Microbiological Contamination Assessment in Higher Education Institutes

: The higher education sector represents a unique environment and it acts as a work environment, a learning environment for students, and frequently, also a home environment. The aim of this study was to determine the microbial contamination ( SARS-CoV-2 , fungi, and bacteria) in Higher Education Facilities (HEI) by using active and passive sampling methods and combining culture-based methods with molecular tools targeting Aspergillus section Fumigati . In addition, the resistance to azole proﬁle was also assessed. Surface samples showed a range of total bacterial contamination between 1 × 10 3 to 3.1 × 10 6 CFU · m − 2 , while Gram-negative bacteria ranged from 0 to 1.9 × 10 4 CFU · m − 2 . Fungal contamination ranged from 2 × 10 3 to 1.8 × 10 5 CFU · m − 2 on MEA, and from 5 × 10 3 to 1.7 × 10 5 CFU · m − 2 on DG18. The most prevalent species found on both media was Cladosporium sp. (47.36% MEA; 32.33% DG18). Aspergillus genera was observed on MEA (3.21%) and DG18 (14.66%), but not in the supplemented media used for the azole screening. Aspergillus section Fumigati was detected in 2 air samples (2.22%, 2 out of 90 samples) by qPCR. When testing for SARS-CoV-2 all results were negative. The present study showed that although cleaning and disinfection procedures are done regularly due to the COVID-19 pandemic, being effective in eliminating SARS-CoV-2 , surfaces were often contaminated with microorganisms other than SARS-CoV-2 . This can be a result of increasing resistance to biocides, and to the wide range of environmental factors that can contribute to the dissemination of microbial contamination indoors.


Introduction
In the last few decades, several studies were conducted to understand the impact of indoor air quality (IAQ) in public health in different environments, including residential building, shopping malls, schools, health care centers, offices, museums, libraries, temples, and churches, among others [1][2][3][4][5][6][7]. It was indicated that decreased IAQ can negatively affect human health as most people spend around 90% of their time indoors, mainly at home or in the workplace [8]. Due to this, IAQ has emerged and received increasing attention from international scientific community, political institutions, and environmental governances [2,8,9]. The indoor air pollution (IAP), that refers to the existence of pollutants, The ten facilities under study are located in the Lisbon district and all presented different core activities depending of their purpose and graduation courses held ( Table 1).
The sampling sites were chosen based on previous selection by areas of facility by the Occupational Health Services, in the scope of SARS-CoV-2 surveillance held during the 2nd pandemic wave in Portugal and before starting the 2nd semester of the academic year 2020/2021. A walkthrough survey and checklist were applied in order to prioritize the most critical workplaces/areas in relation to SARS-CoV-2 contamination. In summary, sampling sites of each facility were selected according to the following criteria: 50-80% workplaces/facilities occupation, activities performed that can lead to higher number of workers per workplace and/or without a mask (cantine) or based on workers positive serologic surveillance results (positive result for IgM+ or IgG+) [37]. All the facilities implemented a contingency plan that included working, whenever possible, in home office even concerning theoretical lectures, wearing a mask indoor and outdoor (when 2 meters distance was not possible to ensure) and workers should remain at home (quarantine) when COVID-19 symptoms arise until further diagnose.
Environmental samples (air and surface samples) were performed in each area and 6 to 25 samples were collected at each location ( Table 1). Most of the sampling sites were common among all the facilities, such as offices (including human and financial resources, academic services, logistics, accounting, and acquisition department), attendance room, reception rooms, auditoriums, meals spaces, bathrooms and libraries. Although there are classrooms in almost every location, the type of classroom analyzed varied widely in all locations, including rooms for music, dance, choir, theatre and multimedia, laboratories, and gyms. In addition, in some locations, there were samples collected from student's social rooms, workshops, changing rooms and professors' room (Table 1). The samples were collected mainly in the morning and during normal activities, except for HEI 1, where the samples were collected in two days, one of them in the afternoon. In all facilities, the cleaning method applied was based on cleaning and disinfection recurring to bactericide and virucide, bleach and multipurpose detergent. Most of the sampling sites (56.8%) registered between 1 to 9 workers to follow the contingency plans. However, the accurate number of workers was not possible to obtain due to workers quarantine in the same day of the assessments to comply with contingency plans.

Samples Collected and Assays Performed
Air samples of 600 L were collected using 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) ( Figure 1). Surface samples were collected by swabbing the areas of each sampling site, using sterile cotton swabs moistened in Buffer NVL (SARS-CoV-2 assessment) or sterilized water (fungi and bacteria assessment). A 10 cm × 10 cm square stencil, disinfected between samplings with a 70% alcohol solution was used (ISO 18593: 2004) to allow quantification. On some surfaces with common characteristics, such as surfaces material and cleaning procedures, composite samples were performed (swabbing different surfaces with the same swab) [38] (Figure 1).
Culture based methods were applied only in surface samples. Every swab was later extracted with 1 ml of 0.1% Tween™ 80 saline solution (NaCl 0.9%) for 30 min at 250 rpm on an orbital laboratory shaker (Edmund Bühler SM-30, Hechingen, Germany) and plated onto the selected media. Four different culture media were used in order to enhance the selectivity for bacterial and fungal growth, as follows: 2% malt extract agar (MEA) with 0.05 g L−1 chloramphenicol media, and dichloran glycerol (DG18) agar based media, for fungal characterization; Tryptic Soy Agar (TSA) with 0.2% nystatin, for total bacteria assessment; and Violet Red Bile Agar (VRBA), for Gram-negative bacteria.
All the inoculated plates were incubated at 27 • C for five days for fungal growth (four days regarding azole resistance screening) or for seven days at 30 • C and 37 • C for bacterial growth and for Gram-negative bacterial growth, respectively. After the incubation period, quantitative (colony-forming units-CFU·m −2 ) results for fungi and bacteria were obtained. When colony overgrowth was observed due to fungi with fast growing rates (Mucorales, Chrysonilia sitophila and Trichoderma sp.), making it impossible to count colonies, the median of all colony values obtained in all locations of the same facility was assumed. Fungal species were also identified microscopically using lactophenol cotton blue mount procedures. Morphological identification was achieved through macro and microscopic characteristics [40].
Atmosphere 2021, 12, x FOR PEER REVIEW 7 of 20 Figure 1. Sampling strategy adopted and assays applied. * Lack of extracts quantities in two samples to perform the assay.
All the inoculated plates were incubated at 27 °C for five days for fungal growth (four days regarding azole resistance screening) or for seven days at 30 °C and 37 °C for bacterial growth and for Gram-negative bacterial growth, respectively. After the incubation period, quantitative (colony-forming units-CFU·m −2 ) results for fungi and bacteria were obtained. When colony overgrowth was observed due to fungi with fast growing rates (Mucorales, Chrysonilia sitophila and Trichoderma sp.), making it impossible to count colonies, the median of all colony values obtained in all locations of the same facility was assumed. Fungal species were also identified microscopically using lactophenol cotton Air samples obtained by impinger were also analyzed by molecular detection of Aspergillus section Fumigati. Procedures and reactions were performed as previously reported [41].
Concerning SARS-CoV-2 assessment on air and surfaces, the procedures used for sample inactivation, RNA extraction and detection were as previously submitted [37]. In each analysis, a positive (a SARS-CoV-2 positive sample) and a negative (water) sample were included. Moreover, in order to detect possible PCR inhibitors, an internal control was added to each PCR reaction (TATAA Universal RNA Spike I).

Statistical Analysis
Data were analyzed using SPSS statistical software, V26.0 for windows. The results were considered significant at the 5% significance level. To test the normality of the data, the Kolmogorov-Smirnov test was used. To characterize the sample, frequency analysis (n, %) was used for qualitative data. To study the relationship between bacterial and fungal counts and resistance to azoles and Aspergillus section, Spearman's correlation coefficient was used, since the assumption of normality was not verified.

Viable Microbial Contamination
Surface samples showed a range of total bacterial contamination between 1 × 10 3 (classroom and attendance room) and 3.1 × 10 6 CFU·m −2 (offices). The highest median value (or total values in the case of classroom, attendance room, computer room, changing and dressing room, gym, study room and multimedia) was 5 × 10 5 CFU·m −2 (study room) and the lowest was 1 × 10 3 CFU·m −2 (classroom and attendance room). Gram-negative bacteria in surface samples ranged from 0 to 1.9 × 10 4 CFU·m −2 (laboratory) with a highest median value of 9.5 × 10 3 CFU·m −2 in the laboratory (Figures 2 and 3).
sample inactivation, RNA extraction and detection were as previously submitted [37]. In each analysis, a positive (a SARS-CoV-2 positive sample) and a negative (water) sample were included. Moreover, in order to detect possible PCR inhibitors, an internal control was added to each PCR reaction (TATAA Universal RNA Spike I).

Statistical Analysis
Data were analyzed using SPSS statistical software, V26.0 for windows. The results were considered significant at the 5% significance level. To test the normality of the data, the Kolmogorov-Smirnov test was used. To characterize the sample, frequency analysis (n, %) was used for qualitative data. To study the relationship between bacterial and fungal counts and resistance to azoles and Aspergillus section, Spearman's correlation coefficient was used, since the assumption of normality was not verified.

Azole Resistance Screening
Positive fungal growth on at least one azole supplemented media was observed for 10 fungal species. Cladosporium sp. was the most abundant species in all three azoles (64.71% on ITR; 65.49% on VOR; 53.33% on POS). Penicillium sp. was also found in all three azoles, being the second most prevalent species in two of three azoles (14.79% on VOR; 22.22% on POS). Chrysosporium sp., Chrysonilia sitophila and Mucor sp. were also presented in all three azoles. Alternaria sp. and Aureobasidium sp. were only observed on ITR and VOR. Acremonium sp. and Rhizopus sp. were only detected on VOR and POS, respectively (Table 3).

Azole Resistance Screening
Positive fungal growth on at least one azole supplemented media was observed for 10 fungal species. Cladosporium sp. was the most abundant species in all three azoles (64.71% on ITR; 65.49% on VOR; 53.33% on POS). Penicillium sp. was also found in all three azoles, being the second most prevalent species in two of three azoles (14.79% on VOR; 22.22% on POS). Chrysosporium sp., Chrysonilia sitophila and Mucor sp. were also presented in all three azoles. Alternaria sp. and Aureobasidium sp. were only observed on ITR and VOR. Acremonium sp. and Rhizopus sp. were only detected on VOR and POS, respectively (Table 3). Table 3. Fungal distribution in azole-supplemented SAB media from surface swab samples.

Correlation Analysis
Regarding bacterial counts in TSA, significant correlations were detected with counts in VRBA (rS = 0.252, p = 0.015), in SAB (rS = 0.354, p = 0.001), in VOR (rS = 0.235, p = 0.033) and in POS (rS = 0.343, p = 0.001) and with the number of workers (rS = 0.287, p = 0.009). These results indicate that higher bacterial counts in TSA are related to higher bacterial counts in VRBA, higher azole resistance counts (either in SAB, VOR or POS) and higher number of workers (Table 5). Considering bacterial counts in VRBA, only significant correlation was found with fungal counts in DG18 (rS = 0.235, p = 0.024), revealing that higher bacterial counts in VRBA are related to higher fungal counts in DG18 (Table 5).
With regard to fungal counts in MEA, significant correlations were detected with fungal counts in DG18 (rS = 0.586, p = 0.000), in SAB (rS = 0.494, p = 0.000), in ITR (rS = 0.362, p = 0.001) and in VOR (rS = 0.485, p = 0.000), with Aspergillus sp. counts, fungi in MEA (rS = 0.265, p = 0.011) and with number of workers (rS = 0.226 p = 0.043). These results reveal that higher fungal counts in MEA are related to higher fungal counts in DG18, higher azole resistance (either in SAB, ITR or VOR), higher values of Aspergillus sp. counts in MEA and higher number of workers (Table 5).
Regarding azole resistance in SAB, significant correlations were detected with azole resistance in ITR (rS = 0.478, p = 0.000), in VOR (rS = 0.638, p = 0.000) and in POS (rS= 0.289, p = 0.008) and with Aspergillus sp. counts-azole resistance in SAB (rS = 0.388, p = 0.000), showing that higher counts in SAB are related to greater resistance to azoles in other media, including in the Aspergillus sp. counts (Table 5).
Concerning azole resistance in ITR, significant correlations were detected with azole resistance in VOR (rS = 0.472, p = 0.000) and in POS (rS = 0.360, p = 0.001), revealing that higher azole resistance in ITR is related with higher azole resistance in VOR and POS (Table 5).
With regard to azole resistance in VOR, significant correlations were detected with azole resistance in POS (rS = 0.308, p = 0.005) and with Aspergillus sp. counts in MEA (rS= 0.243, p = 0.027) and Aspergillus sp. counts in SAB (rS = 0.375, p = 0.000), revealing that higher azole resistance in VOR is related with higher azole resistance in POS, higher values of Aspergillus sp. counts in MEA and higher and Aspergillus sp. counts in SAB (Table 5).
Regarding Aspergillus section, the following significant correlations were found: (i) fungi in MEA and azole resistance in VOR (rS = 0.360, p = 0.001), which indicates that higher values in MEA are related to higher resistance to azoles in VOR; (ii) higher counts in SAB and VOR (rS = 0.294, p = 0.007), which reveals that higher counts in SAB is related to greater azole resistance in VOR (Table 5).

Discussion
The IAQ in HEI is of great importance due to the impact it has on the health and performance of students, professors and staff [22][23][24]. Furthermore, microbiological contamination assessment is one of the main parameters that affect IAQ, since potentially pathogenic microorganisms can be disseminated as bioaerosols and via contact with contaminated surfaces [8,9,42] or through resuspension from air to surfaces [42].
It has already been previously reported that the presence of fungi and bacteria in indoor air is influenced by a wide range of factors, such as human occupancy and their activities, humidity levels, ventilation, environmental characteristics, water infiltrations, construction and decoration materials and outdoor air [11,43,44]. Due to the extensive list of factors that influence IAQ, exposure assessment to microorganisms remains a challenge to every exposure assessor/industrial hygienist [45]. In fact, as in other risk factors exposure assessment studies, the sampling approach is of critical importance to achieve an accurate risk characterization regarding microbiological agents [46]. Active methods, based on air sampling, rely within the most common methods used for samples collection. However, they may not represent the real scenario regarding the inhalation exposure, since they can only reflect the load from a short period of time (mostly minutes), corresponding to the sampling duration, thus representing only a small fraction of the microbiological contamination exposure [47][48][49]. Despite these concerns, the impingement method, also based on air sampling, has been the most used for the SARS-CoV-2 assessment in indoor environments [37], since it allows for longer active sampling times, thus ensuring collection of sufficient airborne viruses for detection by molecular tools [50].
In the case of passive methods, such as the surface swabs used in this study, they allow to characterize the contamination over a longer period of time (after the last cleaning procedure), thus providing a more comprehensive picture of the real exposure [46][47][48][49]. Indeed, they have previously been used in several indoor environments [46,[51][52][53][54][55][56][57][58][59][60][61] generally providing more detailed and complete information regarding fungal species distribution. The use of both sampling methods allows to overcome each method limitation, ensuring a more precise exposure assessment [44]. This is further reinforced with the use of culture based-methods and molecular tools. Indeed, although most of the studies performed in HEI are focused on air quality screenings [28,30,32,33,62,63], surface analyses have also been shown to be relevant, as they may also reflect the contamination in the air by resuspension depending of the activities developed indoors, thus possibly leading to increased levels in airborne concentration [57,64,65].
Previous studies have shown that, besides the sampling approach, culture media applied also influence the results obtained for fungi and bacteria detection in environments [46,56]. Regarding bacterial contamination, culture media allows for the discrimi-nation between total bacteria and Gram-negative bacteria [56]. As expected, in our study, the contamination of Gram-negative bacteria was lower than that of total bacteria, as expected, since the latter represents the number of Gram-positive bacteria and Gram-negative bacteria. The presence of bacteria on surfaces is a common situation, especially in the most frequently touched surfaces, as their main contamination sources are the occupants and their activities [66][67][68]. Thus, it was not surprising that the highest total bacterial counts were found in the offices (31.8%), followed by the auditorium (15.7%) and the meals space (14.4%), which are the places with higher occupancy. In fact, the positive correlation found between higher bacterial counts in TSA and higher number of workers emphasizes this contribution.
Although bacteria are ubiquitous and generally of human origin (from skin and mucous membranes) and not harmful for health, the presence of Gram-negative bacteria is a special concern, as they may have natural resistance to antibiotics and can also produce endotoxins, which can cause respiratory symptoms [11,43,57,68].
Regarding the fungal contamination assessment, besides the use of MEA, as suggested in the Portuguese guidance for IAQ assessment [11], DG18 was also selected to be used, since this media constitutes a better alternative for colony counting, also allowing to obtain higher diversity of genera [47,69]. Contrary to these expectations in this study we have obtained more diversification of fungal species in MEA (10 different species) than in DG18 (9 species) with the same trend being observed for Aspergillus sections (5 on MEA; 3 on DG18).
Interestingly, in the present study, the concentration of fungi on the surfaces was lower than that of bacteria, similarly to a study on surface swabs in university facilities [23]. The most prevalent fungi found in our study were Cladosporium sp., Penicillium sp. and Aspergillus sp., which is in accordance with other studies based on the use of surfaces swabs as sampling approach [64,[70][71][72].
While offices (28.5%), meals space (13.4%) and theatre and choir room (12.8%) were the areas most contaminated by fungi on MEA, on DG18 we detected higher fungal loads in professors' rooms (32.0%), offices (24.0%) and changing and dressing room (11.8%). A possible reason for the higher counts of fungi in these areas could be the fact that, due to the COVID-19 pandemic, all occupants were encouraged to open windows to prevent COVID-19 infection [20]. Indeed, evidence clearly indicated that opening windows increases the levels of fungi and other microorganisms in the air and on the surfaces as a result of the passage of outdoor air into indoors [33,42,67,73]. HEI 6 and HEI 2 have the highest counts of total fungal contamination on MEA and DG18 media, respectively. HEI 2 samples were analyzed in DG18 and are among the areas with higher contamination values, possessing a considerable concentration of fungi. However, in this building only professor's rooms were analyzed by this method. Fungal growth can be promoted in the presence of moisture, and many fungi grow easily on any surface that becomes wet or moistened, such as faucets, which are present, for example, in meals space and changing and dressing room [67,74,75].
The emergence of pathogenic fungi resistant to antifungal agents widely used in the treatment of fungal infections, which can cause therapeutic failure, has been notorious in recent years [80,81]. In this study, the screening of fungal resistance to three medical azoles was conducted. Cladosporium sp. was the fungal species mostly present in all three azoles (64.71% on ITR; 65.49% on VOR; 53.33% on POS), followed by Penicillium sp. Regarding Mucor sp. and Rhizopus sp., they are intrinsically resistant to voriconazole, with itraconazole and posaconazole as first-line therapy [44,82]. However, in this study, we have observed the growth of these fungal species in the presence of itraconazole and posaconazole thus indicating the need to further characterize fungal resistance of those isolates [44,82].
The identification of a fungal species in more than one azole suggests a multi-drug resistance phenotype that must be further evaluated through antifungal susceptibility testing [82] by the reference microdilution method (EUCAST) [83]. A confirmed resistance phenotype would indicate the presence of azole-resistant fungal species in these settings, thus constituting a higher exposure risk, especially for immunocompromised occupants [57]. One limitation to this characterization is that reference values are defined only for Aspergillus sp. and Candida sp. [84].
Emergent antifungal resistance in Aspergillus fumigatus is the main cause of invasive fungal infections [77,85]. In this study, Aspergillus section Nigri was identified in one azole media (0.70% on VOR), whereas Aspergillus sections Candidi, Fumigati, Nidulantes, Nigri, and Circumdati were identified in control Sabouraud, in MEA and/or DG18. These results are in line with previous data from dwellings and hospital environment, where no Aspergillus species were able to grow on azole-media, despite being observed in Sabouraud, MEA or DG18 [48,54,56,57]. Of note, cryptic Aspergillus species might be underestimated in azole-media due to the presence of fast-growing species, such as Chrysonilia sitophila and Mucorales group [44,56,57].
Our culture-based methods allowed the identification of the Aspergillus section Fumigati in a wide number of samples, with molecular tools also detecting this section in different and in a smaller number of samples. Despite this discrepancy, it is of relevance to use both methods, as they provide complimentary information and answer different questions. Indeed molecular tools allow precise, fast, specific and sensitive detection of microorganisms. Importantly, they also can identify dead or dormant microorganisms and can discriminate toxigenic strains from regular strains within some fungal species [86]. Although culture based-methods are selective, revealing only microorganisms able to grow on a particular growth media, therefore, underestimating the total number of microorganisms in samples, these methods are crucial since the viability of bioburden is of critical importance to estimate health risks, as it affects biological mechanisms, such as inflammatory and cytotoxic responses [82,83,87]. This reinforces the idea of combining both molecular and culture-based methods [44].
As previous suggested [88] the sampling approach to assess SARS-CoV-2 included passive and active sampling methods, swabs being the most common found in the literature [88] and with increased detection when compared with other sampling methods [88,89]. Although the sampling volume from the active sampling was the one recommended [90] and the detection technique was the one widely used for SARS-CoV-2 detection [88], all the results were negative indicating the efficacy of the present measures in place on the assessed facilities. Further studies, should include a different sampling approach by using glass-fiber and PTFE filters to be employed in low and high-volume air samplers and applying samples pretreatments allowing obtain an increased virus concentrations [91].

Conclusions
The present study showed that although the regular cleaning and disinfection procedures effectively removed SARS-CoV-2 from surfaces, these remained contaminated with other microorganisms besides SARS-CoV-2. This can be a result of an increased resistance to biocides, and of the wide range of environmental factors that can contribute to the dissemination of microbial contamination indoors.
Therefore, we recommended that corrective measures should be implemented to reduce bacterial and fungal presence in surfaces to avoid contamination in the air due to resuspension. Additional studies aiming at correlating air and surfaces microorganisms' burden can be a valuable tool in finding the contamination sources.