The interactions of diet-induced obesity and organophosphate flame retardant exposure on energy homeostasis in adult male and female mice.

Abstract

Previously, sex-dependent alterations in energy homeostasis were reported in adult mice fed a standard chow attributed to exposure to a mixture of organophosphate flame retardants (OPFRs) via estrogen receptors (ERα). In this study, adult male and female mice (C57BL/6J; Taconic) were treated with the same mixture of OPFRs (1 mg/kg each of tricresyl phosphate (TCP), triphenyl phosphate (TPP), and tris(1–3-dichloro-2propyl)phosphate (TDCPP) for 7 weeks on a low-fat diet (LFD, 10% kcal fat) or a high fat (HFD, 45% kcal fat) in a diet-induced obesity model. Consistent with our previous observations, OPFRs altered weight gain in males, differentially with diet, while females remained unaffected. OPFR treatment also revealed sex-dependent perturbations in metabolic activity. During the night (approximately 0100–0400 hr), males exhibited elevated activity and oxygen consumption, while in females these parameters were decreased, irrespective of diet. OPFR disrupted feeding behavior and abolished diurnal water intake patterns in females, while increasing nighttime fluid consumption in males. Despite no marked effect of OPFRs on glucose or insulin tolerance, OPFR treatment altered circulating insulin and leptin in females and ghrelin in males. Data indicate that adult OPFR exposure might influence, and perhaps exacerbate, the effects of diet-induced obesity in adult mice by altering activity, ingestive behavior, and metabolism.

INTRODUCTION

Metabolic syndrome, a constellation of conditions including obesity, hypertension, dyslipidemia, and pre-diabetes, has emerged as a national health crisis affecting over 90 million adults and costing over $100 billion each year (Boudreau et al. 2009). A major factor underlying the alarming rise in metabolic syndrome is the consumption of western diets high in fat and sugar (Drake et al. 2018; Rodriguez-Monforte et al. 2017; Moreno-Fernandez et al. 2018). However, it is also clear that diet is not the only factor. Other factors include environmental exposure to endocrine-disrupting compounds (EDCs) that perturb nutrient and hormone metabolism through central and peripheral actions (Decherf and Demeneix 2013). Indeed, investigators demonstrated that exposure to EDCs or metabolic disruptors increase sensitivity to the western diet (Brulport et al. 2017; Strakovsky et al. 2015; Grun and Blumberg 2009; Mackay et al. 2013). One group of ubiquitous EDCs are flame retardants used in the production of electronics, furniture, toys, and foodstuffs (Yang et al. 2019; Peng et al. 2020; Young et al. 2018; Li et al. 2019). The flame retardant market used to be dominated by polybrominated diphenyl ethers (PBDE), before being phased out of American and European production due to neurological and metabolic health concerns in 2004 (Zota et al. 2011; Herbstman et al. 2010; Gilbert et al. 2012; Shaw et al. 2010; Dorman et al 2018).

Recently, an alternative class of retardants, organophosphate flame retardants (OPFRs), has emerged as the leading replacement for PBDEs. Despite already existing toxicity data on these flame retardants at high concentrations, OPFRs quickly became widely used among home furnishing manufacturers resulting in widespread human exposure. OPFRs are embedded in household products and released into household and workplace dust through which humans are exposed primarily via inhalation and ingestion, resulting in biologically relevant levels in human serum (680–709 ng/g lipid), urine (1–10 ng/ml), and breast milk samples (1–10 ng/ml) (Ma et al. 2017; 2019; Butt et al. 2014; Meeker et al. 2013; Hoffman et al. 2017). While OPFRs are not yet reported to accumulate within adipose tissue to the same degree as do PBDEs, aryl OPFRs are hydrophobic (Yang et al. 2019) and demonstrate accumulation in other biological tissues (Hou et al. 2017; Ma et al. 2013). Further, several investigators demonstrated OPFRs’ ability to interact with nuclear receptors important in the pathogenesis of metabolic syndrome (Gray et al. 2005; Pap et al. 2016; Belcher et al. 2014; Pillai et al. 2014), leading to concern over potential long-term adverse health effects.

Homeostatic regulation of feeding behaviors and energy balance is a complex system but predominantly controlled via neuroendocrine pathways originating in the hypothalamus (Waye and Trudeau 2011). Briefly, the hypothalamus consists of multiple nuclei in which discrete neuronal subgroups communicate with each other to integrate peripheral indicators of energy state (Williams et al. 2001). With emotional and reward inputs from the limbic forebrain, the hypothalamus synthesizes feeding drive and communicates with the hindbrain for execution (Grill and Hayes 2012; Berthoud 2002). Within the hypothalamus lies the arcuate nucleus (ARC) which sits adjacent to a leaky portion of the blood-brain-barrier, and thus its neurons are in a unique position to directly sense energy state through peripheral signals such as glucose, insulin, leptin, and ghrelin (Schwartz et al. 2000; Saper et al. 2002). ARC neurons express receptors for these molecules, and their combined inputs to the paraventricular nucleus (PVN) and lateral hypothalamus (LH) help dictate food intake (Arora and Anubhuti 2006; Nahon 2006). Because hypothalamic control of energy homeostasis is highly regulated through hormone signaling pathways including estrogen receptors (ER) α and peroxisome proliferator-activated receptor (PPAR) γ (Sarruf et al. 2009; Garretson et al. 2015; Roepke et al. 2011; Mauvais-Jarvis et al. 2013), any EDC, such as OPFRs, that interact with these receptors may disrupt the complex balance of these pathways, sensitizing the system to metabolic disorders such as obesity and diabetes.

Despite the increasing focus of epidemiological and toxicological data surrounding OPFR exposure, physiological influences of OPFRs on energy homeostasis in adult mammals is underexplored. While most endocrine disruption studies are focused on developmental exposure, it is important to understand the mechanistic role of EDC exposures throughout the lifespan, including adulthood. Little is yet known on how adult exposure to OPFR may interact with neuroendocrine control over energy homeostasis; however, in our previously published exploratory study, adult, sub-chronic OPFR exposure decreased body weight gain and energy intake in intact male mice (Krumm et al. 2018). Further, expression of genes central to hypothalamic control of energy homeostasis were markedly altered by OPFR exposure, differentially in intact males and ovariectomized female mice (Krumm et al. 2018). These data indicate a crucial need for further investigation before OPFR may be considered any safer than their PBDE predecessors.

Since ERα and PPARγ receptors are highly expressed in the ARC and hypothalamus as a whole, and because OPFRs are known to interact with these receptors, OPFRs may be disrupting energy homeostasis as a consequence of these interactions. Previously, Krumm et al (2018) used triphenyl phosphate (TPP), tricresyl phosphate (TCP), and tris(1–3-dichloro-2propyl)phosphate (TDCPP) in a mixture of 1 mg/kg of each OPFR. This mixture was selected because of the prevalence of these compounds in human environments and since the parent compounds or their metabolites interact with ERα and PPARγ, which may play a role in pathogenesis of these disorders. Since Krumm et al (2018) reported that exposure of adults to OPFRs elicited sex-specific changes in body weight, peripheral peptide hormone expression, and gene expression, it was of interest to determine whether this may translate to an increased sensitivity to diet-induced obesity, attributed to effects on feeding behavior, fat accumulation, metabolism, and activity patterns. Thus, the aim of this study was to investigate these parameters in intact adult male and female mice with or without a high-fat diet (HFD) challenge for 7 weeks with continuous daily oral dosing of the same 1 mg/kg OPFR mixture. This dose was selected to be consistent with previous observations (Krumm et al. 2018; Patisaul et al. 2013; Wang et al. 2019b), and because Krumm et al (2018) reported murine serum concentrations of TPP, TCP, and TDCPP similar to those detected within human serum samples (Ma et al. 2017).

MATERIALS AND METHODS

Animals

All animal experiments were approved by the Rutgers University Institutional Animal Care and Use Committee and followed guidelines based upon National Institutions of Health standards. Female and male wild-type (WT C57BL/6J; Taconic) mice bred in-house and provided food and water ad libitum under controlled temperature (23 °C) and light cycle (12/12 hr light/dark cycle). At weaning, animals were ear-tagged for identification and fed a standard low-phytoestrogen chow diet (Lab Diets 5V75) until the start of the experiment.

To examine the effects of OPFR exposure on an adult mouse model of diet-induced obesity, mice were fed either a low-fat diet (LFD, 3.85 kcal/g, 10% fat, 20% protein, 70% carbohydrate; D12450H) or high-fat diet (HFD, 4.73 kcal/g, 45% fat, 20% protein, 35% carbohydrate; D12451; Research Diets). Mice were fed LFD or HFD concurrently with OPFR treatment starting at 10 weeks of age and separated into weight-matched groups (male: OIL – 27.4 ± 0.6 g, OPFR – 27.1 ± 0.4 g; female: OIL – 20.6 ± 0.3 g, OPFR – 20.9 ± 0.2 g) and dosed for the duration of the study.

Organophosphate flame retardants (OPFR) Dosing

The OPFR mixture consisted of 100 mg each of tricresyl phosphate (TCP, CAS #1330–78-5; purity = 99%) which was purchased from AccuStandard (New Haven, CT), and triphenyl phosphate (TPP, CAS #115–86-6; purity 99%) and tris(1,3-dichloro-2-propyl)phosphate (TDCPP, CAS #13674–87-8; purity 95.6%) purchased from Sigma-Aldrich (St. Louis, MO). One hundred (100) mg of each OPFR were dissolved together in the same 1 ml of acetone (Sigma-Aldrich) for long term storage. One hundred (100) μl of acetone stock was transferred to 10 ml sesame oil (Sigma-Aldrich) to create a 1 mg/ml mixture of OPFR-oil. The mixture was stirred for 48–72 hr to completely evaporate the acetone from the mixture. For dosing, the OPFR mixture or vehicle (oil:acetone) was mixed with powdered peanut butter (approximately 50 mg) for oral dosing of mice on per body weight basis for a total exposure of 1 mg/kg of each OPFR/day of OPFR-oil or vehicle control-oil. Weekly body weights were recorded and used for dosage calculation. Starting at 10 weeks of age, all mice were dosed at 900–1100 hr daily for an approximate 7 total weeks in a subchronic paradigm.

Experimental Design

Adult male and female mice (n = 16 males, n = 14 females) were pair-housed and weight-matched per group, fed LFD or HFD, and dosed with either vehicle-oil or OPFR-oil for 4 weeks in two sequential batches of mice (8 males/batch; 6–8 females/batch) to ensure sufficient sample size for metabolic and feeding behavior investigations. Body composition (fat and lean mass) was assessed by EchoMRI™ Body Composition (Houston, TX) on the day of first dose. Body weight and food intake were measured weekly. After 4 weeks exposure, body composition was determined followed by Comprehensive Lab Animal Monitoring System (CLAMS, Columbus Instruments, Columbus, OH) to measure oxygen consumption (V.O₂), carbon-dioxide production (V.CO₂), respiratory exchange ratio (RER), energy expenditure, and general locomotor activity in 72 hr trial under constant 25 °C and 12:12 hr light/dark cycle. Mice were single-housed for the duration. The respiratory exchange ratio (RER) is a measurement of substrate utilization (ratio of carbohydrates vs. lipids). General metabolic rate is also determined through mouse heat expenditure. Food and water intake and activity (X, Y, and Z plane and running wheel) were also recorded. After CLAMS, mice were transferred to the Biological Data Acquisition (BioDAQ, Research Diets, New Brunswick, NJ) chambers for 1 week with 72 hr habituation and 96 hr measurement of feeding behaviors (meal size, frequency, duration). LFD or HFD chow were contained in a touch-sensitive hopper and food consumption was measured by decreases in hopper food weight. Whenever the mouse touched the hopper for food, the system denoted that as a “bout.” When the interval between bouts was greater than 300 sec, the food ingested was determined to be a “meal.” A meal could consist of any number of bouts, until the inter-bout interval exceeded 300 sec. Subsequently, all mice were tested for glucose and insulin tolerance. For the glucose tolerance test (GTT), mice were fasted for 5 hr and then intraperitoneally (IP) injected with a bolus of 2 g/kg glucose. Blood-glucose was measured from tail bleeds using an AlphaTrak glucometer (Zoetis, Parsippany, NJ). Glucose measurements were taken at 0, 15, 30, 60, 90, and 120 min post-injection. Four days later, insulin tolerance tests (ITT) were performed using an IP injection of 0.75 U/kg insulin after a 4 hr fast. After insulin injection, glucose was measured in tail-blood at 0, 15, 30, 60, 90, and 120 min. With 1-week recovery from ITT, mice were dosed at 0900 hr, fasted at 1000 hr, and euthanized at 1100 hr by decapitation after sedation with ketamine (100 mg/ml). Female mice were euthanized during diestrus to control for cycling steroid hormone levels. Trunk blood was collected in K⁺-EDTA coated tubes with the addition of proteinase inhibitor 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (1 mg/ml, Sigma-Aldrich) to protect against peptide degradation. Samples were maintained on ice until centrifugation at 1,000g for 15 min at 4 °C. Plasma supernatant was collected and stored at −80 °C until analysis for insulin, leptin, and ghrelin levels, using a multiplex assay (MMHMAG-44 K, EMD Millipore, Billerica, MA).

Data Analysis

All data are presented as mean ± SEM. Data were analyzed using GraphPad Prisim software (GraphPad Software, LA Jolla, CA) by a two-way ANOVA (OPFR and Diet) with a post-hoc Newman-Keuls multiple comparisons test, or with Statistica 7.1 software (StatSoft, Tulsa, OK, USA) by multi-factorial ANOVA or with repeated-measures, three-way ANOVA (Diet, OPFR, Time), followed with post-hoc Newman-Keuls multiple comparisons test. Effects were considered significant at p ≤ 0.05.

RESULTS

Physiological Parameters

Body weight gain and crude food intake of mice fed control-oil and OPFR-oil mixture (1 mg/kg each of TCP, TPP, and TDCPP) were taken over the course of 4 weeks and followed by body composition assessment by EchoMRI™ (Figures 1 and 2). Feeding efficiency was calculated using the ratio of body weight gain to crude food intake and depicted as grams (g) gained to kcal consumed. As expected, bodyweight gain and feeding efficiency were increased in all HFD fed compared to LFD animals. However, OPFR-treated male mice fed HFD exhibited significantly greater weight gain (Figure 1A) after 4 weeks over their oil-treated counterparts. HFD-fed male mice also displayed elevated fat mass and decreased lean mass following OPFR exposure (Figure 1C). While no marked main effect of OPFRs or interactions were observed in female mice, HFD-fed females gained more weight and more fat mass than LFD-fed counterparts. OPFR-treated mice administered HFD resulted in reduced lean mass (Figure 2D), whereas in oil-treated mice no marked difference in diet was noted, indicating a potential influence of OPFR exposure on HFD effects.

Figure 1.

Physiological parameters in WT males orally dosed with an OPFR mixture (1 mg/kg) for ~7 weeks. (A) Percent Body Weight Gain over 4 weeks; (B) Feeding Efficiency; (C) Percent Fat Mass; (D) Percent Lean Mass. Data were analyzed by a two-way ANOVA with post-hoc Newman-Keuls multiple comparisons test. Uppercase letters denote diet effects within treatment group. Lowercase letters denote treatment effects within diet (A/a=P<.05). Data (A, C, D: n=16; B: n=8 (per cage)) are presented as mean ± SEM.

Figure 2.

Physiological parameters in WT females orally dosed with an OPFR mixture (1 mg/kg) for ~7 weeks. (A) % Body Weight Gain over 4 weeks; (B) Feeding Efficiency; (C) Percent Fat Mass; (D) Percent Lean Mass. Data were analyzed by a two-way ANOVA with post-hoc Newman-Keuls multiple comparisons test. Uppercase letters denote diet effects within treatment group. Lowercase letters denote treatment effects within diet (A/a=P<.05). Data (A, C, D: n=14; B: n=7 (per cage)) are presented as mean ± SEM.

Metabolic Parameters

Metabolic parameters such as V.O₂, V.CO₂, RER, and heat were measured in a 72 hr run in the CLAMS system (Figures 3 and 4). While diet-induced patterns were distinct in both males and females, the OPFR mixture effect appeared only in LFD groups and during peak feeding time around 200–300 hr for both sexes. In LFD-fed males, OPFR significantly augmented V.O₂ and V.CO₂ during 200–300 hr (Figure 3A), while in contrast LFD-fed females responded with diminished V.O₂ during 200–300 hr (Figure 3A) and decreased V.CO₂ during 200–400 hr (Figure 3B). It is noteworthy that females fed LFD also exhibited significantly elevated RER from 300–500 hr (Figure. 4C), as well as reduced RER from 1700–1800 hr, right before lights off (Figure 4C). There were no marked main effects in RER for males. Finally, OPFR decreased energy expenditure in females fed both LFD (Figure 4D, 200–400 hr), and females fed HFD (Figure 4D 100–200 hr). Interestingly, a significant rise was found in male mice fed LFD in energy expenditure during 200–300 hr (Figure 3D) indicating contrasting differences in activity dependent upon sex. Diet did not exert a marked effect on energy expenditure, save for an elevated energy expenditure in LFD-fed males following OPFR treatment at a single time point: 200–300 hr (Figure 3D).

Figure 3.

Analysis of metabolism in WT male mice orally dosed with an OPFR mixture (1 mg/kg) for ~5 weeks. (A) V.O₂; (B) V.CO₂; (C) Respiratory Exchange Ratio; (D) Energy Expenditure. Dark line above X-axis represents dark/light hours. Data were analyzed with a repeated-measures three-way ANOVA with post-hoc Newman-Keuls multiple comparisons test. Lowercase letters denote treatment effect within diet group. Uppercase letters denote diet effects within treatment group, or when barred, denote comparisons between day and night (A/a=P<.05). Data (n=16) are presented as mean ± SEM.

Figure 4.

Analysis of metabolism in WT female mice orally dosed with an OPFR mixture (1 mg/kg) for ~5 weeks. (A) V.O2; (B) V.CO2; (C) Respiratory Exchange Ratio; (D) Energy Expenditure. Dark line above X-axis represents dark/light hours. Data were analyzed with a repeated measures three-way ANOVA with post-hoc Newman-Keuls multiple comparisons test. Lowercase letters denote treatment effect within diet group; uppercase letters denote diet effects within treatment group, or when above a capped line, denote comparisons between day and night (A/a=P<.05). Data (n=14) are presented as mean ± SEM.

Activity

In the same CLAMS system, activity levels were also measured over 72 hr (Figure 5). As mice are a nocturnal species, movement and use of the exercise wheel were increased during nighttime in all groups. However, in female mice, OPFR significantly decreased nighttime activity and wheel use in both LFD (Figure 5E and 5F) and HFD groups (Figure 5E and 5F). There was a significant main effect of OPFR on both activity and wheel use (Figure 5E and 5F), as well as an interaction between OPFR and time (Figure 5E and 5F). Male mice did not exhibit this OPFR-induced pattern but did respond to HFD with a reduction in nighttime activity (Figure 5C).

Figure 5.

Analysis of daytime vs. nighttime activity in WT male and female mice orally dosed with an OPFR mixture (1 mg/kg) for ~5 weeks. (A & D) Water Intake; (B & E) Locomotor Activity; and (C & F) Wheel Running. Data were analyzed by a three-way ANOVA with post-hoc Newman-Keuls multiple comparisons test. Lowercase letters denote treatment effect within diet group. Uppercase letters denote diet effects within treatment group, or when above a capped line, denote comparisons between day and night (A/a=P<.05). Data (males: n=16; female: n=14) are presented as mean ± SEM.

Perhaps more interestingly, water intake was augmented in males during the nighttime in both LFD- and HFD-fed groups (Figure 5A). OPFR induced an overall significant effect as well as OPFR interactions with both time and diet (Figure 5A). Further, OPFR exposure induced marked differences between diets and time (Figure 5A), while differences were not observed in oil-treated male mice. Females exhibited the opposite effect, wherein oil-treated groups exhibited typical elevated water intake during the night (Figure 5D), but within OPFR treatment, the differences between daytime and nighttime drinking were abrogated, indicating potential dysregulation of diurnal fluid intake behaviors.

Feeding Behaviors

The BioDAQ apparatus was utilized for reliable analysis of total and hourly food intake, as well as meal size, duration, and frequency over a 96 hr trial period (Figure 6). Overall, OPFR-treated female mice on a HFD ate less food (Figure 6A) and consumed fewer meals per day (Figure 6B) than oil-treated counterparts. When hourly feeding patterns were analyzed, the difference between HFD-fed groups was also evident with OPFR-treated females who consumed less HFD than oil-treated during two periods during the dark cycle. Oil-treated HFD mice displayed a spike in food intake at 0300 hr, whereas OPFR-treated HFD animals food intake was significantly less during this time (Figure 6E). In addition, OPFR exposure decreased consumption of HFD at 2000 hr, respective of either treatment or diet (Figure 6E). These time-specific perturbations suggest OPFR may be dysregulating diurnal feeding patterns in female mice. Surprisingly, no significant effects of OPFRs were observed on meal patterns in males; however, in the analysis of hourly consumption, OPFR treatment significantly altered HFD consumption at two time points 0400 hr and 2100 hr similar to females (data not shown).

Figure 6.

Feeding behaviors in WT females orally dosed with an OPFR mixture (1 mg/kg) for ~5 weeks. (A) Total Food Ingested; (B) Meals Frequency; (C) Meal Duration; (D) Meal Size; and (E) Average Hourly Food Intake. Data were analyzed by a two-way ANOVA (A-D) and a repeated-measures, three-way ANOVA (E) with post-hoc Newman-Keuls multiple comparisons test. Lowercase letters denote treatment effects within diets. Uppercase letters above a capped line denote diet effects within treatment group (A/a=P<.05). Data (n=14 for LFD groups and n=8–10 for HFD groups) are presented as mean ± SEM.

Glucose and Insulin Tolerance

Glucose and insulin tolerance tests (Figures 7 and 8) were performed to determine whether endocrine disruption by OPFRs might compromise the body’s ability to tolerate sudden changes in glucose homeostasis. Overall, there were minimal to no alterations in glucose or insulin tolerance attributed to OPFR exposure. As expected, HFD elevated circulating glucose levels and increased the latency time to return to baseline after glucose or insulin injection, but did not display significant differences between treatments.

Figure 7.

Glucose tolerance tests in WT male and female mice orally dosed with an OPFR mixture (1 mg/kg) for ~6 weeks. (A & D) Fasting Glucose, (B & E) GTT, and (C & F) Area under the curve (AUC) of GTT. Data were analyzed by a two-way ANOVA (A, C, D, F) or a repeated-measures, three-way ANOVA (B, E) with post-hoc Newman-Keuls multiple comparisons test. Uppercase letters above a capped line denote diet effects within treatment group and lowercase letters denote treatment effect within diet (A/a=P<.05). Data was collected only from the first batch (n=8 for all groups) and are presented as mean ± SEM.

Figure 8.

Insulin tolerance tests in WT male and female mice orally dosed with an OPFR mixture (1 mg/kg) for ~6 weeks. (A & C) ITT and (B & D) AUC of ITT. Data were analyzed by a two-way ANOVA (B, D) or a repeated-measures, three-way ANOVA (A, C) with post-hoc Newman-Keuls multiple comparisons test. Capped letters denote diet effects within treatment group and lowercase letters denote treatment effect within diet (A/a=P<.05). Data was collected only from the first batch (n=8 for all groups) and are presented as mean ± SEM.

Peptide Hormones

Terminal plasma hormone levels of insulin, leptin, and ghrelin were measured. Alterations in these hormones indicate perturbed energy homeostatic control in peripheral endocrine organs. Insulin levels were elevated in HFD-fed males as compared to LFD-fed males in the OPFR group (Figure 9A), where there was no significant difference in controls. Interestingly, in female mice OPFR exposure increased insulin in LFD-fed females compared to oil-treated counterparts, producing a HFD-induced fall in insulin (Figure 9D). In males, HFD elevated plasma leptin concentrations in both treatment groups (Figure 9B). In female mice, OPFR induced a rise in leptin in HFD-fed group compared to oil-treated (Figure 9E). OPFR treatment suppressed ghrelin in males (Figure 9C) and eliminated the HFD-induced reduction in ghrelin. In females, HFD increased ghrelin levels in both treatments (Figure 9F). A summary of these results is presented in Table 1.

Figure 9.

Terminal plasma peptide hormone concentrations in WT male and female mice orally dosed with an OPFR mixture (1 mg/kg) for ~7 weeks. (A & D) Insulin; (B & E) Leptin; and (C & F) Ghrelin. Data were analyzed by a two-way ANOVA with post-hoc Newman-Keuls multiple comparisons test. Uppercase letters denote diet effects within treatment group and lowercase letters denote treatment effects within diet (A/a=P<.05). Data was collected only from the first batch (n=8 for all groups) and are presented as mean ± SEM.

Table 1.

Summary of data comparing oil- and OPFR-treated groups.

Endpoint	Males		Females
	LFD	HFD	LFD	HFD
Bodyweight Gain	n.s.	↑↑	n.s.	n.s.
Feeding Efficiency	n.s.	n.s.	n.s.	n.s.
Fat Mass	n.s.	↑↑	n.s.	n.s.
Lean Mass	n.s.	↓↓	n.s.	n.s.
V.O2	↑	n.s.	↓	n.s.
V.CO2	↑	n.s.	↓	n.s.
RER	n.s.	n.s.	↓↑	n.s.
Energy Expenditure	↑	n.s.	↓	↓
Activity	n.s.	n.s.	↑↑	↓↓
Water Intake	↑↑	↑↑	↓↓	↓↑
96 hr Food Intake	n.s.	n.s.	n.s.	↓↓
Hourly Food Intake	n.s.	↓↑	n.s.	↓↑
Meal Frequency	n.s.	n.s.	n.s.	↓↓
Meal Duration	n.s.	n.s.	n.s.	n.s.
Meal Size	n.s.	n.s.	n.s.	n.s.
Fasting Glucose	n.s.	n.s.	n.s.	n.s.
Glucose Tolerance	n.s.	n.s.	n.s.	n.s.
Insulin Tolerance	n.s.	n.s.	n.s.	n.s.
Insulin	n.s.	n.s.	↑↑	n.s.
Leptin	n.s.	n.s.	n.s.	↑↑
Ghrelin	↓↓	n.s.	n.s.	n.s.

Up arrows denote an OPFR-induced increase and down arrows denote an OPFR-induced decrease. One arrow indicates a modest effect and two arrows indicate a strong effect. One up and one down arrow indicates a mixed effect dependent on time of day. N.S = not significant.

DISCUSSION

Our understanding of the influence of OPFR exposure and its actions as an EDC primarily focused on developmental exposure (Wang et al. 2019a; 2019b). Previous perinatal studies with FM550 identified TPP as a candidate for metabolic alterations through actions on PPARγ (Pillai et al. 2014), as well as changes in behavior and glucose homeostasis (Patisaul et al. 2013; Baldwin et al. 2017; Du et al. 2016). In the current study using a subchronic adult exposure model, weight gain in male mice was detected when exposed to OPFR and fed HFD compared to HFD-fed oil-treated males. The rise in body weight was due to enhanced fat accumulation, suggesting OPFR augmentation of fat accumulation occurs on HFD, a function mediated through PPARγ (Janani and Ranjitha Kumari 2015). Conversely, OPFR-treated males fed LFD gained less weight compared to oil-treated males, which is in agreement with our previous findings (Krumm et al. 2018). Interestingly, our current observations in intact wild-type (WT) females are not similar to our earlier results wherein adult OPFR exposure increased weight gain in ovariectomized WT females, which is known to enhance body weight gain in WT females (Krumm et al. 2018). This discrepancy may also be due to differences between a semi-purified LFD in the current investigation and a standard chow diet used in the previous experiment. Regardless, there does appear to be a disruptive capacity of OPFR on weight gain, perhaps differentially dependent upon sex, diet, or ovarian status. Aside from weight gain, a decrease in food intake and meals per day was observed in HFD-fed male mice, suggesting that OPFR treatment alters ingestive behaviors and metabolism. Further, circulating insulin levels were elevated by OPFR in female mice, as was leptin, offering a potential explanation for why female mice did not undergo a weight-gain effect noted in male mice. Conversely, male mice showed no marked change in insulin or leptin, but exhibited lowered ghrelin levels. Taken together, these alterations demonstrated an overall anorexigenic effect of OPFR, and support OPFR’s capacity as an EDC to disrupt energy homeostasis in a sex-dependent fashion.

Adult exposures to other EDC previously reported metabolic effects including and not limited to glucose homeostasis, thyroid hormone levels, and fat metabolism (Ding et al. 2014; Marmugi et al. 2014; Moghaddam et al. 2015; Brulport et al. 2017; Bertuloso et al. 2015; Sharan et al. 2014; Jansen et al 2017). Data indicate that OPFRs disrupt metabolism (Fernie et al. 2015; Wang et al. 2019a; Du et al. 2016), supporting our examination of metabolic parameters including V.O₂, V.CO₂, RER, and energy expenditure. In males, significant perturbations in metabolism were observed during the nighttime. However, in female mice, OPFR treatment consistently decreased V.O₂, V.CO₂, and energy expenditure on a LFD during the night. In addition, RER was significant elevated on LFD. The shift in RER may indicate an enhanced carbohydrate utilization by OPFR on LFD, as opposed to lipids, during the night. However, this rise occurred concomitantly with a fall in carbohydrate usage during the later afternoon/early evening, prior to nighttime when the mice are more active. More notable than subtle substrate utilization alterations, the decreased V.O₂, V.CO_2, and energy expenditure indicates that mice are simply using less energy overall i.e., less active, when dosed with OPFRs. This response may be due to a variety of mechanisms, but the simplest cellular targets are mitochondria. Mitochondria are responsible for respiration and energy production at the cellular level, and if OPFRs are impinging on mitochondrial function, it may result in basal perturbations of metabolic mitochondrial activity. This hypothesis is supported by a recent study in which both TPP and TDCPP were found to decrease basal mitochondrial respiration in zebrafish embryos (Lee et al. 2019). These investigators also reported a reduction in maximal mitochondrial respiration attributed to TDCPP exposure (Lee et al. 2019). Further, in the nematode C. elegans, TPP and TDCPP both disrupted mitochondrial membrane permeability in a similar manner as brominated FR predecessors, indicating mitochondrial toxicity (Behl et al. 2016). These studies establish a precedent to support a hypothesis that respiratory effects that were noted may in part be due to a mitochondrial effect of OPFR. However, regardless of the cellular mechanism, our data suggest that even subchronic exposures to OPFRs altered respiration and suppressed metabolism, which over a lifespan may influence energy homeostasis leading to greater weight gain and adiposity.

Mice are generally nocturnal animals, and thus their activity is greatest during the night. In our experiment on mouse locomotor activity and wheel running, female mice were more susceptible to OPFR exposure than males displaying a marked reduction in both X-plane movement (general locomotor activity) and wheel running during the nighttime on both LFD and HFD. These data indicate that OPFR ingestion is interfering with neurological pathways that control activity, producing reduced locomotor motivation in females, but not males. Many of these pathways express ERα, a steroid receptor that is known to increase activity upon activation (Ogawa et al. 1998; 2003; Shughrue et al. 1997; Hatcher et al. 2018), or PPARγ, a nuclear receptor that is also involved in exploratory behaviors (Domi et al. 2016; Moreno et al. 2004). Interestingly, developmental exposure to the commercial OPFR mixture, FM550, in rats induced hyperactivity in females (Baldwin et al. 2017). The conflicting data between our findings and the FM550 study may be accounted for differences in species, exposure window, and chemical content. Regardless, there is a striking effect of OPFR exposure in adults on activity, and in regards to energy homeostasis, a reduction of activity (an increase sedentary behavior) is associated with an obese phenotype and its sequelae – metabolic syndrome and type 2 diabetes in rodent models and humans (de Rezende et al. 2014; Hamilton et al. 2014).

Locomotor motivation is a complex and multifaceted behavioral characteristic, which is influenced by more than just the search for food. Through the lens of energy homeostasis, activity is associated with energy expenditure, a process tightly under hormonal control and particularly through 17β-estradiol (Rettberg et al. 2014; Lopez and Tena-Sempere 2015). Energy expenditure is controlled, in part, by actions of hypothalamic neurons in the arcuate nucleus (Nahon 2006), and activity of these neurons may be modulated by 17β-estradiol in an energy state-sensitive manner (Roepke et al. 2011; Gao et al. 2007; Xu et al. 2011). Because OPFR interact with ERs, their exposure may be impinging on estrogenic mediation of energy homeostasis, and increasing the risk of metabolic disruption. Further, activity is not only dictated by energy status, motivation and mood may also be involved, and lack of motivation to move is symptomatic and causative for a variety of mood disorders (Zhai et al. 2015; Schwartz et al. 2000). Future studies may benefit from exploring OPFR action on brain regions involved in motivation such as the ventral tegmental area, ventral striatum, prefrontal cortex, amygdala, and dorsal media habenula (Hsu et al. 2014; Kim 2013).

In addition to energy expenditure, water intake was also dysregulated. Males exhibited an increase in water intake during the night on both diets, while females displayed an ablation of typical diurnal drinking patterns. A possible explanation for female mice is that dysregulation of water intake is a direct result of OPFR actions decreasing activity and energy expenditure. If animals are moving less, then the motivation to drink also diminishes. However, there are likely to be additional complexities to this simple explanation. Although female mice are moving less, these animals maintained the nocturnal pattern of enhanced activity. OPFR exposure seems to exert a specific effect of reducing nighttime water intake or increase daytime intake in females, thus eliminating differences between night and day. Further, despite no marked activity alterations in males, OPFR treatment elevated nighttime fluid consumption that indicating an impact of OPFR on control of fluid balance, either centrally or peripherally in the kidneys. Indeed, TDCPP produced cytotoxic effects on cultured renal cells (Killilea et al. 2017), and in human population study of approximately 1,500 patients. TDCPP exposure also correlated with markers of kidney damage and chronic renal disease (Kang et al. 2019).

Centrally, there are also many areas of the brain that control fluid balance including the paraventricular hypothalamus, supraoptic nucleus, median preoptic area, organum vasculosum laminae terminalis, and subfornical organ (Curtis 2009). Many of these nuclei express ERs and are involved in the control of fluid balance in response to 17β-estradiol (Shughrue et al. 1997; Santollo and Daniels 2015a; 2015b; Santollo et al. 2013; Curtis 2009). In hormone replacement therapies, E2 produced a direct effect on water intake (Krause et al. 2003; Santollo et al. 2013), its actions mediated in part through dampening of angiotensin II (AngII) signaling (Kisley et al. 1999; Findlay et al. 1979; Danielsen and Buggy 1980; Jonklaas and Buggy 1984). Potentially, OPFRs interfere with this estrogen-sensitive balance leading to changes in fluid intake. However, like any homeostatic function, thirst is regulated through a multitude of pathways, allowing for alternate avenues of OPFR actions. Thirst is closely related to energy homeostasis, and the powerful “hunger” hormone ghrelin is also known to exert effects on fluid intake, reducing water consumption by inhibiting Ang II (Mietlicki et al. 2009; Hashimoto et al. 2010; Plyler and Daniels 2017), which as previously indicated, is also under the influence of E2. Conversely, intracerebroventricular infusionsof Ang II diminishes food and enhances energy expenditure, establishing an Ang II link between food and fluid intake mediated by ghrelin (Porter and Potratz 2004). In our current study, OPFR decreased circulating ghrelin in male mice on LFD, supporting a ghrelin-mediated hypothesis for the dipsogenic effect of OPFR on male mice. Finally, somatostatin, produced both centrally in the ventromedial nucleus of the hypothalamus, and peripherally by delta cells in the digestive system, is involved in thirst generation and may be a target for OPFR dysregulation. Central action of somatostatin increases food and water intake (Karasawa et al. 2014; Stengel et al. 2010), and was shown to be altered by exposure to bisphenol A, another well-known estrogenic EDC (Facciolo et al. 2002; 2005). Taken together, these data offer a precedented route for OPFR EDC action on fluid regulation.

The collective observations of this study demonstrate a clear sex-dependent effect of OPFR exposure within adult mice. Body composition was only altered in males, while feeding behavior and activity were largely only modified within females. In addition, circulating ghrelin was diminished in males, while leptin and insulin were elevated within female mice. These differences are likely attributed to the innate biological differences between male and female mice. Energy homeostasis is well-documented to be a sexually-dimorphic function (Shi et al. 2009; Woods et al. 2003; Wu and O’Sullivan 2011), through which estrogen plays a substantial role (Rettberg et al. 2014; Nestor et al. 2014; Lopez and Tena-Sempere 2015). The decline in circulating estrogen following menopause is associated with adverse health effects such as weight gain, hot flashes, and increased risk of diabetes and cardiovascular disease (Cignarella and Bolego 2010; Clegg et al. 2017), and estrogen replacement therapies show marked protection against these effects (Warren et al. 2015). Estrogen is typically present at roughly 10-fold higher serum concentration within women than men, therefore OPFR interference with estrogenic signaling is likely to exert differing impacts depending upon sex. Further, Krumm et al., (2018) reported decreased expression of ERα in males, but not in ovariectomized females, demonstrating a sex-specific interaction of OPFR with estrogenic signaling. Expression of the other OPFR-target PPARγ was found to be upregulated, but not dependent upon sex. Sex-differences in PPARγ signaling is not well-known, but few studies implicate differential PPARγ expression and regulation of adipose tissue and immune function (Kadowaki et al. 2007; Park and Choi 2017). Fernandez et al. (2017) found that brain-specific knockout (KO) of PPARγ KO resulted in additional weight gain in female mice, whereas male KO mice gained less than their WT littermates. Taken together, these implicate that OPFR interactions with ERα and PPARγ may contribute to the observed sex-dependent effects noted in the current study.

Part of the reason for growing concern over EDCs is what is termed the “cocktail effect,” wherein exposures to multiple different EDCs might induce additive or synergistic effects, depending upon the period of exposure, developmental or throughout the lifespan. This is of particular concern for adult exposures due to the ability of most EDCs to bioaccumulate, resulting in chronic exposure of many, sub-adverse effect EDC exposures combining to culminate in significant disruptions (Lauretta et al. 2019). Thus, consistent with the multiplicity of EDC exposures in human, our experimental protocols used a mixture of organophosphate EDCs. While a more accurate-to-life model than studying the singular effects of one specific OPFR, the fact that different OPFR are known to exert both agonistic and antagonistic effects on ERs and PPARs (Kojima et al. 2013; Liu et al. 2013) makes identifying a mechanistic resolution of our findings more difficult. Future studies need to examine the effects of OPFR exposures in tissue- or cell-specific ERα and PPARγ knockout mouse models. Regardless, what remains is that OPFR appear to be interfering with estrogenic and/or PPARγ control of energy expenditure through receptor-mediated actions.

CONCLUSIONS

In summary, our findings indicated that a mixture of three common OPFR compounds exerted disruptive actions on energy homeostasis that subsequently interact with and potentially exacerbate diet-induced obesity. There were a multitude of sex-dependent effects on metabolism, energy expenditure, weight-gain, activity, water intake and circulating hormone concentrations. Most notable were OPFR alterations to water intake and behavioral activity. Our findings demonstrate the EDC capacity of OPFR to disrupt energy homeostasis, which increases the risk of metabolic disorders such as diet-induced obesity, diabetes, and metabolic syndrome. Despite the apparent risk, OPFR still continue to be the leading FR in the United States. This reality calls for the need of continued research, if there are to be regulatory actions taken to limit human OPFR exposure. With the multitude of OPFR endpoints, it may be a while yet before one knows the full scope of OPFR-mediated toxicity. However, it would be interesting to investigate the mechanistic roots of the dysregulated fluid homeostasis noted herein, as well as additional behavioral studies to tease out whether the sedentary behavior of OPFR-treated females is a mood, or motivation effect. Further, while this study focused primarily on peripheral and behavioral outcomes, energy homeostasis is tightly regulated through central processes in the hypothalamus. Thus, it will also be important to investigate potential OPFR actions on neuronal subpopulations that regulate feeding and reward pathways in the brain.