust samples as previous reported [8]. Thirtyeight mycotoxins were analyzed by HPL-MS (HPLC) Nexera (Shimadzu, Tokyo, Japan) with a mass spectrometry detector API 4000 (Sciex, Foster City, CA, USA). The Limits of Detection (LOD) obtained for each mycotoxin with the analytical method used are presented in Table S4 (Supplementary Materials).

Cytotoxicity Analyses
The cytotoxic effect of filter and EDC samples was evaluated on samples' extracts using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay at 510 nm, as previously described [48]. The cells used to assess the cytotoxicity were the human lung epithelial (A549) and the HepG2 cells. Briefly, cells were maintained in Eagle's Minimum Essential Medium (MEM) supplemented with 10,000 units penicillin and 10 mg/mL streptomycin in 0.9% NaCl and foetal bovine serum (Sigma-Aldrich, Burlington, MA, USA), and 0.25% (w/v) Trypsin 0.53 mM EDTA was used for cell detachment. After cell counting (Scepter™ 2.0 Cell Counter, Merck, Kenilworth, NJ, USA), 100 µL cell suspension (densities of 2.5 × 10 5 cells/mL) was transferred to a 96-well plate and exposed to diluted samples for 48 h at 5% CO 2 , 37 • C, and humid atmosphere. The cytotoxicity was measured (ELISA LEDETECT 96, biomed Dr. Wieser GmbH; MikroWin 2013SC software), and the lowest sample concentration dropping absorption to <50% of cell metabolic activity (IC50) was considered the threshold toxicity level.

Statistical Analyses
Data obtained were analyzed using SPSS statistical software, version 26.0 for Windows. The results were considered significant at the 5% significance level. To test the normality of the data, the Shapiro-Wilk test was used. To study the relationship between fungal counts (MEA and DG18) and fungal counts on azole resistance screening (SDA ITZ, VCZ, and PSZ) in each sampling method, Spearman correlation coefficient was applied, since the assumption of normality was not observed. To compare the fungal contamination and azole resistance screening between the sampling methods, the Kruskal-Wallis test was applied, since the normality assumption was not observed. To assess species diversity, Simpson and Shannon indices, given by Shannon Index (H) = − s ∑ i=1 p i ln(p i ) and , were used, where p i is the proportion (n i /n) of individuals of one particular species found (n i ) divided by the total number of individuals found (n).

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/toxins14050348/s1, Figure S1-The prevalent fungal genera on MEA and DG18 in samples from different matrices; Table S1-Assessment of species diversity (Shannon's and Simpson's indices) in each sampling method; Table S2-Novel Coronavirus (2019-nCoV) Real-time RT-PCR Panel Primers and Probes.; Table S3-Sequence of primers and TaqMan probes used for real-time PCR; Table S4-LOD values for the samples analysed.