Available via license: CC BY 4.0
Content may be subject to copyright.
1
Vol.:(0123456789)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports
Indoor green wall aects
health‑associated commensal skin
microbiota and enhances immune
regulation: a randomized trial
among urban oce workers
L. Soininen1, M. I. Roslund 1,3, N. Nurminen2, R. Puhakka1, O. H. Laitinen2, H. Hyöty2,
A. Sinkkonen3* & ADELE research group*
Urbanization reduces microbiological abundance and diversity, which has been associated with
immune mediated diseases. Urban greening may be used as a prophylactic method to restore
microbiological diversity in cities and among urbanites. This study evaluated the impact of air‑
circulating green walls on bacterial abundance and diversity on human skin, and on immune responses
determined by blood cytokine measurements. Human subjects working in oces in two Finnish
cities (Lahti and Tampere) participated in a two‑week intervention, where green walls were installed
in the rooms of the experimental group. Control group worked without green walls. Skin and blood
samples were collected before (Day0), during (Day14) and two weeks after (Day28) the intervention.
The relative abundance of genus Lactobacillus and the Shannon diversity of phylum Proteobacteria
and class Gammaproteobacteria increased in the experimental group. Proteobacterial diversity
was connected to the lower proinammatory cytokine IL‑17A level among participants in Lahti. In
addition, the change in TGF‑β1 levels was opposite between the experimental and control group.
As skin Lactobacillus and the diversity of Proteobacteria and Gammaproteobacteria are considered
advantageous for skin health, air‑circulating green walls may induce benecial changes in a human
microbiome. The immunomodulatory potential of air‑circulating green walls deserves further research
attention.
Due to an increased hygiene level1, biodiversity loss and irregular soil contacts2–4 the exposure to environmen-
tal microbes has reduced in Western cities, which is seen as one of the major reasons for the rise in immune-
mediated diseases, such as autoimmune diseases and allergies5. Nature-derived microbes that have a commensal
relationship with humans contribute to the development and regulation of the human immune system1,4,6–8. e
skin microbiome can be altered via skin contact to microbial sources and hands are a common route of micro-
bial transmission9–13. Indoors, humans aect and are exposed to microbial communities by touching indoor
surfaces14. In addition to direct skin contact, humans are exposed to and aected by microbes in the air, for
example, via skin and airways15–17.
Each city has its own unique microbiome18,19 and its composition in the soil20 and in the air21 is aected by
vegetation. Indeed, plant surfaces are a known source of airborne bacteria17,22,23. Additionally, indoor plants
increase the abundance and diversity in bacterial communities on indoor surfaces24. e amount of vegetated area
in the locality aects the odds of developing immune-mediated diseases as a child25–27. Vegetation also aects the
composition of the human gut microbiome, which impacts human immunoregulation28. Previous research has
identied certain bacterial groups that are abundant in soil and vegetation as indicators of a healthy skin microbi-
ome. For example, Proteobacteria belongs to the most common phyla on human skin (relative abundance > 5%);
OPEN
1Ecosystems and Environment Research Programme, Faculty of Biological and Environmental Sciences, University
of Helsinki, Niemenkatu 73, 15140 Lahti, Finland. 2Faculty of Medicine and Health Technology, Tampere University,
Arvo Ylpön katu 34, 33520 Tampere, Finland. 3Natural Resources Institute Finland, Horticulture Technologies,
Turku and Helsinki, Finland. *A list of authors and their aliations appears at the end of the paper. *email:
aki.sinkkonen@luke.
2
Vol:.(1234567890)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
its diversity and relative abundance seems to have a role in human immune regulation4,12,25. e diversity and
relative abundance in bacteria belonging to the class Gammaproteobacteria is an indication of health on human
skin4,7 and on plants29. Additionally, bacteria belonging to the genus Lactobacillus on skin fend o pathogens and
better the integrity of skin30; some Lactobacillus species found on humans also occur on plants31,32.
Due to poor microbiological assemblages in cities, urbanization reduces indoor microbial diversity33–35.
Microbial assemblages can be aected with urban gardening36 and urban greening12,37,38. In previous studies,
the eects of soil and plant-based biodiversity interventions have been observed in human subjects as bacterial
changes in skin and stool microbiome, and altered immunoregulatory cytokine levels in the blood9,11,12. Surpris-
ingly, hardly any studies survey whether indoor greening shapes commensal microbiota and immune response
among urban dwellers24.
e current study explored if bacterial communities in the human subjects spending time indoors can be
altered via vegetated walls that circulate indoor air. For the intervention, vegetated walls (green walls) were
brought into oces of university personnel for two weeks and the impact was investigated via skin and blood
samples. e study subjects were expected to be exposed to the green walls via microbial communities in the
air and on indoor surfaces but not by touching the green walls. e microbial focus was on the bacterial alpha
diversity and the relative abundance of health-associated proteobacterial taxa and Lactobacillus on skin. To
observe possible immune responses, the levels of the anti-inammatory cytokines interleukin 10 (IL-10) and
transforming growth factor– β1 (TGF-β1)39,40, and the proinammatory cytokine interleukin 17A (IL-17A)41
were measured from the blood samples. Immunomodulatory pathways respond to IL-10 concentration in the
blood and IL-10 has been researched for therapeutic use in immunomodulation9,12 and prevention of immune-
mediated diseases, such as inammatory bowel disease and rheumatoid arthritis40. Similarly, TGF-β1 is connected
to several immune-mediated diseases as an inhibitor and has an essential impact on all types of immune cells42–44.
e upregulation of cytokines in the IL-17 family in turn, seem to advance the pathogenesis of immune-mediated
diseases41,45. We hypothesized that the intervention would increase the relative abundance and alpha diversity of
health-associated taxa on the skin and aect the levels of the measured immune system cytokines.
Materials and methods
Green walls. e green walls (size 2m × 1m × 0.3m) used in this study were Naava One (Naava, Jyväskylä,
Finland; www. naava. io) that circulate indoor air. ey rst absorb the indoor air through the plant roots and
soilless substrate, then automated fans circulate the air back to the room. When the indoor air passes through
the green wall, volatile organic compounds (VOC) are eciently removed via bioltration by microbes, plants
and the growing medium46. e watering system is automated and the water circulates within the wall. Each
green wall contains three plant taxa (heartleaf philodendron (Philodendron scandens), dragon tree (Dracaena
sp.) and bird’s nest fern (Asplenium antiquum) growing altogether in 63 units. Each unit consists of two to four
plant individuals.
Treatment groups and sample collection. e study (a randomized controlled trial with parallel
design) was conducted in oces of university personnel in two Finnish cities (Lahti and Tampere). e study
followed the recommendations of Finnish Advisory Board on Research Integrity, and it was approved by the
ethics committee of the local hospital district (Hospital District of Pirkanmaa, Finland). A written informed
consent in accordance with the Declaration of Helsinki was signed by all participants. e study subjects were
healthy adults. e exclusion criteria were age below 18 at the beginning of the study, daily smoking, immune
deciency (e.g., antibody deciency, HIV infection), immunosuppressive medication (e.g., corticosteroids), a
condition aecting immune response (e.g., rheumatoid arthritis, colitis ulcerosa, Crohn’s disease, diabetes, and
Down syndrome), or cancer diagnosis. All volunteers that lled the inclusion criteria were accepted to the study.
e resulting 28 study subjects were randomly divided (intended allocation ratio 1:1; simple randomization
done by an independent researcher at University of Helsinki; mechanism: random number table) into two treat-
ment groups that were the control group (without green wall exposure) and the experimental group exposed to
green walls (Table1). Aer the randomization, it was ensured that age and sex ratio were similar in both groups,
and no changes were needed. e nal allocation ratio was 17:11 in the control and the experimental group. e
study subjects in the green wall group received a green wall in the oce rooms and were exposed to the green
walls only at the oce during their workdays. e study was implemented in two buildings in Lahti and one
building in Tampere, Finland (Table1). All study subjects answered surveys about their living conditions and
history (such as type of housing, pets and land use type in their locality) and their living habits during the experi-
ment on Day14 and Day28 (such as hours spent in nature, travel, medication, illnesses and food supplements).
Depending on the oce room size, 1–2 green walls were installed in the treatment oce rooms in Tampere and
Lahti for two weeks, according to instructions of the manufacturer (www. naava. io). When the room size was
more than 60 m2, two green walls were used as instructed by the manufacturer.
Skin and blood samples were collected from both experimental and control group participants before install-
ing the green walls (Day0), on the last day of the intervention (Day14) and two weeks aer the intervention
(Day28) by trained nurses as described by Roslund etal.12. Briey, skin samples were collected by swabbing an
area of 5cm–by–5cm on the back of the palm for 10s. e swabs were wetted with saline buer (0.1% Tween
20 in 0.15M NaCl) before sample collection, and aer sampling the cotton tips were cut o into sterile poly-
ethene tubes and stored at − 80°C until analysis. Venous blood was collected into Vacutainer CPT Mononuclear
Cell Preparation tubes containing sodium citrate (BD Biosciences, NJ, USA) and centrifuged according to the
manufacturer’s instructions to separate the plasma and the plasma samples were stored at − 80°C until analysis.
3
Vol.:(0123456789)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
Skin and blood sample processing. e skin samples were prepared for bacterial DNA sequencing as
in Roslund etal.47. e bacterial DNA was extracted from the skin swabs with Fast DNA spin kit for soil (MP
biomedicals, Santa Ana, CA) according to the manufacturer’s protocol. e DNA concentration was quanti-
ed by Quant-iTTM PicoGreen® dsDNA reagent kit (ermo Fisher Scientic, Waltham, MA, USA). e DNA
concentration in the samples was adjusted to 0.4ng/ml before polymerase chain reaction (PCR) with which
variable region V3-V4 within the 16S ribosomal RNA (rRNA) gene was amplied. Forward primer was 515F 50-
GTG CCA GCMGCC GCG GTAA-30 and reverse primer 806R 50- GGA CTA CHVGGG TWT CTAAT-30 with
truncated Illumina overhangs as in Hui etal.37. Negative controls for DNA extraction (sterile water) and PCR
(no sample) were sequenced with the samples. Positive control for PCR was made using (Cupriavidus necator
JMP134, DSM 4058). Success of amplication process was conrmed with agarose gel (1.5%) electrophoresis.
e primers were cleaned from the PCR products with Agencourt AMPure XP solution (Beckman Coulter Inc.,
Brea, CA, USA). e samples were sequenced with Illumina MiSeq 16S rRNA gene metabarcoding with read
length 2 × 300 base pairs using a V3-V4 reagent kit at the Institute for Molecular Medicine Finland (FIMM,
Helsinki, Finland).
e concentration of cytokines IL-17A and IL-10 were measured from the plasma samples using Milliplex
MAP high sensitivity T cell panel kit (Merck KGaA, Darmstadt, Germany) with Bio-Plex® 200 system (Bio-Rad
Laboratories, Hercules, CA, USA) and Bio-Plex Manager soware (version 4.1, Bio-Rad Laboratories, Hercules,
CA, USA). TGF-β1 concentration was analyzed using ELISA (BioVendor, Czech Republic).
Bioinformatics. From the skin samples’ sequence data, the bacterial OTUs were identied to the genus level
according to studies by Schloss etal.48 and Kozich etal.49 as in Soininen etal.50. Briey, using Mothur (version
1.44.1) the sequences were aligned with SILVA (version 138)51 as a reference. e sequences were preclustered
to avoid sequencing errors52. Chimeras were searched by UCHIME53 and deleted. e sequences were classied
using Bayesian classier54 with SILVA (version 138)51 with 80% bootstrap threshold. Non-bacterial sequences
were deleted. Unique sequences were clustered to OTUs at 97% sequence similarity. OTUs with 10 sequences or
less were removed. Good’s coverage index (average ± SD: 0.98 ± 0.01) and alpha diversity indices were calculated
for each sample using summary.single command. ese calculations and the subsampling of the samples were
done according to the smallest sequence count (3893) in the samples. Contaminant OTUs were removed as
in Roslund etal.13. Abundant bacterial taxa (relative abundance of > 0.01%) were selected for further analyses.
Alpha diversity indices for phylum Proteobacteria, class Gammaproteobacteria and genus Lactobacillus were
calculated from the subsampled data using R version 3.6.155 function diversity of package vegan56.
Outcome measures and sample size estimation. Primary outcome measure was Alpha diversity of
skin Gammaproteobacteria, since it was associated with environmental biodiversity, andTGF-β in aprevious
study12. Gammaproteobacteria Shannon diversity index was measured at baseline and aer 28-day intervention.
Secondary outcome measure was relative abundance of skin Lactobacillus and cytokine levels measured from
plasma. All the secondary outcome measures were analyzed from baseline to end of intervention. No side eects
were observed.
e primary outcome measure for the power calculation was the dierence between intervention and control
study subjects in the change of Gammaproteobacterial diversity on the skin between baseline and day 28. We
used prior eect estimates from the study that estimates correlations between environmental biodiversity, human
microbiota and immune function12. In this study, the alpha diversity of Gammaproteobacteria was higher among
study subjects in the intervention arm in more biodiverse environment, and Gammaproteabacterial abundance
on skin was associated with TGF-β expression. Generic diversity of Gammaproteobacteria among study subjects
in contact with green materials (intervention arm) had an average of 17 Gammaproteobacterial genera in their
hands and a standard deviation of 5, whereas in the hands of study subjects in the urban control arm they had
an average of 8 and a standard deviation of 5. When the signicance level is set to P ≤ 0.05 and the statistical
Table 1. Characteristics of treatment groups: Ctrl = control, Exp = experimental.
Groups Ctrl Exp
Participants 17 11
Sex Female 13 9
Male 4 2
Age
25–35 3 4
36–45 7 2
36–45 5 5
Average 40 40
SD 9 10
Type of residence
Apartment building 4 4
Rowhouse 2 2
Private house 10 5
Work place Lahti 10 7
Tampere 4 7
4
Vol:.(1234567890)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
force is 0.8 (80.1%), the between-cluster (between cities) coecient of variation is 0.2, the required sample size
for each group is 14.
Statistics. Bacterial diversity and relative abundance (dependent variables) of selected taxa were tested sta-
tistically in contrast to timepoint and treatment (explanatory variables) using linear mixed models (LMMs)
(function lmer in lme4 package in R) with study subject (nested in cities) as the random factor. LMMs are a
good t for analyzing clustered data and by using study subject as the random factor, the fact that one person
is the source for several samples, can be taken into account in the statistical evaluation57,58. Additionally, the
amount of change (between timepoints Day0–Day14 and Day0–Day28)59 in diversity and relative abundance
were calculated and compared using LMMs as in Roslund etal.12. Additionally, the treatments were compared
on each timepoint separately using t-test or Wilcoxon test depending on the Shapiro–Wilk distribution of the
variable. e cytokine levels and their changes (dependent variables) were tested in contrast to bacterial values,
the interaction of timepoint and treatment (explanatory variable) using LMMs with study subject (nested in cit-
ies) as the random factor. e background information and living habits were compared between the treatments
using Chi-Square test for nominal data and t-test or Wilcoxon’s test for quantitative data.
Results
e relative abundance of Lactobacillus spp. (Fig.1 and Supplementary Table1) was higher in the skin samples
of the experimental group than the control group during the treatments, on Day14 (Wilcoxon P = 0.0058).
Additionally, the change (Day14 – Day0) in the relative abundance of Lactobacillus spp. was higher in the experi-
mental group than in the control group. Within the experimental group, the relative abundance of Lactobacillus
spp. increased in six study subjects and decreased in three study subjects. Within the control group, the relative
abundance decreased in 13 study subjects and increased in three study subjects. Importantly, random variation
between individuals explains total variation only partially (LMM All: P < 0.001, R2 = 0.05, R2 random = 0.21).
e signicance of the model did not depend on the city (LMM Lahti P < 0.001, R2 = 0.05, R2random = 0.31;
LMM Tampere P < 0.001, R2 = 0.25, R2random = 0.06).
ere were subtle dierences in the Shannon diversity of Gammaproteobacteria (Fig.2A and Supplemen-
tary Table1) and Proteobacteria (Fig.2B and Supplementary Table1). e change in Shannon diversity (Day28
– Day0) diered between treatment groups in Proteobacteria, plausibly due to high Day28 values in the experi-
mental group (LMM: P = 0.04, R2 = 0.02, R2random = 0.67) and Gammaproteobacteria (LMM: P = 0.02, R2 = 0.03,
R2random = 0.66). Interestingly, the diversity changes in Proteobacteria were dominant among participants in
Lahti but not in Tampere (Supplementary Fig.1).
Among Lahti dwellers, low cytokine IL-17A levels were associated to high Shannon diversity in class Gam-
maproteobacteria and phylum Proteobacteria (Fig.3 and Supplementary Table2) with time as the random
factor. e association with Gammaproteobacteria was observed when both treatment groups were included
in the model (LMM: P = 0.04, R2 = 0.08, R2random = 0.02). e association with Proteobacteria was observed
when both treatment groups were included (LMM: P = 0.017, R2 = 0.10, R2random = 0.029; Fig.3) and within
the experimental group (LMM: P = 0.045, R2 = 0.19, R2random = 0).
e study groups diered signicantly (LMM: P = 0.04, R2 = 0.08, R2random = 0.52) in the level of change
in the anti-inammatory cytokine TGF-β1 on Day28 (Day28 – Day0). e concentration of TGF-β1 (Fig.4 and
Supplementary Table1) increased in the experimental group and lowered in the control group. According to the
R2 –values, location (city) explains the result more (52%) than the treatment (8%) (Supplementary Fig.2). In
IL-10 levels, there were no signicant changes in connection to the treatments or health-associated bacterial taxa.
Figure1. Relative abundance of Lactobacillus spp. (mean ± SE). e relative abundance was calculated by
subsampling to the lowest sequence count in the samples (3893). e relative abundance on Day14 (Wilcoxon
P = 0.0058) and the amount of change (Day14–Day0) was higher in the experimental group.
5
Vol.:(0123456789)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
Regarding the living habits during the experiment and the background information, there were no signicant
dierences found between the treatment groups.
Discussion
e changes observed in this green wall study were connected to Proteobacteria and Lactobacillus that have been
shown to be benecial for human health. As far as we are aware of, this is the rst study that shows a change
in the relative abundance of Lactobacillus spp. on skin in response to green wall exposure. e bacteria from
Lactobacillaceae family (such as Lactobacillus spp.) are known to act against pathogens and inammation on
skin30,60. eir application as a probiotic on skin has been recommended in the treatment of sunburns61, skin
oxidative damage and hyperpigmentation62. erefore, the observed steady and continuous increase in the
Figure2. Shannon diversity index of class Gammaroteobacteria (a) and phylum Proteobacteria (b) on
days 0, 14 and 28 for the experimental group (Exp) and the control group (Ctrl). e change in Shannon
diversity (Day28–Day0) diered between treatment groups in Proteobacteria and (LMM: P = 0.04, R2 = 0.02,
R2random = 0.67) and Gammaproteobacteria (LMM: P = 0.02, R2 = 0.03, R2random = 0.66).
Figure3. Shannon diversity (y-axis) in phylum Proteobacteria on skinassociated with IL-17A concentration
(pg/ml) in blood (x-axis) among Lahti participants. High Shannon diversity of Proteobacteria was associated to
a low levels of proinammatory IL-17A concentration (LMM: P = 0.017, R2 = 0.10, R2random = 0.029).
6
Vol:.(1234567890)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
relative abundance of skin Lactobacillus is an important nding. Spending time in green wall rooms seems to
be related to increasing abundance of health-supporting skin microbiota within a relatively short time period
of two weeks. is support health benets of working in rooms having green walls with air circulation system;
usually green wall are long-term interior design elements.
e diversity of Proteobacteria and Gammaproteobacteria has been observed to be higher among healthy
people compared to people with immune-mediated diseases such as atopy and allergies4,7,25. e diversity of Gam-
maproteobacteria on skin has successfully been altered via biodiversity intervention with an impact to immune
regulation12,13. e elevation in the diversity of proteobacterial taxa on the skin of participants working in the
green wall oces of this study makes sense because Proteobacteria are a common part of plant microbiomes.
However, the elevation was observed only in Lahti (17 study subjects), and according to the R2 values regarding
proteobacterial taxa, city as a factor had a high eect on the results. As seen in Fig.3, the plausible reason for the
dierence in IL-17A level is the increasing proteobacterial abundance in Lahti experimental group. An inter-
esting detail is that graphically even Day0 values were slightly higher in Lahti, though there were no statistical
dierence (Fig.3). In Tampere, all study rooms were situated in the area of Tampere University Hospital, whereas
in Lahti the study rooms were at two separate campus areas (in the city center and between industrial areas)
without a connection to medical sciences. erefore, the daily hygiene practices were probably dierent between
the oce workers at the medical campus in Tampere and the two non-medical campuses in Lahti. Due to the
dierences in the location, the surroundings of the study buildings may also have dierent hygiene levels which
aects microbial diversity1,4. Further, the building in Tampere was built in 2016 whereas the buildings Lahti were
considerably older (built 1993 and 1980); the age of a building aects the indoor microbiome composition26. A
third, potentially parallel explanation is that microbial communities in oces are city-specic18,19; it is tempting
to speculate whether the impact of city is strong enough to mask subtle changes in the relative abundance of
Proteobacteria.
e current study was not designed to explore the mechanisms that lead to changes in skin microbiota. Our
hypothesis is that green walls balance air moisture and release spores or live bacteria that land on skin17. How-
ever, we cannot separate the role of the introduced microbiome from the green walls from the consequences
of the removal of volatile organic compounds (VOC) by the green walls; the green walls used in this study
remove VOCs46. Since VOCs are known to aect the composition and processes of bacterial communities in
the environment63,64 and on skin65, the green walls could have an indirect impact to indoor and skin bacterial
communities. VOCs include pollutants released from materials used in interior decoration46 but they also include
compounds emitted by organisms which may use them for interaction65. For example, skin bacteria may inhibit
one another via VOCs66. erefore, the green walls may remove VOCs that would otherwise impact the bacte-
rial communities indoors and on skin. To distinguish the mechanism responsible for altered skin microbiota,
the microbiome of the green walls should be sampled and the VOC composition in the study rooms should be
analyzed.
Since IL-17A is a proinammatory cytokine associated with adverse health outcomes, like low-grade
inammation67, the association between the high proteobacterial diversity and the low IL-17A concentration
seems benecial. In addition, the change of anti-inammatory cytokine TGF-β1 in the experimental group on
Day28 seems benecial due to the gain of concentration. As with bacterial results, the cytokine results were
impacted by the random factors (city and study subject). Individual dierences are typically large when the
study subjects live outside lab conditions. However, this does not diminish the importance of the observed dif-
ference between the experimental and control group; based on our results, air-circulating green walls change
skin microbial communities among urban dwellers.
Although access to nature outside workhours was permitted, the hours spent in nature was surveyed on Day14
and Day28 and no dierence was found between the experimental and control group. erefore, it seems unlikely
that free time in nature was sucient to overcome the eect of green walls. e access to other study oces was
Figure4. e concentration of the cytokine TGF-β1 (ng/ml) increased in the experimental group and
decreased in the control group on Day28 in comparison to Day0 (LMM: P = 0.04, R2 = 0.08, R2random = 0.52) in
the level of change in the anti-inammatory cytokine TGF-β1 on Day28 (Day28–Day0).
7
Vol.:(0123456789)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
not restricted (contamination) but visits to other oces were either very short or nonexistent; the typical places
of interaction were the coee rooms.
Based on our ndings, air-circulating green walls alter the microbiome and modulate the immune system
among oce workers. Air-circulating green walls have potential in promoting microbiological diversity and
human health in built environments and the topic requires further research attention.
Data availability
Raw sequencing data has been deposited to the Sequence Read Archive (SRA) under BioProject PRJNA757748.
e sensitive data that support the ndings of this study are available from University of Helsinki but restrictions
dened in General Data Protection Regulation (EU 2016/679) and Finnish Data Protection Act 1050/2018 apply
to the availability of these data, and so are not publicly available. Data are, however, available from the authors
upon reasonable request and with permission from the ethical committee of the local hospital district (Ethical
statement number R18026 by Tampereen yliopistollisen sairaalan erityisvastuualue, Pirkanmaa, Finland, the full
trial protocol can be requested from the authors).
Received: 21 September 2021; Accepted: 29 March 2022
References
1. Rook, G. A. W. Review series on helminths, immune modulation and the hygiene hypothesis: e broader implications of the
hygiene hypothesis. Immunology 126, 3–11 (2009).
2. Von Hertzen, L., Hanski, I. & Haahtela, T. Natural immunity. Biodiversity loss and inammatory diseases are two global megatrends
that might be related. EMBO Rep. 12, 1089–1093 (2011).
3. Von Hertzen, L. & Haahtela, T. Disconnection of man and the soil: Reason for the asthma and atopy epidemic?. J. Allergy Clin.
Immunol. 117, 334–344 (2006).
4. Hanski, I. et al. Environmental biodiversity, human microbiota, and allergy are interrelated. Proc. Natl. Acad. Sci. U. S. A. 109,
8334–8339 (2012).
5. Haahtela, T. et al. Immunological resilience and biodiversity for prevention of allergic diseases and asthma. Allergy Eur. J. Allergy
Clin. Immunol. https:// doi. org/ 10. 1111/ all. 14895 (2021).
6. Rook, G. A. W. et al. Mycobacteria and other environmental organisms as immunomodulators for immunoregulatory disorders.
Springer Semin. Immunopathol. 25, 237–255 (2004).
7. Fyhrquist, N. et al. Acinetobacter species in the skin microbiota protect against allergic sensitization and inammation. J. Allergy
Clin. Immunol. 134, 1301-1309.e11 (2014).
8. Ottman, N. et al. Soil exposure modies the gut microbiota and supports immune tolerance in a mouse model. J. Allergy Clin.
Immunol. 143, 1198-1206.e12 (2019).
9. Nurminen, N. et al. Nature-derived microbiota exposure as a novel immunomodulatory approach. Fut. Microbiol. 13, 737–744
(2018).
10. Shaer, M. & Lozupone, C. Prevalence and source of fecal and oral bacteria on infant, child, and adult hands. mSystems 3, 1–12
(2018).
11. Grönroos, M. et al. Short-term direct contact with soil and plant materials leads to an immediate increase in diversity of skin
microbiota. MicrobiologyOpen https:// doi. org/ 10. 1002/ mbo3. 645 (2019).
12. Roslund, M. I. et al. Biodiversity intervention enhances immune regulation and health-associated commensal microbiota among
daycare children. Sci. Adv. 6, 7–105 (2020).
13. Roslund, M. I. et al. Long-term biodiversity intervention shapes health-associated commensal microbiota among urban day-care
children. Environ. Int. 157, 7008 (2021).
14. Lax, S. et al. Longitudinal analysis of microbial interaction between humans and the indoor environment. Science (80-.). 345,
1048–1052 (2014).
15. Flies, E. J., Clarke, L. J., Brook, B. W. & Jones, P. Urbanisation reduces the abundance and diversity of airborne microbes-but what
does that mean for our health? A systematic review. Sci. Total Environ. 738, 140337 (2020).
16. Ege, M. J. et al. Exposure to environmental microorganisms and childhood asthma.. Science 364, 701–709 (2011).
17. L i, H. et al. Spatial and seasonal variation of the airborne microbiome in a rapidly developing city of China. Sci. Total Environ.
665, 61–68 (2019).
18. C hase, J. et al. Geography and location are the primary drivers of oce microbiome composition. mSystems 1, 1–18 (2016).
19. D anko, D. et al. A global metagenomic map of urban microbiomes and antimicrobial resistance. Cell 184, 3376-3393.e17 (2021).
20. Hui, N. et al. Soil microbial communities are shaped by vegetation type and park age in cities under cold climate. Environ. Microbiol.
19, 1281–1295 (2017).
21. Mhuireach, G. et al. Urban greenness inuences airborne bacterial community composition. Sci. Total Environ. 571, 680–687
(2016).
22. Franzetti, A., Gandol, I., Gaspari, E., Ambrosini, R. & Bestetti, G. Seasonal variability of bacteria in ne and coarse urban air
particulate matter. Appl. Microbiol. Biotechnol. 90, 745–753 (2011).
23. Mhuireach, G., Wilson, H. & Johnson, B. R. Urban aerobiomes are inuenced by season, vegetation, and individual site charac-
teristics. EcoHealth 18, 331–344 (2021).
24. Mahnert, A., Moissl-Eichinger, C. & Berg, G. Microbiome interplay: Plants alter microbial abundance and diversity within the
built environment. Front. Microbiol. 6, 1–11 (2015).
25. Ruokolainen, L. et al. Green areas around homes reduce atopic sensitization in children. Allergy Eur. J. Allergy Clin. Immunol. 70,
195–202 (2015).
26. Kirjavainen, P. V. et al. Farm-like indoor microbiota in non-farm homes protects children from asthma development. Nat. Med.
25, 1089–1095 (2019).
27. Nurminen, N. et al. Land cover of early-life environment modulates the risk of type 1 diabetes. Diabetes Care 44, 1506–1514 (2021).
28. Parajuli, A. et al. Yard vegetation is associated with gut microbiota composition. Sci. Total Environ. 713, 136707 (2020).
29. Köberl, M., Dita, M., Martinuz, A., Staver, C. & Berg, G. Members of Gammaproteobacteria as indicator species of healthy banana
plants on Fusarium wilt-infested elds in Central America. Sci. Rep. 7, 1–9 (2017).
30. D elanghe, L. et al. e role of lactobacilli in inhibiting skin pathogens. Biochem. Soc. Trans. 5, 617–627. https:// doi. org/ 10. 1042/
bst20 200329 (2021).
31. G eorge, F. et al. Occurrence and dynamism of lactic acid bacteria in distinct ecological niches: A multifaceted functional health
perspective. Front. Microbiol. 9, 1–15 (2018).
8
Vol:.(1234567890)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
32. Yu, A. O., Leveau, J. H. J. & Marco, M. L. Abundance, diversity and plant-specic adaptations of plant-associated lactic acid bacteria.
Environ. Microbiol. Rep. 12, 16–29 (2020).
33. Parajuli, A. et al. Urbanization reduces transfer of diverse environmental microbiota indoors. Front. Microbiol. 9, 1405 (2018).
34. Parajuli, A. et al. e abundance of health-associated bacteria is altered in PAH polluted soils—Implications for health in urban
areas?. PLoS One 7, 1–18. https:// doi. org/ 10. 1371/ journ al. pone. 01878 52 (2017).
35. Vari, H. K. et al. Associations between land cover categories, gaseous PAH levels in ambient air and endocrine signaling predicted
from gut bacterial metagenome of the elderly. Chemosphere 265, 1559 (2021).
36. Orsini, F., Kahane, R., Nono-Womdim, R. & Gianquinto, G. Urban agriculture in the developing world: A review. Agron. Sustain.
Dev. 33, 695–720 (2013).
37. Hui, N. et al. Diverse environmental microbiota as a tool to augment biodiversity in urban landscaping materials. Front. Microbiol.
10, 1–10 (2019).
38. Pu hakka, R. et al. Greening of daycare yards with biodiverse materials aords well-being, play and environmental relationships.
Int. J. Environ. Res. Public Health 16, 2948 (2019).
39. Burmeister, A. R. & Marriott, I. e interleukin-10 family of cytokines and their role in the CNS. Front. Cell. Neurosci. 12, 1–13
(2018).
40. Opal, S. M. & DePalo, V. A. Anti-inammatory cytokines. Chest 117, 1162–1172 (2000).
41. Kuwabara, T., Ishikawa, F., Kondo, M. & Kakiuchi, T. e role of IL-17 and related cytokines in inammatory autoimmune diseases.
Mediators Inamm. 2017, 4598 (2017).
42. Li, M. O., Wan, Y. Y., Sanjabi, S., Robertson, A. K. L. & Flavell, R. A. Transforming growth factor-β regulation of immune responses.
Annu. Rev. Immunol. 24, 99–146 (2006).
43. Prudhomme, G. J. & Piccirillo, C. A. e inhibitory eects of transforming growth factor-beta-1 (TGF-β1) in autoimmune diseases.
J. Autoimmun. 14, 23–42 (2000).
44. Esebanmen, G. E. & Langridge, W. H. R. e role of TGF-beta signaling in dendritic cell tolerance. Immunol. Res. 65, 987–994
(2017).
45. Honkanen, J. et al. IL-17 immunity in human type 1 diabetes. J. Immunol. 185, 1959–1967 (2010).
46. Torpy, F. et al. Testing the single-pass VOC removal eciency of an active green wall using methyl ethyl ketone (MEK). Air Qual.
Atmos. Heal. 11, 163–170 (2018).
47. Roslund, M. I. et al. Endocrine disruption and commensal bacteria alteration associated with gaseous and soil PAH contamination
among daycare children. Environ. Int. 130, 104894 (2019).
48. Schloss, P. D., Gevers, D. & Westcott, S. L. Reducing the eects of PCR amplication and sequencing Artifacts on 16s rRNA-based
studies. PLoS One 6, 1789 (2011).
49. Kozich, J., Westcott, S., Baxter, N., Highlander, S. & Schloss, P. Development of a dual-index sequencing strategy and curation
pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl. Environ. Microbiol. 79, 5112–5120
(2013).
50. Soininen, L., Grönroos, M., Roslund, M. I. & Sinkkonen, A. Long-term storage aects resource availability and occurrence of
bacterial taxa linked to pollutant degradation and human health in landscaping materials. Urban For. Urban Green. 60, 1789 (2021).
51. Q uast, C. et al. e SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids
Res. 41, 590–596 (2013).
52. Huse, S. M., Welch, D. M., Morrison, H. G. & Sogin, M. L. Ironing out the wrinkles in the rare biosphere through improved OTU
clustering. Environ. Microbiol. 12, 1889–1898 (2010).
53. Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C. & Knight, R. UCHIME improves sensitivity and speed of chimera detection.
Bioinformatics 27, 2194–2200 (2011).
54. Wang, Q., Garrity, G. M., Tiedje, J. M., Cole, J. R. & Al, W. E. T. Naive Bayesian classier for rapid assignment of rRNA sequences
into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261–5267 (2007).
55. R Core Team. R: A Language and Environment for Statistical Computing. (2020).
56. Oksanen, J. et al. vegan: Community Ecology Package. (2019).
57. Huang, F. L. Alternatives to multilevel modeling for the analysis of clustered data. J. Exp. Educ. 84, 175–196 (2016).
58. Moen, E. L., Fricano-Kugler, C. J., Luikart, B. W. & O’Malley, A. J. Analyzing clustered data: Why and how to account for multiple
observations nested within a study participant?. PLoS ONE 11, 1–17 (2016).
59. Twisk, J. et al. Dierent ways to estimate treatment eects in randomised controlled trials. Contemp. Clin. Trials Commun. 10,
80–85 (2018).
60. Chapat, L., Chemin, K., Dubois, B., Bourdet-Sicard, R. & Kaiserlian, D. Lactobacillus casei reduces CD8+ T cell-mediated skin
inammation. Eur. J. Immunol. 34, 2520–2528 (2004).
61. Kaur, K. & Rath, G. Formulation and evaluation of UV protective synbiotic skin care topical formulation. J. Cosmet. Laser er.
21, 332–342 (2019).
62. Rong, J. et al. Skin resistance to UVB-induced oxidative stress and hyperpigmentation by the topical use of Lactobacillus helveticus
NS8-fermented milk supernatant. J. Appl. Microbiol. 123, 511–523 (2017).
63. Yuan, J. et al. Microbial volatile compounds alter the soil microbial community. Environ. Sci. Pollut. Res. 24, 22485–22493 (2017).
64. Abis, L. et al. Reduced microbial diversity induces larger volatile organic compound emissions from soils. Sci. Rep. 10, 1–15 (2020).
65. Duy, E. & Morrin, A. Endogenous and microbial volatile organic compounds in cutaneous health and disease. TrAC Trends Anal.
Chem. 111, 163–172 (2019).
66. Lemfack, M. C. et al. Novel volatiles of skin-borne bacteria inhibit the growth of Gram-positive bacteria and aect quorum-sensing
controlled phenotypes of Gram-negative bacteria. Syst. Appl. Microbiol. 39, 503–515 (2016).
67. Ahmed, M. & Gaen, S. L. IL-17 in obesity and adipogenesis. Cytokine Growth Factor Rev. 21, 449–453 (2010).
Acknowledgements
We thank the participants of this study for their time and input. is study was funded by Business Finland (grant
numbers 6766/31/2017 and 7941/31/2017) (grant to A.S. and H.H) and Doctoral Programme in Interdisciplinary
Environmental Sciences (DENVI) in the University of Helsinki, and supported by Naava who provided the green
walls for this study. We thank Environmental Laboratory at University of Helsinki, CSC – IT Center for Science,
Finland, for computational resources, and Institute for Molecular Medicine Finland (FIMM) for their work.
Author contributions
A.S., H.H., O.H.L., R.P., M.I.R., and N.N. wrote the ethical application for the study. L.S. wrote the rst dra of
the manuscript. N.N. performed the cytokine analyses. L.S and M.I.R performed the bioinformatic and statistical
analyses and prepared the gures. L.S., R.P., M.I.R., A.S., N.N., and O.H.L. implemented the study. L.S., M.I.R.,
A.S., R.P., N.N., H.H., and O.H.L. wrote the nal version of the manuscript. A.S. and H.H. were the principal
investigators of the project.
9
Vol.:(0123456789)
Scientic Reports | (2022) 12:6518 | https://doi.org/10.1038/s41598-022-10432-4
www.nature.com/scientificreports/
Competing interests
A.S., H.H. and O.H.L are members of the board of Uute scientic LtD which develops topical immunomodula-
tory treatments.
Additional information
Supplementary Information e online version contains supplementary material available at https:// doi. org/
10. 1038/ s41598- 022- 10432-4.
Correspondence and requests for materials should be addressed to A.S.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access is article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
© e Author(s) 2022
ADELE research group
Damiano Cerrone2, Mira Grönroos1, Nan Hui1, Anna Luukkonen3, Iida Mäkelä1,
Noora Nurminen2, Sami Oikarinen2, Anirudra Parajuli1, Riikka Puhakka1,3, Marja I. Roslund1,
Mika Saarenpää1, Laura Soininen1, Yan Sun1, Heli K. Vari1, Olli H. Laitinen2, Juho Rajaniemi2,
Heikki Hyoty2 & Aki Sinkkonen3
