Experiment protocol used in this experiment to investigate the effects of the Mycobacterium vaccae soil on the physiological responses of humans during the soil-mixing activity.
(A) Principal component analysis and (B) partial least-square discriminant analysis score plot derived from solid-phase microextraction gas chromatography–time-of-flight–mass spectrometry datasets for various soil samples; soil treated with distilled water [S (λ)]; soil treated with culture media [SM (λ)]; soil inoculated with Mycobacterium vaccae [SMV (λ)]. (C) Heat map analysis for the relative abundance of different volatile organic compounds (variable importance in the projection > 1.0, P < 0.05) derived from the gas chromatography–time-of-flight–mass spectrometry analysis. The colored squares (blue to red) indicate fold changes that are normalized by the average of each metabolite; N.I. = nonidentified.
(A) Principal component analysis and (B) partial least-square discriminant analysis score plot derived from the gas chromatography–time-of-flight–mass spectrometry datasets for serum samples; control [soil treated with culture media (λ)]; treatment [soil treated with Mycobacterium vaccae (λ)]. (C) Heat map analysis for the relative abundance of different serum metabolites (variable importance in the projection > 1.0) derived from the gas chromatography–time-of-flight–mass spectrometry analysis. The colored squares (blue to red) indicate fold changes that were normalized by the average of each metabolite. *Significantly different metabolite between control and treatment groups (P < 0.05, Student’s t test); N.I. = nonidentified.
Correlation analysis between physiological measurement and significantly altered serum metabolites. Each square indicates the Pearson’s correlation coefficient values (r). The red and blue colors represent positive (0.0 < r < 0.3) and negative (−0.3 < r < 0.0) correlations, respectively. *P < 0.05; RFA = relative fast alpha power; ASEF50 = spectral edge frequency 50% of alpha; bps = beats per minute; LF = low frequency band; HF = high frequency band; SDNN = standard deviation of RR intervals (where R is a point corresponding to the peak of the QRS complex of the electrocardiography wave); BDNF = brain-derived neurotrophic factor.
Results of brain-derived neurotrophic factor, according to the presence and absence of Mycobacterium vaccae in the soil during the soil-mixing activity.
Click on author name to view affiliation information
Mycobacterium vaccae is a species of nonpathogenic bacterium that lives naturally in soil. This study compared the physiological effects at a metabolomic level with autonomic nervous system responses in adults during soil-mixing activities, based on the presence or absence of M. vaccae in the soil. Twenty-nine adult participants performed soil-mixing activities for 5 minutes using sterilized soil with culture media and M. vaccae, respectively. Blood samples were drawn twice from each participant after each activity. Electroencephalograms and electrocardiograms were measured during the activity. Serum metabolites underwent metabolite profiling by gas chromatography, followed by multivariate analyses. Soil-emitted volatile organic compounds were identified using the solid-phase microextraction and gas chromatography-mass spectroscopy, followed by multivariate analyses. The volatile compound analysis revealed that the metabolites related to esters and sulfur-containing compounds are greater in soil with M. vaccae. Serum metabolomics revealed that the treatment group (soil inoculated by M. vaccae) possesses relatively higher levels of inter-alia organic and amino acids compared with the control group (soil mixed with culture media). In the treatment group, the electroencephalogram and electrocardiogram revealed that alpha band activity of the occipital lobe increases, while heart rate decreases. This study concludes that M. vaccae soil contact can affect human metabolic and autonomic reactions.
By 2050, ≈70% of humans are expected to live in urban areas (Dye, 2008). Urbanization has exposed humans to more artificial elements. Continuous exposure to these environments exacerbates human stress levels. Increases in stress levels caused by urban life could lead to mental health problems, such as mood and anxiety disorders (Peen et al., 2010) and schizophrenia (Krabbendam and van Os, 2005; Pedersen and Mortensen, 2001). Owing to the negative effects of urban life, people’s interests in human health, well-being, and nature have recently increased.
Humans feel an instinctive preference for nature, and Orians’ Savannah theory and Wilson’s biophilia hypothesis support this (Orians, 1986; Wilson, 1984). Ulrich (1983) reported that stress from the external environment can be reduced through exposure to nature.
Previous studies have shown that staring at green plants stabilizes the human autonomic nervous system and activates the alpha wave frequency of brain waves, leading to physiological and psychological relaxation (Ikei et al., 2014; Park et al., 2016). In addition, when performing work activities in an environment with plants, the sympathetic nerve activity and oxyhemoglobin of the worker’s left frontal cortex is reduced, resulting in physiological relaxation (Park et al., 2017). Park et al. (2020) performed metabolite profiling to investigate the effects of horticultural interventions on cognitive improvements in older adults; they reported that horticultural activities have psychological and cognitive health effects by enhancing tryptophan and serotonin in the serum of elderly individuals.
As such, various effects in physiological and psychological aspects that are obtained from interactions with nature have been reported, but prior studies have focused on green plants, predominantly in the form of holistic and synergistic approaches that combine plants and soil. To fully understand the healing mechanism of horticultural activities, we must now study not only the effectiveness through a complex approach but also the role of each plant and soil. There are currently few studies related to the mechanism of healing on the human body through soil contact.
Examining the effects of urban environmental biodiversity on commensal microbiome and immunoregulation in children, biodiversity interventions such as contact with plants and soil enhanced immunoregulatory pathways, as were associated with increased plasma transforming growth factor beta 1 levels and proportion of regulatory T cells (Roslund et al., 2020). In addition, with long-term follow-up of more than 1 year, biodiversity interventions enriched the commensal microbiome and suppressed the potentially pathogenic bacteria on the skin, including taxa related to immune regulation (Roslund et al., 2021).
Mycobacterium vaccae belongs to the Actinomycetales and are a nonpathogenic bacterial species that live naturally in the soil. They are predominantly found in soil, water, and mud (Hoisington et al., 2015), and are reported to have a strong immunomodulatory effect, according to the hygiene hypothesis in which the immune system was developed by exposure to various microorganisms (Rook et al., 2004). In rodents (Rodentia), exposure to M. vaccae has been reported to activate their immune response, thereby reducing inflammation and stress-induced behavioral disorders and improving learning abilities (Fonken et al., 2018; Frank et al., 2018; Reber et al., 2016; Smith et al., 2019). Furthermore, this immune activation promoted the activation of serotonin neurons, thereby increasing the level of serotonin and reducing anxiety in rodents (Lowry et al., 2007). With similar results, O’Brien et al. (2004) reported that when M. vaccae is injected into patients with lung cancer, the patients’ sense of happiness and vitality increases. However, there are few studies examining the effect of M. vaccae in clinical trials other than intracorporeal injections or oral administration on humans.
To understand the therapeutic mechanism of horticultural activities and nature-based therapy clearly, studies on the effects of interactions with soil, soil microorganisms, and plants are necessary. Therefore, this study measures the effect of a soil-mixing activity on the psychological and physiological responses of humans, according to the presence or absence of M. vaccae microorganisms in the soil. Depending on the presence or absence of M. vaccae microorganisms in the soil, the effects of a soil-mixing activity on human metabolic and autonomic reactions will be different.
Twenty-nine adults from 20 to 59 years old (8 men, 21 women; average age 28.6 ± 9.8 years) participated in this study. Participants were recruited using a convenience sampling method. A flyer that included study information was distributed to apartments and churches in Gwangjin-gu, Seoul, Republic of Korea. Participants were recruited according to the inclusion and exclusion criteria shown in Table 1, so as not to influence other physiological data. Before conducting the experiment, participants were informed of the research contents and precautions, and written consents were obtained before participation in the research. To collect participants’ demographic information, age, sex, height, weight, and body mass index using a body composition analyzer (ioi 353; Jawon Medical, Gyeongsan, Republic of Korea) were recorded. Participants received the equivalent of $10 as an incentive to complete the experiment. This study was approved by the Bioethics Committee of Konkuk University, Seoul, Republic of Korea (7001355-201911-HR-345).
M. vaccae KCTC 19087 was obtained from the Korean Collection for Type Cultures (KCTC, Jeongeup, Republic of Korea). M. vaccae was cultivated on tryptic soy broth for 4 d at 37 °C in the dark and by shaking (200 rpm). For soil sample preparation, soil samples were autoclaved at 121 °C for 15 min. The sterile soil (1.5 g) was mixed with 2.5 mL of sterile water, a tryptic soy broth, and a 4-d cultured M. vaccae strain [1.35 × 109 colony-forming units (cfu)/mL], for 2 d at 37 °C to obtain various types of volatile organic compounds (VOCs). After incubation, the samples were transferred to a gas chromatography–time-of-flight–mass spectrometry (GC-TOF-MS) instrument to analyze VOCs.
Microbial-treated soil was mixed with sterilized peatmoss (2000 mL), perlite (800 mL), and water (200 mL), with a 50 mL M. vaccae medium that was cultured for 4 d (1.35 × 109 cfu/mL). The control soil was mixed with 50 mL of cultured medium that did not contain microorganisms.
This study was conducted in an experimental space (180 cm × 200 cm) in Konkuk University. In the experimental space, there was a desk (180 cm × 90 cm) on which a basin of soil could be placed, and the distance between the basin and the participants was set at 50 cm. To minimize external visual stimulation, white hardboard paper was placed before the desk, and ivory-colored curtains were installed on either side of it. The environmental conditions measured by thermo-hygrometer (O-257; DRETEC Co., Kawaguchi, Japan) of the space during the experiment were as follows: temperature 26.5 ± 1.8 °C, relative humidity 43.3% ± 15.8%.
This study was conducted through a single-blinded experiment and randomized crossover study method. To investigate the effect of the M. vaccae soil on the physiological responses of humans during the soil-mixing activity, an experiment was performed according to the procedure shown in Fig. 1. Before performing the soil-mixing activity, participants were asked to look at the white wall in front of them for 5 min to encourage relaxation. Thereafter, they mixed the soil in the basin with both hands for 5 min. After the soil-mixing activity, 7 mL of blood was collected for the metabolome analysis. After the first trial, participants performed the other activity with the procedure mentioned previously following a 5-min break, and the trial order was randomly assigned. The duration of the entire experiment was ≈40 min per participant.
Citation: Journal of the American Society for Horticultural Science 147, 3; 10.21273/JASHS05146-21
To compare the physiological responses of participants when performing soil-mixing activities according to the presence or absence of M. vaccae, electroencephalography (EEG) and electrocardiography (ECG) were measured using a wireless dry EEG device (Quick-20; Cognionics, San Diego, CA) and a medical electrode (HP-OP42; Hurev, Wonju, Republic of Korea), respectively. The dry wireless EEG device used in this study minimizes the risk of electrical stimulation by using a dry electrode. It also can be detached quickly if the participant feels uncomfortable.
Data were collected by amplifying and processing electrical signals measured through dry electrodes applied to the scalp. The device is safety certified by the European Commission and Federal Communications Commission (Kim et al., 2020). The electrode application complied with the international 10- to 20-electrode arrangement system (Jasper, 1958). The reference electrode was attached to the left earlobe (A1). According to the international electrode method, this study performed EEG monitoring at the left occipital cortex (O1) and right occipital cortex (O2). Previous studies have reported that EEG could improve our understanding of brain activity and human central nervous system activity through olfactory stimulation (Lorig, 1989). Furthermore, it has been reported that cortical activity can be enhanced by a significant increase in fast alpha activity in bilateral posterior regions of the brain through olfactory stimulation (Iijima et al., 2009). Masago et al. (2000) reported that the occipital and temporal lobes are involved in integrated sensory information processing, including the sense of smell, and are associated with complex and integrated neural activity related to odors thought to occur in these areas. In addition, as serotonin is a major neurotransmitter involved in the occipital lobe (UKEssays, 2020), which also regulates human emotions and mood, the occipital lobe was selected as the measurement cortex to investigate the effect of the soil-mixing activity on human mood, based on the presence or absence of M. vaccae. The ECG electrodes were placed at the end of the right collarbones and left rib bones of participants.
Blood samples were collected to measure variations in the brain-derived neurotrophic factor. The participant’s specimens were collected by a skilled sampler, each of whom was a professional nurse. The blood samples were collected in a clot activator tube (Vacutainer 367896; BD Diagnostics, Franklin Lakes, NJ). The collected blood was kept at room temperature for 20 min and centrifuged for 10 min at 1000 gn to separate the serum samples. Thereafter, the aliquot was stored at –70 °C in a deep freezer. The enzyme-linked immunosorbent assay kits were used to measure brain-derived neurotrophic factor (BDNF; AbCAM, Cambridge, UK) according to the manufacturer’s instructions.
The headspace solid-phase microextraction (HS-SPME) was performed for three biological replicates of soil-treated distilled water (S), culture media (SM), and soil inoculated with M. vaccae strain (SMV) samples to obtain VOCs. Each soil sample (1.5 g) was mixed with 2.5 mL of distilled water, culture media, and the M. vaccae strain. The sample mixtures were transferred into a 20-mL SPME vial and pre-incubated for 2 h. The HS-SPME of VOCs from soil samples were immediately performed using carboxen/polydimethylsiloxane/divinylbenzene (CAR/PDMS/DVB)-coated SPME fibers [75 μm (Supelco; Sigma-Aldrich, St. Louis, MO)] and collected by exposing the fiber to soil samples for 50 min at 37 °C with 250 rpm. Following the extraction, the fiber was inserted into the GC injector (230 °C) for the desorption procedure for 1 min. We followed the method of Singh and Lee (2018) for the extraction for VOCs with few modifications.
Each human serum (200 μL) was extracted with cold methanol (1 mL) and 10 μL of an internal standard (1 mg·mL−1 2-chlorophenylalanine) using a mixer mill (MM400; Retsch, Haan, Germany) at a frequency of 30 Hz for 10 min, with sonication for another 10 min. After homogenization, the suspension was stored at 20 °C for 60 min. It was then centrifuged at 13,250 gn for 10 min at 4 °C (Centrifuge Universal 320; Hettich, Tuttlingen, Germany). The supernatant was filtered through a 0.2-μm polytetrafluoroethylene filter (Chromdisc, Daegu, Republic of Korea). The filtered samples were dried completely using a speed vacuum concentrator (Biotron, Seoul, Republic of Korea). The final concentration of each sample was adjusted to 10 mg·mL−1 for the MS analysis. The serum extraction procedure was based on our previous research (Park et al., 2020).
The GC-TOF-MS instrument involved in the analysis of VOCs was identical to our previous study (Park et al., 2020). The GC analytical program for VOCs was programmed as follows: initially maintained at 33 °C for 3 min and elevated to 180 °C at a rate of 10 °C·min−1. Finally, the temperature was raised to 240 °C at a rate of 40 °C·min−1, for a duration of 4 min. The flow rate of helium was 1.5 mL·min−1. The mass spectrum was collected with the mass range of 50–500 m/z at a rate of 10 scans per second.
The GC-TOF-MS analysis was performed as described previously by Park et al. (2020). The derivatized samples (1 μL) were injected into the GC-TOF-MS instrument in the splitless mode. The analytical program and parameter setting for analysis were adapted from our previous study (Park et al., 2020). The analytical samples were randomized in each block to reduce the effects of systematic errors.
The measured EEG and ECG data were analyzed using the Bio-scan analysis program (Bio-Tech, Daejeon, Republic of Korea). The collected EEG raw data were analyzed using power spectrum analysis to identify the relative fast alpha (RFA) power spectrum and spectral edge frequency as 50% of alpha (ASEF50) (Sowndhararajan et al., 2015). ECG data were converted to heart rate variability (HRV) after high-pass filter preprocessing to obtain the average heart rate, low frequency band (LF), high frequency band (HF), and standard deviation of NN interval (SDNN).
The MS data processing and multivariate statistical analysis were performed as previously described (Park et al., 2020). For MS data processing, raw data derived from GC-TOF-MS were converted into a netCDF (*.cdf) format using an MS data system (ChromaTOF ver. 4.44; LECO Corp., St. Joseph, MI). Subsequently, the peak alignment, peak detection, peak normalization, and retention time were determined using MetAlign software (RIKILT-Institute of Food Safety, Wageningen, The Netherlands). The results of alignment data were exported to spreadsheet files (Microsoft Excel, Office 2007; Microsoft, Redmond, WA). Multivariate statistical analysis was performed using software (SIMCA-P+ version 12.0; Umetrics, Urea, Sweden). Principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), and orthogonal partial least squares-discriminant analysis (OPLS-DA) were performed to compare different VOCs and serum metabolites. The significance of the PLS-DA and OPLS-DA models were defined by an analysis of variance testing of cross-validated predictive residuals (CV-ANOVA) using the SIMCA-P+ program. Different metabolites were selected by variable importance in the projection (VIP) value of the PLS-DA and OPLS-DA models. The discriminated metabolites were identified through comparing their retention time and mass fragment data (through MS) to the available databases, such as the Human Metabolome Database (2021), the National Institute of Standards and Technology database (version 2.0, 2011; FairCom, Gaithersburg, MD), Wiley 9 database (Wiley-VCH, Weinheim, Germany), and our in-house library of standard compounds.
Processed EEG and ECG data were analyzed by the Wilcoxon signed-ranks test, which was performed using statistical analysis software (IBM SPSS Statistics version 25 for Windows; IBM Corp., Armonk, NY). The significance level was set at P < 0.05. To analyze demographic information, descriptive statistics were performed on the mean, standard deviation, and percentage of each collection item using spreadsheet software (Microsoft Excel, Office 2007). Furthermore, the significantly different metabolites from the experimental groups were evaluated through Student’s t test and one-way ANOVA, coupled with the Pearson’s correlation coefficient between serum metabolites and the phenotypes, using statistical analysis software (IBM SPSS Statistics version 18 for Windows).
The characteristics of participants of this study are summarized in Table 2. Twenty-nine adults from 20 to 59 years old (8 men and 21 women; mean age 28.6 ± 9.8 years) participated.
From the comparison of the EEG during the soil-mixing activity (based on the presence or absence of M. vaccae in the soil), the RFA and ASEF50 of the right occipital lobe were significantly higher during the soil-mixing activity that included M. vaccae (P < 0.05) (Table 3). As a result of comparing the HRV of participants during the soil-mixing activity (according to the presence or absence of M. vaccae in the soil), heart rate was significantly lower during the soil-mixing activity that included M. vaccae (P < 0.05) (Table 4). There were no significant differences in LF, HF, and SDNN between the two conditions (P > 0.05).
The disparities of VOCs in various soil samples were identified, including the soil of S, SM, and SMV. These were evaluated using multivariate analysis of SPME-GC-TOF-MS datasets. As shown in Fig. 2, the PCA score plot based on SPME-GC-TOF-MS data showed a marked distinguishment with different soil samples by PC1 (45.48%) and PC2 (24.31%) (Fig. 2A). Moreover, the PLS-DA showed a similar pattern to the PCA score plot (Fig. 2B). The statistical parameters of PLS-DA models were evaluated by R2X (0.697), R2Y (0.987), Q2 (0.959), and CV-ANOVA P value (P < 0.05). This indicated the model validation, fitness, and prediction accuracy, as shown in the Fig. 2. Based on the PLS-DA model, significantly different VOCs among the soil samples were selected by the VIP value (> 0.7) and the P value (< 0.05), as evaluated by analysis of variance for statistical significance. A total of 44 VOCs were identified (six aldehydes, six benzenoids, two sulfur-containing compounds, five alkanes and alkenes, six esters, three furans, four ketones, three terpenes, three others, and six unknowns). These possessed significantly different VOCs among various soil samples (Supplemental Table 1). The relative contents of discriminant metabolites were shown in the heat map (Fig. 2C). According to the heat map analysis, most of the benzenoids, alkanes and alkenes, ketones, and esters were relatively high in the SMV samples. Collectively, based on the heatmap analysis, we selected SM and SMV samples to understand the effect of the soil-mixing activity in human physiology.
Citation: Journal of the American Society for Horticultural Science 147, 3; 10.21273/JASHS05146-21
Based on our results of soil variation with VOCs, we performed the soil-mixing activity with both the SM and SMV. We subjected the metabolite profiling of serum samples to identify the metabolite levels and clarify how these were affected by the soil-mixing activity. The OPLS-DA score plot for serum datasets showed a clearly distinct pattern between the control group (using SM) and treatment group (using SMV) (Fig. 3B). However, the PCA score plots showed an unclear cluster between experimental groups compared with OPLS-DA score plots (Fig. 3A). According to the OPLS-DA model, discriminant metabolites between the control and treatment groups were selected by the VIP value (> 1.0). A total of 59 metabolites were identified [5 organic acids, 10 amino acids, 11 carbohydrates, 17 fatty acids and lipids, 5 others, and 11 unknowns (Supplemental Table 2)]. For visualization of different metabolites, all were displayed on a heat map (Fig. 3C). Based on this, the organic acids, amino acids, and others, except for the citric acid, proline, serine, phenylalanine, tryptophan, and uric acid, were relatively higher in the treatment group than the control group. Collectively, most of the metabolites were higher in SMV. However, most of the fatty acids and lipids were relatively higher in the control group than in the treatment group. Furthermore, we conducted a correlation analysis between physiological measurements and significantly altered serum metabolites (Supplemental Fig. 1). Most of the amino acids, carbohydrates, and fatty acids/lipids were positively correlated with RFA and BDNF. Particularly, tryptophan and phenylalanine significantly correlated with RFA (O2) and BDNF, respectively.
Citation: Journal of the American Society for Horticultural Science 147, 3; 10.21273/JASHS05146-21
The role of soil VOCs is poorly understood, and the effects of soil-mixing activities also require further investigation. Therefore, this study investigated VOC profiling of soil and the effect of the soil-mixing activity using the metabolomic approach. The VOC analysis of various soil samples (i.e., S, SM, and SMV) were performed to identify the VOCs and then select the soil for the soil-mixing activity. Moreover, we analyzed the serum metabolites, EEG, and HRV between the experimental group (control vs treatment) to understand the effects of the soil-mixing activity. As a result, the effects of soil-mixing activity on human metabolic and autonomic reactions was different, depending on the presence or absence of M. vaccae microorganisms in the soil.
The comparative VOC analysis of different soil samples demonstrated a marked distinction (Fig. 2). Particularly, most VOCs (including the benzenoids, sulfur-containing compounds, alkanes and alkenes, esters, and ketones) were found to be higher in SMV samples than in other soil samples. The effects of soil influences on human health included food production, nutrient supply, enhancement of the immune system, and as a source of medications (Brevik et al., 2020). The VOCs include aldehydes, alcohols, alkenes, benzenoids, esters, furans, ketones, sulfur-containing compounds, and terpenes, which were detected from the natural soil and forest soil (Antonelli et al., 2020; Perrault et al., 2015; Tassi et al., 2015; Wheatley et al., 1996).
Moreover, the exposure to soil VOCs has an effect on human health, such as alleviating inflammation and stress, assisting with sleep, and psychological behaviors (e.g., anxiolytic and antidepressive properties) (Antonelli et al., 2020). Furthermore, antibiotic characteristics, antimicrobial properties that were found in soil, and the exposure to soil-borne microorganisms all played a crucial role in regulation of the human immune systems (Brevik et al., 2020). Notably, exposure to M. vaccae in immune system activation and serotonin pathways could influence behavioral and emotional responses (Brevik et al., 2020). Some research demonstrated that injection with heat-killed M. vaccae to mice (Mus musculus) could influence immunocompetence through gastrointestinal tract interaction, immune activation of serotonergic neurons located in related parts of the brain, upregulation of serotonin metabolism in the prefrontal cortex, and the production of metabolites such as lipids (Foxx et al., 2021; Matthews and Jenks, 2013). However, a limited number of VOCs and metabolism that are derived from M. vaccae studies have been reported (McNerney et al., 2012; Nawrath et al., 2012). To address the gaps in previous studies, we examined the serum metabolomic difference between soil-mixing experimental groups and revealed the synergetic effect of soil-mixing activities and VOCs.
The results of the EEG showed that the RFA and ASEF50 of the right occipital lobe were significantly higher in the treatment group compared with the control group (P < 0.05) (Table 3). The RFA means neural oscillations in the frequency range of 11 to 13 Hz, and the ASEF50 is the frequency of the point on the power spectrum graph, in which the area from 8 to 13 Hz occupies 50% of the entire frequency range (Choi et al., 2012). Both indicators are related to alpha waves (8–13 Hz). As the size of fast frequency band among the alpha wave bands increases, the values of the two indicators appear larger. Increased cortical alpha activity was associated with states of relaxation and calmness in the brain (Basar, 2012; Iijima et al., 2009; Sayorwan et al., 2012) and was attenuated under conditions of emotional tension and stress (Kim et al., 2017; Lorig and Schwartz, 1988). In particular, an increase in the fast alpha band indicated that emotional anxiety becomes stable and that the brain is awake (Choi et al., 2012).
A previous study revealed that when 2-methylisoborneol (a major odor molecule of soil) was inhaled, the participant’s fast alpha band increased significantly (Kim et al., 2017). Lorig (2000) reported that pleasant odors promote increased alpha waves, whereas unpleasant odors cause a decrease in alpha waves. It is thought that olfactory stimulation through increased soil-derived VOCs, which are caused by the addition of M. vaccae, made the participants feel pleasant and caused an increase in calmness and arousal in the brain. These EEG changes may be the result of the incense component in M. vaccae that acts as a ligand for mammalian receptors, such as transient receptor potential vanilloid (TRPV) 3, that induce rapid changes in neurophysiology. In the previous study, insensol acetate, a incense component, was found to influence changes in emotional regulation by activating TRPV3 channels in the brain (Moussaieff et al., 2008).
The HRV measured by the ECG showed that heart rates in the treatment group were lower than in the control group [P < 0.05 (Table 4)]. Heart rate is controlled by the action of the autonomic nervous system, and a decreased heart rate can indicate an increase in parasympathetic activity or decrease in sympathetic outflow (Carter, 2009). Wichrowski et al. (2005) reported that when cardiac rehabilitation inpatients performed horticultural activities, the heart rate of the patients decreased significantly. As an elevated heart rate is an indicator of a stress response (Todd, 2014), the decrease in heart rate through horticultural activity implies a stabilization of the autonomic nervous system and reduced stress levels.
Serum metabolomics showed that the treatment group possessed relatively higher levels of organic acids, amino acids, and others compared with the control group (Fig. 3). However, most of the fatty acids and lipids were relatively higher in the control group. Particularly, lactic acid was known as a crucial metabolite for brain functioning and an energy source used by neurons during activity (Todd, 2014). In addition, lactic acid optimized functioning of the gamma-aminobutyric acid receptor, ensuring that central nervous system inhibition is effectually recognized (Todd, 2014). Lactic acid could be considered vital to the maintenance of cognitive functions and protection from neuron damage (Todd, 2014). Moreover, there is limited research on the effects of pyruvic acid on cognitive functions. Research has demonstrated that pyruvate, when infused into the hippocampus and the medial septum, showed a reversal in the memory-impairing effects of morphine or septal muscimol, a GABAA receptor agonist (Owen and Sunram-Lea, 2011).
Tryptophan is an essential amino acid, derived from various protein-based foods and dietary proteins. It is the precursor of various physiologically essential metabolites, such as kynurenine and serotonin (Jenkins et al., 2016). Serotonin has been correlated with motivational and emotional aspects of human behavior, such as depression, anxiety disorders, and compulsive disorders (Silber and Schmitt, 2010). Psychological disorders of depression and emotions such as loneliness have been correlated with serotonin levels, which showed decreasing patterns (Park et al., 2020). Furthermore, serotonin-related drugs have been commonly used in the treatment of psychological abnormalities (Meneses and Liy-Salmeron, 2012). Notably, most fatty acids and lipids showed relatively lower levels in treatment groups (Fig. 3). An excessive level of fat in the diet might influence neuropsychiatric functions negatively, through cognition, anxiety, and emotional levels (Moon et al., 2014). Elevated contents of palmitic and linoleic acid were negatively linked to anxiety-like behavior and cognitive functions (Bernard et al., 2015; Moon et al., 2014). Moreover, elevation of these metabolites in plasma were correlated with diseases such as metabolic syndrome, obesity, and poor clinical outcomes (Moon et al., 2014). The decreased fatty acids, such as palmitic acid in the SMV group, could be due to metabolism of lipids by M. vaccae. Previous studies have shown that mycobacteria synthesize cis-10-hexadecenoic acid from palmitic acid (Scheuerbrandt and Bloch, 1962), which in turn binds to the host peroxisome proliferator-activated receptor-alpha receptor and has been shown to have anti-inflammatory effects (Smith et al., 2019).
In this study, the levels of organic acids (lactic and pyruvic acids and serotonin) were shown to increase, whereas most fatty acids and lipids (including palmitic and linoleic acid) decreased in soil-mixing activities inoculated by the M. vaccae strain. These results can be beneficial in explaining the effect of M. vaccae soil-mixing activities on physiological and psychiatric disorders. Furthermore, the correlation analysis showed that tryptophan and serotonin were positively correlated with RFA (O2), ASEF50 (O2), and BDNF (Supplemental Fig. 2). These results were similar to our previous research on gardening intervention (Park et al., 2020). Components of incenses could potentially act not only on the primary olfactory cortex but also on the mesocorticolimbic dopaminergic and serotoninergic systems (Iijima et al., 2009). In particular, several 5-hydroxytryptamine (serotonin) receptors were expressed highly by most excitatory neurons in the occipital cortex (Beliveau et al., 2017; Watakabe et al., 2009). Therefore, this positive correlation may have appeared as an association between the activity of the occipital cortex and the activity of the serotonin system.
In conclusion, the soil-mixing activity with M. vaccae showed an increased level of organic acids (including lactic and pyruvic acids and serotonin) and decreased levels of fatty acids and lipids (including palmitic and linoleic acids). This could contribute to improvements in psychological health. In addition, the alpha wave of the occipital cortex increased, and the heart rate decreased, resulting in a stabilizing effect on the autonomic nervous system. This study suggests that human contact with and exposure to soil and microbes correlates with human health for both psychological and therapeutic aspects. This physiological response appeared immediately, and it could be predicted that this is because the M. vaccae was introduced through the oral cavity and upper respiratory tract, which are the main ports for microorganisms to enter the human body (Macovei et al., 2015).
The main limitations of this study are the small sample size and lack of representativeness of the sample because the participants were mostly healthy women. However, this study is valuable because it provided new data that could fill the void presented in previous studies by measuring the effects of soil-mixing activity on human psycho-physiological aspects, depending on the presence or absence of soil microorganisms. This study contributes to propose the role of soil microorganisms in the therapeutic effects of contact with soil and humans.
In the future, the impact of various soil microorganisms on human physiological health should be explored. In addition, further studies should be conducted to develop soil with therapeutic functions for relevant participants, such as older adults and people with disabilities, and to develop long-term programs using soil for maintaining and promoting the health of healthy participants.
Experiment protocol used in this experiment to investigate the effects of the Mycobacterium vaccae soil on the physiological responses of humans during the soil-mixing activity.
(A) Principal component analysis and (B) partial least-square discriminant analysis score plot derived from solid-phase microextraction gas chromatography–time-of-flight–mass spectrometry datasets for various soil samples; soil treated with distilled water [S (λ)]; soil treated with culture media [SM (λ)]; soil inoculated with Mycobacterium vaccae [SMV (λ)]. (C) Heat map analysis for the relative abundance of different volatile organic compounds (variable importance in the projection > 1.0, P < 0.05) derived from the gas chromatography–time-of-flight–mass spectrometry analysis. The colored squares (blue to red) indicate fold changes that are normalized by the average of each metabolite; N.I. = nonidentified.
(A) Principal component analysis and (B) partial least-square discriminant analysis score plot derived from the gas chromatography–time-of-flight–mass spectrometry datasets for serum samples; control [soil treated with culture media (λ)]; treatment [soil treated with Mycobacterium vaccae (λ)]. (C) Heat map analysis for the relative abundance of different serum metabolites (variable importance in the projection > 1.0) derived from the gas chromatography–time-of-flight–mass spectrometry analysis. The colored squares (blue to red) indicate fold changes that were normalized by the average of each metabolite. *Significantly different metabolite between control and treatment groups (P < 0.05, Student’s t test); N.I. = nonidentified.
Correlation analysis between physiological measurement and significantly altered serum metabolites. Each square indicates the Pearson’s correlation coefficient values (r). The red and blue colors represent positive (0.0 < r < 0.3) and negative (−0.3 < r < 0.0) correlations, respectively. *P < 0.05; RFA = relative fast alpha power; ASEF50 = spectral edge frequency 50% of alpha; bps = beats per minute; LF = low frequency band; HF = high frequency band; SDNN = standard deviation of RR intervals (where R is a point corresponding to the peak of the QRS complex of the electrocardiography wave); BDNF = brain-derived neurotrophic factor.
Results of brain-derived neurotrophic factor, according to the presence and absence of Mycobacterium vaccae in the soil during the soil-mixing activity.
Contributor Notes
This work was performed with the support of the National Research Foundation of Korea grant funded by the Korean government (Project title: Psychological and emotional control effects and mechanisms of plant and soil-derived metabolites in horticultural activity; Project No. 2019R1F1A105441113). This paper was supported by the KU Research Professor Program of Konkuk University.
S.-O.K. and S.Y.S. contributed equally to this work.
S.-A.P. is the corresponding author. E-mail: sapark42@konkuk.ac.kr.
Experiment protocol used in this experiment to investigate the effects of the Mycobacterium vaccae soil on the physiological responses of humans during the soil-mixing activity.
(A) Principal component analysis and (B) partial least-square discriminant analysis score plot derived from solid-phase microextraction gas chromatography–time-of-flight–mass spectrometry datasets for various soil samples; soil treated with distilled water [S (λ)]; soil treated with culture media [SM (λ)]; soil inoculated with Mycobacterium vaccae [SMV (λ)]. (C) Heat map analysis for the relative abundance of different volatile organic compounds (variable importance in the projection > 1.0, P < 0.05) derived from the gas chromatography–time-of-flight–mass spectrometry analysis. The colored squares (blue to red) indicate fold changes that are normalized by the average of each metabolite; N.I. = nonidentified.
(A) Principal component analysis and (B) partial least-square discriminant analysis score plot derived from the gas chromatography–time-of-flight–mass spectrometry datasets for serum samples; control [soil treated with culture media (λ)]; treatment [soil treated with Mycobacterium vaccae (λ)]. (C) Heat map analysis for the relative abundance of different serum metabolites (variable importance in the projection > 1.0) derived from the gas chromatography–time-of-flight–mass spectrometry analysis. The colored squares (blue to red) indicate fold changes that were normalized by the average of each metabolite. *Significantly different metabolite between control and treatment groups (P < 0.05, Student’s t test); N.I. = nonidentified.
Correlation analysis between physiological measurement and significantly altered serum metabolites. Each square indicates the Pearson’s correlation coefficient values (r). The red and blue colors represent positive (0.0 < r < 0.3) and negative (−0.3 < r < 0.0) correlations, respectively. *P < 0.05; RFA = relative fast alpha power; ASEF50 = spectral edge frequency 50% of alpha; bps = beats per minute; LF = low frequency band; HF = high frequency band; SDNN = standard deviation of RR intervals (where R is a point corresponding to the peak of the QRS complex of the electrocardiography wave); BDNF = brain-derived neurotrophic factor.
Results of brain-derived neurotrophic factor, according to the presence and absence of Mycobacterium vaccae in the soil during the soil-mixing activity.
