Social isolation exacerbates acute ozone inhalation induced pulmonary and systemic health outcomes

Abstract

Psychosocially-stressed individuals might have exacerbated responses to air pollution exposure. Acute ozone exposure activates the neuroendocrine stress response leading to systemic metabolic and lung inflammatory changes. We hypothesized chronic mild stress (CS) and/or social isolation (SI) would cause neuroendocrine, inflammatory, and metabolic phenotypes that would be exacerbated by an acute ozone exposure. Male 5-week-old Wistar-Kyoto rats were randomly assigned into 3 groups: no stress (NS) (pair-housed, regular-handling); SI (single-housed, minimal-handling); CS (single-housed, subjected to mild unpredicted-randomized stressors [restraint-1h, tilted cage-1h, shaking-1h, intermittent noise-6h, and predator odor-1h], 1-stressor/day*5-days/week*8-weeks. All animals then 13-week-old were subsequently exposed to filtered-air or ozone (0.8-ppm) for 4h and immediately necropsied. CS, but not SI animals had increased adrenal weights. However, relative to NS, both CS and SI had lower circulating luteinizing hormone, prolactin, and follicle-stimulating hormone regardless of exposure (SI>CS), and only CS demonstrated lower thyroid-stimulating hormone levels. SI caused more severe systemic inflammation than CS, as evidenced by higher circulating cytokines and cholesterol. Ozone exposure increased urine corticosterone and catecholamine metabolites with no significant stressor effect. Ozone-induced lung injury/inflammation, and increases in lavage-fluid neutrophils and IL-6, were exacerbated by SI. Ozone severely lowered circulating thyroid-stimulating hormone, prolactin, and luteinizing hormone in all groups and exacerbated systemic inflammation in SI. Ozone-induced increases in serum glucose, leptin, and triglycerides were consistent across stressors; however, increases in cholesterol were exacerbated by SI. Collectively, psychosocial stressors, especially SI, affected the neuroendocrine system and induced adverse metabolic and inflammatory effects that were exacerbated by ozone exposure.

1. Introduction

Environmental exposures exacerbate preexistent disease in vulnerable populations. Air pollution accounts for nearly 70% of all environmental causes of human mortality and morbidity (Landrigan et al., 2018) and now is the 4^th leading risk factor for death (HEI, 2020) worldwide. The health effects of air pollutants are not restricted to cardiopulmonary outcomes; but also associated with chronic diseases including diabetes (McAlexander et al., 2021), exacerbation of atherosclerosis (Bevan et al, 2021), and neurological abnormalities such as Parkinson’s disease (Levesque et al., 2011), and Alzheimer’s diseases (Rhew et al., 2021). Further, they have also been associated with psychiatric conditions such as depression (Deng et al., 2021), post-traumatic stress disorder (PTSD) (Remch et al., 2018), cognitive decline, and even criminal behavior (Ailshire and Crimmins, 2014; Berman et al., 2019). More importantly, individuals who suffer from these diseases have exacerbated peripheral disease outcomes when exposed to air pollutants (Paul et al., 2018; Biessels and Despa, 2019). Understanding how air pollutant exposure exacerbates stress-related mental health crises in vulnerable populations with psychosocial stresses, and especially those with chronic metabolic disease comorbidities will be critical for regulatory decision making and health interventions.

Psychosocial stress has been a major contributor to mental health crises and peripheral diseases involving cardiometabolic and systemic inflammatory conditions and reproductive abnormalities (Herman et al., 2020). Incidentally, communities with poor socioeconomic conditions are more likely to be chronically stressed and be in areas of increased air pollution such as near major roadways or industrial sites, leading to elevated exposure to air pollution (Ou et al., 2018). These communities encounter poor housing conditions, increased physical work-related demands, uncomfortable social encounters, increased exposure to noise, fear, and overall exposure to stressors. Higher levels of psychosocial stresses have been associated with air pollution in epidemiological studies (Hajat et al., 2019; Laratta et al., 2020). Because of socioeconomic hardship, these communities suffer from increased incidence of mental health crises including anxiety, depression and cognitive decline, and greater chronic diseases such as hypertension, diabetes, asthma, chronic pulmonary disease and even reproductive abnormalities (Oraka et al., 2009; Ailshire et al., 2017).

Social isolation (SI), one of the contributing factors to chronic stress, has become a norm for day-to-day life for a large fraction of the population and is linked to mental health crises in young and old generations, especially during the COVID-19 pandemic (Gordon et al., 2021; Klyne et al., 2021; Wilkialis et al., 2021; Smith et al., 2020; Śniadach et al., 2021; O’Regan et al., 2021). COVID-related SI has led to increased incidence of Alzheimer’s disease, dementia, depression, poor sleep quality, and exacerbation of many psychiatric conditions, such as schizophrenia and psychosis (Varchmin et al., 2021; Attademo et al., 2021). Mechanistic evidence points to an interactive network linking stress-related disorders and central and peripheral disease susceptibility (Herman et al., 2020; McEwen and Akil, 2020). The limbic system, alongside other critical brain regions, perceive emotion-related stress signals, which are relayed to various brain structures including hypothalamus to orchestrate peripheral changes through neuroendocrine system (Wilkialis et al., 2021). This is the same central neuroendocrine system involved in responding to air pollution stressors (Kodavanti 2019). Thus, the interactive influence of multiple stressors (perceived and environmental) may amplify peripheral responses and pathologies.

Although association studies discussed earlier have clearly established links between poor socioeconomic conditions, stress, neuropsychiatric disorders and chronic peripheral diseases, the biological mechanisms of interactive effects of psychosocial stressors and air pollutants are not well understood. The contribution of the hypothalamic stress response system and circulating levels of stress hormones are implicated in psychosocial stresses (Dziurkowska and Wesolowski, 2021; McEwen and Akil, 2020). Animal models of chronic stress and social isolation are used to understand the mechanistic link between stress and disease susceptibility (Mumtaz et al., 2018; Dziedzicka-Wasylewska et al., 2021; Willner et al., 2019). The most frequently employed animal model for chronic mild stress include daily random application of variable stressors such as restraint, placement in tilted cages, shaking, exposure to noise, fox-urine chemical exposure for 8 to 12 weeks in rodents (Willner, 2005; Monteiro, 2015; Bondi et al 2007). However, there are no standardized stress protocols developed for creating a specific neurological deficit. The social isolation condition in animals can be modeled through single-housing animals without providing enrichment and avoidance of significant handling during the protocol to understand its impact on disease susceptibility (Mumtaz et al., 2018).

We have shown that the air pollutant ozone induces a classical acute stress response characterized by the activation of sympathetic adrenal medullary (SAM) and hypothalamic pituitary adrenal (HPA) axes to release catecholamines and glucocorticoids into the circulation and mediate peripheral metabolic and immunological effects (Snow et al., 2018; Kodavanti, 2019). Thus, ozone exposure allows us to study the interaction between non-chemical (psychosocial) stresses and environmental exposures (chemical stressor) affecting susceptibility to central and peripheral metabolic and inflammatory conditions. In this study, we hypothesized that underlying chronic mild stress and/or social isolation in rat models would be associated with exacerbated response to acute ozone exposure leading to stress-associated changes in peripheral pathological indices. Using Wistar-Kyoto rats that possess an underlying depression phenotype (Raghavan et al., 2018; Willner et al., 2019), we employed mild chronic stress and social isolation protocols for 8 weeks starting at 5-week of age, and then at 13 weeks of age to examine how underlying stressors effects such as, neuroendocrine, pulmonary, and systemic might be exacerbated by acute ozone exposure.

2. Materials and Methods

2.1. Animals

Male Wistar-Kyoto (WKY) rats 12-week-old for pilot study and 4-week-old for main study were purchased from Charles River Laboratories, Inc, Raleigh, NC. During 1-week acclimation, animals were double-housed in cages with hardwood chip bedding and EnviroDry^R enrichment material. The animal rooms were maintained at 21-22°C, ~55% relative humidity, and 12h light/dark cycle. Throughout acclimation and experimentation period, animals were fed Purina 5001 rat chow (Brentwood, MO) and drank tap water, ad libitum unless noted. At the beginning of experimentation, rats were randomized by body weight and allocated to each stress condition. EPA animal facility is approved by the Association for Assessment and Accreditation of Laboratory Animal Care. All procedures were approved by U.S. EPA’s Institutional Animal Care and Use Committee following the experimental protocol and guidelines of the National Institutes of Health for the care and use of rats (NIH Publications No. 8023).

2.2. Pilot study assessing the impact of single application of each stressor

A pilot study was performed using a separate cohort of 12–13-week-old rats to determine effects of individual stressors on WKY male rats and standardize the chronic stress model for subsequent experimentation. Following one week acclimation period in our animal facility, each group of animals were subjected to each of the following stressors and, within 15 min following stress application, were euthanized for sample collection as described below. During the (1 hour) stressor protocol no water and food were provided except for the ( 6 hour) exposure to noise.

1. Restraint: Animals were placed in nose-only inhalation exposure tubes that were arranged on a rack for 1h. To ensure the rats were immobilized and unable to turn around in the tube, foam pieces were added in the posterior side to decrease the length of the tube when needed.

2. Tilted cage: Rats (1/cage) were placed in a cage tilted at 45° for 1h. The cages had a wire mesh bottom for added grip but without bedding.

3. Shaking: Rats (4/cage) were placed in a cage with a clear plastic divider that separates the cage into 4 equal quadrants. Three cages (n= 12 total rats) were anchored on a modified orbital plate shaker set at 100 rpm for 1h.

4. Noise: Intermittent white noise of 85 dB was broadcasted from speakers located above each individual cage. A timer was set to vary and on-off times of the noise, from 5 to 25 min with 5 to 45 min between noise bursts for a total time of 6h.

5. Predator odor: Rats were exposed to a predator odor (2,5-dihydro-2,4,5-trimethylthiazoline, a chemical isolated from fox urine; used for inducing fear in rodents; Srq Bio Inc., Tallevast, FL) for 1h. Since this chemical is so volatile, a small amount can be placed on gauze pads in an open petri dish placed in the middle of the animal room. Filtertops placed on the rack were removed to avoid filtering of odor and passage of unfiltered room air into the cages. The room was maintained at negative pressure to avoid spreading of odor into other areas.

2.3. No Stress (NS; control group) and Chronic Stress (CS) Models

NS and CS protocols began at five-weeks of age in male WKY rats. Rats in the Chronic Stress (CS) group were single-housed with no enrichment provided (Figure 2). These CS animals were subjected once a day to the above listed mild stressors 5 days per week Monday-Friday for 8 weeks. On the final week, stressor applications were continued over the weekend until the day before air or ozone exposure and necropsy. The order in which these stressors applied was randomized on weekly basis to decrease the chance of adaptation to each individual stressor. CS animals were handled daily for stressor manipulations and periodic urine collection as well as body composition assessments during 8-week protocol. Control group (no stress, NS) animals were double-housed with no stress protocol employed but animals were weighed weekly and handled periodically for urine collection and body composition assessments during 8-week normal housing. These animals were provided EnviroDry^R enrichment material for nesting (crinkled paper).

Figure 2.

Experimental protocol for stressors application and temporal assessment of body weight, body composition as well as urinary corticosterone until prior to air or ozone exposure. A. Experimental design. Variable unpredicted stresses were applied over 8 weeks (from 5 to 13 weeks of age) in male WKY rats. Responsiveness to a single ozone inhalation exposure was evaluated upon the termination of stress protocols. Necropsy and tissue collection were carried out immediately after air or ozone exposure (within 2 hr). No stress (NS) animals were double housed with crinkled paper provided in cages for environmental enrichment and were handled regularly for body weights, body composition and urine sample collection, but no stressors were applied. For chronic stress (CS), animals were single-housed without environmental enrichment, and each stressor was applied in a random manner Monday-Friday, 5 days/week for 8 weeks. These rats were handled daily. Social isolation (SI) animals were single-housed without environmental enrichment with limited handling only for weighing and cage changes. Temporal changes in body weight (B), body fat % assessed through body composition analysis (C) and urine corticosterone levels (D) during 8-week of stress maneuver in NS, CS and SI groups. Body weights were measured each week for all animals (B). For NS and CS body composition was assessed at week 4 and prior to week 8 (C). Note that body composition was not assessed for SI. Urine corticosterone was measured in NS and CS after the 4^th week of stress protocol and after the last stress application on the 7^th week of the protocol (Friday of week 7), and finally after the weekend (Monday at the start of week 8) on Monday to determine the reversibility of stressor effect observed on Friday. Data were normalized to creatinine levels (D). Curves values indicate means ± SEM for n=24 animals/group. An “*” indicates significant ozone effect relative to matched air group. A (“†” indicates significant stressor effect relative to matched NS group (p ≤ 0.05).

2.4. Social isolation (SI) model

A separate cohort of 5-week-old male rats was single-housed for SI model. No stressors were applied to the SI group except for single-housing and no enrichment added to the cages to develop an exclusively “psychosocial” stress model. The handling of these animals was avoided except for cage changes and periodic weighing. Therefore, no urine samples were collected for stress hormone assessment during isolation period of 8 weeks.

2.5. Body composition analysis

Body composition and urine collection were assessed only for NS and CS groups during stress protocol. To limit animal handling SI group was not assessed. Body composition was analyzed prior to starting the stress regimen (~4 weeks of age), at 4 weeks of stress protocol (~8 weeks of age) and few days prior to completion of 8-week stress protocol (rounded to 8 weeks, ~12-13 weeks of age). For body composition analysis we used a Bruker Minispec LF90 II TD-NMR body composition analyzer (Bruker Optics, Inc., Billerica, MA) and relative percentages of lean body mass, body fat, and body fluid were assessed as previously described (Gordon et al., 2017).

2.6. Urine collection

For NS and CS groups urine samples were collected immediately following stressor application at the end of 4-week of stress protocol on Friday. Urine samples were also collected at the end of week 7 immediately following stressor application as well as at the beginning of week 8 on Monday prior to stressor application to determine the chronicity of corticosterone changes with no stressor application for 2 days during weekend. To collect urine samples animals were placed in a clean cage without the bedding and allowed to urinate. The urine samples were transferred in a clean microcentrifuge tube and stored at −80 °C until analysis.

2.7. Ozone generation and animal exposures

Because ozone is an air pollutant known to induce a classical stress response, ozone exposure was employed to determine the interactive effects of psychosocial stressors and air pollution. NS, CS and SI animals were exposed to air or ozone after 8 weeks of stress protocol. A silent arc discharge generator (OREC, Phoenix, AZ) generated ozone from oxygen. Mass flow controllers were used to regulate the entry of ozone into the Rochester style “Hinners” chambers. Photometric ozone analyzers (API Model 400) were used to monitor the ozone concentrations in the chambers. Rats were exposed to filtered air or ozone (0.8 ppm) for 4h followed immediately by a necropsy to collect tissues. This concentration of ozone is an order of magnitude higher than what is generally achieved in non-attainment areas in the United States (de Foy et al., 2020), however this concentration in resting rats is comparable to the concentrations used in human clinical studies done using intermittent exercise (Hatch et al., 1994). Exposure chamber conditions and actual ozone concentrations achieved were monitored during each exposure. Actual average chamber ozone concentration achieved was 0.804±0.006 ppm (mean ± standard deviation). The chamber temperature (°C), relative humidity (RH) and airflow (liters/minute) for air control chamber were 23.73 ± 0.17, 45.34 ± 0.89, and 331.03 ± 0.15 and for ozone chamber were 23.64 ±0.02, 43.93 ± 0.20, and 335.38 ± 0.39, respectively. During exposure, all rats were individually housed in wire-mesh cages. Food and water were not available during exposures.

2.8. Necropsy and samples collection

To assure consistent diurnal timing for stress sensitive endpoints, necropsies were performed at the same circadian time for each study. Following application of a single stress protocol in a pilot study or a single 4h air or ozone exposure for NS, CS and SI groups (n=12 animals/group), necropsies were performed. Animals were fasted for 4-6 hours before necropsy (only during the duration of exposure and until necropsies were completed). Exposures were done in a staggered manner over several days to accommodate for necropsies which were carried out in no more than 2 hours after exposure ended. Intraperitoneal injections of sodium pentobarbital (Fatal Plus, Virbac AH, Inc., Fort Worth, TX; >200 mg/kg, intraperitoneal) was used for euthanasia. Blood samples were collected in glass EDTA and serum separator vacutainer tubes from the abdominal aorta. Urine samples were collected directly from the bladder using a syringe and stored at −80 °C for stress hormone analysis. Blood glucose levels were assessed using a Bayer Contour glucometer (Bayer Corp., Leverkusen, Germany). Complete blood count was performed using EDTA blood tube on a Beckman-Coulter AcT blood analyzer (Beckman-Coulter Inc., Fullerton, California). EDTA-containing and serum separator tubes were centrifuged (3500 x g for 10 min) and plasma/serum samples were aliquoted and stored at −80 °C . As described previously (Miller et al., 2015; Henriquez et al., 2018), right lung was lavaged with Ca²⁺ and Mg²⁺ free PBS at 37 °C to collect lavage fluid. Adrenal, thymus, and spleen weights were recorded, and several tissue samples including brain regions were collected and stored for subsequent analysis. The data for blood lymphocyte count, serum adrenocorticotropic hormone, serum corticosterone, and brain regions are reported in a companion paper (Valdez et al., 2022).

2.9. Assessment of lung injury and inflammation using bronchoalveolar lavage fluid (BALF)

The number of cells present in the whole BALF were counted using Z1 Coulter Counter (Coulter, Inc., Miami, FL). The cell differentials were performed by preparing cytospin slides using Shandon cytocentrifuge (Thermo Fisher Scientific, Waltham, MA) from whole BALF and when dry, staining with Diff-Quik (Thermo Fisher Scientific). Neutrophils, macrophages, and other cells were identified and differentially counted under light microscopy (>300 cells/slide). Whole BALF was then spun and the cell-free BALF used for determination of injury markers. Total protein was assessed using Coomassie plus Protein Assay Kit (Pierce, Rockford, IL); albumin was assessed using kits from (DiaSorin (Stillwater, MN), and N-acetyl-β-D-glucosaminidase (NAG) activity was assessed using kit from Roche Diagnostics (Indianapolis IN). These clinical assays were adapted for use on a clinical analyzer, Konelab Arena 30 (Thermo Chemical Lab Systems, Espoo, Finland).

2.10. Assessment of BALF, serum and urine markers

BALF IL-6 and TNF-α were analyzed using rat custom V-PLEX kit (Mesoscale Discovery Inc., Rockville, MD). Cytokines in serum were analyzed using the V-PLEX proinflammatory panel 2 (rat) from Mesoscale Discovery using the manufacturers protocol (Rockville, MD). The MESO QuickPlex SQ 120 platform was used to detect electrochemiluminescence signals (Mesoscale Discovery Inc., Rockville, MD). For samples where IL-1β or IL-5 were below the limit of detection, values were substituted with the lowest quantified value for the given cytokine across all stressors/exposures. Serum levels of anterior pituitary hormones including thyroid stimulating hormone (TSH); gonadotrophin releasing hormone (GnRH), prolactin (PRL); luteinizing hormone (LH); and follicle stimulating hormone (FSH) were measured using MILLIPLEX MAP Rat Pituitary Magnetic Bead Panel (Merck-Millipore, Burlington, MA). Urine corticosterone levels were measured using kits from Arbor Assays (Ann Arbor, MI) following manufacturer’s protocol. Urine levels of normetanephrine and metanephrine were determined using immunoassay kits from Eagle Biosciences (Amhurst, NH) following manufacturer’s protocols. Serum insulin levels were determined using a rat-specific chemiluminescence assay kit (Millipore, Billerica, MA) following the manufacturer’s instructions. Serum leptin levels were analyzed using kits and protocols from Mesoscale Discovery® (Rockville, MD). Branched chain amino acid levels in serum were analyzed using ELISA kit from Abcam (Cambridge, MA) following manufacturers protocol. ELISA plates were read on a SpectraMax i3x Multi-Mode Microplate Reader (Molecular Devices, San Jose, CA). Total cholesterol and triglycerides were analyzed in the serum samples using kits from TECO Diagnostics (Anaheim, CA) and these assays were modified for use on the Konelab Arena 30 system (Thermo LabSystems, Espoo, Finland). Urine creatinine was analyzed following manufactures protocol using kit from Arbor Assays (Ann Arbor, MI). Urine corticosterone, metanephrine and nor-metanephrine data were normalized to creatinine levels.

2.11. Statistics

Normality was assessed using Shapiro-Wilk test and, where necessary, data were log-transformed before analysis of variance. All transformed data were subsequently normally distributed. For some endpoints, outliers were identified and removed using robust regression and outlier removal (ROUT) with a false discovery rate of 1%. For the pilot study, significant effects of stressors were determined using 1-way ANOVA, comparisons between stressed and non-stressed group were carried out using the Holm-Sidak multiple comparison test. For the chronic stressor study, body composition data were compared using unpaired t-test and urine corticosterone was analyzed using mixed-effects analysis (stressor, time) with multiple comparisons corrected for using Sidak’s method. For all other analyses in the chronic stressor study, significant differences were determined using a 2-way ANOVA (exposure [air, ozone], stressor [NS, CS, SI]). Multiple comparisons were carried out using the Tukey’s multiple comparison test. For all tests, group differences were considered significant when p values were ≤ 0.05. GraphPad prism 9.1.2 software was used for statistical analysis and graph design.

3. Results

3.1. Effects of a single application of each stressor on white blood cells, glucose, stress hormones and pituitary neurohormones.

To assess the acute effect of each stressor without habituation when applied for the first time, we measured anterior pituitary, posterior pituitary, and adrenal-derived hormones in the serum along with circulating white blood cells and lymphocytes (Figure 1) as markers of acute stress (Henriquez et al., 2021). Each of the stress conditions we employed has been previously validated in rodent studies (Monteiro et al., 2015; Willner, 2005; Willner et al., 2019). Because each stressor could cause differential effects (Figure 1A), this assessment was intended to determine appropriate stressor application for the mild chronic stress model using male WKY rats. WKY rats have been used as a model of treatment resistant depression with high baseline levels of circulating corticosterone (Willner et al., 2019).

Figure 1.

Changes in stress associated biomarkers immediately after a single application of each individual stressor for the pilot study in 12-13 week old male WKY rats. A) experimental protocol; B) body weight; C) plasma epinephrine; D) plasma corticosterone; E) plasma norepinephrine; F) serum oxytocin; G) serum arginine-vasopressin; H) circulating white blood cells; I) circulating lymphocytes; J) blood glucose. Each group of rats were subjected to one of the stressors, and immediately following stressor application, rats were euthanized for biomarkers assessment. The groups included were: 1) No stress- control, 2) restraint: rats are placed in nose-only inhalation tube for 1-hr; 3) tilted cage: each rat is placed in a 45° tilted in wire-mesh bottom cage for 1-hr (C); 4) noise: rats are exposed to white noise, of 85 dB through speakers above each cage for 5-25 min, with 5-45 min between noise bursts for 6-hrs; 5) predator odor: rats are housed in an isolated room where fox urine odorous ingredient, 2,5-dihydro-2,4,5-trimethylthiazoline is allowed to dissipate through an open petri dish, 1-hr; and 6) shaking: each rat is placed in cage with 4 dividers, placed on a modified orbital plate shaker set at 100 rpm for 1-hr. Each graph shows median, 25^th-75^th percentiles and min-max for n=10 animals/group for no-stress controls and n=8 for each stressor. A “†” indicates significant stressor effect relative to no stress control (p ≤ 0.05). One urine sample was not collected for CS group. One animal each from control, restraint and noise group for epinephrine and one from control group for oxytocin were removed by outlier test.

A single application of stressors such as restraint, tilted cage, noise, predator odor, or shaking did not affect body weight (Figure 1B). No changes in circulating epinephrine (Figure 1C) or corticosterone (Figure 1D) were noted with any of the stressor applications. Noise exposure occurred for six hours as opposed to 1 hour for all other stressors; however, the onset of stressor application was adjusted such that all animals were necropsied at the same time of the day to avoid any differences in hormone levels due to diurnal differences. Plasma levels of nor-epinephrine were lower in animals with restraint and shaking (Figure 1E), whereas levels of oxytocin and arginine-vasopressin were variable and not consistently changed by any stressor type (Figure 1F-G). Circulating white blood cells and lymphocytes were depleted after restraint, tilted cage and shaking maneuvers, with no changes in blood glucose noted (Figure 1H-J). While changes in many of the assessed markers might require more than an hour to change, and stress hormones can be variable based on the handling stress and activity status, these results collectively demonstrate that these stressors had a variable influence on markers of stress response previously shown to be altered by ozone exposure (Henriquez et al., 2021).

3.2. Body weight, composition and urine corticosterone during 8-week stressor application

The protocol for randomized application of 5 stressors for 8 weeks is depicted in Figure 2A. Prior to air or ozone exposure, and during chronic stress application, animals were weighed weekly, and body composition and urine corticosterone were assessed at week 4, week 8. No animal handling was done for SI cohort of rats for these 8 weeks except for body weight assessment and cage changes. Analysis of body weight gain found a small but significant reduction in the CS group subjected to random daily stressor compared to NS or SI during week 8 prior to final week of stressor application (Figure 2B). However, body fat % did not change significantly due to CS relative to NS (Figure 2C). Urine corticosterone was assessed immediately post-stressor on Friday at the end of 4 and 7, and after the weekend recovery at the beginning of week 8 (Monday). CS animals had higher levels of corticosterone at week 4 and 8 (Week 8 > week 4). After weekend of no stressor application, the levels of urine corticosterone were still slightly elevated in the CS group relative to NS (Figure 2D).

3.3. Body and organ weights at necropsy

Although body weights of CS animals were slightly lower by 5 weeks of stressor application and continuing until the end of 8-week stress protocol (Figure 2B), when measured after 4-hour air or ozone exposure prior to necropsy, weights did not differ significantly between stress or exposure groups (Figure 3A). Weights of adrenals and lymphatic organs were assessed to determine stress impact on relevant organs. Adrenal weights were significantly higher in CS relative to NS and SI groups regardless of ozone exposure (Figure 3B). Thymus weight was not changed by any prior stressor application; however, in ozone-exposed SI animals, the weight was significantly lower when compared to air-exposed SI group (Figure 3C). For spleen, weights were not changed by exposure nor by any stress conditions (Figure 3D).

Figure 3.

Body and organ weights at necropsy in no-stress (NS), chronic stress (CS) and social isolation (SI) groups after exposure to air or ozone. A) body weights at necropsy; B) adrenal weights; C) thymus weights; D) spleen weights. Data are shown as mean ± SEM for n=12 animals per group. An “*” indicates significant ozone effect relative to matched air group. A “†” indicates significant stressor effect relative to matched exposure NS group (p ≤ 0.05).

3.4. Effects of stressors and ozone on urinary catecholamine metabolites and corticosterone

Urine samples collected from the bladder during necropsy were analyzed to determine the levels of corticosterone as well as metanephrine and normetanephrine as an index of catecholaminergic changes. While ozone exposure increased levels of urine corticosterone/creatinine ratio as expected, the air-exposed CS did not differ significantly from air-exposed NS. Marked increases were noted in levels of urine corticosterone in all ozone-exposed animals regardless of stressor application (Figure 4A). The levels of metanephrine only increased in ozone-exposed animals in the CS and SI groups (Figure 4B). For normetanephrine, a similar trend was observed, but significant increases were observed after ozone exposure only in the CS group (Figure 4C).

Figure 4.

Acute exposure to ozone increased urine levels of stress hormones in no-stress (NS), chronic stress (CS) and social isolation (SI) groups. A) urine corticosterone; B) urine metanephrine; C) urine normetanephrine. Urine samples were collected directly from urinary bladder at the time of necropsy. Data were normalized to urine creatinine levels. Data are shown as mean ± SEM for n=12 animals per group. An “*” indicates significant ozone effect relative to matched air group (p ≤ 0.05). Corticosterone data for one animal in SI air group; and metanephrine data for one animal in NS air group and one in NS ozone group were removed by outlier test.

3.5. Lung injury, inflammation and proinflammatory cytokines markers in the bronchoalveolar lavage fluid

Lung injury markers were analyzed in BALF immediately following 4h air or ozone exposure. BALF protein and albumin were increased significantly in all ozone-exposed animals regardless of stress protocol; the ozone-induced increases in these markers were significantly enhanced in SI relative to CS or NS groups (Figure 5A and B). BALF N-acetyl-β-D-glucosaminidase (NAG) activity reflective of macrophage activation was significantly increased in all ozone exposed groups; this increase was exacerbated in SI when compared to NS and CS groups (Figure 5C). BALF neutrophils assessed as an innate immune response were increased after ozone exposure in all stress conditions; and exacerbated in SI animals relative to NS and CS (Figure 5D). Proinflammatory cytokines assessed in BALF, IL-6 and TNF-α were increased in BAF after ozone exposure regardless of stress condition (Figure 5E and F). The ozone-induced increase of IL-6 in SI group was significantly greater when compared to ozone-exposed NS and CS groups These data showed that SI was associated with exacerbated ozone-induced lung injury and inflammation when compared to CS and NS groups. While ozone exposure was associated with lower circulating white blood cells relative to respective air controls, this effect was independent of either stress condition (Figure 5G).

Figure 5.

BALF markers of lung injury, inflammation and circulating white blood cells after a single acute ozone exposure in no-stress (NS), chronic stress (CS) and social isolation (SI) groups of animals. Lung injury and inflammation were determined in the bronchoalveolar lavage fluid (BALF) collected at the time of necropsy. A) BALF protein; B) BALF albumin; C) BALF NAG (N-acetyl-β-D-glucosaminidase) activity; D) BALF neutrophils; E) BALF IL-6; F) BALF TNF-α; G) circulating white blood cells. Data are shown as mean ± SEM for n=12 animals per group. An “*” indicates significant ozone effect relative to matched air group. A “†” indicates significant stress effect relative to matched exposure NS group. A “φ” indicates significant SI effect relative to matched exposure CS group (p ≤ 0.05). BALF protein and albumin data for two animals in NS air group; one in CS air group; one in SI air group; one for BALF protein in NS ozone group, one for BALF NAG in NS ozone group; one BALF IL-6 in CS ozone group and one for TNF-α in SI ozone group were removed by outlier test.

3.6. Systemic inflammation

To determine if the application of CS and/or SI protocols during young age was associated with systemic inflammation or enhancement of systemic inflammatory responses induced by ozone, we next analyzed a panel of proinflammatory cytokines in the serum of animals subjected to NS, CS or SI and exposed to air or ozone. IL-1β, involved in the activation of inflammasome cascade was significantly higher in air as well as ozone-exposed SI animals relative to NS; ozone did not exacerbate this increase in SI. Ozone exposure did not increase this cytokine in NS or any other stress condition (Figure 6A). Small but significantly higher levels of circulating IL-6 were observed in air-exposed SI group relative to NS and CS (Figure 6B). Ozone exposure significantly exacerbated the serum IL-6 increase but only in SI group. IL-4 levels were higher in ozone-exposed SI group relative to NS and CS (Figure 6C). IL-5 levels were higher in ozone-exposed CS group when compared to ozone-exposed NS group. IL-5 levels were also higher in SI animals relative to NS (regardless of air or ozone) (Figure 6D). IL-10, a cytokine involved in repair phase of acute injury and IL-13 involved in allergic phenotype and mucus hypersecretion changed in a similar manner (Figure 6E and F). Both these cytokines were higher in ozone-exposed CS group relative to air-exposed CS group. Also, the levels of these cytokines were higher in both air and ozone-exposed SI animals relative to air or ozone-exposed NS animals. Ozone-exposure exacerbated this effect in SI animals (Figure 6E and F). In SI group IFN-g levels were higher in both air-and ozone-exposed animals when compared to matched NS and CS groups. Furthermore, the effect of ozone in SI was significantly exacerbated relative to air SI group (Figure 6G). The serum levels of KC-GRO were only increased in SI group exposed to ozone compared to all other matching groups (Figure 6H). Consistent with our previous findings (Henriquez et al., 2021), serum levels of TNF-a tended to be lower in ozone-exposed NS and CS groups relative to respective air groups, however, the levels were significantly higher in ozone-exposed SI relative to air group (Figure 6I). These data indicate an elevated systemic inflammation in SI group that is generally exacerbated after an acute ozone exposure. Taken together, these data suggest that social isolation stress, compared to NS controls, results in increased baseline levels of pro-inflammatory cytokines compared, with a further dysregulation effect observed from ozone exposure.

Figure 6.

Circulating cytokine changes indicating systemic inflammation following ozone exposure in no-stress (NS), mild chronic stress (CS) and social isolation (SI) groups. Cytokines were determined in the serum samples collected during necropsy. A) IL-1β; B) IL-6; C) IL-4; D) IL-5; E) IL-10; F) IL-13; G) IFN-γ; H) KC-GRO; I) TNF-α. Data are shown as mean ± SEM for n=12 animals per group. An “*” indicates significant ozone effect relative to matched air group. A “†” indicates significant stress effect relative to matched exposure NS group. A “φ” indicates significant SI effect relative to matched exposure CS group (p ≤ 0.05). Serum IL-1β data for one animal in NS air group; and IL-5 data for two animals in NS air group were removed by outlier test.

3.7. Circulating neuroendocrine hormones

Based on prior evidence that ozone mediates its effects through neuroendocrine activation (Kodavanti, 2019), we next measured pituitary-derived and other relevant hormones in the serum to determine if underlying CS or SI may modulate ozone-induced neuroendocrine response. Circulating levels of TSH were significantly lower in air-exposed CS and SI animals when compared to air-exposed NS. Ozone exposure markedly lowered TSH in all animals including NS, CS and SI (Figure 7A). Serum levels of GnRH were variable and increased only in ozone-exposed SI group when compared to matched NS group (Figure 7B). Serum levels of PRL were depleted by CS and SI in air-exposed animals. Ozone exposure caused >90% depletion of PRL in all groups regardless of stress status (Figure 7C). For LH levels, ozone exposure resulted in marked depletion in all groups regardless of stress status (Figure 7D). Serum levels of FSH was decreased in air-exposed and ozone exposed CS and SI animals when compared to matched NS air group. FSH levels tended to decrease after ozone exposure in NS animals and remained low in CS and SI groups (Figure 7E). These data suggest that underlying CS and SI are associated with altered neuroendocrine activity and exposure to ozone had an inhibitory effect on these hormones.

Figure 7.

Changes in circulating pituitary hormones following ozone exposure in no-stress (NS), mild chronic stress (CS) and social isolation (SI) groups. Pituitary hormones were analyzed in serum samples collected during necropsy. A) Thyroid stimulating hormone (TSH); B), gonadotropin releasing hormone (GnRH); C) prolactin (PRL); D) luteinizing hormone (LH); E) follicle stimulating hormone (FSH). Data are shown as mean ± SEM for n=12 animals per group. An “*” indicates significant ozone effect relative to matched air group. A “†” indicates significant stress effect relative to matched exposure NS group (p ≤ 0.05). Serum PRL data for two animals in each, NS ozone, CS ozone and SI ozone groups; and FSH data for one animal in NS air group and one animal in CS air group were removed by outlier test.

3.8. Circulating metabolites and metabolic hormones

To understand the link between systemic inflammation and chronic disease susceptibility to metabolic disease in stress models, we analyzed several circulating metabolites and metabolic hormones. Blood glucose levels were not changed by any stressor application in air-exposed animals; however, as observed in several of our previous studies (Miller et al., 2015), ozone exposure was associated with significant increases in blood glucose levels in all three stress conditions (Figure 8A). Serum insulin levels were neither changed by any stressor application nor ozone exposure (Figure 8B). Serum levels of leptin were not changed by CS or SI in air-exposed animals, however ozone exposure significantly increased leptin to a similar extent in NS, CS and SI animals (Figure 8C). Since circulating lipids are associated with chronic metabolic disease, we determined circulating cholesterol and triglycerides in all animals. Circulating cholesterol was higher in air-exposed SI animals when compared to air-exposed NS and CS animals. Exposure to ozone led to higher levels of serum cholesterol in all stress conditions and this effect of ozone was significantly exacerbated in SI animals relative to NS and CS groups regardless of exposure (Figure 8D). Circulating triglycerides while not changed by SI or CS relative to NS in air-exposed animals, were significantly higher after ozone exposure in NS and CS but not SI groups (Figure 8E). Finally, since ozone exposure has been shown increase circulating branched-chain amino acids, we determined ozone effects in animals with different stress conditions. Neither CS nor SI air-exposed animals had higher levels of branched-chain amino acids relative to NS group, however significantly higher levels were noted in all ozone exposed animals regardless of stress status (Figure 8F). The magnitude of this response was smaller in CS and SI groups.

Figure 8.

Circulating metabolites and metabolic hormones following ozone exposure in no-stress (NS), mild chronic stress (CS) and social isolation (SI) groups. Serum levels of hormones and metabolites were determined in the samples collected during necropsy. A) blood glucose; B) serum insulin; C) serum leptin; D) serum cholesterol; E) serum triglycerides; F) serum branched chain amino acids (BCAA). Data are shown as mean ± SEM for n=12 animals per group. An “*” indicates significant ozone effect relative to matched air group. A “†” indicates significant stress effect relative to matched exposure NS group. A “φ” indicates significant SI effect relative to matched exposure CS group (p ≤ 0.05). Serum insulin and leptin data for one animal in NS air group; and cholesterol data for one animal in SI air group were removed by outlier test.

4. Discussion

A large proportion of the human population suffers from various mental health conditions attributable to stressful lifestyles (Guidi et al., 2021). Rodent models of CS and SI are employed to understand psychosocial stress-associated human pathobiology and vulnerabilities to subsequent stressor-induced chronic diseases. Since oxidant air pollutants such as ozone also activate the anabolic neuroendocrine stress pathways, we hypothesized that underlying CS (daily mild stress application, frequent handling plus single housing without enrichment) and SI (only single housing without enrichment and no frequent handling) would modify ozone-induced stress responses and associated pulmonary and systemic health outcomes. We show that ozone-induced pulmonary inflammation and injury were exacerbated in SI, but not in CS, relative to NS. Importantly, CS had relatively smaller stress-related systemic changes than SI. SI in air-exposed animals was associated with systemic inflammation as evidenced by higher levels of circulating cytokines, and this effect was exacerbated by ozone. Circulating PRL, FSH and LH were depleted by prior stress (SI>CS) in air-exposed rats. Ozone exposure caused major depletion of TSH, PRL, and LH levels in all groups of animals. Furthermore, ozone-induced increases in glucose, branched chain amino acids, leptin, and triglycerides were not different between stress conditions; however, the SI-induced increase in cholesterol was exacerbated by ozone. These data demonstrate that SI could be a risk factor for systemic inflammation and heightened pulmonary response to ozone likely involving neuroendocrine pathways (Table 1).

Table 1.

Summary of findings related to CS and SI effects in air- or ozone-exposed rats.

Biological process	Endpoint affected by stress condition or ozone	CS effects in air- expose d rats	SI effects in air- expose d rats	Ozone effects in No Stress rats	Ozone effects in CS rats	Ozone effects in SI rats
Stress response	Relative adrenals weights	↑	---	---	↑	---
	Relative thymus weights	---	---	---	---	↓
	Urine corticosterone	---	---	↑	↑	↑
	Urine metanephrine	---	---	---	↑	↑
	Urine normetanephrine	---	---	---	↑	---
Lung injury/inflammati on (BALF changes)	Protein and albumin	---	---	↑	↑	↑↑
	NAG activity, neutrophils count and IL-6	---	---	↑	↑	↑↑
	TNF-α	---	---	↑	↑	↑
Systemic Inflammation (serum WBC and cytokines)	WBC numbers	---	---	↓	↓	↓
	IL-1β	---	↑	---	---	↑
	IL-6	---	↑	---	---	↑↑
	IL-4	---	↑	---	---	↑↑
	IL-5	---	↑	---	↑	↑
	IL-10 and IL-13	---	↑	---	↑	↑↑
	IFN-γ	---	↑	---	---	↑↑
	KC-GRO	---	---	---	---	↑
	TNF-α	---	---	---	↓	↑
Serum pituitary hormones changes	TSH	↓	---	↓	↓	↓
	GnRH	---	---	---	---	↑
	PRL	↓	↓	↓	↓	↓
	LH	---	---	↓	↓	↓
	FSH	↓	↓	---	↓	↓
Metabolic changes	Blood glucose	---	---	↑	↑	↑
	Serum insulin	---	---	---	---	---
	Serum leptin	---	---	↑	↑	↑
	Serum cholesterol	---	↑	↑	↑	↑↑
	Serum triglycerides	---	---	↑	↑	---
	Serum BCAA	---	---	↑	↑	↑

This summary table illustrates changes observed in biological endpoints reflective of stress and pituitary hormone changes, lung injury/inflammation and systemic immune as well as metabolic alterations. No significant effects due to stressor or ozone relative to no stress (NS) air-group are shown by a broken straight line (---), significant increases due to stressor in air-exposed group or due to ozone relative to matched air-exposed group are shown as a single arrow pointing up (↑) for increases and pointing down for decreases (↓). Interactive effects of ozone and stressor showing significantly exacerbated increases are shown as two arrows pointing up (↑↑) and decreases are shown as two arrows pointing down (↓↓). CS = single housed animals without enrichment subjected to daily chronic mild stresses with frequent handling. SI = single housed animals without enrichment or frequent handling. BALF = bronchoalveolar lavage fluid. NAG = N-acetyl-β-D-glucosaminidase. WBC = white blood cells. BCAA = branched chain amino acids. TSH = thyroid stimulating hormone. GnRH = gonadotrophin releasing hormone. PRL = prolactin. LH = luteinizing hormone. FSH = follicle stimulating hormone. Note that the ozone-induced increases in BCAA were significantly smaller in CS and SS groups relative to NS that is not reflective from arrows shown in the table.

Stress hormones, specifically corticosterone in rodents (cortisol in humans), increase in the circulation upon application of a stressor. Restraint stress is the most commonly used physical stressor in animals, but rodent strains differ in their corticosterone response to any given stressor, and pituitary hormones, in addition to ACTH, might be changed upon application of a single stressor. WKY rats, a strain known to possess underlying depression phenotype (Will et al., 2003), did not respond to any acute stressors applied in this study by increasing circulating corticosterone. Interestingly, depletion of circulating lymphocytes and white blood cells, indicators of ozone-induced acute stress effect (Miller et al., 2016a) were significantly depleted by stressors such as restraint, tilted cage, and shaking, indicating that these stressors indeed were effective in producing a mild stress response in WKY rats. However, none of the stressors employed increased circulating catecholamines at necropsy; rather, restraint and shaking-related stresses decreased circulating levels of norepinephrine, suggesting that stressor-specific changes are likely temporal in nature and vary by stressor. Considering the temporality of the stress response (Henriquez et al., 2021, 2022), it is likely that, depending on stressor type and longevity of its application, a single assessment of hormone levels at any given time might not reveal expected changes in stress hormones. Moreover, the lack of increases in restraint-related corticosterone has been reported previously in male WKY rats (Djordjevic et al., 2007) and could be due to an underlying strain-specific abnormality related to specific trigger-induced HPA activation. Since adolescent WKY rats are also susceptible to SI-induced HPA activation and brain monoamine alterations (Shetty and Sadananda, 2017), we further compared CS with only SI to differentiate the role of social isolation from CS.

Rat models of CS vary by type of stressor, intensity, and duration of stressor application, and perhaps most importantly, the rat strain being used; thus, the reported outcomes of CS, in general vary greatly between studies (Atrooz et al., 2021; Willner, 2005). We selected commonly used stressors for rodents, and considering the known depression phenotype inherent to WKY rats (Will et al., 2003), stressors were selected to mimic a mild chronic stress scenario while avoiding overt trauma. The presence of stress in CS and SI was confirmed through increases in urinary corticosterone in rats together with increases in adrenal weights relative to NS. Understanding that CS models might present complex neurological sequalae and stress system alterations in rats might not be consistent with human pathobiology of depression and stress, we exposed rats to ozone to determine how the response to this well-characterized inhaled pollutant that produces pulmonary and peripheral effects through SAM and HPA activation (Henriquez et al., 2018, 2019, 2021, 2022) might be altered by CS or SI in male WKY rats.

Socioeconomically deprived communities in rural areas are likely to suffer hectic lifestyle with chronic physical hardships together with social deprivation and isolation. We attempted to mimic some of those conditions in male WKY rats to better understand how environmental exposure, such as ozone, might exacerbate CS together with SI (single housing) or SI by itself to determine the contribution of each. We hypothesized that CS plus associated SI would cause neuroendocrine, inflammatory, and metabolic dysregulation, and SI would be a contributing factor. Further, we hypothesized that ozone exposure, which causes neuroendocrine stress pathway activation (Snow et al., 2018), would exacerbate stress-associated health outcomes. However, despite increases in urinary corticosterone and adrenal weights, CS in air-exposed animals caused fewer of the systemic inflammatory and metabolic changes compared to air-exposed SI. Relative to NS, CS was evident by small increases in circulating cytokines and cholesterol but these changes were much lower than the effects induced in air-exposed SI. There were trends of increases in circulating IL-6, IL-4, IL-10, and IL-13 in air-exposed CS rats, consistent with prior studies in rodent models of CS (López-López et al., 2017; Demirtaş et al., 2014). It is possible that frequent animal handling and application of the chronic mild stressor paradigm negated the effect of isolation and lack of enrichment. This further emphasizes the complexities of stress response system that also regulates habituation, learning and the response to other environmental cues, and necessitates caution in generalizing interpretations from a single exposure condition or strain.

Health impacts of underlying CS or SI might be unmasked by a single application of a stronger acute stressor. We hypothesized that using a challenge stressor ozone and examining biomarkers of stress-associated immune and metabolic changes could provide useful information on the interactive effects of acute ozone exposure in animals with preexistent CS and SI. As anticipated, and consistent with previous results, ozone exposure in NS group was associated with increases in markers of lung injury and inflammation including increases in cytokines in the BALF and causing systemic lymphopenia (Henriquez et al., 2018; 2021) all these effects were exacerbated in SI animals, but not CS, suggesting that SI was associated with increased susceptibility to developing lung injury and inflammation from exposure to an oxidant air pollutant. Relative to SI, reduced exacerbation of baseline and ozone-induced immune and metabolic health outcomes in CS (also subjected to SI but involving regular handling and application of mild stressors) could imply that reduced social contacts and activity might be specifically injurious to health. It is also likely that SI could have reduced resiliency when compared to CS. The mechanisms for these differences between CS and SI could be highly complex and might involve central neuroendocrine networks regulating learning, resiliency, and habituation. Because social isolation is linked to depression phenotype in humans (Śniadach et al., 2021; Qirtas et al., 2022), the use of WKY animals that are predisposed to neuroendocrine changes linked to depression phenotype (Shetty and Sadananda, 2017) might be relevant to human SI-related psychiatric disorders. Ozone effects are mediated though the activation of neuroendocrine system (Henriquez et al., 2019); therefore, the mechanism by which ozone might exacerbate lung injury, inflammation and systemic alterations in SI could involve depressed corticosterone response. Glucocorticoids are known to regulate the plasticity at the central and peripheral levels (Herman, 2022). However, the precise mechanisms of SI in WKY and ozone inhalation involving diverse neural networks will need to be further examined.

SI and chronic psychosocial stresses have been associated with systemic inflammation characterized by increases in circulating cytokines in humans (Smith et al., 2020; Zilioli and Jiang, 2021; Miller et al., 2019), and hippocampal and systemic inflammation co-occur in animal models of SI and psychiatric disorders (Lappizo et al., 2021; Shetty et al., 2017). Our data with SI in WKY rats demonstrated consistent increases in several proinflammatory and anti-inflammatory circulating cytokines, demonstrating that 8-week of adolescent SI was sufficient to induce systemic inflammation. SI-induced systemic inflammation might be linked to brain inflammation in stress responsive regions including the hippocampus, and more information on these neurological stress-responsive regions can be found in our companion paper (Valdez et al., 2022). It is likely that reduced peripheral effectiveness of glucocorticoid-induced immunosuppression could be a contributing factor, but these mechanisms will need to be explored further using targeted animal studies.

We explored the hypothesis that prior exposure to SI could exacerbate the effects of subsequent stressor response. We noted that ozone exposure exacerbated pulmonary injury and inflammation in animals subjected to SI. The mechanism by which this response is exacerbated could involve the effect of SI on SAM- and HPA-mediated peripheral stress response, since adrenal-derived stress hormones modulate pulmonary injury and inflammation (Henriquez et al., 2018, 2019, 2021). The availability of glucocorticoids was postulated to be involved in modulating eosinophilic inflammation in sensitized mice challenged with ovalbumin (Kumlien et al., 2008), indicating that SI modulation of circulating glucocorticoids could modify the subsequent response to a challenge stressor.

Because systemic inflammation has been often linked with dyslipidemia (Collado et al., 2021), we examined if systemic inflammation in CS and SI rats was associated with changes in circulating lipids. We noted that, relative to NS, SI and, to a smaller extent CS, demonstrated increases in levels of circulating cholesterol but not triglycerides. Previous studies have shown that SI exacerbates diet-induced LDL cholesterol increases in prepubertal and adolescent Wistar rats (Arcego et al., 2014). Rats subjected to CS also have shown increases in circulating cholesterol (Neves et al., 2009). However, it should be noted that the dietary composition in different studies might impact the circulating levels of lipids and likely interactively influence stressor effects. Whereas acute exposure to ozone as a stressor is shown to increase circulating metabolites including glucose, leptin, cholesterol, triglycerides, and BCAA reflective of acute stressor effects (Miller et al., 2015; 2016a, 2016b 2016c), it is conceivable that exacerbated increases in cholesterol are mediated through common mechanistic pathways between ozone and SI. The differential exacerbation of SI on ozone-induced increases in cholesterol might suggest neural mechanisms regulating metabolic processes, in addition to contribution of adrenal hormones on peripheral tissues.

Based on the evidence that acute ozone-induced neuroendocrine stress response also inhibits the hormones associated with hypothalamus-pituitary-thyroid (HPT) and hypothalamus-pituitary-gonadal (HPG) axes (Henriquez et al., 2019; 2022), we explored the contribution of CS and SI in modulating HPT and HPG-related hormone levels. Generally, we noted that the levels of circulating PRL, LH, and FSH (hormones related to gonadal axis) were decreased in SI and CS animals exposed to air. These observation are consistent with studies showing impaired spermatogenesis in male rats exposed to unpredicted stress (Zou et al., 2019) and decreased LH in rats exposed to stress (Kirby et al., 2009). The altered sensitivity of regulatory corticotropin-releasing hormone on HPA-mediated effects on HPG and HPT axes could be responsible for the inhibitory effects of SI and CS (Raftogianni et al., 2018; Sterrenburg et al., 2011). Acute ozone exposure severely depleted circulating TSH, PRL and LH in this study, consistent with prior studies (Henriquez et al., 2019; 2022), regardless of CS or SI. Further, long-term SI and CS (SI>CS)-induced depletion at baseline might be involved in increased GnRH in ozone-exposed SI rats. Increased GnRH in ozone-exposed SI rats could result from central influence of collective depletion of HPG-related hormones. These neuroendocrine effects of CS, SI, and acute ozone, and the resultant changes in circulating hormones might interactively contribute to immune and metabolic response to stress.

The mechanisms of how stressors and subsequent environmental exposures impact physiological processes are likely sex-dependent (Iqbal et al., 2010; Stephens et al., 2016); however, we only considered male rats for the current experiment. Although male WKY rats are more susceptible to second-hand tobacco smoke-induced lung pathology than females (Shen et al. 2016), the future analyses should include females as sex hormones will likely modulate stressor effects. We chose to employ CS with a shorter duration of application for 4 stressors (1 hour) to ensure a mild stress protocol, but longer application could likely have resulted in more significant changes in stress biomarkers. Animal handling is important to consider for subsequent follow-up experiments as the single-housed CS group without enrichment, when handled more regularly than the SI and exposed to more activity throughout the study period caused milder stressor effects. It is important to note that handling of SI animals prior to necropsy for application of anesthesia could not be avoided. We did not include behavioral testing paradigm for assessing stress-related physiological impairment status since we were focused on peripheral effects and not mechanistic neural outcomes. Finally, some of the effects of stressor interaction might be masked by a strong acute stressor response of ozone, such as inhibition of TSH, PRL, and LH. The temporality and reversibility of stressor effects upon termination of stress applications was not considered in the present study. Mechanistic studies are needed to understand how sympathetic mediators and steroidal hormones interactively orchestrate peripheral and central stress responses and how chronic stressors and ozone in chronic exposures exacerbate psychiatric disorders.

In conclusion, the use of CS and/or SI provided the opportunity to examine interactive effects of psychosocial stressors on ozone-induced immune and metabolic alterations, and allowed investigation of the potential link between successive stressors and neuroendocrine pathways (Table 1). The neuroendocrine stress effects of ozone were apparent, as urinary catecholamines and corticosterone increased regardless of stress condition. Ozone-induced lung injury, inflammation and cytokine responses were exacerbated but only in SI, and not CS, with coexistent SI where animals were handled frequently together with mild stress applications. SI and, to a limited extent CS, resulted in systemic inflammation as evidenced by increases in proinflammatory and anti-inflammatory cytokines and associated increases in circulating cholesterol. These systemic effects were exacerbated by acute ozone exposure in SI but not CS (Table 1). SI and CS without ozone exposure decreased levels of circulating hormones regulating HPG axis (SI>CS). Interestingly, these stress effects were exacerbated in rats exposed to ozone. Thus, immune and metabolic disease phenotypes of SI and CS might be exacerbated by exposure to air pollutants through changes in neuroendocrine activity.

Highlights.

• Social isolation but not mild chronic stress in rats caused systemic inflammation

• Social isolation decreased pituitary-gonadal and chronic stress thyroid hormones

• Ozone-induced lung injury and inflammation were exacerbated in socially isolated rats

• Ozone exposure exacerbated systemic inflammation observed in socially isolated rats

• Ozone increased epinephrine and corticosterone but decreased other pituitary hormones